diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,129855 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 7636,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0,
+      "loss_partial": 0.349609375,
+      "step": 0
+    },
+    {
+      "epoch": 0,
+      "loss_partial": 0.458984375,
+      "step": 0
+    },
+    {
+      "epoch": 0.0005238344683080147,
+      "grad_norm": 201.01861572265625,
+      "learning_rate": 0.0,
+      "loss": 0.7756,
+      "step": 1
+    },
+    {
+      "epoch": 0.0005238344683080147,
+      "loss_partial": 0.376953125,
+      "step": 1
+    },
+    {
+      "epoch": 0.0005238344683080147,
+      "loss_partial": 0.388671875,
+      "step": 1
+    },
+    {
+      "epoch": 0.0010476689366160294,
+      "grad_norm": 200.69735717773438,
+      "learning_rate": 2.617801047120419e-08,
+      "loss": 0.7822,
+      "step": 2
+    },
+    {
+      "epoch": 0.0010476689366160294,
+      "loss_partial": 0.3671875,
+      "step": 2
+    },
+    {
+      "epoch": 0.0010476689366160294,
+      "loss_partial": 0.40234375,
+      "step": 2
+    },
+    {
+      "epoch": 0.001571503404924044,
+      "grad_norm": 200.6500244140625,
+      "learning_rate": 5.235602094240838e-08,
+      "loss": 0.7666,
+      "step": 3
+    },
+    {
+      "epoch": 0.001571503404924044,
+      "loss_partial": 0.400390625,
+      "step": 3
+    },
+    {
+      "epoch": 0.001571503404924044,
+      "loss_partial": 0.3828125,
+      "step": 3
+    },
+    {
+      "epoch": 0.0020953378732320588,
+      "grad_norm": 199.184326171875,
+      "learning_rate": 7.853403141361257e-08,
+      "loss": 0.7771,
+      "step": 4
+    },
+    {
+      "epoch": 0.0020953378732320588,
+      "loss_partial": 0.392578125,
+      "step": 4
+    },
+    {
+      "epoch": 0.0020953378732320588,
+      "loss_partial": 0.38671875,
+      "step": 4
+    },
+    {
+      "epoch": 0.0026191723415400735,
+      "grad_norm": 199.8030242919922,
+      "learning_rate": 1.0471204188481677e-07,
+      "loss": 0.793,
+      "step": 5
+    },
+    {
+      "epoch": 0.0026191723415400735,
+      "loss_partial": 0.400390625,
+      "step": 5
+    },
+    {
+      "epoch": 0.0026191723415400735,
+      "loss_partial": 0.412109375,
+      "step": 5
+    },
+    {
+      "epoch": 0.003143006809848088,
+      "grad_norm": 202.49392700195312,
+      "learning_rate": 1.3089005235602095e-07,
+      "loss": 0.7729,
+      "step": 6
+    },
+    {
+      "epoch": 0.003143006809848088,
+      "loss_partial": 0.447265625,
+      "step": 6
+    },
+    {
+      "epoch": 0.003143006809848088,
+      "loss_partial": 0.373046875,
+      "step": 6
+    },
+    {
+      "epoch": 0.003666841278156103,
+      "grad_norm": 197.7106170654297,
+      "learning_rate": 1.5706806282722514e-07,
+      "loss": 0.7666,
+      "step": 7
+    },
+    {
+      "epoch": 0.003666841278156103,
+      "loss_partial": 0.373046875,
+      "step": 7
+    },
+    {
+      "epoch": 0.003666841278156103,
+      "loss_partial": 0.365234375,
+      "step": 7
+    },
+    {
+      "epoch": 0.0041906757464641176,
+      "grad_norm": 212.06765747070312,
+      "learning_rate": 1.8324607329842932e-07,
+      "loss": 0.7666,
+      "step": 8
+    },
+    {
+      "epoch": 0.0041906757464641176,
+      "loss_partial": 0.376953125,
+      "step": 8
+    },
+    {
+      "epoch": 0.0041906757464641176,
+      "loss_partial": 0.38671875,
+      "step": 8
+    },
+    {
+      "epoch": 0.004714510214772132,
+      "grad_norm": 201.1160125732422,
+      "learning_rate": 2.0942408376963353e-07,
+      "loss": 0.7668,
+      "step": 9
+    },
+    {
+      "epoch": 0.004714510214772132,
+      "loss_partial": 0.408203125,
+      "step": 9
+    },
+    {
+      "epoch": 0.004714510214772132,
+      "loss_partial": 0.37890625,
+      "step": 9
+    },
+    {
+      "epoch": 0.005238344683080147,
+      "grad_norm": 200.5970458984375,
+      "learning_rate": 2.3560209424083772e-07,
+      "loss": 0.7571,
+      "step": 10
+    },
+    {
+      "epoch": 0.005238344683080147,
+      "loss_partial": 0.392578125,
+      "step": 10
+    },
+    {
+      "epoch": 0.005238344683080147,
+      "loss_partial": 0.34765625,
+      "step": 10
+    },
+    {
+      "epoch": 0.005762179151388162,
+      "grad_norm": 214.21119689941406,
+      "learning_rate": 2.617801047120419e-07,
+      "loss": 0.7454,
+      "step": 11
+    },
+    {
+      "epoch": 0.005762179151388162,
+      "loss_partial": 0.404296875,
+      "step": 11
+    },
+    {
+      "epoch": 0.005762179151388162,
+      "loss_partial": 0.318359375,
+      "step": 11
+    },
+    {
+      "epoch": 0.006286013619696176,
+      "grad_norm": 229.40335083007812,
+      "learning_rate": 2.879581151832461e-07,
+      "loss": 0.7302,
+      "step": 12
+    },
+    {
+      "epoch": 0.006286013619696176,
+      "loss_partial": 0.38671875,
+      "step": 12
+    },
+    {
+      "epoch": 0.006286013619696176,
+      "loss_partial": 0.375,
+      "step": 12
+    },
+    {
+      "epoch": 0.006809848088004191,
+      "grad_norm": 227.31634521484375,
+      "learning_rate": 3.1413612565445027e-07,
+      "loss": 0.7407,
+      "step": 13
+    },
+    {
+      "epoch": 0.006809848088004191,
+      "loss_partial": 0.27734375,
+      "step": 13
+    },
+    {
+      "epoch": 0.006809848088004191,
+      "loss_partial": 0.29296875,
+      "step": 13
+    },
+    {
+      "epoch": 0.007333682556312206,
+      "grad_norm": 250.37820434570312,
+      "learning_rate": 3.403141361256545e-07,
+      "loss": 0.603,
+      "step": 14
+    },
+    {
+      "epoch": 0.007333682556312206,
+      "loss_partial": 0.275390625,
+      "step": 14
+    },
+    {
+      "epoch": 0.007333682556312206,
+      "loss_partial": 0.267578125,
+      "step": 14
+    },
+    {
+      "epoch": 0.00785751702462022,
+      "grad_norm": 261.68060302734375,
+      "learning_rate": 3.6649214659685864e-07,
+      "loss": 0.605,
+      "step": 15
+    },
+    {
+      "epoch": 0.00785751702462022,
+      "loss_partial": 0.318359375,
+      "step": 15
+    },
+    {
+      "epoch": 0.00785751702462022,
+      "loss_partial": 0.294921875,
+      "step": 15
+    },
+    {
+      "epoch": 0.008381351492928235,
+      "grad_norm": 267.8633117675781,
+      "learning_rate": 3.926701570680629e-07,
+      "loss": 0.6191,
+      "step": 16
+    },
+    {
+      "epoch": 0.008381351492928235,
+      "loss_partial": 0.298828125,
+      "step": 16
+    },
+    {
+      "epoch": 0.008381351492928235,
+      "loss_partial": 0.25,
+      "step": 16
+    },
+    {
+      "epoch": 0.00890518596123625,
+      "grad_norm": 264.9583435058594,
+      "learning_rate": 4.1884816753926706e-07,
+      "loss": 0.583,
+      "step": 17
+    },
+    {
+      "epoch": 0.00890518596123625,
+      "loss_partial": 0.255859375,
+      "step": 17
+    },
+    {
+      "epoch": 0.00890518596123625,
+      "loss_partial": 0.30859375,
+      "step": 17
+    },
+    {
+      "epoch": 0.009429020429544265,
+      "grad_norm": 259.4761657714844,
+      "learning_rate": 4.4502617801047125e-07,
+      "loss": 0.575,
+      "step": 18
+    },
+    {
+      "epoch": 0.009429020429544265,
+      "loss_partial": 0.126953125,
+      "step": 18
+    },
+    {
+      "epoch": 0.009429020429544265,
+      "loss_partial": 0.1279296875,
+      "step": 18
+    },
+    {
+      "epoch": 0.00995285489785228,
+      "grad_norm": 87.19998168945312,
+      "learning_rate": 4.7120418848167543e-07,
+      "loss": 0.2389,
+      "step": 19
+    },
+    {
+      "epoch": 0.00995285489785228,
+      "loss_partial": 0.12255859375,
+      "step": 19
+    },
+    {
+      "epoch": 0.00995285489785228,
+      "loss_partial": 0.11474609375,
+      "step": 19
+    },
+    {
+      "epoch": 0.010476689366160294,
+      "grad_norm": 83.0079116821289,
+      "learning_rate": 4.973821989528796e-07,
+      "loss": 0.239,
+      "step": 20
+    },
+    {
+      "epoch": 0.010476689366160294,
+      "loss_partial": 0.109375,
+      "step": 20
+    },
+    {
+      "epoch": 0.010476689366160294,
+      "loss_partial": 0.119140625,
+      "step": 20
+    },
+    {
+      "epoch": 0.011000523834468309,
+      "grad_norm": 71.53241729736328,
+      "learning_rate": 5.235602094240838e-07,
+      "loss": 0.2306,
+      "step": 21
+    },
+    {
+      "epoch": 0.011000523834468309,
+      "loss_partial": 0.1162109375,
+      "step": 21
+    },
+    {
+      "epoch": 0.011000523834468309,
+      "loss_partial": 0.1025390625,
+      "step": 21
+    },
+    {
+      "epoch": 0.011524358302776323,
+      "grad_norm": 63.62755584716797,
+      "learning_rate": 5.49738219895288e-07,
+      "loss": 0.2201,
+      "step": 22
+    },
+    {
+      "epoch": 0.011524358302776323,
+      "loss_partial": 0.0888671875,
+      "step": 22
+    },
+    {
+      "epoch": 0.011524358302776323,
+      "loss_partial": 0.08740234375,
+      "step": 22
+    },
+    {
+      "epoch": 0.012048192771084338,
+      "grad_norm": 38.03909683227539,
+      "learning_rate": 5.759162303664922e-07,
+      "loss": 0.2015,
+      "step": 23
+    },
+    {
+      "epoch": 0.012048192771084338,
+      "loss_partial": 0.10009765625,
+      "step": 23
+    },
+    {
+      "epoch": 0.012048192771084338,
+      "loss_partial": 0.09619140625,
+      "step": 23
+    },
+    {
+      "epoch": 0.012572027239392353,
+      "grad_norm": 41.71686553955078,
+      "learning_rate": 6.020942408376964e-07,
+      "loss": 0.2048,
+      "step": 24
+    },
+    {
+      "epoch": 0.012572027239392353,
+      "loss_partial": 0.095703125,
+      "step": 24
+    },
+    {
+      "epoch": 0.012572027239392353,
+      "loss_partial": 0.0859375,
+      "step": 24
+    },
+    {
+      "epoch": 0.013095861707700367,
+      "grad_norm": 28.76995086669922,
+      "learning_rate": 6.282722513089005e-07,
+      "loss": 0.189,
+      "step": 25
+    },
+    {
+      "epoch": 0.013095861707700367,
+      "loss_partial": 0.09619140625,
+      "step": 25
+    },
+    {
+      "epoch": 0.013095861707700367,
+      "loss_partial": 0.099609375,
+      "step": 25
+    },
+    {
+      "epoch": 0.013619696176008382,
+      "grad_norm": 12.129317283630371,
+      "learning_rate": 6.544502617801048e-07,
+      "loss": 0.1727,
+      "step": 26
+    },
+    {
+      "epoch": 0.013619696176008382,
+      "loss_partial": 0.0810546875,
+      "step": 26
+    },
+    {
+      "epoch": 0.013619696176008382,
+      "loss_partial": 0.091796875,
+      "step": 26
+    },
+    {
+      "epoch": 0.014143530644316397,
+      "grad_norm": 2.552935838699341,
+      "learning_rate": 6.80628272251309e-07,
+      "loss": 0.1575,
+      "step": 27
+    },
+    {
+      "epoch": 0.014143530644316397,
+      "loss_partial": 0.0712890625,
+      "step": 27
+    },
+    {
+      "epoch": 0.014143530644316397,
+      "loss_partial": 0.095703125,
+      "step": 27
+    },
+    {
+      "epoch": 0.014667365112624411,
+      "grad_norm": 2.301330804824829,
+      "learning_rate": 7.068062827225131e-07,
+      "loss": 0.1557,
+      "step": 28
+    },
+    {
+      "epoch": 0.014667365112624411,
+      "loss_partial": 0.08447265625,
+      "step": 28
+    },
+    {
+      "epoch": 0.014667365112624411,
+      "loss_partial": 0.07373046875,
+      "step": 28
+    },
+    {
+      "epoch": 0.015191199580932426,
+      "grad_norm": 2.3691599369049072,
+      "learning_rate": 7.329842931937173e-07,
+      "loss": 0.1574,
+      "step": 29
+    },
+    {
+      "epoch": 0.015191199580932426,
+      "loss_partial": 0.08251953125,
+      "step": 29
+    },
+    {
+      "epoch": 0.015191199580932426,
+      "loss_partial": 0.06494140625,
+      "step": 29
+    },
+    {
+      "epoch": 0.01571503404924044,
+      "grad_norm": 2.4412319660186768,
+      "learning_rate": 7.591623036649215e-07,
+      "loss": 0.1546,
+      "step": 30
+    },
+    {
+      "epoch": 0.01571503404924044,
+      "loss_partial": 0.072265625,
+      "step": 30
+    },
+    {
+      "epoch": 0.01571503404924044,
+      "loss_partial": 0.0771484375,
+      "step": 30
+    },
+    {
+      "epoch": 0.016238868517548456,
+      "grad_norm": 2.4062652587890625,
+      "learning_rate": 7.853403141361258e-07,
+      "loss": 0.1505,
+      "step": 31
+    },
+    {
+      "epoch": 0.016238868517548456,
+      "loss_partial": 0.06591796875,
+      "step": 31
+    },
+    {
+      "epoch": 0.016238868517548456,
+      "loss_partial": 0.072265625,
+      "step": 31
+    },
+    {
+      "epoch": 0.01676270298585647,
+      "grad_norm": 2.221773147583008,
+      "learning_rate": 8.115183246073299e-07,
+      "loss": 0.1556,
+      "step": 32
+    },
+    {
+      "epoch": 0.01676270298585647,
+      "loss_partial": 0.064453125,
+      "step": 32
+    },
+    {
+      "epoch": 0.01676270298585647,
+      "loss_partial": 0.0751953125,
+      "step": 32
+    },
+    {
+      "epoch": 0.017286537454164485,
+      "grad_norm": 2.212191343307495,
+      "learning_rate": 8.376963350785341e-07,
+      "loss": 0.1527,
+      "step": 33
+    },
+    {
+      "epoch": 0.017286537454164485,
+      "loss_partial": 0.07373046875,
+      "step": 33
+    },
+    {
+      "epoch": 0.017286537454164485,
+      "loss_partial": 0.0947265625,
+      "step": 33
+    },
+    {
+      "epoch": 0.0178103719224725,
+      "grad_norm": 2.1011016368865967,
+      "learning_rate": 8.638743455497383e-07,
+      "loss": 0.1533,
+      "step": 34
+    },
+    {
+      "epoch": 0.0178103719224725,
+      "loss_partial": 0.06640625,
+      "step": 34
+    },
+    {
+      "epoch": 0.0178103719224725,
+      "loss_partial": 0.0849609375,
+      "step": 34
+    },
+    {
+      "epoch": 0.018334206390780514,
+      "grad_norm": 2.0751523971557617,
+      "learning_rate": 8.900523560209425e-07,
+      "loss": 0.152,
+      "step": 35
+    },
+    {
+      "epoch": 0.018334206390780514,
+      "loss_partial": 0.07861328125,
+      "step": 35
+    },
+    {
+      "epoch": 0.018334206390780514,
+      "loss_partial": 0.078125,
+      "step": 35
+    },
+    {
+      "epoch": 0.01885804085908853,
+      "grad_norm": 1.994758129119873,
+      "learning_rate": 9.162303664921466e-07,
+      "loss": 0.151,
+      "step": 36
+    },
+    {
+      "epoch": 0.01885804085908853,
+      "loss_partial": 0.076171875,
+      "step": 36
+    },
+    {
+      "epoch": 0.01885804085908853,
+      "loss_partial": 0.08056640625,
+      "step": 36
+    },
+    {
+      "epoch": 0.019381875327396544,
+      "grad_norm": 1.8580564260482788,
+      "learning_rate": 9.424083769633509e-07,
+      "loss": 0.1516,
+      "step": 37
+    },
+    {
+      "epoch": 0.019381875327396544,
+      "loss_partial": 0.07470703125,
+      "step": 37
+    },
+    {
+      "epoch": 0.019381875327396544,
+      "loss_partial": 0.087890625,
+      "step": 37
+    },
+    {
+      "epoch": 0.01990570979570456,
+      "grad_norm": 1.7582027912139893,
+      "learning_rate": 9.685863874345552e-07,
+      "loss": 0.1545,
+      "step": 38
+    },
+    {
+      "epoch": 0.01990570979570456,
+      "loss_partial": 0.06005859375,
+      "step": 38
+    },
+    {
+      "epoch": 0.01990570979570456,
+      "loss_partial": 0.07958984375,
+      "step": 38
+    },
+    {
+      "epoch": 0.020429544264012573,
+      "grad_norm": 1.580087423324585,
+      "learning_rate": 9.947643979057591e-07,
+      "loss": 0.1526,
+      "step": 39
+    },
+    {
+      "epoch": 0.020429544264012573,
+      "loss_partial": 0.07275390625,
+      "step": 39
+    },
+    {
+      "epoch": 0.020429544264012573,
+      "loss_partial": 0.07666015625,
+      "step": 39
+    },
+    {
+      "epoch": 0.020953378732320588,
+      "grad_norm": 1.3927592039108276,
+      "learning_rate": 1.0209424083769635e-06,
+      "loss": 0.1528,
+      "step": 40
+    },
+    {
+      "epoch": 0.020953378732320588,
+      "loss_partial": 0.06982421875,
+      "step": 40
+    },
+    {
+      "epoch": 0.020953378732320588,
+      "loss_partial": 0.0859375,
+      "step": 40
+    },
+    {
+      "epoch": 0.021477213200628602,
+      "grad_norm": 1.2891699075698853,
+      "learning_rate": 1.0471204188481676e-06,
+      "loss": 0.1473,
+      "step": 41
+    },
+    {
+      "epoch": 0.021477213200628602,
+      "loss_partial": 0.07373046875,
+      "step": 41
+    },
+    {
+      "epoch": 0.021477213200628602,
+      "loss_partial": 0.07080078125,
+      "step": 41
+    },
+    {
+      "epoch": 0.022001047668936617,
+      "grad_norm": 1.2740473747253418,
+      "learning_rate": 1.0732984293193717e-06,
+      "loss": 0.1515,
+      "step": 42
+    },
+    {
+      "epoch": 0.022001047668936617,
+      "loss_partial": 0.08837890625,
+      "step": 42
+    },
+    {
+      "epoch": 0.022001047668936617,
+      "loss_partial": 0.072265625,
+      "step": 42
+    },
+    {
+      "epoch": 0.022524882137244632,
+      "grad_norm": 1.1968275308609009,
+      "learning_rate": 1.099476439790576e-06,
+      "loss": 0.1496,
+      "step": 43
+    },
+    {
+      "epoch": 0.022524882137244632,
+      "loss_partial": 0.0703125,
+      "step": 43
+    },
+    {
+      "epoch": 0.022524882137244632,
+      "loss_partial": 0.076171875,
+      "step": 43
+    },
+    {
+      "epoch": 0.023048716605552647,
+      "grad_norm": 1.140224575996399,
+      "learning_rate": 1.1256544502617802e-06,
+      "loss": 0.1498,
+      "step": 44
+    },
+    {
+      "epoch": 0.023048716605552647,
+      "loss_partial": 0.0703125,
+      "step": 44
+    },
+    {
+      "epoch": 0.023048716605552647,
+      "loss_partial": 0.08251953125,
+      "step": 44
+    },
+    {
+      "epoch": 0.02357255107386066,
+      "grad_norm": 1.0370817184448242,
+      "learning_rate": 1.1518324607329843e-06,
+      "loss": 0.1478,
+      "step": 45
+    },
+    {
+      "epoch": 0.02357255107386066,
+      "loss_partial": 0.0625,
+      "step": 45
+    },
+    {
+      "epoch": 0.02357255107386066,
+      "loss_partial": 0.07080078125,
+      "step": 45
+    },
+    {
+      "epoch": 0.024096385542168676,
+      "grad_norm": 1.0084102153778076,
+      "learning_rate": 1.1780104712041885e-06,
+      "loss": 0.1436,
+      "step": 46
+    },
+    {
+      "epoch": 0.024096385542168676,
+      "loss_partial": 0.06884765625,
+      "step": 46
+    },
+    {
+      "epoch": 0.024096385542168676,
+      "loss_partial": 0.0673828125,
+      "step": 46
+    },
+    {
+      "epoch": 0.02462022001047669,
+      "grad_norm": 0.9838353395462036,
+      "learning_rate": 1.2041884816753928e-06,
+      "loss": 0.1453,
+      "step": 47
+    },
+    {
+      "epoch": 0.02462022001047669,
+      "loss_partial": 0.0771484375,
+      "step": 47
+    },
+    {
+      "epoch": 0.02462022001047669,
+      "loss_partial": 0.078125,
+      "step": 47
+    },
+    {
+      "epoch": 0.025144054478784705,
+      "grad_norm": 0.9764026999473572,
+      "learning_rate": 1.230366492146597e-06,
+      "loss": 0.1489,
+      "step": 48
+    },
+    {
+      "epoch": 0.025144054478784705,
+      "loss_partial": 0.07470703125,
+      "step": 48
+    },
+    {
+      "epoch": 0.025144054478784705,
+      "loss_partial": 0.0654296875,
+      "step": 48
+    },
+    {
+      "epoch": 0.02566788894709272,
+      "grad_norm": 0.996488630771637,
+      "learning_rate": 1.256544502617801e-06,
+      "loss": 0.1511,
+      "step": 49
+    },
+    {
+      "epoch": 0.02566788894709272,
+      "loss_partial": 0.08056640625,
+      "step": 49
+    },
+    {
+      "epoch": 0.02566788894709272,
+      "loss_partial": 0.072265625,
+      "step": 49
+    },
+    {
+      "epoch": 0.026191723415400735,
+      "grad_norm": 0.9501944184303284,
+      "learning_rate": 1.2827225130890052e-06,
+      "loss": 0.1481,
+      "step": 50
+    },
+    {
+      "epoch": 0.026191723415400735,
+      "loss_partial": 0.0732421875,
+      "step": 50
+    },
+    {
+      "epoch": 0.026191723415400735,
+      "loss_partial": 0.0791015625,
+      "step": 50
+    },
+    {
+      "epoch": 0.02671555788370875,
+      "grad_norm": 0.9404588341712952,
+      "learning_rate": 1.3089005235602096e-06,
+      "loss": 0.1491,
+      "step": 51
+    },
+    {
+      "epoch": 0.02671555788370875,
+      "loss_partial": 0.064453125,
+      "step": 51
+    },
+    {
+      "epoch": 0.02671555788370875,
+      "loss_partial": 0.08203125,
+      "step": 51
+    },
+    {
+      "epoch": 0.027239392352016764,
+      "grad_norm": 0.8905503153800964,
+      "learning_rate": 1.3350785340314137e-06,
+      "loss": 0.1472,
+      "step": 52
+    },
+    {
+      "epoch": 0.027239392352016764,
+      "loss_partial": 0.09619140625,
+      "step": 52
+    },
+    {
+      "epoch": 0.027239392352016764,
+      "loss_partial": 0.0693359375,
+      "step": 52
+    },
+    {
+      "epoch": 0.02776322682032478,
+      "grad_norm": 0.9073057174682617,
+      "learning_rate": 1.361256544502618e-06,
+      "loss": 0.1523,
+      "step": 53
+    },
+    {
+      "epoch": 0.02776322682032478,
+      "loss_partial": 0.0791015625,
+      "step": 53
+    },
+    {
+      "epoch": 0.02776322682032478,
+      "loss_partial": 0.06884765625,
+      "step": 53
+    },
+    {
+      "epoch": 0.028287061288632794,
+      "grad_norm": 0.8603799343109131,
+      "learning_rate": 1.3874345549738222e-06,
+      "loss": 0.1471,
+      "step": 54
+    },
+    {
+      "epoch": 0.028287061288632794,
+      "loss_partial": 0.0830078125,
+      "step": 54
+    },
+    {
+      "epoch": 0.028287061288632794,
+      "loss_partial": 0.064453125,
+      "step": 54
+    },
+    {
+      "epoch": 0.028810895756940808,
+      "grad_norm": 0.8530461192131042,
+      "learning_rate": 1.4136125654450263e-06,
+      "loss": 0.1446,
+      "step": 55
+    },
+    {
+      "epoch": 0.028810895756940808,
+      "loss_partial": 0.080078125,
+      "step": 55
+    },
+    {
+      "epoch": 0.028810895756940808,
+      "loss_partial": 0.07080078125,
+      "step": 55
+    },
+    {
+      "epoch": 0.029334730225248823,
+      "grad_norm": 0.8335225582122803,
+      "learning_rate": 1.4397905759162306e-06,
+      "loss": 0.1477,
+      "step": 56
+    },
+    {
+      "epoch": 0.029334730225248823,
+      "loss_partial": 0.078125,
+      "step": 56
+    },
+    {
+      "epoch": 0.029334730225248823,
+      "loss_partial": 0.06640625,
+      "step": 56
+    },
+    {
+      "epoch": 0.029858564693556838,
+      "grad_norm": 0.8138421177864075,
+      "learning_rate": 1.4659685863874346e-06,
+      "loss": 0.1507,
+      "step": 57
+    },
+    {
+      "epoch": 0.029858564693556838,
+      "loss_partial": 0.0849609375,
+      "step": 57
+    },
+    {
+      "epoch": 0.029858564693556838,
+      "loss_partial": 0.07421875,
+      "step": 57
+    },
+    {
+      "epoch": 0.030382399161864852,
+      "grad_norm": 0.8134170174598694,
+      "learning_rate": 1.4921465968586387e-06,
+      "loss": 0.1489,
+      "step": 58
+    },
+    {
+      "epoch": 0.030382399161864852,
+      "loss_partial": 0.0751953125,
+      "step": 58
+    },
+    {
+      "epoch": 0.030382399161864852,
+      "loss_partial": 0.06884765625,
+      "step": 58
+    },
+    {
+      "epoch": 0.030906233630172867,
+      "grad_norm": 0.8158202767372131,
+      "learning_rate": 1.518324607329843e-06,
+      "loss": 0.1505,
+      "step": 59
+    },
+    {
+      "epoch": 0.030906233630172867,
+      "loss_partial": 0.076171875,
+      "step": 59
+    },
+    {
+      "epoch": 0.030906233630172867,
+      "loss_partial": 0.07421875,
+      "step": 59
+    },
+    {
+      "epoch": 0.03143006809848088,
+      "grad_norm": 0.767124354839325,
+      "learning_rate": 1.5445026178010472e-06,
+      "loss": 0.1481,
+      "step": 60
+    },
+    {
+      "epoch": 0.03143006809848088,
+      "loss_partial": 0.0693359375,
+      "step": 60
+    },
+    {
+      "epoch": 0.03143006809848088,
+      "loss_partial": 0.064453125,
+      "step": 60
+    },
+    {
+      "epoch": 0.03195390256678889,
+      "grad_norm": 0.7568233609199524,
+      "learning_rate": 1.5706806282722515e-06,
+      "loss": 0.1446,
+      "step": 61
+    },
+    {
+      "epoch": 0.03195390256678889,
+      "loss_partial": 0.06640625,
+      "step": 61
+    },
+    {
+      "epoch": 0.03195390256678889,
+      "loss_partial": 0.06689453125,
+      "step": 61
+    },
+    {
+      "epoch": 0.03247773703509691,
+      "grad_norm": 0.7397628426551819,
+      "learning_rate": 1.5968586387434556e-06,
+      "loss": 0.1448,
+      "step": 62
+    },
+    {
+      "epoch": 0.03247773703509691,
+      "loss_partial": 0.0751953125,
+      "step": 62
+    },
+    {
+      "epoch": 0.03247773703509691,
+      "loss_partial": 0.0771484375,
+      "step": 62
+    },
+    {
+      "epoch": 0.03300157150340492,
+      "grad_norm": 0.7285221815109253,
+      "learning_rate": 1.6230366492146598e-06,
+      "loss": 0.1493,
+      "step": 63
+    },
+    {
+      "epoch": 0.03300157150340492,
+      "loss_partial": 0.0771484375,
+      "step": 63
+    },
+    {
+      "epoch": 0.03300157150340492,
+      "loss_partial": 0.07373046875,
+      "step": 63
+    },
+    {
+      "epoch": 0.03352540597171294,
+      "grad_norm": 0.679966151714325,
+      "learning_rate": 1.6492146596858641e-06,
+      "loss": 0.1451,
+      "step": 64
+    },
+    {
+      "epoch": 0.03352540597171294,
+      "loss_partial": 0.0791015625,
+      "step": 64
+    },
+    {
+      "epoch": 0.03352540597171294,
+      "loss_partial": 0.07861328125,
+      "step": 64
+    },
+    {
+      "epoch": 0.03404924044002095,
+      "grad_norm": 0.6869559288024902,
+      "learning_rate": 1.6753926701570683e-06,
+      "loss": 0.1442,
+      "step": 65
+    },
+    {
+      "epoch": 0.03404924044002095,
+      "loss_partial": 0.06884765625,
+      "step": 65
+    },
+    {
+      "epoch": 0.03404924044002095,
+      "loss_partial": 0.0830078125,
+      "step": 65
+    },
+    {
+      "epoch": 0.03457307490832897,
+      "grad_norm": 0.67109614610672,
+      "learning_rate": 1.7015706806282726e-06,
+      "loss": 0.1459,
+      "step": 66
+    },
+    {
+      "epoch": 0.03457307490832897,
+      "loss_partial": 0.064453125,
+      "step": 66
+    },
+    {
+      "epoch": 0.03457307490832897,
+      "loss_partial": 0.0634765625,
+      "step": 66
+    },
+    {
+      "epoch": 0.03509690937663698,
+      "grad_norm": 0.6107531189918518,
+      "learning_rate": 1.7277486910994765e-06,
+      "loss": 0.1434,
+      "step": 67
+    },
+    {
+      "epoch": 0.03509690937663698,
+      "loss_partial": 0.0732421875,
+      "step": 67
+    },
+    {
+      "epoch": 0.03509690937663698,
+      "loss_partial": 0.0654296875,
+      "step": 67
+    },
+    {
+      "epoch": 0.035620743844945,
+      "grad_norm": 0.6340813040733337,
+      "learning_rate": 1.7539267015706806e-06,
+      "loss": 0.1469,
+      "step": 68
+    },
+    {
+      "epoch": 0.035620743844945,
+      "loss_partial": 0.072265625,
+      "step": 68
+    },
+    {
+      "epoch": 0.035620743844945,
+      "loss_partial": 0.083984375,
+      "step": 68
+    },
+    {
+      "epoch": 0.03614457831325301,
+      "grad_norm": 0.612872302532196,
+      "learning_rate": 1.780104712041885e-06,
+      "loss": 0.1526,
+      "step": 69
+    },
+    {
+      "epoch": 0.03614457831325301,
+      "loss_partial": 0.06396484375,
+      "step": 69
+    },
+    {
+      "epoch": 0.03614457831325301,
+      "loss_partial": 0.060546875,
+      "step": 69
+    },
+    {
+      "epoch": 0.03666841278156103,
+      "grad_norm": 0.6016877889633179,
+      "learning_rate": 1.8062827225130891e-06,
+      "loss": 0.1411,
+      "step": 70
+    },
+    {
+      "epoch": 0.03666841278156103,
+      "loss_partial": 0.06591796875,
+      "step": 70
+    },
+    {
+      "epoch": 0.03666841278156103,
+      "loss_partial": 0.0673828125,
+      "step": 70
+    },
+    {
+      "epoch": 0.03719224724986904,
+      "grad_norm": 0.5817959308624268,
+      "learning_rate": 1.8324607329842933e-06,
+      "loss": 0.1436,
+      "step": 71
+    },
+    {
+      "epoch": 0.03719224724986904,
+      "loss_partial": 0.0810546875,
+      "step": 71
+    },
+    {
+      "epoch": 0.03719224724986904,
+      "loss_partial": 0.0732421875,
+      "step": 71
+    },
+    {
+      "epoch": 0.03771608171817706,
+      "grad_norm": 0.5475994944572449,
+      "learning_rate": 1.8586387434554976e-06,
+      "loss": 0.1447,
+      "step": 72
+    },
+    {
+      "epoch": 0.03771608171817706,
+      "loss_partial": 0.0693359375,
+      "step": 72
+    },
+    {
+      "epoch": 0.03771608171817706,
+      "loss_partial": 0.06884765625,
+      "step": 72
+    },
+    {
+      "epoch": 0.03823991618648507,
+      "grad_norm": 0.5594205260276794,
+      "learning_rate": 1.8848167539267017e-06,
+      "loss": 0.1454,
+      "step": 73
+    },
+    {
+      "epoch": 0.03823991618648507,
+      "loss_partial": 0.07861328125,
+      "step": 73
+    },
+    {
+      "epoch": 0.03823991618648507,
+      "loss_partial": 0.07470703125,
+      "step": 73
+    },
+    {
+      "epoch": 0.03876375065479309,
+      "grad_norm": 0.5500811338424683,
+      "learning_rate": 1.910994764397906e-06,
+      "loss": 0.1449,
+      "step": 74
+    },
+    {
+      "epoch": 0.03876375065479309,
+      "loss_partial": 0.0703125,
+      "step": 74
+    },
+    {
+      "epoch": 0.03876375065479309,
+      "loss_partial": 0.08349609375,
+      "step": 74
+    },
+    {
+      "epoch": 0.0392875851231011,
+      "grad_norm": 0.5343813300132751,
+      "learning_rate": 1.9371727748691104e-06,
+      "loss": 0.1442,
+      "step": 75
+    },
+    {
+      "epoch": 0.0392875851231011,
+      "loss_partial": 0.0791015625,
+      "step": 75
+    },
+    {
+      "epoch": 0.0392875851231011,
+      "loss_partial": 0.0654296875,
+      "step": 75
+    },
+    {
+      "epoch": 0.03981141959140912,
+      "grad_norm": 0.5083532929420471,
+      "learning_rate": 1.9633507853403143e-06,
+      "loss": 0.145,
+      "step": 76
+    },
+    {
+      "epoch": 0.03981141959140912,
+      "loss_partial": 0.06396484375,
+      "step": 76
+    },
+    {
+      "epoch": 0.03981141959140912,
+      "loss_partial": 0.0732421875,
+      "step": 76
+    },
+    {
+      "epoch": 0.04033525405971713,
+      "grad_norm": 0.5143418908119202,
+      "learning_rate": 1.9895287958115183e-06,
+      "loss": 0.1417,
+      "step": 77
+    },
+    {
+      "epoch": 0.04033525405971713,
+      "loss_partial": 0.0625,
+      "step": 77
+    },
+    {
+      "epoch": 0.04033525405971713,
+      "loss_partial": 0.07763671875,
+      "step": 77
+    },
+    {
+      "epoch": 0.040859088528025146,
+      "grad_norm": 0.4876868724822998,
+      "learning_rate": 2.0157068062827226e-06,
+      "loss": 0.143,
+      "step": 78
+    },
+    {
+      "epoch": 0.040859088528025146,
+      "loss_partial": 0.06640625,
+      "step": 78
+    },
+    {
+      "epoch": 0.040859088528025146,
+      "loss_partial": 0.08154296875,
+      "step": 78
+    },
+    {
+      "epoch": 0.04138292299633316,
+      "grad_norm": 0.5035227537155151,
+      "learning_rate": 2.041884816753927e-06,
+      "loss": 0.1443,
+      "step": 79
+    },
+    {
+      "epoch": 0.04138292299633316,
+      "loss_partial": 0.0751953125,
+      "step": 79
+    },
+    {
+      "epoch": 0.04138292299633316,
+      "loss_partial": 0.07373046875,
+      "step": 79
+    },
+    {
+      "epoch": 0.041906757464641176,
+      "grad_norm": 0.48566165566444397,
+      "learning_rate": 2.068062827225131e-06,
+      "loss": 0.1458,
+      "step": 80
+    },
+    {
+      "epoch": 0.041906757464641176,
+      "loss_partial": 0.07275390625,
+      "step": 80
+    },
+    {
+      "epoch": 0.041906757464641176,
+      "loss_partial": 0.07373046875,
+      "step": 80
+    },
+    {
+      "epoch": 0.04243059193294919,
+      "grad_norm": 0.4721881151199341,
+      "learning_rate": 2.094240837696335e-06,
+      "loss": 0.1447,
+      "step": 81
+    },
+    {
+      "epoch": 0.04243059193294919,
+      "loss_partial": 0.09326171875,
+      "step": 81
+    },
+    {
+      "epoch": 0.04243059193294919,
+      "loss_partial": 0.078125,
+      "step": 81
+    },
+    {
+      "epoch": 0.042954426401257205,
+      "grad_norm": 0.46265748143196106,
+      "learning_rate": 2.1204188481675396e-06,
+      "loss": 0.1475,
+      "step": 82
+    },
+    {
+      "epoch": 0.042954426401257205,
+      "loss_partial": 0.06884765625,
+      "step": 82
+    },
+    {
+      "epoch": 0.042954426401257205,
+      "loss_partial": 0.07275390625,
+      "step": 82
+    },
+    {
+      "epoch": 0.043478260869565216,
+      "grad_norm": 0.4572789967060089,
+      "learning_rate": 2.1465968586387435e-06,
+      "loss": 0.1471,
+      "step": 83
+    },
+    {
+      "epoch": 0.043478260869565216,
+      "loss_partial": 0.07421875,
+      "step": 83
+    },
+    {
+      "epoch": 0.043478260869565216,
+      "loss_partial": 0.064453125,
+      "step": 83
+    },
+    {
+      "epoch": 0.044002095337873234,
+      "grad_norm": 0.4532865583896637,
+      "learning_rate": 2.172774869109948e-06,
+      "loss": 0.1435,
+      "step": 84
+    },
+    {
+      "epoch": 0.044002095337873234,
+      "loss_partial": 0.078125,
+      "step": 84
+    },
+    {
+      "epoch": 0.044002095337873234,
+      "loss_partial": 0.07470703125,
+      "step": 84
+    },
+    {
+      "epoch": 0.044525929806181246,
+      "grad_norm": 0.4346810579299927,
+      "learning_rate": 2.198952879581152e-06,
+      "loss": 0.1443,
+      "step": 85
+    },
+    {
+      "epoch": 0.044525929806181246,
+      "loss_partial": 0.07177734375,
+      "step": 85
+    },
+    {
+      "epoch": 0.044525929806181246,
+      "loss_partial": 0.0634765625,
+      "step": 85
+    },
+    {
+      "epoch": 0.045049764274489264,
+      "grad_norm": 0.43043193221092224,
+      "learning_rate": 2.2251308900523565e-06,
+      "loss": 0.1445,
+      "step": 86
+    },
+    {
+      "epoch": 0.045049764274489264,
+      "loss_partial": 0.0712890625,
+      "step": 86
+    },
+    {
+      "epoch": 0.045049764274489264,
+      "loss_partial": 0.0712890625,
+      "step": 86
+    },
+    {
+      "epoch": 0.045573598742797275,
+      "grad_norm": 0.42547914385795593,
+      "learning_rate": 2.2513089005235604e-06,
+      "loss": 0.144,
+      "step": 87
+    },
+    {
+      "epoch": 0.045573598742797275,
+      "loss_partial": 0.08544921875,
+      "step": 87
+    },
+    {
+      "epoch": 0.045573598742797275,
+      "loss_partial": 0.076171875,
+      "step": 87
+    },
+    {
+      "epoch": 0.04609743321110529,
+      "grad_norm": 0.414821982383728,
+      "learning_rate": 2.2774869109947643e-06,
+      "loss": 0.1445,
+      "step": 88
+    },
+    {
+      "epoch": 0.04609743321110529,
+      "loss_partial": 0.06787109375,
+      "step": 88
+    },
+    {
+      "epoch": 0.04609743321110529,
+      "loss_partial": 0.07470703125,
+      "step": 88
+    },
+    {
+      "epoch": 0.046621267679413304,
+      "grad_norm": 0.4278256297111511,
+      "learning_rate": 2.3036649214659687e-06,
+      "loss": 0.1424,
+      "step": 89
+    },
+    {
+      "epoch": 0.046621267679413304,
+      "loss_partial": 0.0703125,
+      "step": 89
+    },
+    {
+      "epoch": 0.046621267679413304,
+      "loss_partial": 0.06396484375,
+      "step": 89
+    },
+    {
+      "epoch": 0.04714510214772132,
+      "grad_norm": 0.412080854177475,
+      "learning_rate": 2.329842931937173e-06,
+      "loss": 0.1418,
+      "step": 90
+    },
+    {
+      "epoch": 0.04714510214772132,
+      "loss_partial": 0.07470703125,
+      "step": 90
+    },
+    {
+      "epoch": 0.04714510214772132,
+      "loss_partial": 0.072265625,
+      "step": 90
+    },
+    {
+      "epoch": 0.047668936616029334,
+      "grad_norm": 0.4302641451358795,
+      "learning_rate": 2.356020942408377e-06,
+      "loss": 0.144,
+      "step": 91
+    },
+    {
+      "epoch": 0.047668936616029334,
+      "loss_partial": 0.06982421875,
+      "step": 91
+    },
+    {
+      "epoch": 0.047668936616029334,
+      "loss_partial": 0.07421875,
+      "step": 91
+    },
+    {
+      "epoch": 0.04819277108433735,
+      "grad_norm": 0.4115174114704132,
+      "learning_rate": 2.3821989528795813e-06,
+      "loss": 0.1383,
+      "step": 92
+    },
+    {
+      "epoch": 0.04819277108433735,
+      "loss_partial": 0.07421875,
+      "step": 92
+    },
+    {
+      "epoch": 0.04819277108433735,
+      "loss_partial": 0.06396484375,
+      "step": 92
+    },
+    {
+      "epoch": 0.04871660555264536,
+      "grad_norm": 0.41746917366981506,
+      "learning_rate": 2.4083769633507856e-06,
+      "loss": 0.1431,
+      "step": 93
+    },
+    {
+      "epoch": 0.04871660555264536,
+      "loss_partial": 0.0712890625,
+      "step": 93
+    },
+    {
+      "epoch": 0.04871660555264536,
+      "loss_partial": 0.0625,
+      "step": 93
+    },
+    {
+      "epoch": 0.04924044002095338,
+      "grad_norm": 0.39474278688430786,
+      "learning_rate": 2.43455497382199e-06,
+      "loss": 0.1427,
+      "step": 94
+    },
+    {
+      "epoch": 0.04924044002095338,
+      "loss_partial": 0.06494140625,
+      "step": 94
+    },
+    {
+      "epoch": 0.04924044002095338,
+      "loss_partial": 0.0703125,
+      "step": 94
+    },
+    {
+      "epoch": 0.04976427448926139,
+      "grad_norm": 0.3907967805862427,
+      "learning_rate": 2.460732984293194e-06,
+      "loss": 0.1432,
+      "step": 95
+    },
+    {
+      "epoch": 0.04976427448926139,
+      "loss_partial": 0.07080078125,
+      "step": 95
+    },
+    {
+      "epoch": 0.04976427448926139,
+      "loss_partial": 0.06396484375,
+      "step": 95
+    },
+    {
+      "epoch": 0.05028810895756941,
+      "grad_norm": 0.3945806324481964,
+      "learning_rate": 2.4869109947643982e-06,
+      "loss": 0.1448,
+      "step": 96
+    },
+    {
+      "epoch": 0.05028810895756941,
+      "loss_partial": 0.06640625,
+      "step": 96
+    },
+    {
+      "epoch": 0.05028810895756941,
+      "loss_partial": 0.07861328125,
+      "step": 96
+    },
+    {
+      "epoch": 0.05081194342587742,
+      "grad_norm": 0.38689950108528137,
+      "learning_rate": 2.513089005235602e-06,
+      "loss": 0.1404,
+      "step": 97
+    },
+    {
+      "epoch": 0.05081194342587742,
+      "loss_partial": 0.0673828125,
+      "step": 97
+    },
+    {
+      "epoch": 0.05081194342587742,
+      "loss_partial": 0.06396484375,
+      "step": 97
+    },
+    {
+      "epoch": 0.05133577789418544,
+      "grad_norm": 0.3990933299064636,
+      "learning_rate": 2.5392670157068065e-06,
+      "loss": 0.1435,
+      "step": 98
+    },
+    {
+      "epoch": 0.05133577789418544,
+      "loss_partial": 0.061279296875,
+      "step": 98
+    },
+    {
+      "epoch": 0.05133577789418544,
+      "loss_partial": 0.06640625,
+      "step": 98
+    },
+    {
+      "epoch": 0.05185961236249345,
+      "grad_norm": 0.40120232105255127,
+      "learning_rate": 2.5654450261780104e-06,
+      "loss": 0.1419,
+      "step": 99
+    },
+    {
+      "epoch": 0.05185961236249345,
+      "loss_partial": 0.07373046875,
+      "step": 99
+    },
+    {
+      "epoch": 0.05185961236249345,
+      "loss_partial": 0.07080078125,
+      "step": 99
+    },
+    {
+      "epoch": 0.05238344683080147,
+      "grad_norm": 0.39076119661331177,
+      "learning_rate": 2.591623036649215e-06,
+      "loss": 0.1447,
+      "step": 100
+    },
+    {
+      "epoch": 0.05238344683080147,
+      "loss_partial": 0.06298828125,
+      "step": 100
+    },
+    {
+      "epoch": 0.05238344683080147,
+      "loss_partial": 0.07177734375,
+      "step": 100
+    },
+    {
+      "epoch": 0.05290728129910948,
+      "grad_norm": 0.3818507492542267,
+      "learning_rate": 2.617801047120419e-06,
+      "loss": 0.1426,
+      "step": 101
+    },
+    {
+      "epoch": 0.05290728129910948,
+      "loss_partial": 0.07666015625,
+      "step": 101
+    },
+    {
+      "epoch": 0.05290728129910948,
+      "loss_partial": 0.07958984375,
+      "step": 101
+    },
+    {
+      "epoch": 0.0534311157674175,
+      "grad_norm": 0.3883614242076874,
+      "learning_rate": 2.643979057591623e-06,
+      "loss": 0.145,
+      "step": 102
+    },
+    {
+      "epoch": 0.0534311157674175,
+      "loss_partial": 0.07177734375,
+      "step": 102
+    },
+    {
+      "epoch": 0.0534311157674175,
+      "loss_partial": 0.0634765625,
+      "step": 102
+    },
+    {
+      "epoch": 0.05395495023572551,
+      "grad_norm": 0.38329410552978516,
+      "learning_rate": 2.6701570680628274e-06,
+      "loss": 0.1399,
+      "step": 103
+    },
+    {
+      "epoch": 0.05395495023572551,
+      "loss_partial": 0.0693359375,
+      "step": 103
+    },
+    {
+      "epoch": 0.05395495023572551,
+      "loss_partial": 0.07177734375,
+      "step": 103
+    },
+    {
+      "epoch": 0.05447878470403353,
+      "grad_norm": 0.37754353880882263,
+      "learning_rate": 2.6963350785340313e-06,
+      "loss": 0.1442,
+      "step": 104
+    },
+    {
+      "epoch": 0.05447878470403353,
+      "loss_partial": 0.0673828125,
+      "step": 104
+    },
+    {
+      "epoch": 0.05447878470403353,
+      "loss_partial": 0.08056640625,
+      "step": 104
+    },
+    {
+      "epoch": 0.05500261917234154,
+      "grad_norm": 0.3691037893295288,
+      "learning_rate": 2.722513089005236e-06,
+      "loss": 0.1416,
+      "step": 105
+    },
+    {
+      "epoch": 0.05500261917234154,
+      "loss_partial": 0.0654296875,
+      "step": 105
+    },
+    {
+      "epoch": 0.05500261917234154,
+      "loss_partial": 0.0654296875,
+      "step": 105
+    },
+    {
+      "epoch": 0.05552645364064956,
+      "grad_norm": 0.3641149401664734,
+      "learning_rate": 2.74869109947644e-06,
+      "loss": 0.137,
+      "step": 106
+    },
+    {
+      "epoch": 0.05552645364064956,
+      "loss_partial": 0.06787109375,
+      "step": 106
+    },
+    {
+      "epoch": 0.05552645364064956,
+      "loss_partial": 0.07177734375,
+      "step": 106
+    },
+    {
+      "epoch": 0.05605028810895757,
+      "grad_norm": 0.3739701509475708,
+      "learning_rate": 2.7748691099476443e-06,
+      "loss": 0.1426,
+      "step": 107
+    },
+    {
+      "epoch": 0.05605028810895757,
+      "loss_partial": 0.07275390625,
+      "step": 107
+    },
+    {
+      "epoch": 0.05605028810895757,
+      "loss_partial": 0.0712890625,
+      "step": 107
+    },
+    {
+      "epoch": 0.05657412257726559,
+      "grad_norm": 0.3921554386615753,
+      "learning_rate": 2.8010471204188483e-06,
+      "loss": 0.1467,
+      "step": 108
+    },
+    {
+      "epoch": 0.05657412257726559,
+      "loss_partial": 0.08642578125,
+      "step": 108
+    },
+    {
+      "epoch": 0.05657412257726559,
+      "loss_partial": 0.0693359375,
+      "step": 108
+    },
+    {
+      "epoch": 0.0570979570455736,
+      "grad_norm": 0.36871716380119324,
+      "learning_rate": 2.8272251308900526e-06,
+      "loss": 0.1455,
+      "step": 109
+    },
+    {
+      "epoch": 0.0570979570455736,
+      "loss_partial": 0.07763671875,
+      "step": 109
+    },
+    {
+      "epoch": 0.0570979570455736,
+      "loss_partial": 0.08837890625,
+      "step": 109
+    },
+    {
+      "epoch": 0.057621791513881616,
+      "grad_norm": 0.3827666938304901,
+      "learning_rate": 2.853403141361257e-06,
+      "loss": 0.1442,
+      "step": 110
+    },
+    {
+      "epoch": 0.057621791513881616,
+      "loss_partial": 0.072265625,
+      "step": 110
+    },
+    {
+      "epoch": 0.057621791513881616,
+      "loss_partial": 0.06396484375,
+      "step": 110
+    },
+    {
+      "epoch": 0.05814562598218963,
+      "grad_norm": 0.3689327836036682,
+      "learning_rate": 2.8795811518324613e-06,
+      "loss": 0.1431,
+      "step": 111
+    },
+    {
+      "epoch": 0.05814562598218963,
+      "loss_partial": 0.0634765625,
+      "step": 111
+    },
+    {
+      "epoch": 0.05814562598218963,
+      "loss_partial": 0.07275390625,
+      "step": 111
+    },
+    {
+      "epoch": 0.058669460450497646,
+      "grad_norm": 0.3583448827266693,
+      "learning_rate": 2.905759162303665e-06,
+      "loss": 0.1445,
+      "step": 112
+    },
+    {
+      "epoch": 0.058669460450497646,
+      "loss_partial": 0.078125,
+      "step": 112
+    },
+    {
+      "epoch": 0.058669460450497646,
+      "loss_partial": 0.068359375,
+      "step": 112
+    },
+    {
+      "epoch": 0.05919329491880566,
+      "grad_norm": 0.38085445761680603,
+      "learning_rate": 2.931937172774869e-06,
+      "loss": 0.1487,
+      "step": 113
+    },
+    {
+      "epoch": 0.05919329491880566,
+      "loss_partial": 0.07373046875,
+      "step": 113
+    },
+    {
+      "epoch": 0.05919329491880566,
+      "loss_partial": 0.07470703125,
+      "step": 113
+    },
+    {
+      "epoch": 0.059717129387113675,
+      "grad_norm": 0.3677493929862976,
+      "learning_rate": 2.9581151832460735e-06,
+      "loss": 0.1393,
+      "step": 114
+    },
+    {
+      "epoch": 0.059717129387113675,
+      "loss_partial": 0.076171875,
+      "step": 114
+    },
+    {
+      "epoch": 0.059717129387113675,
+      "loss_partial": 0.07470703125,
+      "step": 114
+    },
+    {
+      "epoch": 0.060240963855421686,
+      "grad_norm": 0.3653370440006256,
+      "learning_rate": 2.9842931937172774e-06,
+      "loss": 0.1407,
+      "step": 115
+    },
+    {
+      "epoch": 0.060240963855421686,
+      "loss_partial": 0.0693359375,
+      "step": 115
+    },
+    {
+      "epoch": 0.060240963855421686,
+      "loss_partial": 0.0693359375,
+      "step": 115
+    },
+    {
+      "epoch": 0.060764798323729705,
+      "grad_norm": 0.36011698842048645,
+      "learning_rate": 3.010471204188482e-06,
+      "loss": 0.14,
+      "step": 116
+    },
+    {
+      "epoch": 0.060764798323729705,
+      "loss_partial": 0.06884765625,
+      "step": 116
+    },
+    {
+      "epoch": 0.060764798323729705,
+      "loss_partial": 0.0732421875,
+      "step": 116
+    },
+    {
+      "epoch": 0.061288632792037716,
+      "grad_norm": 0.37809130549430847,
+      "learning_rate": 3.036649214659686e-06,
+      "loss": 0.1428,
+      "step": 117
+    },
+    {
+      "epoch": 0.061288632792037716,
+      "loss_partial": 0.0830078125,
+      "step": 117
+    },
+    {
+      "epoch": 0.061288632792037716,
+      "loss_partial": 0.06787109375,
+      "step": 117
+    },
+    {
+      "epoch": 0.061812467260345734,
+      "grad_norm": 0.37813907861709595,
+      "learning_rate": 3.0628272251308904e-06,
+      "loss": 0.1386,
+      "step": 118
+    },
+    {
+      "epoch": 0.061812467260345734,
+      "loss_partial": 0.06884765625,
+      "step": 118
+    },
+    {
+      "epoch": 0.061812467260345734,
+      "loss_partial": 0.07421875,
+      "step": 118
+    },
+    {
+      "epoch": 0.062336301728653745,
+      "grad_norm": 0.36767861247062683,
+      "learning_rate": 3.0890052356020943e-06,
+      "loss": 0.1455,
+      "step": 119
+    },
+    {
+      "epoch": 0.062336301728653745,
+      "loss_partial": 0.0634765625,
+      "step": 119
+    },
+    {
+      "epoch": 0.062336301728653745,
+      "loss_partial": 0.06640625,
+      "step": 119
+    },
+    {
+      "epoch": 0.06286013619696176,
+      "grad_norm": 0.3605897128582001,
+      "learning_rate": 3.115183246073299e-06,
+      "loss": 0.141,
+      "step": 120
+    },
+    {
+      "epoch": 0.06286013619696176,
+      "loss_partial": 0.0703125,
+      "step": 120
+    },
+    {
+      "epoch": 0.06286013619696176,
+      "loss_partial": 0.07568359375,
+      "step": 120
+    },
+    {
+      "epoch": 0.06338397066526978,
+      "grad_norm": 0.3637034595012665,
+      "learning_rate": 3.141361256544503e-06,
+      "loss": 0.1421,
+      "step": 121
+    },
+    {
+      "epoch": 0.06338397066526978,
+      "loss_partial": 0.062255859375,
+      "step": 121
+    },
+    {
+      "epoch": 0.06338397066526978,
+      "loss_partial": 0.059814453125,
+      "step": 121
+    },
+    {
+      "epoch": 0.06390780513357779,
+      "grad_norm": 0.3617505133152008,
+      "learning_rate": 3.167539267015707e-06,
+      "loss": 0.1407,
+      "step": 122
+    },
+    {
+      "epoch": 0.06390780513357779,
+      "loss_partial": 0.07275390625,
+      "step": 122
+    },
+    {
+      "epoch": 0.06390780513357779,
+      "loss_partial": 0.061279296875,
+      "step": 122
+    },
+    {
+      "epoch": 0.0644316396018858,
+      "grad_norm": 0.37682801485061646,
+      "learning_rate": 3.1937172774869113e-06,
+      "loss": 0.1409,
+      "step": 123
+    },
+    {
+      "epoch": 0.0644316396018858,
+      "loss_partial": 0.06591796875,
+      "step": 123
+    },
+    {
+      "epoch": 0.0644316396018858,
+      "loss_partial": 0.0751953125,
+      "step": 123
+    },
+    {
+      "epoch": 0.06495547407019382,
+      "grad_norm": 0.37003788352012634,
+      "learning_rate": 3.219895287958115e-06,
+      "loss": 0.1398,
+      "step": 124
+    },
+    {
+      "epoch": 0.06495547407019382,
+      "loss_partial": 0.0771484375,
+      "step": 124
+    },
+    {
+      "epoch": 0.06495547407019382,
+      "loss_partial": 0.064453125,
+      "step": 124
+    },
+    {
+      "epoch": 0.06547930853850184,
+      "grad_norm": 0.39087042212486267,
+      "learning_rate": 3.2460732984293196e-06,
+      "loss": 0.1399,
+      "step": 125
+    },
+    {
+      "epoch": 0.06547930853850184,
+      "loss_partial": 0.07080078125,
+      "step": 125
+    },
+    {
+      "epoch": 0.06547930853850184,
+      "loss_partial": 0.06201171875,
+      "step": 125
+    },
+    {
+      "epoch": 0.06600314300680984,
+      "grad_norm": 0.3683389127254486,
+      "learning_rate": 3.2722513089005235e-06,
+      "loss": 0.1424,
+      "step": 126
+    },
+    {
+      "epoch": 0.06600314300680984,
+      "loss_partial": 0.0712890625,
+      "step": 126
+    },
+    {
+      "epoch": 0.06600314300680984,
+      "loss_partial": 0.07421875,
+      "step": 126
+    },
+    {
+      "epoch": 0.06652697747511786,
+      "grad_norm": 0.3923996090888977,
+      "learning_rate": 3.2984293193717282e-06,
+      "loss": 0.1407,
+      "step": 127
+    },
+    {
+      "epoch": 0.06652697747511786,
+      "loss_partial": 0.076171875,
+      "step": 127
+    },
+    {
+      "epoch": 0.06652697747511786,
+      "loss_partial": 0.07421875,
+      "step": 127
+    },
+    {
+      "epoch": 0.06705081194342588,
+      "grad_norm": 0.3800947368144989,
+      "learning_rate": 3.324607329842932e-06,
+      "loss": 0.1437,
+      "step": 128
+    },
+    {
+      "epoch": 0.06705081194342588,
+      "loss_partial": 0.07275390625,
+      "step": 128
+    },
+    {
+      "epoch": 0.06705081194342588,
+      "loss_partial": 0.07080078125,
+      "step": 128
+    },
+    {
+      "epoch": 0.06757464641173389,
+      "grad_norm": 0.3816847801208496,
+      "learning_rate": 3.3507853403141365e-06,
+      "loss": 0.1428,
+      "step": 129
+    },
+    {
+      "epoch": 0.06757464641173389,
+      "loss_partial": 0.0595703125,
+      "step": 129
+    },
+    {
+      "epoch": 0.06757464641173389,
+      "loss_partial": 0.0771484375,
+      "step": 129
+    },
+    {
+      "epoch": 0.0680984808800419,
+      "grad_norm": 0.3732149302959442,
+      "learning_rate": 3.3769633507853404e-06,
+      "loss": 0.136,
+      "step": 130
+    },
+    {
+      "epoch": 0.0680984808800419,
+      "loss_partial": 0.057861328125,
+      "step": 130
+    },
+    {
+      "epoch": 0.0680984808800419,
+      "loss_partial": 0.076171875,
+      "step": 130
+    },
+    {
+      "epoch": 0.06862231534834992,
+      "grad_norm": 0.37888339161872864,
+      "learning_rate": 3.403141361256545e-06,
+      "loss": 0.1427,
+      "step": 131
+    },
+    {
+      "epoch": 0.06862231534834992,
+      "loss_partial": 0.08056640625,
+      "step": 131
+    },
+    {
+      "epoch": 0.06862231534834992,
+      "loss_partial": 0.0654296875,
+      "step": 131
+    },
+    {
+      "epoch": 0.06914614981665794,
+      "grad_norm": 0.3805946409702301,
+      "learning_rate": 3.429319371727749e-06,
+      "loss": 0.1391,
+      "step": 132
+    },
+    {
+      "epoch": 0.06914614981665794,
+      "loss_partial": 0.07568359375,
+      "step": 132
+    },
+    {
+      "epoch": 0.06914614981665794,
+      "loss_partial": 0.0625,
+      "step": 132
+    },
+    {
+      "epoch": 0.06966998428496594,
+      "grad_norm": 0.3996601104736328,
+      "learning_rate": 3.455497382198953e-06,
+      "loss": 0.1399,
+      "step": 133
+    },
+    {
+      "epoch": 0.06966998428496594,
+      "loss_partial": 0.07470703125,
+      "step": 133
+    },
+    {
+      "epoch": 0.06966998428496594,
+      "loss_partial": 0.06396484375,
+      "step": 133
+    },
+    {
+      "epoch": 0.07019381875327396,
+      "grad_norm": 0.39929482340812683,
+      "learning_rate": 3.4816753926701574e-06,
+      "loss": 0.139,
+      "step": 134
+    },
+    {
+      "epoch": 0.07019381875327396,
+      "loss_partial": 0.0703125,
+      "step": 134
+    },
+    {
+      "epoch": 0.07019381875327396,
+      "loss_partial": 0.07958984375,
+      "step": 134
+    },
+    {
+      "epoch": 0.07071765322158198,
+      "grad_norm": 0.40471184253692627,
+      "learning_rate": 3.5078534031413613e-06,
+      "loss": 0.1414,
+      "step": 135
+    },
+    {
+      "epoch": 0.07071765322158198,
+      "loss_partial": 0.0751953125,
+      "step": 135
+    },
+    {
+      "epoch": 0.07071765322158198,
+      "loss_partial": 0.06591796875,
+      "step": 135
+    },
+    {
+      "epoch": 0.07124148768989,
+      "grad_norm": 0.3969804048538208,
+      "learning_rate": 3.534031413612566e-06,
+      "loss": 0.1377,
+      "step": 136
+    },
+    {
+      "epoch": 0.07124148768989,
+      "loss_partial": 0.06884765625,
+      "step": 136
+    },
+    {
+      "epoch": 0.07124148768989,
+      "loss_partial": 0.060791015625,
+      "step": 136
+    },
+    {
+      "epoch": 0.071765322158198,
+      "grad_norm": 0.38891926407814026,
+      "learning_rate": 3.56020942408377e-06,
+      "loss": 0.1378,
+      "step": 137
+    },
+    {
+      "epoch": 0.071765322158198,
+      "loss_partial": 0.0703125,
+      "step": 137
+    },
+    {
+      "epoch": 0.071765322158198,
+      "loss_partial": 0.060546875,
+      "step": 137
+    },
+    {
+      "epoch": 0.07228915662650602,
+      "grad_norm": 0.41542863845825195,
+      "learning_rate": 3.5863874345549743e-06,
+      "loss": 0.1448,
+      "step": 138
+    },
+    {
+      "epoch": 0.07228915662650602,
+      "loss_partial": 0.072265625,
+      "step": 138
+    },
+    {
+      "epoch": 0.07228915662650602,
+      "loss_partial": 0.06298828125,
+      "step": 138
+    },
+    {
+      "epoch": 0.07281299109481404,
+      "grad_norm": 0.42096996307373047,
+      "learning_rate": 3.6125654450261782e-06,
+      "loss": 0.1418,
+      "step": 139
+    },
+    {
+      "epoch": 0.07281299109481404,
+      "loss_partial": 0.07568359375,
+      "step": 139
+    },
+    {
+      "epoch": 0.07281299109481404,
+      "loss_partial": 0.07470703125,
+      "step": 139
+    },
+    {
+      "epoch": 0.07333682556312206,
+      "grad_norm": 0.3953329920768738,
+      "learning_rate": 3.6387434554973826e-06,
+      "loss": 0.1394,
+      "step": 140
+    },
+    {
+      "epoch": 0.07333682556312206,
+      "loss_partial": 0.07080078125,
+      "step": 140
+    },
+    {
+      "epoch": 0.07333682556312206,
+      "loss_partial": 0.06640625,
+      "step": 140
+    },
+    {
+      "epoch": 0.07386066003143006,
+      "grad_norm": 0.40706074237823486,
+      "learning_rate": 3.6649214659685865e-06,
+      "loss": 0.138,
+      "step": 141
+    },
+    {
+      "epoch": 0.07386066003143006,
+      "loss_partial": 0.07080078125,
+      "step": 141
+    },
+    {
+      "epoch": 0.07386066003143006,
+      "loss_partial": 0.05908203125,
+      "step": 141
+    },
+    {
+      "epoch": 0.07438449449973808,
+      "grad_norm": 0.42318087816238403,
+      "learning_rate": 3.6910994764397904e-06,
+      "loss": 0.1353,
+      "step": 142
+    },
+    {
+      "epoch": 0.07438449449973808,
+      "loss_partial": 0.072265625,
+      "step": 142
+    },
+    {
+      "epoch": 0.07438449449973808,
+      "loss_partial": 0.06298828125,
+      "step": 142
+    },
+    {
+      "epoch": 0.0749083289680461,
+      "grad_norm": 0.41577693819999695,
+      "learning_rate": 3.717277486910995e-06,
+      "loss": 0.1381,
+      "step": 143
+    },
+    {
+      "epoch": 0.0749083289680461,
+      "loss_partial": 0.080078125,
+      "step": 143
+    },
+    {
+      "epoch": 0.0749083289680461,
+      "loss_partial": 0.072265625,
+      "step": 143
+    },
+    {
+      "epoch": 0.07543216343635412,
+      "grad_norm": 0.4162280559539795,
+      "learning_rate": 3.743455497382199e-06,
+      "loss": 0.1393,
+      "step": 144
+    },
+    {
+      "epoch": 0.07543216343635412,
+      "loss_partial": 0.0673828125,
+      "step": 144
+    },
+    {
+      "epoch": 0.07543216343635412,
+      "loss_partial": 0.0732421875,
+      "step": 144
+    },
+    {
+      "epoch": 0.07595599790466212,
+      "grad_norm": 0.41159018874168396,
+      "learning_rate": 3.7696335078534035e-06,
+      "loss": 0.1376,
+      "step": 145
+    },
+    {
+      "epoch": 0.07595599790466212,
+      "loss_partial": 0.06298828125,
+      "step": 145
+    },
+    {
+      "epoch": 0.07595599790466212,
+      "loss_partial": 0.07421875,
+      "step": 145
+    },
+    {
+      "epoch": 0.07647983237297014,
+      "grad_norm": 0.42202210426330566,
+      "learning_rate": 3.7958115183246074e-06,
+      "loss": 0.1383,
+      "step": 146
+    },
+    {
+      "epoch": 0.07647983237297014,
+      "loss_partial": 0.06982421875,
+      "step": 146
+    },
+    {
+      "epoch": 0.07647983237297014,
+      "loss_partial": 0.0634765625,
+      "step": 146
+    },
+    {
+      "epoch": 0.07700366684127816,
+      "grad_norm": 0.40507227182388306,
+      "learning_rate": 3.821989528795812e-06,
+      "loss": 0.1354,
+      "step": 147
+    },
+    {
+      "epoch": 0.07700366684127816,
+      "loss_partial": 0.07421875,
+      "step": 147
+    },
+    {
+      "epoch": 0.07700366684127816,
+      "loss_partial": 0.076171875,
+      "step": 147
+    },
+    {
+      "epoch": 0.07752750130958617,
+      "grad_norm": 0.45268356800079346,
+      "learning_rate": 3.848167539267016e-06,
+      "loss": 0.1429,
+      "step": 148
+    },
+    {
+      "epoch": 0.07752750130958617,
+      "loss_partial": 0.06787109375,
+      "step": 148
+    },
+    {
+      "epoch": 0.07752750130958617,
+      "loss_partial": 0.06494140625,
+      "step": 148
+    },
+    {
+      "epoch": 0.07805133577789418,
+      "grad_norm": 0.414663702249527,
+      "learning_rate": 3.874345549738221e-06,
+      "loss": 0.1357,
+      "step": 149
+    },
+    {
+      "epoch": 0.07805133577789418,
+      "loss_partial": 0.07568359375,
+      "step": 149
+    },
+    {
+      "epoch": 0.07805133577789418,
+      "loss_partial": 0.08056640625,
+      "step": 149
+    },
+    {
+      "epoch": 0.0785751702462022,
+      "grad_norm": 0.4210567772388458,
+      "learning_rate": 3.900523560209425e-06,
+      "loss": 0.1383,
+      "step": 150
+    },
+    {
+      "epoch": 0.0785751702462022,
+      "loss_partial": 0.07275390625,
+      "step": 150
+    },
+    {
+      "epoch": 0.0785751702462022,
+      "loss_partial": 0.06689453125,
+      "step": 150
+    },
+    {
+      "epoch": 0.07909900471451022,
+      "grad_norm": 0.39930424094200134,
+      "learning_rate": 3.926701570680629e-06,
+      "loss": 0.138,
+      "step": 151
+    },
+    {
+      "epoch": 0.07909900471451022,
+      "loss_partial": 0.06591796875,
+      "step": 151
+    },
+    {
+      "epoch": 0.07909900471451022,
+      "loss_partial": 0.0634765625,
+      "step": 151
+    },
+    {
+      "epoch": 0.07962283918281823,
+      "grad_norm": 0.41718700528144836,
+      "learning_rate": 3.952879581151833e-06,
+      "loss": 0.1347,
+      "step": 152
+    },
+    {
+      "epoch": 0.07962283918281823,
+      "loss_partial": 0.06640625,
+      "step": 152
+    },
+    {
+      "epoch": 0.07962283918281823,
+      "loss_partial": 0.06689453125,
+      "step": 152
+    },
+    {
+      "epoch": 0.08014667365112624,
+      "grad_norm": 0.4196344017982483,
+      "learning_rate": 3.9790575916230365e-06,
+      "loss": 0.1412,
+      "step": 153
+    },
+    {
+      "epoch": 0.08014667365112624,
+      "loss_partial": 0.0625,
+      "step": 153
+    },
+    {
+      "epoch": 0.08014667365112624,
+      "loss_partial": 0.0625,
+      "step": 153
+    },
+    {
+      "epoch": 0.08067050811943426,
+      "grad_norm": 0.42472606897354126,
+      "learning_rate": 4.005235602094241e-06,
+      "loss": 0.1372,
+      "step": 154
+    },
+    {
+      "epoch": 0.08067050811943426,
+      "loss_partial": 0.0654296875,
+      "step": 154
+    },
+    {
+      "epoch": 0.08067050811943426,
+      "loss_partial": 0.0712890625,
+      "step": 154
+    },
+    {
+      "epoch": 0.08119434258774227,
+      "grad_norm": 0.4298214316368103,
+      "learning_rate": 4.031413612565445e-06,
+      "loss": 0.1418,
+      "step": 155
+    },
+    {
+      "epoch": 0.08119434258774227,
+      "loss_partial": 0.07080078125,
+      "step": 155
+    },
+    {
+      "epoch": 0.08119434258774227,
+      "loss_partial": 0.072265625,
+      "step": 155
+    },
+    {
+      "epoch": 0.08171817705605029,
+      "grad_norm": 0.4288150668144226,
+      "learning_rate": 4.05759162303665e-06,
+      "loss": 0.1381,
+      "step": 156
+    },
+    {
+      "epoch": 0.08171817705605029,
+      "loss_partial": 0.06884765625,
+      "step": 156
+    },
+    {
+      "epoch": 0.08171817705605029,
+      "loss_partial": 0.0634765625,
+      "step": 156
+    },
+    {
+      "epoch": 0.0822420115243583,
+      "grad_norm": 0.43042105436325073,
+      "learning_rate": 4.083769633507854e-06,
+      "loss": 0.1369,
+      "step": 157
+    },
+    {
+      "epoch": 0.0822420115243583,
+      "loss_partial": 0.06494140625,
+      "step": 157
+    },
+    {
+      "epoch": 0.0822420115243583,
+      "loss_partial": 0.07373046875,
+      "step": 157
+    },
+    {
+      "epoch": 0.08276584599266631,
+      "grad_norm": 0.4310128390789032,
+      "learning_rate": 4.109947643979058e-06,
+      "loss": 0.1359,
+      "step": 158
+    },
+    {
+      "epoch": 0.08276584599266631,
+      "loss_partial": 0.07080078125,
+      "step": 158
+    },
+    {
+      "epoch": 0.08276584599266631,
+      "loss_partial": 0.06689453125,
+      "step": 158
+    },
+    {
+      "epoch": 0.08328968046097433,
+      "grad_norm": 0.42166629433631897,
+      "learning_rate": 4.136125654450262e-06,
+      "loss": 0.1396,
+      "step": 159
+    },
+    {
+      "epoch": 0.08328968046097433,
+      "loss_partial": 0.06494140625,
+      "step": 159
+    },
+    {
+      "epoch": 0.08328968046097433,
+      "loss_partial": 0.06298828125,
+      "step": 159
+    },
+    {
+      "epoch": 0.08381351492928235,
+      "grad_norm": 0.44633281230926514,
+      "learning_rate": 4.1623036649214665e-06,
+      "loss": 0.136,
+      "step": 160
+    },
+    {
+      "epoch": 0.08381351492928235,
+      "loss_partial": 0.0576171875,
+      "step": 160
+    },
+    {
+      "epoch": 0.08381351492928235,
+      "loss_partial": 0.057373046875,
+      "step": 160
+    },
+    {
+      "epoch": 0.08433734939759036,
+      "grad_norm": 0.43694010376930237,
+      "learning_rate": 4.18848167539267e-06,
+      "loss": 0.1385,
+      "step": 161
+    },
+    {
+      "epoch": 0.08433734939759036,
+      "loss_partial": 0.0693359375,
+      "step": 161
+    },
+    {
+      "epoch": 0.08433734939759036,
+      "loss_partial": 0.0712890625,
+      "step": 161
+    },
+    {
+      "epoch": 0.08486118386589837,
+      "grad_norm": 0.4263884425163269,
+      "learning_rate": 4.214659685863874e-06,
+      "loss": 0.141,
+      "step": 162
+    },
+    {
+      "epoch": 0.08486118386589837,
+      "loss_partial": 0.0693359375,
+      "step": 162
+    },
+    {
+      "epoch": 0.08486118386589837,
+      "loss_partial": 0.059326171875,
+      "step": 162
+    },
+    {
+      "epoch": 0.08538501833420639,
+      "grad_norm": 0.42644524574279785,
+      "learning_rate": 4.240837696335079e-06,
+      "loss": 0.1377,
+      "step": 163
+    },
+    {
+      "epoch": 0.08538501833420639,
+      "loss_partial": 0.06396484375,
+      "step": 163
+    },
+    {
+      "epoch": 0.08538501833420639,
+      "loss_partial": 0.0771484375,
+      "step": 163
+    },
+    {
+      "epoch": 0.08590885280251441,
+      "grad_norm": 0.4281913936138153,
+      "learning_rate": 4.267015706806283e-06,
+      "loss": 0.1383,
+      "step": 164
+    },
+    {
+      "epoch": 0.08590885280251441,
+      "loss_partial": 0.07275390625,
+      "step": 164
+    },
+    {
+      "epoch": 0.08590885280251441,
+      "loss_partial": 0.068359375,
+      "step": 164
+    },
+    {
+      "epoch": 0.08643268727082241,
+      "grad_norm": 0.4207417964935303,
+      "learning_rate": 4.293193717277487e-06,
+      "loss": 0.1404,
+      "step": 165
+    },
+    {
+      "epoch": 0.08643268727082241,
+      "loss_partial": 0.07763671875,
+      "step": 165
+    },
+    {
+      "epoch": 0.08643268727082241,
+      "loss_partial": 0.055419921875,
+      "step": 165
+    },
+    {
+      "epoch": 0.08695652173913043,
+      "grad_norm": 0.41510045528411865,
+      "learning_rate": 4.319371727748692e-06,
+      "loss": 0.1357,
+      "step": 166
+    },
+    {
+      "epoch": 0.08695652173913043,
+      "loss_partial": 0.06640625,
+      "step": 166
+    },
+    {
+      "epoch": 0.08695652173913043,
+      "loss_partial": 0.06201171875,
+      "step": 166
+    },
+    {
+      "epoch": 0.08748035620743845,
+      "grad_norm": 0.3964497745037079,
+      "learning_rate": 4.345549738219896e-06,
+      "loss": 0.1361,
+      "step": 167
+    },
+    {
+      "epoch": 0.08748035620743845,
+      "loss_partial": 0.0625,
+      "step": 167
+    },
+    {
+      "epoch": 0.08748035620743845,
+      "loss_partial": 0.05517578125,
+      "step": 167
+    },
+    {
+      "epoch": 0.08800419067574647,
+      "grad_norm": 0.3660443425178528,
+      "learning_rate": 4.3717277486910996e-06,
+      "loss": 0.134,
+      "step": 168
+    },
+    {
+      "epoch": 0.08800419067574647,
+      "loss_partial": 0.0771484375,
+      "step": 168
+    },
+    {
+      "epoch": 0.08800419067574647,
+      "loss_partial": 0.06201171875,
+      "step": 168
+    },
+    {
+      "epoch": 0.08852802514405447,
+      "grad_norm": 0.39950138330459595,
+      "learning_rate": 4.397905759162304e-06,
+      "loss": 0.1376,
+      "step": 169
+    },
+    {
+      "epoch": 0.08852802514405447,
+      "loss_partial": 0.068359375,
+      "step": 169
+    },
+    {
+      "epoch": 0.08852802514405447,
+      "loss_partial": 0.06640625,
+      "step": 169
+    },
+    {
+      "epoch": 0.08905185961236249,
+      "grad_norm": 0.4078963100910187,
+      "learning_rate": 4.424083769633508e-06,
+      "loss": 0.1384,
+      "step": 170
+    },
+    {
+      "epoch": 0.08905185961236249,
+      "loss_partial": 0.0771484375,
+      "step": 170
+    },
+    {
+      "epoch": 0.08905185961236249,
+      "loss_partial": 0.06982421875,
+      "step": 170
+    },
+    {
+      "epoch": 0.08957569408067051,
+      "grad_norm": 0.367013543844223,
+      "learning_rate": 4.450261780104713e-06,
+      "loss": 0.1338,
+      "step": 171
+    },
+    {
+      "epoch": 0.08957569408067051,
+      "loss_partial": 0.0703125,
+      "step": 171
+    },
+    {
+      "epoch": 0.08957569408067051,
+      "loss_partial": 0.0693359375,
+      "step": 171
+    },
+    {
+      "epoch": 0.09009952854897853,
+      "grad_norm": 0.3751515746116638,
+      "learning_rate": 4.476439790575917e-06,
+      "loss": 0.1373,
+      "step": 172
+    },
+    {
+      "epoch": 0.09009952854897853,
+      "loss_partial": 0.06640625,
+      "step": 172
+    },
+    {
+      "epoch": 0.09009952854897853,
+      "loss_partial": 0.06884765625,
+      "step": 172
+    },
+    {
+      "epoch": 0.09062336301728653,
+      "grad_norm": 0.37735965847969055,
+      "learning_rate": 4.502617801047121e-06,
+      "loss": 0.1357,
+      "step": 173
+    },
+    {
+      "epoch": 0.09062336301728653,
+      "loss_partial": 0.07080078125,
+      "step": 173
+    },
+    {
+      "epoch": 0.09062336301728653,
+      "loss_partial": 0.06201171875,
+      "step": 173
+    },
+    {
+      "epoch": 0.09114719748559455,
+      "grad_norm": 0.3667598366737366,
+      "learning_rate": 4.528795811518325e-06,
+      "loss": 0.1346,
+      "step": 174
+    },
+    {
+      "epoch": 0.09114719748559455,
+      "loss_partial": 0.06494140625,
+      "step": 174
+    },
+    {
+      "epoch": 0.09114719748559455,
+      "loss_partial": 0.06494140625,
+      "step": 174
+    },
+    {
+      "epoch": 0.09167103195390257,
+      "grad_norm": 0.3580823838710785,
+      "learning_rate": 4.554973821989529e-06,
+      "loss": 0.1321,
+      "step": 175
+    },
+    {
+      "epoch": 0.09167103195390257,
+      "loss_partial": 0.0732421875,
+      "step": 175
+    },
+    {
+      "epoch": 0.09167103195390257,
+      "loss_partial": 0.0712890625,
+      "step": 175
+    },
+    {
+      "epoch": 0.09219486642221059,
+      "grad_norm": 0.3762848973274231,
+      "learning_rate": 4.5811518324607335e-06,
+      "loss": 0.1366,
+      "step": 176
+    },
+    {
+      "epoch": 0.09219486642221059,
+      "loss_partial": 0.06787109375,
+      "step": 176
+    },
+    {
+      "epoch": 0.09219486642221059,
+      "loss_partial": 0.06494140625,
+      "step": 176
+    },
+    {
+      "epoch": 0.09271870089051859,
+      "grad_norm": 0.36078372597694397,
+      "learning_rate": 4.607329842931937e-06,
+      "loss": 0.1377,
+      "step": 177
+    },
+    {
+      "epoch": 0.09271870089051859,
+      "loss_partial": 0.06494140625,
+      "step": 177
+    },
+    {
+      "epoch": 0.09271870089051859,
+      "loss_partial": 0.0654296875,
+      "step": 177
+    },
+    {
+      "epoch": 0.09324253535882661,
+      "grad_norm": 0.3635613024234772,
+      "learning_rate": 4.633507853403142e-06,
+      "loss": 0.1393,
+      "step": 178
+    },
+    {
+      "epoch": 0.09324253535882661,
+      "loss_partial": 0.06689453125,
+      "step": 178
+    },
+    {
+      "epoch": 0.09324253535882661,
+      "loss_partial": 0.059326171875,
+      "step": 178
+    },
+    {
+      "epoch": 0.09376636982713463,
+      "grad_norm": 0.3441106379032135,
+      "learning_rate": 4.659685863874346e-06,
+      "loss": 0.1349,
+      "step": 179
+    },
+    {
+      "epoch": 0.09376636982713463,
+      "loss_partial": 0.07373046875,
+      "step": 179
+    },
+    {
+      "epoch": 0.09376636982713463,
+      "loss_partial": 0.0654296875,
+      "step": 179
+    },
+    {
+      "epoch": 0.09429020429544265,
+      "grad_norm": 0.3488267958164215,
+      "learning_rate": 4.68586387434555e-06,
+      "loss": 0.1312,
+      "step": 180
+    },
+    {
+      "epoch": 0.09429020429544265,
+      "loss_partial": 0.0693359375,
+      "step": 180
+    },
+    {
+      "epoch": 0.09429020429544265,
+      "loss_partial": 0.0654296875,
+      "step": 180
+    },
+    {
+      "epoch": 0.09481403876375065,
+      "grad_norm": 0.34671565890312195,
+      "learning_rate": 4.712041884816754e-06,
+      "loss": 0.1363,
+      "step": 181
+    },
+    {
+      "epoch": 0.09481403876375065,
+      "loss_partial": 0.06494140625,
+      "step": 181
+    },
+    {
+      "epoch": 0.09481403876375065,
+      "loss_partial": 0.06396484375,
+      "step": 181
+    },
+    {
+      "epoch": 0.09533787323205867,
+      "grad_norm": 0.34436556696891785,
+      "learning_rate": 4.738219895287958e-06,
+      "loss": 0.1346,
+      "step": 182
+    },
+    {
+      "epoch": 0.09533787323205867,
+      "loss_partial": 0.07373046875,
+      "step": 182
+    },
+    {
+      "epoch": 0.09533787323205867,
+      "loss_partial": 0.07177734375,
+      "step": 182
+    },
+    {
+      "epoch": 0.09586170770036669,
+      "grad_norm": 0.3462006747722626,
+      "learning_rate": 4.764397905759163e-06,
+      "loss": 0.1346,
+      "step": 183
+    },
+    {
+      "epoch": 0.09586170770036669,
+      "loss_partial": 0.0703125,
+      "step": 183
+    },
+    {
+      "epoch": 0.09586170770036669,
+      "loss_partial": 0.0673828125,
+      "step": 183
+    },
+    {
+      "epoch": 0.0963855421686747,
+      "grad_norm": 0.3395676910877228,
+      "learning_rate": 4.7905759162303665e-06,
+      "loss": 0.134,
+      "step": 184
+    },
+    {
+      "epoch": 0.0963855421686747,
+      "loss_partial": 0.06494140625,
+      "step": 184
+    },
+    {
+      "epoch": 0.0963855421686747,
+      "loss_partial": 0.062255859375,
+      "step": 184
+    },
+    {
+      "epoch": 0.09690937663698271,
+      "grad_norm": 0.3469320237636566,
+      "learning_rate": 4.816753926701571e-06,
+      "loss": 0.1304,
+      "step": 185
+    },
+    {
+      "epoch": 0.09690937663698271,
+      "loss_partial": 0.06787109375,
+      "step": 185
+    },
+    {
+      "epoch": 0.09690937663698271,
+      "loss_partial": 0.06494140625,
+      "step": 185
+    },
+    {
+      "epoch": 0.09743321110529073,
+      "grad_norm": 0.33478790521621704,
+      "learning_rate": 4.842931937172775e-06,
+      "loss": 0.1369,
+      "step": 186
+    },
+    {
+      "epoch": 0.09743321110529073,
+      "loss_partial": 0.06298828125,
+      "step": 186
+    },
+    {
+      "epoch": 0.09743321110529073,
+      "loss_partial": 0.0712890625,
+      "step": 186
+    },
+    {
+      "epoch": 0.09795704557359874,
+      "grad_norm": 0.3486548662185669,
+      "learning_rate": 4.86910994764398e-06,
+      "loss": 0.136,
+      "step": 187
+    },
+    {
+      "epoch": 0.09795704557359874,
+      "loss_partial": 0.0693359375,
+      "step": 187
+    },
+    {
+      "epoch": 0.09795704557359874,
+      "loss_partial": 0.06494140625,
+      "step": 187
+    },
+    {
+      "epoch": 0.09848088004190676,
+      "grad_norm": 0.3477793335914612,
+      "learning_rate": 4.895287958115184e-06,
+      "loss": 0.1304,
+      "step": 188
+    },
+    {
+      "epoch": 0.09848088004190676,
+      "loss_partial": 0.0693359375,
+      "step": 188
+    },
+    {
+      "epoch": 0.09848088004190676,
+      "loss_partial": 0.06298828125,
+      "step": 188
+    },
+    {
+      "epoch": 0.09900471451021477,
+      "grad_norm": 0.35378918051719666,
+      "learning_rate": 4.921465968586388e-06,
+      "loss": 0.1361,
+      "step": 189
+    },
+    {
+      "epoch": 0.09900471451021477,
+      "loss_partial": 0.07421875,
+      "step": 189
+    },
+    {
+      "epoch": 0.09900471451021477,
+      "loss_partial": 0.06640625,
+      "step": 189
+    },
+    {
+      "epoch": 0.09952854897852279,
+      "grad_norm": 0.32317793369293213,
+      "learning_rate": 4.947643979057592e-06,
+      "loss": 0.129,
+      "step": 190
+    },
+    {
+      "epoch": 0.09952854897852279,
+      "loss_partial": 0.0693359375,
+      "step": 190
+    },
+    {
+      "epoch": 0.09952854897852279,
+      "loss_partial": 0.07275390625,
+      "step": 190
+    },
+    {
+      "epoch": 0.1000523834468308,
+      "grad_norm": 0.3380870223045349,
+      "learning_rate": 4.9738219895287965e-06,
+      "loss": 0.1345,
+      "step": 191
+    },
+    {
+      "epoch": 0.1000523834468308,
+      "loss_partial": 0.0751953125,
+      "step": 191
+    },
+    {
+      "epoch": 0.1000523834468308,
+      "loss_partial": 0.064453125,
+      "step": 191
+    },
+    {
+      "epoch": 0.10057621791513882,
+      "grad_norm": 0.3331289291381836,
+      "learning_rate": 5e-06,
+      "loss": 0.1323,
+      "step": 192
+    },
+    {
+      "epoch": 0.10057621791513882,
+      "loss_partial": 0.068359375,
+      "step": 192
+    },
+    {
+      "epoch": 0.10057621791513882,
+      "loss_partial": 0.07373046875,
+      "step": 192
+    },
+    {
+      "epoch": 0.10110005238344683,
+      "grad_norm": 0.32727527618408203,
+      "learning_rate": 5.026178010471204e-06,
+      "loss": 0.1354,
+      "step": 193
+    },
+    {
+      "epoch": 0.10110005238344683,
+      "loss_partial": 0.06884765625,
+      "step": 193
+    },
+    {
+      "epoch": 0.10110005238344683,
+      "loss_partial": 0.0615234375,
+      "step": 193
+    },
+    {
+      "epoch": 0.10162388685175484,
+      "grad_norm": 0.3294677734375,
+      "learning_rate": 5.052356020942408e-06,
+      "loss": 0.1314,
+      "step": 194
+    },
+    {
+      "epoch": 0.10162388685175484,
+      "loss_partial": 0.059814453125,
+      "step": 194
+    },
+    {
+      "epoch": 0.10162388685175484,
+      "loss_partial": 0.06298828125,
+      "step": 194
+    },
+    {
+      "epoch": 0.10214772132006286,
+      "grad_norm": 0.32721561193466187,
+      "learning_rate": 5.078534031413613e-06,
+      "loss": 0.1309,
+      "step": 195
+    },
+    {
+      "epoch": 0.10214772132006286,
+      "loss_partial": 0.06689453125,
+      "step": 195
+    },
+    {
+      "epoch": 0.10214772132006286,
+      "loss_partial": 0.0625,
+      "step": 195
+    },
+    {
+      "epoch": 0.10267155578837088,
+      "grad_norm": 0.33825433254241943,
+      "learning_rate": 5.104712041884817e-06,
+      "loss": 0.131,
+      "step": 196
+    },
+    {
+      "epoch": 0.10267155578837088,
+      "loss_partial": 0.0712890625,
+      "step": 196
+    },
+    {
+      "epoch": 0.10267155578837088,
+      "loss_partial": 0.064453125,
+      "step": 196
+    },
+    {
+      "epoch": 0.10319539025667888,
+      "grad_norm": 0.31203576922416687,
+      "learning_rate": 5.130890052356021e-06,
+      "loss": 0.1284,
+      "step": 197
+    },
+    {
+      "epoch": 0.10319539025667888,
+      "loss_partial": 0.06689453125,
+      "step": 197
+    },
+    {
+      "epoch": 0.10319539025667888,
+      "loss_partial": 0.06201171875,
+      "step": 197
+    },
+    {
+      "epoch": 0.1037192247249869,
+      "grad_norm": 0.3228139281272888,
+      "learning_rate": 5.157068062827225e-06,
+      "loss": 0.1357,
+      "step": 198
+    },
+    {
+      "epoch": 0.1037192247249869,
+      "loss_partial": 0.06201171875,
+      "step": 198
+    },
+    {
+      "epoch": 0.1037192247249869,
+      "loss_partial": 0.060791015625,
+      "step": 198
+    },
+    {
+      "epoch": 0.10424305919329492,
+      "grad_norm": 0.32399246096611023,
+      "learning_rate": 5.18324607329843e-06,
+      "loss": 0.131,
+      "step": 199
+    },
+    {
+      "epoch": 0.10424305919329492,
+      "loss_partial": 0.06103515625,
+      "step": 199
+    },
+    {
+      "epoch": 0.10424305919329492,
+      "loss_partial": 0.0634765625,
+      "step": 199
+    },
+    {
+      "epoch": 0.10476689366160294,
+      "grad_norm": 0.33202290534973145,
+      "learning_rate": 5.209424083769634e-06,
+      "loss": 0.1338,
+      "step": 200
+    },
+    {
+      "epoch": 0.10476689366160294,
+      "loss_partial": 0.0703125,
+      "step": 200
+    },
+    {
+      "epoch": 0.10476689366160294,
+      "loss_partial": 0.06884765625,
+      "step": 200
+    },
+    {
+      "epoch": 0.10529072812991094,
+      "grad_norm": 0.3111969828605652,
+      "learning_rate": 5.235602094240838e-06,
+      "loss": 0.1343,
+      "step": 201
+    },
+    {
+      "epoch": 0.10529072812991094,
+      "loss_partial": 0.068359375,
+      "step": 201
+    },
+    {
+      "epoch": 0.10529072812991094,
+      "loss_partial": 0.0654296875,
+      "step": 201
+    },
+    {
+      "epoch": 0.10581456259821896,
+      "grad_norm": 0.3213214576244354,
+      "learning_rate": 5.261780104712042e-06,
+      "loss": 0.1347,
+      "step": 202
+    },
+    {
+      "epoch": 0.10581456259821896,
+      "loss_partial": 0.059326171875,
+      "step": 202
+    },
+    {
+      "epoch": 0.10581456259821896,
+      "loss_partial": 0.0673828125,
+      "step": 202
+    },
+    {
+      "epoch": 0.10633839706652698,
+      "grad_norm": 0.30267563462257385,
+      "learning_rate": 5.287958115183246e-06,
+      "loss": 0.1277,
+      "step": 203
+    },
+    {
+      "epoch": 0.10633839706652698,
+      "loss_partial": 0.0703125,
+      "step": 203
+    },
+    {
+      "epoch": 0.10633839706652698,
+      "loss_partial": 0.0634765625,
+      "step": 203
+    },
+    {
+      "epoch": 0.106862231534835,
+      "grad_norm": 0.303792804479599,
+      "learning_rate": 5.314136125654451e-06,
+      "loss": 0.1298,
+      "step": 204
+    },
+    {
+      "epoch": 0.106862231534835,
+      "loss_partial": 0.05908203125,
+      "step": 204
+    },
+    {
+      "epoch": 0.106862231534835,
+      "loss_partial": 0.06982421875,
+      "step": 204
+    },
+    {
+      "epoch": 0.107386066003143,
+      "grad_norm": 0.31618326902389526,
+      "learning_rate": 5.340314136125655e-06,
+      "loss": 0.1313,
+      "step": 205
+    },
+    {
+      "epoch": 0.107386066003143,
+      "loss_partial": 0.07275390625,
+      "step": 205
+    },
+    {
+      "epoch": 0.107386066003143,
+      "loss_partial": 0.0712890625,
+      "step": 205
+    },
+    {
+      "epoch": 0.10790990047145102,
+      "grad_norm": 0.2919757068157196,
+      "learning_rate": 5.366492146596859e-06,
+      "loss": 0.1306,
+      "step": 206
+    },
+    {
+      "epoch": 0.10790990047145102,
+      "loss_partial": 0.07275390625,
+      "step": 206
+    },
+    {
+      "epoch": 0.10790990047145102,
+      "loss_partial": 0.0673828125,
+      "step": 206
+    },
+    {
+      "epoch": 0.10843373493975904,
+      "grad_norm": 0.31756073236465454,
+      "learning_rate": 5.392670157068063e-06,
+      "loss": 0.1313,
+      "step": 207
+    },
+    {
+      "epoch": 0.10843373493975904,
+      "loss_partial": 0.0712890625,
+      "step": 207
+    },
+    {
+      "epoch": 0.10843373493975904,
+      "loss_partial": 0.054443359375,
+      "step": 207
+    },
+    {
+      "epoch": 0.10895756940806706,
+      "grad_norm": 0.3006095290184021,
+      "learning_rate": 5.418848167539268e-06,
+      "loss": 0.1332,
+      "step": 208
+    },
+    {
+      "epoch": 0.10895756940806706,
+      "loss_partial": 0.05810546875,
+      "step": 208
+    },
+    {
+      "epoch": 0.10895756940806706,
+      "loss_partial": 0.06591796875,
+      "step": 208
+    },
+    {
+      "epoch": 0.10948140387637506,
+      "grad_norm": 0.2915012240409851,
+      "learning_rate": 5.445026178010472e-06,
+      "loss": 0.1316,
+      "step": 209
+    },
+    {
+      "epoch": 0.10948140387637506,
+      "loss_partial": 0.068359375,
+      "step": 209
+    },
+    {
+      "epoch": 0.10948140387637506,
+      "loss_partial": 0.064453125,
+      "step": 209
+    },
+    {
+      "epoch": 0.11000523834468308,
+      "grad_norm": 0.2950975000858307,
+      "learning_rate": 5.471204188481676e-06,
+      "loss": 0.1298,
+      "step": 210
+    },
+    {
+      "epoch": 0.11000523834468308,
+      "loss_partial": 0.056884765625,
+      "step": 210
+    },
+    {
+      "epoch": 0.11000523834468308,
+      "loss_partial": 0.064453125,
+      "step": 210
+    },
+    {
+      "epoch": 0.1105290728129911,
+      "grad_norm": 0.26776283979415894,
+      "learning_rate": 5.49738219895288e-06,
+      "loss": 0.1275,
+      "step": 211
+    },
+    {
+      "epoch": 0.1105290728129911,
+      "loss_partial": 0.06396484375,
+      "step": 211
+    },
+    {
+      "epoch": 0.1105290728129911,
+      "loss_partial": 0.06005859375,
+      "step": 211
+    },
+    {
+      "epoch": 0.11105290728129912,
+      "grad_norm": 0.2805287539958954,
+      "learning_rate": 5.523560209424085e-06,
+      "loss": 0.1311,
+      "step": 212
+    },
+    {
+      "epoch": 0.11105290728129912,
+      "loss_partial": 0.06787109375,
+      "step": 212
+    },
+    {
+      "epoch": 0.11105290728129912,
+      "loss_partial": 0.07568359375,
+      "step": 212
+    },
+    {
+      "epoch": 0.11157674174960712,
+      "grad_norm": 0.2672548294067383,
+      "learning_rate": 5.549738219895289e-06,
+      "loss": 0.1308,
+      "step": 213
+    },
+    {
+      "epoch": 0.11157674174960712,
+      "loss_partial": 0.0673828125,
+      "step": 213
+    },
+    {
+      "epoch": 0.11157674174960712,
+      "loss_partial": 0.07080078125,
+      "step": 213
+    },
+    {
+      "epoch": 0.11210057621791514,
+      "grad_norm": 0.2751561999320984,
+      "learning_rate": 5.575916230366493e-06,
+      "loss": 0.1301,
+      "step": 214
+    },
+    {
+      "epoch": 0.11210057621791514,
+      "loss_partial": 0.06982421875,
+      "step": 214
+    },
+    {
+      "epoch": 0.11210057621791514,
+      "loss_partial": 0.06298828125,
+      "step": 214
+    },
+    {
+      "epoch": 0.11262441068622316,
+      "grad_norm": 0.26281580328941345,
+      "learning_rate": 5.6020942408376965e-06,
+      "loss": 0.1294,
+      "step": 215
+    },
+    {
+      "epoch": 0.11262441068622316,
+      "loss_partial": 0.068359375,
+      "step": 215
+    },
+    {
+      "epoch": 0.11262441068622316,
+      "loss_partial": 0.06396484375,
+      "step": 215
+    },
+    {
+      "epoch": 0.11314824515453117,
+      "grad_norm": 0.2621370851993561,
+      "learning_rate": 5.6282722513089e-06,
+      "loss": 0.132,
+      "step": 216
+    },
+    {
+      "epoch": 0.11314824515453117,
+      "loss_partial": 0.06494140625,
+      "step": 216
+    },
+    {
+      "epoch": 0.11314824515453117,
+      "loss_partial": 0.0693359375,
+      "step": 216
+    },
+    {
+      "epoch": 0.11367207962283918,
+      "grad_norm": 0.24804644286632538,
+      "learning_rate": 5.654450261780105e-06,
+      "loss": 0.1248,
+      "step": 217
+    },
+    {
+      "epoch": 0.11367207962283918,
+      "loss_partial": 0.07275390625,
+      "step": 217
+    },
+    {
+      "epoch": 0.11367207962283918,
+      "loss_partial": 0.057861328125,
+      "step": 217
+    },
+    {
+      "epoch": 0.1141959140911472,
+      "grad_norm": 0.2556990087032318,
+      "learning_rate": 5.680628272251309e-06,
+      "loss": 0.1295,
+      "step": 218
+    },
+    {
+      "epoch": 0.1141959140911472,
+      "loss_partial": 0.06396484375,
+      "step": 218
+    },
+    {
+      "epoch": 0.1141959140911472,
+      "loss_partial": 0.05908203125,
+      "step": 218
+    },
+    {
+      "epoch": 0.11471974855945521,
+      "grad_norm": 0.2596417963504791,
+      "learning_rate": 5.706806282722514e-06,
+      "loss": 0.1281,
+      "step": 219
+    },
+    {
+      "epoch": 0.11471974855945521,
+      "loss_partial": 0.059326171875,
+      "step": 219
+    },
+    {
+      "epoch": 0.11471974855945521,
+      "loss_partial": 0.0849609375,
+      "step": 219
+    },
+    {
+      "epoch": 0.11524358302776323,
+      "grad_norm": 0.2656058967113495,
+      "learning_rate": 5.732984293193718e-06,
+      "loss": 0.1335,
+      "step": 220
+    },
+    {
+      "epoch": 0.11524358302776323,
+      "loss_partial": 0.072265625,
+      "step": 220
+    },
+    {
+      "epoch": 0.11524358302776323,
+      "loss_partial": 0.06396484375,
+      "step": 220
+    },
+    {
+      "epoch": 0.11576741749607124,
+      "grad_norm": 0.2639547884464264,
+      "learning_rate": 5.7591623036649226e-06,
+      "loss": 0.1304,
+      "step": 221
+    },
+    {
+      "epoch": 0.11576741749607124,
+      "loss_partial": 0.0673828125,
+      "step": 221
+    },
+    {
+      "epoch": 0.11576741749607124,
+      "loss_partial": 0.06591796875,
+      "step": 221
+    },
+    {
+      "epoch": 0.11629125196437926,
+      "grad_norm": 0.25407183170318604,
+      "learning_rate": 5.7853403141361265e-06,
+      "loss": 0.1294,
+      "step": 222
+    },
+    {
+      "epoch": 0.11629125196437926,
+      "loss_partial": 0.076171875,
+      "step": 222
+    },
+    {
+      "epoch": 0.11629125196437926,
+      "loss_partial": 0.060791015625,
+      "step": 222
+    },
+    {
+      "epoch": 0.11681508643268727,
+      "grad_norm": 0.23891626298427582,
+      "learning_rate": 5.81151832460733e-06,
+      "loss": 0.127,
+      "step": 223
+    },
+    {
+      "epoch": 0.11681508643268727,
+      "loss_partial": 0.06396484375,
+      "step": 223
+    },
+    {
+      "epoch": 0.11681508643268727,
+      "loss_partial": 0.05908203125,
+      "step": 223
+    },
+    {
+      "epoch": 0.11733892090099529,
+      "grad_norm": 0.2478952407836914,
+      "learning_rate": 5.837696335078534e-06,
+      "loss": 0.1311,
+      "step": 224
+    },
+    {
+      "epoch": 0.11733892090099529,
+      "loss_partial": 0.0595703125,
+      "step": 224
+    },
+    {
+      "epoch": 0.11733892090099529,
+      "loss_partial": 0.055419921875,
+      "step": 224
+    },
+    {
+      "epoch": 0.1178627553693033,
+      "grad_norm": 0.2414340078830719,
+      "learning_rate": 5.863874345549738e-06,
+      "loss": 0.1328,
+      "step": 225
+    },
+    {
+      "epoch": 0.1178627553693033,
+      "loss_partial": 0.06396484375,
+      "step": 225
+    },
+    {
+      "epoch": 0.1178627553693033,
+      "loss_partial": 0.06396484375,
+      "step": 225
+    },
+    {
+      "epoch": 0.11838658983761131,
+      "grad_norm": 0.23522719740867615,
+      "learning_rate": 5.890052356020943e-06,
+      "loss": 0.1313,
+      "step": 226
+    },
+    {
+      "epoch": 0.11838658983761131,
+      "loss_partial": 0.07275390625,
+      "step": 226
+    },
+    {
+      "epoch": 0.11838658983761131,
+      "loss_partial": 0.0849609375,
+      "step": 226
+    },
+    {
+      "epoch": 0.11891042430591933,
+      "grad_norm": 0.23278149962425232,
+      "learning_rate": 5.916230366492147e-06,
+      "loss": 0.1364,
+      "step": 227
+    },
+    {
+      "epoch": 0.11891042430591933,
+      "loss_partial": 0.064453125,
+      "step": 227
+    },
+    {
+      "epoch": 0.11891042430591933,
+      "loss_partial": 0.06396484375,
+      "step": 227
+    },
+    {
+      "epoch": 0.11943425877422735,
+      "grad_norm": 0.2235010266304016,
+      "learning_rate": 5.942408376963351e-06,
+      "loss": 0.1269,
+      "step": 228
+    },
+    {
+      "epoch": 0.11943425877422735,
+      "loss_partial": 0.05908203125,
+      "step": 228
+    },
+    {
+      "epoch": 0.11943425877422735,
+      "loss_partial": 0.059326171875,
+      "step": 228
+    },
+    {
+      "epoch": 0.11995809324253535,
+      "grad_norm": 0.21635471284389496,
+      "learning_rate": 5.968586387434555e-06,
+      "loss": 0.13,
+      "step": 229
+    },
+    {
+      "epoch": 0.11995809324253535,
+      "loss_partial": 0.0712890625,
+      "step": 229
+    },
+    {
+      "epoch": 0.11995809324253535,
+      "loss_partial": 0.072265625,
+      "step": 229
+    },
+    {
+      "epoch": 0.12048192771084337,
+      "grad_norm": 0.22030894458293915,
+      "learning_rate": 5.99476439790576e-06,
+      "loss": 0.1329,
+      "step": 230
+    },
+    {
+      "epoch": 0.12048192771084337,
+      "loss_partial": 0.068359375,
+      "step": 230
+    },
+    {
+      "epoch": 0.12048192771084337,
+      "loss_partial": 0.06494140625,
+      "step": 230
+    },
+    {
+      "epoch": 0.12100576217915139,
+      "grad_norm": 0.22704921662807465,
+      "learning_rate": 6.020942408376964e-06,
+      "loss": 0.1296,
+      "step": 231
+    },
+    {
+      "epoch": 0.12100576217915139,
+      "loss_partial": 0.056884765625,
+      "step": 231
+    },
+    {
+      "epoch": 0.12100576217915139,
+      "loss_partial": 0.06298828125,
+      "step": 231
+    },
+    {
+      "epoch": 0.12152959664745941,
+      "grad_norm": 0.21254359185695648,
+      "learning_rate": 6.047120418848168e-06,
+      "loss": 0.1237,
+      "step": 232
+    },
+    {
+      "epoch": 0.12152959664745941,
+      "loss_partial": 0.06298828125,
+      "step": 232
+    },
+    {
+      "epoch": 0.12152959664745941,
+      "loss_partial": 0.051025390625,
+      "step": 232
+    },
+    {
+      "epoch": 0.12205343111576741,
+      "grad_norm": 0.20440198481082916,
+      "learning_rate": 6.073298429319372e-06,
+      "loss": 0.1322,
+      "step": 233
+    },
+    {
+      "epoch": 0.12205343111576741,
+      "loss_partial": 0.0654296875,
+      "step": 233
+    },
+    {
+      "epoch": 0.12205343111576741,
+      "loss_partial": 0.05859375,
+      "step": 233
+    },
+    {
+      "epoch": 0.12257726558407543,
+      "grad_norm": 0.20669147372245789,
+      "learning_rate": 6.099476439790576e-06,
+      "loss": 0.1295,
+      "step": 234
+    },
+    {
+      "epoch": 0.12257726558407543,
+      "loss_partial": 0.0634765625,
+      "step": 234
+    },
+    {
+      "epoch": 0.12257726558407543,
+      "loss_partial": 0.068359375,
+      "step": 234
+    },
+    {
+      "epoch": 0.12310110005238345,
+      "grad_norm": 0.2108694463968277,
+      "learning_rate": 6.125654450261781e-06,
+      "loss": 0.1267,
+      "step": 235
+    },
+    {
+      "epoch": 0.12310110005238345,
+      "loss_partial": 0.056396484375,
+      "step": 235
+    },
+    {
+      "epoch": 0.12310110005238345,
+      "loss_partial": 0.064453125,
+      "step": 235
+    },
+    {
+      "epoch": 0.12362493452069147,
+      "grad_norm": 0.20883959531784058,
+      "learning_rate": 6.151832460732985e-06,
+      "loss": 0.1249,
+      "step": 236
+    },
+    {
+      "epoch": 0.12362493452069147,
+      "loss_partial": 0.05810546875,
+      "step": 236
+    },
+    {
+      "epoch": 0.12362493452069147,
+      "loss_partial": 0.0693359375,
+      "step": 236
+    },
+    {
+      "epoch": 0.12414876898899947,
+      "grad_norm": 0.2052280753850937,
+      "learning_rate": 6.178010471204189e-06,
+      "loss": 0.1277,
+      "step": 237
+    },
+    {
+      "epoch": 0.12414876898899947,
+      "loss_partial": 0.06591796875,
+      "step": 237
+    },
+    {
+      "epoch": 0.12414876898899947,
+      "loss_partial": 0.058837890625,
+      "step": 237
+    },
+    {
+      "epoch": 0.12467260345730749,
+      "grad_norm": 0.18247245252132416,
+      "learning_rate": 6.204188481675393e-06,
+      "loss": 0.1281,
+      "step": 238
+    },
+    {
+      "epoch": 0.12467260345730749,
+      "loss_partial": 0.072265625,
+      "step": 238
+    },
+    {
+      "epoch": 0.12467260345730749,
+      "loss_partial": 0.06787109375,
+      "step": 238
+    },
+    {
+      "epoch": 0.1251964379256155,
+      "grad_norm": 0.18815632164478302,
+      "learning_rate": 6.230366492146598e-06,
+      "loss": 0.1313,
+      "step": 239
+    },
+    {
+      "epoch": 0.1251964379256155,
+      "loss_partial": 0.06494140625,
+      "step": 239
+    },
+    {
+      "epoch": 0.1251964379256155,
+      "loss_partial": 0.06298828125,
+      "step": 239
+    },
+    {
+      "epoch": 0.12572027239392353,
+      "grad_norm": 0.18276932835578918,
+      "learning_rate": 6.256544502617802e-06,
+      "loss": 0.1313,
+      "step": 240
+    },
+    {
+      "epoch": 0.12572027239392353,
+      "loss_partial": 0.0654296875,
+      "step": 240
+    },
+    {
+      "epoch": 0.12572027239392353,
+      "loss_partial": 0.072265625,
+      "step": 240
+    },
+    {
+      "epoch": 0.12624410686223153,
+      "grad_norm": 0.18943659961223602,
+      "learning_rate": 6.282722513089006e-06,
+      "loss": 0.1266,
+      "step": 241
+    },
+    {
+      "epoch": 0.12624410686223153,
+      "loss_partial": 0.062255859375,
+      "step": 241
+    },
+    {
+      "epoch": 0.12624410686223153,
+      "loss_partial": 0.0634765625,
+      "step": 241
+    },
+    {
+      "epoch": 0.12676794133053956,
+      "grad_norm": 0.1848103553056717,
+      "learning_rate": 6.30890052356021e-06,
+      "loss": 0.1273,
+      "step": 242
+    },
+    {
+      "epoch": 0.12676794133053956,
+      "loss_partial": 0.07080078125,
+      "step": 242
+    },
+    {
+      "epoch": 0.12676794133053956,
+      "loss_partial": 0.06494140625,
+      "step": 242
+    },
+    {
+      "epoch": 0.12729177579884757,
+      "grad_norm": 0.17900794744491577,
+      "learning_rate": 6.335078534031414e-06,
+      "loss": 0.1277,
+      "step": 243
+    },
+    {
+      "epoch": 0.12729177579884757,
+      "loss_partial": 0.06298828125,
+      "step": 243
+    },
+    {
+      "epoch": 0.12729177579884757,
+      "loss_partial": 0.06298828125,
+      "step": 243
+    },
+    {
+      "epoch": 0.12781561026715557,
+      "grad_norm": 0.16777996718883514,
+      "learning_rate": 6.361256544502619e-06,
+      "loss": 0.131,
+      "step": 244
+    },
+    {
+      "epoch": 0.12781561026715557,
+      "loss_partial": 0.0595703125,
+      "step": 244
+    },
+    {
+      "epoch": 0.12781561026715557,
+      "loss_partial": 0.060546875,
+      "step": 244
+    },
+    {
+      "epoch": 0.1283394447354636,
+      "grad_norm": 0.1715957224369049,
+      "learning_rate": 6.3874345549738226e-06,
+      "loss": 0.1287,
+      "step": 245
+    },
+    {
+      "epoch": 0.1283394447354636,
+      "loss_partial": 0.0673828125,
+      "step": 245
+    },
+    {
+      "epoch": 0.1283394447354636,
+      "loss_partial": 0.0693359375,
+      "step": 245
+    },
+    {
+      "epoch": 0.1288632792037716,
+      "grad_norm": 0.16495846211910248,
+      "learning_rate": 6.4136125654450265e-06,
+      "loss": 0.1258,
+      "step": 246
+    },
+    {
+      "epoch": 0.1288632792037716,
+      "loss_partial": 0.061767578125,
+      "step": 246
+    },
+    {
+      "epoch": 0.1288632792037716,
+      "loss_partial": 0.060302734375,
+      "step": 246
+    },
+    {
+      "epoch": 0.1293871136720796,
+      "grad_norm": 0.16605092585086823,
+      "learning_rate": 6.43979057591623e-06,
+      "loss": 0.1242,
+      "step": 247
+    },
+    {
+      "epoch": 0.1293871136720796,
+      "loss_partial": 0.06396484375,
+      "step": 247
+    },
+    {
+      "epoch": 0.1293871136720796,
+      "loss_partial": 0.072265625,
+      "step": 247
+    },
+    {
+      "epoch": 0.12991094814038764,
+      "grad_norm": 0.17033381760120392,
+      "learning_rate": 6.465968586387435e-06,
+      "loss": 0.1285,
+      "step": 248
+    },
+    {
+      "epoch": 0.12991094814038764,
+      "loss_partial": 0.06298828125,
+      "step": 248
+    },
+    {
+      "epoch": 0.12991094814038764,
+      "loss_partial": 0.06787109375,
+      "step": 248
+    },
+    {
+      "epoch": 0.13043478260869565,
+      "grad_norm": 0.15971876680850983,
+      "learning_rate": 6.492146596858639e-06,
+      "loss": 0.1281,
+      "step": 249
+    },
+    {
+      "epoch": 0.13043478260869565,
+      "loss_partial": 0.0693359375,
+      "step": 249
+    },
+    {
+      "epoch": 0.13043478260869565,
+      "loss_partial": 0.06494140625,
+      "step": 249
+    },
+    {
+      "epoch": 0.13095861707700368,
+      "grad_norm": 0.16109918057918549,
+      "learning_rate": 6.518324607329843e-06,
+      "loss": 0.126,
+      "step": 250
+    },
+    {
+      "epoch": 0.13095861707700368,
+      "loss_partial": 0.06494140625,
+      "step": 250
+    },
+    {
+      "epoch": 0.13095861707700368,
+      "loss_partial": 0.06787109375,
+      "step": 250
+    },
+    {
+      "epoch": 0.13148245154531168,
+      "grad_norm": 0.15467949211597443,
+      "learning_rate": 6.544502617801047e-06,
+      "loss": 0.127,
+      "step": 251
+    },
+    {
+      "epoch": 0.13148245154531168,
+      "loss_partial": 0.0654296875,
+      "step": 251
+    },
+    {
+      "epoch": 0.13148245154531168,
+      "loss_partial": 0.05712890625,
+      "step": 251
+    },
+    {
+      "epoch": 0.1320062860136197,
+      "grad_norm": 0.1527758538722992,
+      "learning_rate": 6.5706806282722526e-06,
+      "loss": 0.1284,
+      "step": 252
+    },
+    {
+      "epoch": 0.1320062860136197,
+      "loss_partial": 0.060546875,
+      "step": 252
+    },
+    {
+      "epoch": 0.1320062860136197,
+      "loss_partial": 0.06298828125,
+      "step": 252
+    },
+    {
+      "epoch": 0.13253012048192772,
+      "grad_norm": 0.1509239226579666,
+      "learning_rate": 6.5968586387434565e-06,
+      "loss": 0.1267,
+      "step": 253
+    },
+    {
+      "epoch": 0.13253012048192772,
+      "loss_partial": 0.06103515625,
+      "step": 253
+    },
+    {
+      "epoch": 0.13253012048192772,
+      "loss_partial": 0.06494140625,
+      "step": 253
+    },
+    {
+      "epoch": 0.13305395495023573,
+      "grad_norm": 0.1574198603630066,
+      "learning_rate": 6.62303664921466e-06,
+      "loss": 0.1285,
+      "step": 254
+    },
+    {
+      "epoch": 0.13305395495023573,
+      "loss_partial": 0.05810546875,
+      "step": 254
+    },
+    {
+      "epoch": 0.13305395495023573,
+      "loss_partial": 0.06591796875,
+      "step": 254
+    },
+    {
+      "epoch": 0.13357778941854373,
+      "grad_norm": 0.15486107766628265,
+      "learning_rate": 6.649214659685864e-06,
+      "loss": 0.13,
+      "step": 255
+    },
+    {
+      "epoch": 0.13357778941854373,
+      "loss_partial": 0.0693359375,
+      "step": 255
+    },
+    {
+      "epoch": 0.13357778941854373,
+      "loss_partial": 0.06640625,
+      "step": 255
+    },
+    {
+      "epoch": 0.13410162388685176,
+      "grad_norm": 0.15332625806331635,
+      "learning_rate": 6.675392670157068e-06,
+      "loss": 0.1301,
+      "step": 256
+    },
+    {
+      "epoch": 0.13410162388685176,
+      "loss_partial": 0.055908203125,
+      "step": 256
+    },
+    {
+      "epoch": 0.13410162388685176,
+      "loss_partial": 0.068359375,
+      "step": 256
+    },
+    {
+      "epoch": 0.13462545835515977,
+      "grad_norm": 0.146895632147789,
+      "learning_rate": 6.701570680628273e-06,
+      "loss": 0.1266,
+      "step": 257
+    },
+    {
+      "epoch": 0.13462545835515977,
+      "loss_partial": 0.06787109375,
+      "step": 257
+    },
+    {
+      "epoch": 0.13462545835515977,
+      "loss_partial": 0.064453125,
+      "step": 257
+    },
+    {
+      "epoch": 0.13514929282346777,
+      "grad_norm": 0.13986945152282715,
+      "learning_rate": 6.727748691099477e-06,
+      "loss": 0.1263,
+      "step": 258
+    },
+    {
+      "epoch": 0.13514929282346777,
+      "loss_partial": 0.054931640625,
+      "step": 258
+    },
+    {
+      "epoch": 0.13514929282346777,
+      "loss_partial": 0.058837890625,
+      "step": 258
+    },
+    {
+      "epoch": 0.1356731272917758,
+      "grad_norm": 0.1384933739900589,
+      "learning_rate": 6.753926701570681e-06,
+      "loss": 0.1284,
+      "step": 259
+    },
+    {
+      "epoch": 0.1356731272917758,
+      "loss_partial": 0.0625,
+      "step": 259
+    },
+    {
+      "epoch": 0.1356731272917758,
+      "loss_partial": 0.064453125,
+      "step": 259
+    },
+    {
+      "epoch": 0.1361969617600838,
+      "grad_norm": 0.14610351622104645,
+      "learning_rate": 6.780104712041885e-06,
+      "loss": 0.1274,
+      "step": 260
+    },
+    {
+      "epoch": 0.1361969617600838,
+      "loss_partial": 0.05908203125,
+      "step": 260
+    },
+    {
+      "epoch": 0.1361969617600838,
+      "loss_partial": 0.07177734375,
+      "step": 260
+    },
+    {
+      "epoch": 0.13672079622839184,
+      "grad_norm": 0.14630573987960815,
+      "learning_rate": 6.80628272251309e-06,
+      "loss": 0.1261,
+      "step": 261
+    },
+    {
+      "epoch": 0.13672079622839184,
+      "loss_partial": 0.06103515625,
+      "step": 261
+    },
+    {
+      "epoch": 0.13672079622839184,
+      "loss_partial": 0.07177734375,
+      "step": 261
+    },
+    {
+      "epoch": 0.13724463069669984,
+      "grad_norm": 0.13782291114330292,
+      "learning_rate": 6.832460732984294e-06,
+      "loss": 0.1252,
+      "step": 262
+    },
+    {
+      "epoch": 0.13724463069669984,
+      "loss_partial": 0.0615234375,
+      "step": 262
+    },
+    {
+      "epoch": 0.13724463069669984,
+      "loss_partial": 0.0556640625,
+      "step": 262
+    },
+    {
+      "epoch": 0.13776846516500785,
+      "grad_norm": 0.15570396184921265,
+      "learning_rate": 6.858638743455498e-06,
+      "loss": 0.1267,
+      "step": 263
+    },
+    {
+      "epoch": 0.13776846516500785,
+      "loss_partial": 0.06591796875,
+      "step": 263
+    },
+    {
+      "epoch": 0.13776846516500785,
+      "loss_partial": 0.061767578125,
+      "step": 263
+    },
+    {
+      "epoch": 0.13829229963331588,
+      "grad_norm": 0.13707728683948517,
+      "learning_rate": 6.884816753926702e-06,
+      "loss": 0.1284,
+      "step": 264
+    },
+    {
+      "epoch": 0.13829229963331588,
+      "loss_partial": 0.068359375,
+      "step": 264
+    },
+    {
+      "epoch": 0.13829229963331588,
+      "loss_partial": 0.06640625,
+      "step": 264
+    },
+    {
+      "epoch": 0.13881613410162388,
+      "grad_norm": 0.14523907005786896,
+      "learning_rate": 6.910994764397906e-06,
+      "loss": 0.1303,
+      "step": 265
+    },
+    {
+      "epoch": 0.13881613410162388,
+      "loss_partial": 0.06396484375,
+      "step": 265
+    },
+    {
+      "epoch": 0.13881613410162388,
+      "loss_partial": 0.0634765625,
+      "step": 265
+    },
+    {
+      "epoch": 0.1393399685699319,
+      "grad_norm": 0.14172007143497467,
+      "learning_rate": 6.937172774869111e-06,
+      "loss": 0.1259,
+      "step": 266
+    },
+    {
+      "epoch": 0.1393399685699319,
+      "loss_partial": 0.05419921875,
+      "step": 266
+    },
+    {
+      "epoch": 0.1393399685699319,
+      "loss_partial": 0.0615234375,
+      "step": 266
+    },
+    {
+      "epoch": 0.13986380303823992,
+      "grad_norm": 0.13336606323719025,
+      "learning_rate": 6.963350785340315e-06,
+      "loss": 0.1244,
+      "step": 267
+    },
+    {
+      "epoch": 0.13986380303823992,
+      "loss_partial": 0.06103515625,
+      "step": 267
+    },
+    {
+      "epoch": 0.13986380303823992,
+      "loss_partial": 0.054443359375,
+      "step": 267
+    },
+    {
+      "epoch": 0.14038763750654792,
+      "grad_norm": 0.1307591050863266,
+      "learning_rate": 6.989528795811519e-06,
+      "loss": 0.1281,
+      "step": 268
+    },
+    {
+      "epoch": 0.14038763750654792,
+      "loss_partial": 0.06298828125,
+      "step": 268
+    },
+    {
+      "epoch": 0.14038763750654792,
+      "loss_partial": 0.059814453125,
+      "step": 268
+    },
+    {
+      "epoch": 0.14091147197485596,
+      "grad_norm": 0.13535946607589722,
+      "learning_rate": 7.015706806282723e-06,
+      "loss": 0.1276,
+      "step": 269
+    },
+    {
+      "epoch": 0.14091147197485596,
+      "loss_partial": 0.0771484375,
+      "step": 269
+    },
+    {
+      "epoch": 0.14091147197485596,
+      "loss_partial": 0.0537109375,
+      "step": 269
+    },
+    {
+      "epoch": 0.14143530644316396,
+      "grad_norm": 0.12640510499477386,
+      "learning_rate": 7.041884816753927e-06,
+      "loss": 0.1276,
+      "step": 270
+    },
+    {
+      "epoch": 0.14143530644316396,
+      "loss_partial": 0.05517578125,
+      "step": 270
+    },
+    {
+      "epoch": 0.14143530644316396,
+      "loss_partial": 0.06201171875,
+      "step": 270
+    },
+    {
+      "epoch": 0.14195914091147196,
+      "grad_norm": 0.13004356622695923,
+      "learning_rate": 7.068062827225132e-06,
+      "loss": 0.125,
+      "step": 271
+    },
+    {
+      "epoch": 0.14195914091147196,
+      "loss_partial": 0.06298828125,
+      "step": 271
+    },
+    {
+      "epoch": 0.14195914091147196,
+      "loss_partial": 0.0693359375,
+      "step": 271
+    },
+    {
+      "epoch": 0.14248297537978,
+      "grad_norm": 0.12260151654481888,
+      "learning_rate": 7.094240837696336e-06,
+      "loss": 0.1256,
+      "step": 272
+    },
+    {
+      "epoch": 0.14248297537978,
+      "loss_partial": 0.06787109375,
+      "step": 272
+    },
+    {
+      "epoch": 0.14248297537978,
+      "loss_partial": 0.0712890625,
+      "step": 272
+    },
+    {
+      "epoch": 0.143006809848088,
+      "grad_norm": 0.12187730520963669,
+      "learning_rate": 7.12041884816754e-06,
+      "loss": 0.1313,
+      "step": 273
+    },
+    {
+      "epoch": 0.143006809848088,
+      "loss_partial": 0.057373046875,
+      "step": 273
+    },
+    {
+      "epoch": 0.143006809848088,
+      "loss_partial": 0.055419921875,
+      "step": 273
+    },
+    {
+      "epoch": 0.143530644316396,
+      "grad_norm": 0.12773172557353973,
+      "learning_rate": 7.146596858638744e-06,
+      "loss": 0.1211,
+      "step": 274
+    },
+    {
+      "epoch": 0.143530644316396,
+      "loss_partial": 0.0615234375,
+      "step": 274
+    },
+    {
+      "epoch": 0.143530644316396,
+      "loss_partial": 0.059814453125,
+      "step": 274
+    },
+    {
+      "epoch": 0.14405447878470404,
+      "grad_norm": 0.1362551897764206,
+      "learning_rate": 7.172774869109949e-06,
+      "loss": 0.1265,
+      "step": 275
+    },
+    {
+      "epoch": 0.14405447878470404,
+      "loss_partial": 0.053955078125,
+      "step": 275
+    },
+    {
+      "epoch": 0.14405447878470404,
+      "loss_partial": 0.06298828125,
+      "step": 275
+    },
+    {
+      "epoch": 0.14457831325301204,
+      "grad_norm": 0.11896193772554398,
+      "learning_rate": 7.1989528795811526e-06,
+      "loss": 0.1255,
+      "step": 276
+    },
+    {
+      "epoch": 0.14457831325301204,
+      "loss_partial": 0.064453125,
+      "step": 276
+    },
+    {
+      "epoch": 0.14457831325301204,
+      "loss_partial": 0.06689453125,
+      "step": 276
+    },
+    {
+      "epoch": 0.14510214772132007,
+      "grad_norm": 0.12356580048799515,
+      "learning_rate": 7.2251308900523565e-06,
+      "loss": 0.1258,
+      "step": 277
+    },
+    {
+      "epoch": 0.14510214772132007,
+      "loss_partial": 0.052978515625,
+      "step": 277
+    },
+    {
+      "epoch": 0.14510214772132007,
+      "loss_partial": 0.05615234375,
+      "step": 277
+    },
+    {
+      "epoch": 0.14562598218962808,
+      "grad_norm": 0.12001247704029083,
+      "learning_rate": 7.25130890052356e-06,
+      "loss": 0.1253,
+      "step": 278
+    },
+    {
+      "epoch": 0.14562598218962808,
+      "loss_partial": 0.0673828125,
+      "step": 278
+    },
+    {
+      "epoch": 0.14562598218962808,
+      "loss_partial": 0.060546875,
+      "step": 278
+    },
+    {
+      "epoch": 0.14614981665793608,
+      "grad_norm": 0.12279926985502243,
+      "learning_rate": 7.277486910994765e-06,
+      "loss": 0.1259,
+      "step": 279
+    },
+    {
+      "epoch": 0.14614981665793608,
+      "loss_partial": 0.0791015625,
+      "step": 279
+    },
+    {
+      "epoch": 0.14614981665793608,
+      "loss_partial": 0.060302734375,
+      "step": 279
+    },
+    {
+      "epoch": 0.14667365112624411,
+      "grad_norm": 0.12339012324810028,
+      "learning_rate": 7.303664921465969e-06,
+      "loss": 0.1274,
+      "step": 280
+    },
+    {
+      "epoch": 0.14667365112624411,
+      "loss_partial": 0.0693359375,
+      "step": 280
+    },
+    {
+      "epoch": 0.14667365112624411,
+      "loss_partial": 0.06298828125,
+      "step": 280
+    },
+    {
+      "epoch": 0.14719748559455212,
+      "grad_norm": 0.11625789105892181,
+      "learning_rate": 7.329842931937173e-06,
+      "loss": 0.1282,
+      "step": 281
+    },
+    {
+      "epoch": 0.14719748559455212,
+      "loss_partial": 0.058349609375,
+      "step": 281
+    },
+    {
+      "epoch": 0.14719748559455212,
+      "loss_partial": 0.061279296875,
+      "step": 281
+    },
+    {
+      "epoch": 0.14772132006286012,
+      "grad_norm": 0.1174723282456398,
+      "learning_rate": 7.356020942408377e-06,
+      "loss": 0.1247,
+      "step": 282
+    },
+    {
+      "epoch": 0.14772132006286012,
+      "loss_partial": 0.05908203125,
+      "step": 282
+    },
+    {
+      "epoch": 0.14772132006286012,
+      "loss_partial": 0.061279296875,
+      "step": 282
+    },
+    {
+      "epoch": 0.14824515453116816,
+      "grad_norm": 0.1141645684838295,
+      "learning_rate": 7.382198952879581e-06,
+      "loss": 0.1232,
+      "step": 283
+    },
+    {
+      "epoch": 0.14824515453116816,
+      "loss_partial": 0.07861328125,
+      "step": 283
+    },
+    {
+      "epoch": 0.14824515453116816,
+      "loss_partial": 0.06298828125,
+      "step": 283
+    },
+    {
+      "epoch": 0.14876898899947616,
+      "grad_norm": 0.12124475091695786,
+      "learning_rate": 7.4083769633507865e-06,
+      "loss": 0.1297,
+      "step": 284
+    },
+    {
+      "epoch": 0.14876898899947616,
+      "loss_partial": 0.05126953125,
+      "step": 284
+    },
+    {
+      "epoch": 0.14876898899947616,
+      "loss_partial": 0.061279296875,
+      "step": 284
+    },
+    {
+      "epoch": 0.1492928234677842,
+      "grad_norm": 0.10918641090393066,
+      "learning_rate": 7.43455497382199e-06,
+      "loss": 0.1299,
+      "step": 285
+    },
+    {
+      "epoch": 0.1492928234677842,
+      "loss_partial": 0.0703125,
+      "step": 285
+    },
+    {
+      "epoch": 0.1492928234677842,
+      "loss_partial": 0.0693359375,
+      "step": 285
+    },
+    {
+      "epoch": 0.1498166579360922,
+      "grad_norm": 0.10577938705682755,
+      "learning_rate": 7.460732984293194e-06,
+      "loss": 0.1254,
+      "step": 286
+    },
+    {
+      "epoch": 0.1498166579360922,
+      "loss_partial": 0.0849609375,
+      "step": 286
+    },
+    {
+      "epoch": 0.1498166579360922,
+      "loss_partial": 0.0595703125,
+      "step": 286
+    },
+    {
+      "epoch": 0.1503404924044002,
+      "grad_norm": 0.10764198750257492,
+      "learning_rate": 7.486910994764398e-06,
+      "loss": 0.1272,
+      "step": 287
+    },
+    {
+      "epoch": 0.1503404924044002,
+      "loss_partial": 0.060302734375,
+      "step": 287
+    },
+    {
+      "epoch": 0.1503404924044002,
+      "loss_partial": 0.05615234375,
+      "step": 287
+    },
+    {
+      "epoch": 0.15086432687270823,
+      "grad_norm": 0.10993047058582306,
+      "learning_rate": 7.513089005235603e-06,
+      "loss": 0.1221,
+      "step": 288
+    },
+    {
+      "epoch": 0.15086432687270823,
+      "loss_partial": 0.052490234375,
+      "step": 288
+    },
+    {
+      "epoch": 0.15086432687270823,
+      "loss_partial": 0.062255859375,
+      "step": 288
+    },
+    {
+      "epoch": 0.15138816134101624,
+      "grad_norm": 0.1134459376335144,
+      "learning_rate": 7.539267015706807e-06,
+      "loss": 0.1225,
+      "step": 289
+    },
+    {
+      "epoch": 0.15138816134101624,
+      "loss_partial": 0.06298828125,
+      "step": 289
+    },
+    {
+      "epoch": 0.15138816134101624,
+      "loss_partial": 0.051513671875,
+      "step": 289
+    },
+    {
+      "epoch": 0.15191199580932424,
+      "grad_norm": 0.10504646599292755,
+      "learning_rate": 7.565445026178011e-06,
+      "loss": 0.1269,
+      "step": 290
+    },
+    {
+      "epoch": 0.15191199580932424,
+      "loss_partial": 0.0654296875,
+      "step": 290
+    },
+    {
+      "epoch": 0.15191199580932424,
+      "loss_partial": 0.0556640625,
+      "step": 290
+    },
+    {
+      "epoch": 0.15243583027763227,
+      "grad_norm": 0.11143013089895248,
+      "learning_rate": 7.591623036649215e-06,
+      "loss": 0.1291,
+      "step": 291
+    },
+    {
+      "epoch": 0.15243583027763227,
+      "loss_partial": 0.0654296875,
+      "step": 291
+    },
+    {
+      "epoch": 0.15243583027763227,
+      "loss_partial": 0.06689453125,
+      "step": 291
+    },
+    {
+      "epoch": 0.15295966474594028,
+      "grad_norm": 0.1072806715965271,
+      "learning_rate": 7.61780104712042e-06,
+      "loss": 0.123,
+      "step": 292
+    },
+    {
+      "epoch": 0.15295966474594028,
+      "loss_partial": 0.06640625,
+      "step": 292
+    },
+    {
+      "epoch": 0.15295966474594028,
+      "loss_partial": 0.062255859375,
+      "step": 292
+    },
+    {
+      "epoch": 0.1534834992142483,
+      "grad_norm": 0.10931787639856339,
+      "learning_rate": 7.643979057591624e-06,
+      "loss": 0.1246,
+      "step": 293
+    },
+    {
+      "epoch": 0.1534834992142483,
+      "loss_partial": 0.0546875,
+      "step": 293
+    },
+    {
+      "epoch": 0.1534834992142483,
+      "loss_partial": 0.0595703125,
+      "step": 293
+    },
+    {
+      "epoch": 0.1540073336825563,
+      "grad_norm": 0.11012646555900574,
+      "learning_rate": 7.670157068062828e-06,
+      "loss": 0.1284,
+      "step": 294
+    },
+    {
+      "epoch": 0.1540073336825563,
+      "loss_partial": 0.053466796875,
+      "step": 294
+    },
+    {
+      "epoch": 0.1540073336825563,
+      "loss_partial": 0.064453125,
+      "step": 294
+    },
+    {
+      "epoch": 0.15453116815086432,
+      "grad_norm": 0.10471461713314056,
+      "learning_rate": 7.696335078534032e-06,
+      "loss": 0.1249,
+      "step": 295
+    },
+    {
+      "epoch": 0.15453116815086432,
+      "loss_partial": 0.05517578125,
+      "step": 295
+    },
+    {
+      "epoch": 0.15453116815086432,
+      "loss_partial": 0.0654296875,
+      "step": 295
+    },
+    {
+      "epoch": 0.15505500261917235,
+      "grad_norm": 0.10276065766811371,
+      "learning_rate": 7.722513089005236e-06,
+      "loss": 0.1254,
+      "step": 296
+    },
+    {
+      "epoch": 0.15505500261917235,
+      "loss_partial": 0.0634765625,
+      "step": 296
+    },
+    {
+      "epoch": 0.15505500261917235,
+      "loss_partial": 0.051025390625,
+      "step": 296
+    },
+    {
+      "epoch": 0.15557883708748035,
+      "grad_norm": 0.10413376241922379,
+      "learning_rate": 7.748691099476442e-06,
+      "loss": 0.125,
+      "step": 297
+    },
+    {
+      "epoch": 0.15557883708748035,
+      "loss_partial": 0.053466796875,
+      "step": 297
+    },
+    {
+      "epoch": 0.15557883708748035,
+      "loss_partial": 0.06396484375,
+      "step": 297
+    },
+    {
+      "epoch": 0.15610267155578836,
+      "grad_norm": 0.1039932370185852,
+      "learning_rate": 7.774869109947646e-06,
+      "loss": 0.1241,
+      "step": 298
+    },
+    {
+      "epoch": 0.15610267155578836,
+      "loss_partial": 0.06298828125,
+      "step": 298
+    },
+    {
+      "epoch": 0.15610267155578836,
+      "loss_partial": 0.0625,
+      "step": 298
+    },
+    {
+      "epoch": 0.1566265060240964,
+      "grad_norm": 0.10825380682945251,
+      "learning_rate": 7.80104712041885e-06,
+      "loss": 0.1263,
+      "step": 299
+    },
+    {
+      "epoch": 0.1566265060240964,
+      "loss_partial": 0.06396484375,
+      "step": 299
+    },
+    {
+      "epoch": 0.1566265060240964,
+      "loss_partial": 0.0595703125,
+      "step": 299
+    },
+    {
+      "epoch": 0.1571503404924044,
+      "grad_norm": 0.10472093522548676,
+      "learning_rate": 7.827225130890053e-06,
+      "loss": 0.1282,
+      "step": 300
+    },
+    {
+      "epoch": 0.1571503404924044,
+      "loss_partial": 0.059326171875,
+      "step": 300
+    },
+    {
+      "epoch": 0.1571503404924044,
+      "loss_partial": 0.059814453125,
+      "step": 300
+    },
+    {
+      "epoch": 0.15767417496071243,
+      "grad_norm": 0.11641892790794373,
+      "learning_rate": 7.853403141361257e-06,
+      "loss": 0.123,
+      "step": 301
+    },
+    {
+      "epoch": 0.15767417496071243,
+      "loss_partial": 0.060791015625,
+      "step": 301
+    },
+    {
+      "epoch": 0.15767417496071243,
+      "loss_partial": 0.06591796875,
+      "step": 301
+    },
+    {
+      "epoch": 0.15819800942902043,
+      "grad_norm": 0.10560357570648193,
+      "learning_rate": 7.879581151832461e-06,
+      "loss": 0.1253,
+      "step": 302
+    },
+    {
+      "epoch": 0.15819800942902043,
+      "loss_partial": 0.0625,
+      "step": 302
+    },
+    {
+      "epoch": 0.15819800942902043,
+      "loss_partial": 0.0546875,
+      "step": 302
+    },
+    {
+      "epoch": 0.15872184389732844,
+      "grad_norm": 0.09606872498989105,
+      "learning_rate": 7.905759162303665e-06,
+      "loss": 0.1227,
+      "step": 303
+    },
+    {
+      "epoch": 0.15872184389732844,
+      "loss_partial": 0.07421875,
+      "step": 303
+    },
+    {
+      "epoch": 0.15872184389732844,
+      "loss_partial": 0.06298828125,
+      "step": 303
+    },
+    {
+      "epoch": 0.15924567836563647,
+      "grad_norm": 0.10215487331151962,
+      "learning_rate": 7.931937172774869e-06,
+      "loss": 0.126,
+      "step": 304
+    },
+    {
+      "epoch": 0.15924567836563647,
+      "loss_partial": 0.07080078125,
+      "step": 304
+    },
+    {
+      "epoch": 0.15924567836563647,
+      "loss_partial": 0.06494140625,
+      "step": 304
+    },
+    {
+      "epoch": 0.15976951283394447,
+      "grad_norm": 0.09981030225753784,
+      "learning_rate": 7.958115183246073e-06,
+      "loss": 0.1278,
+      "step": 305
+    },
+    {
+      "epoch": 0.15976951283394447,
+      "loss_partial": 0.06396484375,
+      "step": 305
+    },
+    {
+      "epoch": 0.15976951283394447,
+      "loss_partial": 0.0625,
+      "step": 305
+    },
+    {
+      "epoch": 0.16029334730225248,
+      "grad_norm": 0.09993433952331543,
+      "learning_rate": 7.984293193717279e-06,
+      "loss": 0.1247,
+      "step": 306
+    },
+    {
+      "epoch": 0.16029334730225248,
+      "loss_partial": 0.0625,
+      "step": 306
+    },
+    {
+      "epoch": 0.16029334730225248,
+      "loss_partial": 0.057373046875,
+      "step": 306
+    },
+    {
+      "epoch": 0.1608171817705605,
+      "grad_norm": 0.09227127581834793,
+      "learning_rate": 8.010471204188483e-06,
+      "loss": 0.1221,
+      "step": 307
+    },
+    {
+      "epoch": 0.1608171817705605,
+      "loss_partial": 0.06494140625,
+      "step": 307
+    },
+    {
+      "epoch": 0.1608171817705605,
+      "loss_partial": 0.05712890625,
+      "step": 307
+    },
+    {
+      "epoch": 0.1613410162388685,
+      "grad_norm": 0.10311154276132584,
+      "learning_rate": 8.036649214659686e-06,
+      "loss": 0.1231,
+      "step": 308
+    },
+    {
+      "epoch": 0.1613410162388685,
+      "loss_partial": 0.05322265625,
+      "step": 308
+    },
+    {
+      "epoch": 0.1613410162388685,
+      "loss_partial": 0.0634765625,
+      "step": 308
+    },
+    {
+      "epoch": 0.16186485070717654,
+      "grad_norm": 0.10721491277217865,
+      "learning_rate": 8.06282722513089e-06,
+      "loss": 0.1216,
+      "step": 309
+    },
+    {
+      "epoch": 0.16186485070717654,
+      "loss_partial": 0.0654296875,
+      "step": 309
+    },
+    {
+      "epoch": 0.16186485070717654,
+      "loss_partial": 0.059326171875,
+      "step": 309
+    },
+    {
+      "epoch": 0.16238868517548455,
+      "grad_norm": 0.10199577361345291,
+      "learning_rate": 8.089005235602096e-06,
+      "loss": 0.1258,
+      "step": 310
+    },
+    {
+      "epoch": 0.16238868517548455,
+      "loss_partial": 0.06884765625,
+      "step": 310
+    },
+    {
+      "epoch": 0.16238868517548455,
+      "loss_partial": 0.0732421875,
+      "step": 310
+    },
+    {
+      "epoch": 0.16291251964379255,
+      "grad_norm": 0.09550805389881134,
+      "learning_rate": 8.1151832460733e-06,
+      "loss": 0.1263,
+      "step": 311
+    },
+    {
+      "epoch": 0.16291251964379255,
+      "loss_partial": 0.06103515625,
+      "step": 311
+    },
+    {
+      "epoch": 0.16291251964379255,
+      "loss_partial": 0.062255859375,
+      "step": 311
+    },
+    {
+      "epoch": 0.16343635411210058,
+      "grad_norm": 0.09088976681232452,
+      "learning_rate": 8.141361256544504e-06,
+      "loss": 0.1264,
+      "step": 312
+    },
+    {
+      "epoch": 0.16343635411210058,
+      "loss_partial": 0.064453125,
+      "step": 312
+    },
+    {
+      "epoch": 0.16343635411210058,
+      "loss_partial": 0.0625,
+      "step": 312
+    },
+    {
+      "epoch": 0.1639601885804086,
+      "grad_norm": 0.09322135150432587,
+      "learning_rate": 8.167539267015708e-06,
+      "loss": 0.1219,
+      "step": 313
+    },
+    {
+      "epoch": 0.1639601885804086,
+      "loss_partial": 0.05859375,
+      "step": 313
+    },
+    {
+      "epoch": 0.1639601885804086,
+      "loss_partial": 0.057373046875,
+      "step": 313
+    },
+    {
+      "epoch": 0.1644840230487166,
+      "grad_norm": 0.11350459605455399,
+      "learning_rate": 8.193717277486912e-06,
+      "loss": 0.123,
+      "step": 314
+    },
+    {
+      "epoch": 0.1644840230487166,
+      "loss_partial": 0.055908203125,
+      "step": 314
+    },
+    {
+      "epoch": 0.1644840230487166,
+      "loss_partial": 0.06689453125,
+      "step": 314
+    },
+    {
+      "epoch": 0.16500785751702463,
+      "grad_norm": 0.08779610693454742,
+      "learning_rate": 8.219895287958116e-06,
+      "loss": 0.1245,
+      "step": 315
+    },
+    {
+      "epoch": 0.16500785751702463,
+      "loss_partial": 0.05859375,
+      "step": 315
+    },
+    {
+      "epoch": 0.16500785751702463,
+      "loss_partial": 0.05224609375,
+      "step": 315
+    },
+    {
+      "epoch": 0.16553169198533263,
+      "grad_norm": 0.10314323008060455,
+      "learning_rate": 8.24607329842932e-06,
+      "loss": 0.1241,
+      "step": 316
+    },
+    {
+      "epoch": 0.16553169198533263,
+      "loss_partial": 0.060546875,
+      "step": 316
+    },
+    {
+      "epoch": 0.16553169198533263,
+      "loss_partial": 0.056640625,
+      "step": 316
+    },
+    {
+      "epoch": 0.16605552645364066,
+      "grad_norm": 0.10564667731523514,
+      "learning_rate": 8.272251308900523e-06,
+      "loss": 0.1248,
+      "step": 317
+    },
+    {
+      "epoch": 0.16605552645364066,
+      "loss_partial": 0.07421875,
+      "step": 317
+    },
+    {
+      "epoch": 0.16605552645364066,
+      "loss_partial": 0.05322265625,
+      "step": 317
+    },
+    {
+      "epoch": 0.16657936092194867,
+      "grad_norm": 0.10108835995197296,
+      "learning_rate": 8.298429319371727e-06,
+      "loss": 0.1246,
+      "step": 318
+    },
+    {
+      "epoch": 0.16657936092194867,
+      "loss_partial": 0.0546875,
+      "step": 318
+    },
+    {
+      "epoch": 0.16657936092194867,
+      "loss_partial": 0.061279296875,
+      "step": 318
+    },
+    {
+      "epoch": 0.16710319539025667,
+      "grad_norm": 0.10099860280752182,
+      "learning_rate": 8.324607329842933e-06,
+      "loss": 0.127,
+      "step": 319
+    },
+    {
+      "epoch": 0.16710319539025667,
+      "loss_partial": 0.056640625,
+      "step": 319
+    },
+    {
+      "epoch": 0.16710319539025667,
+      "loss_partial": 0.0595703125,
+      "step": 319
+    },
+    {
+      "epoch": 0.1676270298585647,
+      "grad_norm": 0.09249000251293182,
+      "learning_rate": 8.350785340314137e-06,
+      "loss": 0.126,
+      "step": 320
+    },
+    {
+      "epoch": 0.1676270298585647,
+      "loss_partial": 0.0673828125,
+      "step": 320
+    },
+    {
+      "epoch": 0.1676270298585647,
+      "loss_partial": 0.0732421875,
+      "step": 320
+    },
+    {
+      "epoch": 0.1681508643268727,
+      "grad_norm": 0.09697876870632172,
+      "learning_rate": 8.37696335078534e-06,
+      "loss": 0.124,
+      "step": 321
+    },
+    {
+      "epoch": 0.1681508643268727,
+      "loss_partial": 0.0654296875,
+      "step": 321
+    },
+    {
+      "epoch": 0.1681508643268727,
+      "loss_partial": 0.059814453125,
+      "step": 321
+    },
+    {
+      "epoch": 0.1686746987951807,
+      "grad_norm": 0.10463929921388626,
+      "learning_rate": 8.403141361256545e-06,
+      "loss": 0.1244,
+      "step": 322
+    },
+    {
+      "epoch": 0.1686746987951807,
+      "loss_partial": 0.064453125,
+      "step": 322
+    },
+    {
+      "epoch": 0.1686746987951807,
+      "loss_partial": 0.06103515625,
+      "step": 322
+    },
+    {
+      "epoch": 0.16919853326348874,
+      "grad_norm": 0.09309922158718109,
+      "learning_rate": 8.429319371727749e-06,
+      "loss": 0.12,
+      "step": 323
+    },
+    {
+      "epoch": 0.16919853326348874,
+      "loss_partial": 0.0634765625,
+      "step": 323
+    },
+    {
+      "epoch": 0.16919853326348874,
+      "loss_partial": 0.059814453125,
+      "step": 323
+    },
+    {
+      "epoch": 0.16972236773179675,
+      "grad_norm": 0.08914816379547119,
+      "learning_rate": 8.455497382198954e-06,
+      "loss": 0.1213,
+      "step": 324
+    },
+    {
+      "epoch": 0.16972236773179675,
+      "loss_partial": 0.0703125,
+      "step": 324
+    },
+    {
+      "epoch": 0.16972236773179675,
+      "loss_partial": 0.064453125,
+      "step": 324
+    },
+    {
+      "epoch": 0.17024620220010478,
+      "grad_norm": 0.09371978044509888,
+      "learning_rate": 8.481675392670158e-06,
+      "loss": 0.125,
+      "step": 325
+    },
+    {
+      "epoch": 0.17024620220010478,
+      "loss_partial": 0.05322265625,
+      "step": 325
+    },
+    {
+      "epoch": 0.17024620220010478,
+      "loss_partial": 0.064453125,
+      "step": 325
+    },
+    {
+      "epoch": 0.17077003666841278,
+      "grad_norm": 0.09782564640045166,
+      "learning_rate": 8.507853403141362e-06,
+      "loss": 0.1241,
+      "step": 326
+    },
+    {
+      "epoch": 0.17077003666841278,
+      "loss_partial": 0.07421875,
+      "step": 326
+    },
+    {
+      "epoch": 0.17077003666841278,
+      "loss_partial": 0.059326171875,
+      "step": 326
+    },
+    {
+      "epoch": 0.1712938711367208,
+      "grad_norm": 0.10060156881809235,
+      "learning_rate": 8.534031413612566e-06,
+      "loss": 0.1258,
+      "step": 327
+    },
+    {
+      "epoch": 0.1712938711367208,
+      "loss_partial": 0.06884765625,
+      "step": 327
+    },
+    {
+      "epoch": 0.1712938711367208,
+      "loss_partial": 0.0693359375,
+      "step": 327
+    },
+    {
+      "epoch": 0.17181770560502882,
+      "grad_norm": 0.09494626522064209,
+      "learning_rate": 8.56020942408377e-06,
+      "loss": 0.126,
+      "step": 328
+    },
+    {
+      "epoch": 0.17181770560502882,
+      "loss_partial": 0.059814453125,
+      "step": 328
+    },
+    {
+      "epoch": 0.17181770560502882,
+      "loss_partial": 0.05224609375,
+      "step": 328
+    },
+    {
+      "epoch": 0.17234154007333682,
+      "grad_norm": 0.10345195233821869,
+      "learning_rate": 8.586387434554974e-06,
+      "loss": 0.1204,
+      "step": 329
+    },
+    {
+      "epoch": 0.17234154007333682,
+      "loss_partial": 0.059326171875,
+      "step": 329
+    },
+    {
+      "epoch": 0.17234154007333682,
+      "loss_partial": 0.06640625,
+      "step": 329
+    },
+    {
+      "epoch": 0.17286537454164483,
+      "grad_norm": 0.12483423948287964,
+      "learning_rate": 8.612565445026178e-06,
+      "loss": 0.1203,
+      "step": 330
+    },
+    {
+      "epoch": 0.17286537454164483,
+      "loss_partial": 0.0634765625,
+      "step": 330
+    },
+    {
+      "epoch": 0.17286537454164483,
+      "loss_partial": 0.06787109375,
+      "step": 330
+    },
+    {
+      "epoch": 0.17338920900995286,
+      "grad_norm": 0.09235789626836777,
+      "learning_rate": 8.638743455497383e-06,
+      "loss": 0.1212,
+      "step": 331
+    },
+    {
+      "epoch": 0.17338920900995286,
+      "loss_partial": 0.054931640625,
+      "step": 331
+    },
+    {
+      "epoch": 0.17338920900995286,
+      "loss_partial": 0.0576171875,
+      "step": 331
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "grad_norm": 0.10272180289030075,
+      "learning_rate": 8.664921465968587e-06,
+      "loss": 0.1245,
+      "step": 332
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "loss_partial": 0.0673828125,
+      "step": 332
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "loss_partial": 0.054931640625,
+      "step": 332
+    },
+    {
+      "epoch": 0.1744368779465689,
+      "grad_norm": 0.09875684231519699,
+      "learning_rate": 8.691099476439791e-06,
+      "loss": 0.1229,
+      "step": 333
+    },
+    {
+      "epoch": 0.1744368779465689,
+      "loss_partial": 0.062255859375,
+      "step": 333
+    },
+    {
+      "epoch": 0.1744368779465689,
+      "loss_partial": 0.07470703125,
+      "step": 333
+    },
+    {
+      "epoch": 0.1749607124148769,
+      "grad_norm": 0.12296989560127258,
+      "learning_rate": 8.717277486910995e-06,
+      "loss": 0.128,
+      "step": 334
+    },
+    {
+      "epoch": 0.1749607124148769,
+      "loss_partial": 0.06884765625,
+      "step": 334
+    },
+    {
+      "epoch": 0.1749607124148769,
+      "loss_partial": 0.06640625,
+      "step": 334
+    },
+    {
+      "epoch": 0.1754845468831849,
+      "grad_norm": 0.10168777406215668,
+      "learning_rate": 8.743455497382199e-06,
+      "loss": 0.1248,
+      "step": 335
+    },
+    {
+      "epoch": 0.1754845468831849,
+      "loss_partial": 0.06689453125,
+      "step": 335
+    },
+    {
+      "epoch": 0.1754845468831849,
+      "loss_partial": 0.058837890625,
+      "step": 335
+    },
+    {
+      "epoch": 0.17600838135149294,
+      "grad_norm": 0.11815474182367325,
+      "learning_rate": 8.769633507853403e-06,
+      "loss": 0.1259,
+      "step": 336
+    },
+    {
+      "epoch": 0.17600838135149294,
+      "loss_partial": 0.05615234375,
+      "step": 336
+    },
+    {
+      "epoch": 0.17600838135149294,
+      "loss_partial": 0.055419921875,
+      "step": 336
+    },
+    {
+      "epoch": 0.17653221581980094,
+      "grad_norm": 0.09097845107316971,
+      "learning_rate": 8.795811518324609e-06,
+      "loss": 0.1218,
+      "step": 337
+    },
+    {
+      "epoch": 0.17653221581980094,
+      "loss_partial": 0.05322265625,
+      "step": 337
+    },
+    {
+      "epoch": 0.17653221581980094,
+      "loss_partial": 0.0625,
+      "step": 337
+    },
+    {
+      "epoch": 0.17705605028810895,
+      "grad_norm": 0.10610512644052505,
+      "learning_rate": 8.821989528795813e-06,
+      "loss": 0.1213,
+      "step": 338
+    },
+    {
+      "epoch": 0.17705605028810895,
+      "loss_partial": 0.0888671875,
+      "step": 338
+    },
+    {
+      "epoch": 0.17705605028810895,
+      "loss_partial": 0.052490234375,
+      "step": 338
+    },
+    {
+      "epoch": 0.17757988475641698,
+      "grad_norm": 0.09790844470262527,
+      "learning_rate": 8.848167539267016e-06,
+      "loss": 0.1234,
+      "step": 339
+    },
+    {
+      "epoch": 0.17757988475641698,
+      "loss_partial": 0.052978515625,
+      "step": 339
+    },
+    {
+      "epoch": 0.17757988475641698,
+      "loss_partial": 0.060791015625,
+      "step": 339
+    },
+    {
+      "epoch": 0.17810371922472498,
+      "grad_norm": 0.12197770178318024,
+      "learning_rate": 8.87434554973822e-06,
+      "loss": 0.1252,
+      "step": 340
+    },
+    {
+      "epoch": 0.17810371922472498,
+      "loss_partial": 0.05712890625,
+      "step": 340
+    },
+    {
+      "epoch": 0.17810371922472498,
+      "loss_partial": 0.0625,
+      "step": 340
+    },
+    {
+      "epoch": 0.17862755369303301,
+      "grad_norm": 0.09331309050321579,
+      "learning_rate": 8.900523560209426e-06,
+      "loss": 0.126,
+      "step": 341
+    },
+    {
+      "epoch": 0.17862755369303301,
+      "loss_partial": 0.068359375,
+      "step": 341
+    },
+    {
+      "epoch": 0.17862755369303301,
+      "loss_partial": 0.0654296875,
+      "step": 341
+    },
+    {
+      "epoch": 0.17915138816134102,
+      "grad_norm": 0.1140226274728775,
+      "learning_rate": 8.92670157068063e-06,
+      "loss": 0.1232,
+      "step": 342
+    },
+    {
+      "epoch": 0.17915138816134102,
+      "loss_partial": 0.05908203125,
+      "step": 342
+    },
+    {
+      "epoch": 0.17915138816134102,
+      "loss_partial": 0.0595703125,
+      "step": 342
+    },
+    {
+      "epoch": 0.17967522262964902,
+      "grad_norm": 0.13891302049160004,
+      "learning_rate": 8.952879581151834e-06,
+      "loss": 0.1233,
+      "step": 343
+    },
+    {
+      "epoch": 0.17967522262964902,
+      "loss_partial": 0.0625,
+      "step": 343
+    },
+    {
+      "epoch": 0.17967522262964902,
+      "loss_partial": 0.05908203125,
+      "step": 343
+    },
+    {
+      "epoch": 0.18019905709795705,
+      "grad_norm": 0.11912474036216736,
+      "learning_rate": 8.979057591623038e-06,
+      "loss": 0.1188,
+      "step": 344
+    },
+    {
+      "epoch": 0.18019905709795705,
+      "loss_partial": 0.051025390625,
+      "step": 344
+    },
+    {
+      "epoch": 0.18019905709795705,
+      "loss_partial": 0.060302734375,
+      "step": 344
+    },
+    {
+      "epoch": 0.18072289156626506,
+      "grad_norm": 0.1336831897497177,
+      "learning_rate": 9.005235602094242e-06,
+      "loss": 0.1213,
+      "step": 345
+    },
+    {
+      "epoch": 0.18072289156626506,
+      "loss_partial": 0.0615234375,
+      "step": 345
+    },
+    {
+      "epoch": 0.18072289156626506,
+      "loss_partial": 0.06005859375,
+      "step": 345
+    },
+    {
+      "epoch": 0.18124672603457306,
+      "grad_norm": 0.09774051606655121,
+      "learning_rate": 9.031413612565446e-06,
+      "loss": 0.1225,
+      "step": 346
+    },
+    {
+      "epoch": 0.18124672603457306,
+      "loss_partial": 0.064453125,
+      "step": 346
+    },
+    {
+      "epoch": 0.18124672603457306,
+      "loss_partial": 0.06396484375,
+      "step": 346
+    },
+    {
+      "epoch": 0.1817705605028811,
+      "grad_norm": 0.14156687259674072,
+      "learning_rate": 9.05759162303665e-06,
+      "loss": 0.1272,
+      "step": 347
+    },
+    {
+      "epoch": 0.1817705605028811,
+      "loss_partial": 0.055908203125,
+      "step": 347
+    },
+    {
+      "epoch": 0.1817705605028811,
+      "loss_partial": 0.06689453125,
+      "step": 347
+    },
+    {
+      "epoch": 0.1822943949711891,
+      "grad_norm": 0.1506284475326538,
+      "learning_rate": 9.083769633507853e-06,
+      "loss": 0.1234,
+      "step": 348
+    },
+    {
+      "epoch": 0.1822943949711891,
+      "loss_partial": 0.06591796875,
+      "step": 348
+    },
+    {
+      "epoch": 0.1822943949711891,
+      "loss_partial": 0.0654296875,
+      "step": 348
+    },
+    {
+      "epoch": 0.18281822943949713,
+      "grad_norm": 0.15449652075767517,
+      "learning_rate": 9.109947643979057e-06,
+      "loss": 0.1215,
+      "step": 349
+    },
+    {
+      "epoch": 0.18281822943949713,
+      "loss_partial": 0.052001953125,
+      "step": 349
+    },
+    {
+      "epoch": 0.18281822943949713,
+      "loss_partial": 0.057861328125,
+      "step": 349
+    },
+    {
+      "epoch": 0.18334206390780514,
+      "grad_norm": 0.11602424085140228,
+      "learning_rate": 9.136125654450263e-06,
+      "loss": 0.1225,
+      "step": 350
+    },
+    {
+      "epoch": 0.18334206390780514,
+      "loss_partial": 0.06298828125,
+      "step": 350
+    },
+    {
+      "epoch": 0.18334206390780514,
+      "loss_partial": 0.059814453125,
+      "step": 350
+    },
+    {
+      "epoch": 0.18386589837611314,
+      "grad_norm": 0.13446538150310516,
+      "learning_rate": 9.162303664921467e-06,
+      "loss": 0.121,
+      "step": 351
+    },
+    {
+      "epoch": 0.18386589837611314,
+      "loss_partial": 0.06396484375,
+      "step": 351
+    },
+    {
+      "epoch": 0.18386589837611314,
+      "loss_partial": 0.06201171875,
+      "step": 351
+    },
+    {
+      "epoch": 0.18438973284442117,
+      "grad_norm": 0.15640699863433838,
+      "learning_rate": 9.18848167539267e-06,
+      "loss": 0.1243,
+      "step": 352
+    },
+    {
+      "epoch": 0.18438973284442117,
+      "loss_partial": 0.06884765625,
+      "step": 352
+    },
+    {
+      "epoch": 0.18438973284442117,
+      "loss_partial": 0.0517578125,
+      "step": 352
+    },
+    {
+      "epoch": 0.18491356731272918,
+      "grad_norm": 0.1092001423239708,
+      "learning_rate": 9.214659685863875e-06,
+      "loss": 0.1216,
+      "step": 353
+    },
+    {
+      "epoch": 0.18491356731272918,
+      "loss_partial": 0.060791015625,
+      "step": 353
+    },
+    {
+      "epoch": 0.18491356731272918,
+      "loss_partial": 0.06103515625,
+      "step": 353
+    },
+    {
+      "epoch": 0.18543740178103718,
+      "grad_norm": 0.09094612300395966,
+      "learning_rate": 9.240837696335079e-06,
+      "loss": 0.1263,
+      "step": 354
+    },
+    {
+      "epoch": 0.18543740178103718,
+      "loss_partial": 0.060302734375,
+      "step": 354
+    },
+    {
+      "epoch": 0.18543740178103718,
+      "loss_partial": 0.052978515625,
+      "step": 354
+    },
+    {
+      "epoch": 0.1859612362493452,
+      "grad_norm": 0.17791736125946045,
+      "learning_rate": 9.267015706806284e-06,
+      "loss": 0.1246,
+      "step": 355
+    },
+    {
+      "epoch": 0.1859612362493452,
+      "loss_partial": 0.056884765625,
+      "step": 355
+    },
+    {
+      "epoch": 0.1859612362493452,
+      "loss_partial": 0.0595703125,
+      "step": 355
+    },
+    {
+      "epoch": 0.18648507071765322,
+      "grad_norm": 0.13169234991073608,
+      "learning_rate": 9.293193717277488e-06,
+      "loss": 0.1218,
+      "step": 356
+    },
+    {
+      "epoch": 0.18648507071765322,
+      "loss_partial": 0.064453125,
+      "step": 356
+    },
+    {
+      "epoch": 0.18648507071765322,
+      "loss_partial": 0.051025390625,
+      "step": 356
+    },
+    {
+      "epoch": 0.18700890518596125,
+      "grad_norm": 0.09410545229911804,
+      "learning_rate": 9.319371727748692e-06,
+      "loss": 0.1216,
+      "step": 357
+    },
+    {
+      "epoch": 0.18700890518596125,
+      "loss_partial": 0.050048828125,
+      "step": 357
+    },
+    {
+      "epoch": 0.18700890518596125,
+      "loss_partial": 0.05810546875,
+      "step": 357
+    },
+    {
+      "epoch": 0.18753273965426925,
+      "grad_norm": 0.13292153179645538,
+      "learning_rate": 9.345549738219896e-06,
+      "loss": 0.1202,
+      "step": 358
+    },
+    {
+      "epoch": 0.18753273965426925,
+      "loss_partial": 0.060791015625,
+      "step": 358
+    },
+    {
+      "epoch": 0.18753273965426925,
+      "loss_partial": 0.062255859375,
+      "step": 358
+    },
+    {
+      "epoch": 0.18805657412257726,
+      "grad_norm": 0.1737011969089508,
+      "learning_rate": 9.3717277486911e-06,
+      "loss": 0.1244,
+      "step": 359
+    },
+    {
+      "epoch": 0.18805657412257726,
+      "loss_partial": 0.056640625,
+      "step": 359
+    },
+    {
+      "epoch": 0.18805657412257726,
+      "loss_partial": 0.0791015625,
+      "step": 359
+    },
+    {
+      "epoch": 0.1885804085908853,
+      "grad_norm": 0.10132495313882828,
+      "learning_rate": 9.397905759162304e-06,
+      "loss": 0.1255,
+      "step": 360
+    },
+    {
+      "epoch": 0.1885804085908853,
+      "loss_partial": 0.06298828125,
+      "step": 360
+    },
+    {
+      "epoch": 0.1885804085908853,
+      "loss_partial": 0.05078125,
+      "step": 360
+    },
+    {
+      "epoch": 0.1891042430591933,
+      "grad_norm": 0.12255586683750153,
+      "learning_rate": 9.424083769633508e-06,
+      "loss": 0.1242,
+      "step": 361
+    },
+    {
+      "epoch": 0.1891042430591933,
+      "loss_partial": 0.0654296875,
+      "step": 361
+    },
+    {
+      "epoch": 0.1891042430591933,
+      "loss_partial": 0.053466796875,
+      "step": 361
+    },
+    {
+      "epoch": 0.1896280775275013,
+      "grad_norm": 0.11022427678108215,
+      "learning_rate": 9.450261780104712e-06,
+      "loss": 0.1208,
+      "step": 362
+    },
+    {
+      "epoch": 0.1896280775275013,
+      "loss_partial": 0.060546875,
+      "step": 362
+    },
+    {
+      "epoch": 0.1896280775275013,
+      "loss_partial": 0.06640625,
+      "step": 362
+    },
+    {
+      "epoch": 0.19015191199580933,
+      "grad_norm": 0.10149624198675156,
+      "learning_rate": 9.476439790575916e-06,
+      "loss": 0.1198,
+      "step": 363
+    },
+    {
+      "epoch": 0.19015191199580933,
+      "loss_partial": 0.06005859375,
+      "step": 363
+    },
+    {
+      "epoch": 0.19015191199580933,
+      "loss_partial": 0.056396484375,
+      "step": 363
+    },
+    {
+      "epoch": 0.19067574646411733,
+      "grad_norm": 0.13758216798305511,
+      "learning_rate": 9.502617801047121e-06,
+      "loss": 0.1281,
+      "step": 364
+    },
+    {
+      "epoch": 0.19067574646411733,
+      "loss_partial": 0.060302734375,
+      "step": 364
+    },
+    {
+      "epoch": 0.19067574646411733,
+      "loss_partial": 0.05078125,
+      "step": 364
+    },
+    {
+      "epoch": 0.19119958093242537,
+      "grad_norm": 0.13273443281650543,
+      "learning_rate": 9.528795811518325e-06,
+      "loss": 0.1234,
+      "step": 365
+    },
+    {
+      "epoch": 0.19119958093242537,
+      "loss_partial": 0.047119140625,
+      "step": 365
+    },
+    {
+      "epoch": 0.19119958093242537,
+      "loss_partial": 0.06005859375,
+      "step": 365
+    },
+    {
+      "epoch": 0.19172341540073337,
+      "grad_norm": 0.10869482904672623,
+      "learning_rate": 9.554973821989529e-06,
+      "loss": 0.1264,
+      "step": 366
+    },
+    {
+      "epoch": 0.19172341540073337,
+      "loss_partial": 0.0732421875,
+      "step": 366
+    },
+    {
+      "epoch": 0.19172341540073337,
+      "loss_partial": 0.064453125,
+      "step": 366
+    },
+    {
+      "epoch": 0.19224724986904138,
+      "grad_norm": 0.11448775976896286,
+      "learning_rate": 9.581151832460733e-06,
+      "loss": 0.1266,
+      "step": 367
+    },
+    {
+      "epoch": 0.19224724986904138,
+      "loss_partial": 0.060791015625,
+      "step": 367
+    },
+    {
+      "epoch": 0.19224724986904138,
+      "loss_partial": 0.0615234375,
+      "step": 367
+    },
+    {
+      "epoch": 0.1927710843373494,
+      "grad_norm": 0.1455133706331253,
+      "learning_rate": 9.607329842931939e-06,
+      "loss": 0.1236,
+      "step": 368
+    },
+    {
+      "epoch": 0.1927710843373494,
+      "loss_partial": 0.07177734375,
+      "step": 368
+    },
+    {
+      "epoch": 0.1927710843373494,
+      "loss_partial": 0.06201171875,
+      "step": 368
+    },
+    {
+      "epoch": 0.1932949188056574,
+      "grad_norm": 0.09707994759082794,
+      "learning_rate": 9.633507853403143e-06,
+      "loss": 0.1219,
+      "step": 369
+    },
+    {
+      "epoch": 0.1932949188056574,
+      "loss_partial": 0.068359375,
+      "step": 369
+    },
+    {
+      "epoch": 0.1932949188056574,
+      "loss_partial": 0.05712890625,
+      "step": 369
+    },
+    {
+      "epoch": 0.19381875327396542,
+      "grad_norm": 0.10645634680986404,
+      "learning_rate": 9.659685863874346e-06,
+      "loss": 0.1181,
+      "step": 370
+    },
+    {
+      "epoch": 0.19381875327396542,
+      "loss_partial": 0.06787109375,
+      "step": 370
+    },
+    {
+      "epoch": 0.19381875327396542,
+      "loss_partial": 0.06787109375,
+      "step": 370
+    },
+    {
+      "epoch": 0.19434258774227345,
+      "grad_norm": 0.1146468073129654,
+      "learning_rate": 9.68586387434555e-06,
+      "loss": 0.1252,
+      "step": 371
+    },
+    {
+      "epoch": 0.19434258774227345,
+      "loss_partial": 0.0595703125,
+      "step": 371
+    },
+    {
+      "epoch": 0.19434258774227345,
+      "loss_partial": 0.057373046875,
+      "step": 371
+    },
+    {
+      "epoch": 0.19486642221058145,
+      "grad_norm": 0.09309422969818115,
+      "learning_rate": 9.712041884816756e-06,
+      "loss": 0.127,
+      "step": 372
+    },
+    {
+      "epoch": 0.19486642221058145,
+      "loss_partial": 0.05859375,
+      "step": 372
+    },
+    {
+      "epoch": 0.19486642221058145,
+      "loss_partial": 0.06591796875,
+      "step": 372
+    },
+    {
+      "epoch": 0.19539025667888948,
+      "grad_norm": 0.1388443112373352,
+      "learning_rate": 9.73821989528796e-06,
+      "loss": 0.1208,
+      "step": 373
+    },
+    {
+      "epoch": 0.19539025667888948,
+      "loss_partial": 0.06640625,
+      "step": 373
+    },
+    {
+      "epoch": 0.19539025667888948,
+      "loss_partial": 0.05322265625,
+      "step": 373
+    },
+    {
+      "epoch": 0.1959140911471975,
+      "grad_norm": 0.14103345572948456,
+      "learning_rate": 9.764397905759164e-06,
+      "loss": 0.1235,
+      "step": 374
+    },
+    {
+      "epoch": 0.1959140911471975,
+      "loss_partial": 0.06201171875,
+      "step": 374
+    },
+    {
+      "epoch": 0.1959140911471975,
+      "loss_partial": 0.07080078125,
+      "step": 374
+    },
+    {
+      "epoch": 0.1964379256155055,
+      "grad_norm": 0.1206183135509491,
+      "learning_rate": 9.790575916230368e-06,
+      "loss": 0.1245,
+      "step": 375
+    },
+    {
+      "epoch": 0.1964379256155055,
+      "loss_partial": 0.0556640625,
+      "step": 375
+    },
+    {
+      "epoch": 0.1964379256155055,
+      "loss_partial": 0.0595703125,
+      "step": 375
+    },
+    {
+      "epoch": 0.19696176008381353,
+      "grad_norm": 0.10004851967096329,
+      "learning_rate": 9.816753926701572e-06,
+      "loss": 0.1215,
+      "step": 376
+    },
+    {
+      "epoch": 0.19696176008381353,
+      "loss_partial": 0.0576171875,
+      "step": 376
+    },
+    {
+      "epoch": 0.19696176008381353,
+      "loss_partial": 0.05810546875,
+      "step": 376
+    },
+    {
+      "epoch": 0.19748559455212153,
+      "grad_norm": 0.24206110835075378,
+      "learning_rate": 9.842931937172776e-06,
+      "loss": 0.1173,
+      "step": 377
+    },
+    {
+      "epoch": 0.19748559455212153,
+      "loss_partial": 0.055419921875,
+      "step": 377
+    },
+    {
+      "epoch": 0.19748559455212153,
+      "loss_partial": 0.058349609375,
+      "step": 377
+    },
+    {
+      "epoch": 0.19800942902042953,
+      "grad_norm": 0.11311155557632446,
+      "learning_rate": 9.86910994764398e-06,
+      "loss": 0.1219,
+      "step": 378
+    },
+    {
+      "epoch": 0.19800942902042953,
+      "loss_partial": 0.057861328125,
+      "step": 378
+    },
+    {
+      "epoch": 0.19800942902042953,
+      "loss_partial": 0.059814453125,
+      "step": 378
+    },
+    {
+      "epoch": 0.19853326348873757,
+      "grad_norm": 0.2662658095359802,
+      "learning_rate": 9.895287958115183e-06,
+      "loss": 0.1208,
+      "step": 379
+    },
+    {
+      "epoch": 0.19853326348873757,
+      "loss_partial": 0.056396484375,
+      "step": 379
+    },
+    {
+      "epoch": 0.19853326348873757,
+      "loss_partial": 0.0615234375,
+      "step": 379
+    },
+    {
+      "epoch": 0.19905709795704557,
+      "grad_norm": 0.13019026815891266,
+      "learning_rate": 9.921465968586387e-06,
+      "loss": 0.1222,
+      "step": 380
+    },
+    {
+      "epoch": 0.19905709795704557,
+      "loss_partial": 0.055419921875,
+      "step": 380
+    },
+    {
+      "epoch": 0.19905709795704557,
+      "loss_partial": 0.0634765625,
+      "step": 380
+    },
+    {
+      "epoch": 0.1995809324253536,
+      "grad_norm": 0.12759250402450562,
+      "learning_rate": 9.947643979057593e-06,
+      "loss": 0.12,
+      "step": 381
+    },
+    {
+      "epoch": 0.1995809324253536,
+      "loss_partial": 0.055419921875,
+      "step": 381
+    },
+    {
+      "epoch": 0.1995809324253536,
+      "loss_partial": 0.06201171875,
+      "step": 381
+    },
+    {
+      "epoch": 0.2001047668936616,
+      "grad_norm": 0.16119660437107086,
+      "learning_rate": 9.973821989528797e-06,
+      "loss": 0.1212,
+      "step": 382
+    },
+    {
+      "epoch": 0.2001047668936616,
+      "loss_partial": 0.06689453125,
+      "step": 382
+    },
+    {
+      "epoch": 0.2001047668936616,
+      "loss_partial": 0.06494140625,
+      "step": 382
+    },
+    {
+      "epoch": 0.2006286013619696,
+      "grad_norm": 0.1874997764825821,
+      "learning_rate": 1e-05,
+      "loss": 0.1226,
+      "step": 383
+    },
+    {
+      "epoch": 0.2006286013619696,
+      "loss_partial": 0.05322265625,
+      "step": 383
+    },
+    {
+      "epoch": 0.2006286013619696,
+      "loss_partial": 0.072265625,
+      "step": 383
+    },
+    {
+      "epoch": 0.20115243583027764,
+      "grad_norm": 0.11407534033060074,
+      "learning_rate": 1.0026178010471206e-05,
+      "loss": 0.1248,
+      "step": 384
+    },
+    {
+      "epoch": 0.20115243583027764,
+      "loss_partial": 0.0654296875,
+      "step": 384
+    },
+    {
+      "epoch": 0.20115243583027764,
+      "loss_partial": 0.056396484375,
+      "step": 384
+    },
+    {
+      "epoch": 0.20167627029858565,
+      "grad_norm": 0.10196191072463989,
+      "learning_rate": 1.0052356020942409e-05,
+      "loss": 0.1156,
+      "step": 385
+    },
+    {
+      "epoch": 0.20167627029858565,
+      "loss_partial": 0.0625,
+      "step": 385
+    },
+    {
+      "epoch": 0.20167627029858565,
+      "loss_partial": 0.060302734375,
+      "step": 385
+    },
+    {
+      "epoch": 0.20220010476689365,
+      "grad_norm": 0.13412357866764069,
+      "learning_rate": 1.0078534031413614e-05,
+      "loss": 0.1253,
+      "step": 386
+    },
+    {
+      "epoch": 0.20220010476689365,
+      "loss_partial": 0.052001953125,
+      "step": 386
+    },
+    {
+      "epoch": 0.20220010476689365,
+      "loss_partial": 0.058837890625,
+      "step": 386
+    },
+    {
+      "epoch": 0.20272393923520168,
+      "grad_norm": 0.1392621099948883,
+      "learning_rate": 1.0104712041884816e-05,
+      "loss": 0.1202,
+      "step": 387
+    },
+    {
+      "epoch": 0.20272393923520168,
+      "loss_partial": 0.068359375,
+      "step": 387
+    },
+    {
+      "epoch": 0.20272393923520168,
+      "loss_partial": 0.05322265625,
+      "step": 387
+    },
+    {
+      "epoch": 0.2032477737035097,
+      "grad_norm": 0.23472721874713898,
+      "learning_rate": 1.0130890052356022e-05,
+      "loss": 0.1212,
+      "step": 388
+    },
+    {
+      "epoch": 0.2032477737035097,
+      "loss_partial": 0.0625,
+      "step": 388
+    },
+    {
+      "epoch": 0.2032477737035097,
+      "loss_partial": 0.07373046875,
+      "step": 388
+    },
+    {
+      "epoch": 0.2037716081718177,
+      "grad_norm": 0.18173450231552124,
+      "learning_rate": 1.0157068062827226e-05,
+      "loss": 0.123,
+      "step": 389
+    },
+    {
+      "epoch": 0.2037716081718177,
+      "loss_partial": 0.06396484375,
+      "step": 389
+    },
+    {
+      "epoch": 0.2037716081718177,
+      "loss_partial": 0.059814453125,
+      "step": 389
+    },
+    {
+      "epoch": 0.20429544264012572,
+      "grad_norm": 0.14787188172340393,
+      "learning_rate": 1.018324607329843e-05,
+      "loss": 0.1216,
+      "step": 390
+    },
+    {
+      "epoch": 0.20429544264012572,
+      "loss_partial": 0.06591796875,
+      "step": 390
+    },
+    {
+      "epoch": 0.20429544264012572,
+      "loss_partial": 0.055908203125,
+      "step": 390
+    },
+    {
+      "epoch": 0.20481927710843373,
+      "grad_norm": 0.18292255699634552,
+      "learning_rate": 1.0209424083769634e-05,
+      "loss": 0.124,
+      "step": 391
+    },
+    {
+      "epoch": 0.20481927710843373,
+      "loss_partial": 0.056396484375,
+      "step": 391
+    },
+    {
+      "epoch": 0.20481927710843373,
+      "loss_partial": 0.06884765625,
+      "step": 391
+    },
+    {
+      "epoch": 0.20534311157674176,
+      "grad_norm": 0.163756862282753,
+      "learning_rate": 1.023560209424084e-05,
+      "loss": 0.1227,
+      "step": 392
+    },
+    {
+      "epoch": 0.20534311157674176,
+      "loss_partial": 0.061767578125,
+      "step": 392
+    },
+    {
+      "epoch": 0.20534311157674176,
+      "loss_partial": 0.06494140625,
+      "step": 392
+    },
+    {
+      "epoch": 0.20586694604504976,
+      "grad_norm": 0.12682513892650604,
+      "learning_rate": 1.0261780104712042e-05,
+      "loss": 0.1231,
+      "step": 393
+    },
+    {
+      "epoch": 0.20586694604504976,
+      "loss_partial": 0.05859375,
+      "step": 393
+    },
+    {
+      "epoch": 0.20586694604504976,
+      "loss_partial": 0.057373046875,
+      "step": 393
+    },
+    {
+      "epoch": 0.20639078051335777,
+      "grad_norm": 0.1256488710641861,
+      "learning_rate": 1.0287958115183247e-05,
+      "loss": 0.1222,
+      "step": 394
+    },
+    {
+      "epoch": 0.20639078051335777,
+      "loss_partial": 0.06298828125,
+      "step": 394
+    },
+    {
+      "epoch": 0.20639078051335777,
+      "loss_partial": 0.0634765625,
+      "step": 394
+    },
+    {
+      "epoch": 0.2069146149816658,
+      "grad_norm": 0.1216011792421341,
+      "learning_rate": 1.031413612565445e-05,
+      "loss": 0.1232,
+      "step": 395
+    },
+    {
+      "epoch": 0.2069146149816658,
+      "loss_partial": 0.0556640625,
+      "step": 395
+    },
+    {
+      "epoch": 0.2069146149816658,
+      "loss_partial": 0.0576171875,
+      "step": 395
+    },
+    {
+      "epoch": 0.2074384494499738,
+      "grad_norm": 0.20447595417499542,
+      "learning_rate": 1.0340314136125655e-05,
+      "loss": 0.119,
+      "step": 396
+    },
+    {
+      "epoch": 0.2074384494499738,
+      "loss_partial": 0.06396484375,
+      "step": 396
+    },
+    {
+      "epoch": 0.2074384494499738,
+      "loss_partial": 0.0556640625,
+      "step": 396
+    },
+    {
+      "epoch": 0.2079622839182818,
+      "grad_norm": 0.17105357348918915,
+      "learning_rate": 1.036649214659686e-05,
+      "loss": 0.1221,
+      "step": 397
+    },
+    {
+      "epoch": 0.2079622839182818,
+      "loss_partial": 0.055419921875,
+      "step": 397
+    },
+    {
+      "epoch": 0.2079622839182818,
+      "loss_partial": 0.059814453125,
+      "step": 397
+    },
+    {
+      "epoch": 0.20848611838658984,
+      "grad_norm": 0.30538350343704224,
+      "learning_rate": 1.0392670157068063e-05,
+      "loss": 0.1221,
+      "step": 398
+    },
+    {
+      "epoch": 0.20848611838658984,
+      "loss_partial": 0.05419921875,
+      "step": 398
+    },
+    {
+      "epoch": 0.20848611838658984,
+      "loss_partial": 0.0556640625,
+      "step": 398
+    },
+    {
+      "epoch": 0.20900995285489785,
+      "grad_norm": 0.24677774310112,
+      "learning_rate": 1.0418848167539269e-05,
+      "loss": 0.1176,
+      "step": 399
+    },
+    {
+      "epoch": 0.20900995285489785,
+      "loss_partial": 0.0654296875,
+      "step": 399
+    },
+    {
+      "epoch": 0.20900995285489785,
+      "loss_partial": 0.060546875,
+      "step": 399
+    },
+    {
+      "epoch": 0.20953378732320588,
+      "grad_norm": 0.1336302012205124,
+      "learning_rate": 1.0445026178010471e-05,
+      "loss": 0.1216,
+      "step": 400
+    },
+    {
+      "epoch": 0.20953378732320588,
+      "loss_partial": 0.056396484375,
+      "step": 400
+    },
+    {
+      "epoch": 0.20953378732320588,
+      "loss_partial": 0.061279296875,
+      "step": 400
+    },
+    {
+      "epoch": 0.21005762179151388,
+      "grad_norm": 0.18161992728710175,
+      "learning_rate": 1.0471204188481676e-05,
+      "loss": 0.1199,
+      "step": 401
+    },
+    {
+      "epoch": 0.21005762179151388,
+      "loss_partial": 0.0654296875,
+      "step": 401
+    },
+    {
+      "epoch": 0.21005762179151388,
+      "loss_partial": 0.06982421875,
+      "step": 401
+    },
+    {
+      "epoch": 0.2105814562598219,
+      "grad_norm": 0.36211898922920227,
+      "learning_rate": 1.0497382198952882e-05,
+      "loss": 0.1226,
+      "step": 402
+    },
+    {
+      "epoch": 0.2105814562598219,
+      "loss_partial": 0.058837890625,
+      "step": 402
+    },
+    {
+      "epoch": 0.2105814562598219,
+      "loss_partial": 0.0634765625,
+      "step": 402
+    },
+    {
+      "epoch": 0.21110529072812992,
+      "grad_norm": 0.2673801779747009,
+      "learning_rate": 1.0523560209424084e-05,
+      "loss": 0.1236,
+      "step": 403
+    },
+    {
+      "epoch": 0.21110529072812992,
+      "loss_partial": 0.0712890625,
+      "step": 403
+    },
+    {
+      "epoch": 0.21110529072812992,
+      "loss_partial": 0.0654296875,
+      "step": 403
+    },
+    {
+      "epoch": 0.21162912519643792,
+      "grad_norm": 0.5919686555862427,
+      "learning_rate": 1.054973821989529e-05,
+      "loss": 0.1247,
+      "step": 404
+    },
+    {
+      "epoch": 0.21162912519643792,
+      "loss_partial": 0.0576171875,
+      "step": 404
+    },
+    {
+      "epoch": 0.21162912519643792,
+      "loss_partial": 0.06494140625,
+      "step": 404
+    },
+    {
+      "epoch": 0.21215295966474593,
+      "grad_norm": 0.7325088381767273,
+      "learning_rate": 1.0575916230366492e-05,
+      "loss": 0.1256,
+      "step": 405
+    },
+    {
+      "epoch": 0.21215295966474593,
+      "loss_partial": 0.07177734375,
+      "step": 405
+    },
+    {
+      "epoch": 0.21215295966474593,
+      "loss_partial": 0.058837890625,
+      "step": 405
+    },
+    {
+      "epoch": 0.21267679413305396,
+      "grad_norm": 0.32472002506256104,
+      "learning_rate": 1.0602094240837698e-05,
+      "loss": 0.1252,
+      "step": 406
+    },
+    {
+      "epoch": 0.21267679413305396,
+      "loss_partial": 0.062255859375,
+      "step": 406
+    },
+    {
+      "epoch": 0.21267679413305396,
+      "loss_partial": 0.0546875,
+      "step": 406
+    },
+    {
+      "epoch": 0.21320062860136196,
+      "grad_norm": 0.4535265564918518,
+      "learning_rate": 1.0628272251308902e-05,
+      "loss": 0.1191,
+      "step": 407
+    },
+    {
+      "epoch": 0.21320062860136196,
+      "loss_partial": 0.06787109375,
+      "step": 407
+    },
+    {
+      "epoch": 0.21320062860136196,
+      "loss_partial": 0.05615234375,
+      "step": 407
+    },
+    {
+      "epoch": 0.21372446306967,
+      "grad_norm": 0.36535021662712097,
+      "learning_rate": 1.0654450261780106e-05,
+      "loss": 0.1233,
+      "step": 408
+    },
+    {
+      "epoch": 0.21372446306967,
+      "loss_partial": 0.06103515625,
+      "step": 408
+    },
+    {
+      "epoch": 0.21372446306967,
+      "loss_partial": 0.06494140625,
+      "step": 408
+    },
+    {
+      "epoch": 0.214248297537978,
+      "grad_norm": 0.22524353861808777,
+      "learning_rate": 1.068062827225131e-05,
+      "loss": 0.1215,
+      "step": 409
+    },
+    {
+      "epoch": 0.214248297537978,
+      "loss_partial": 0.0556640625,
+      "step": 409
+    },
+    {
+      "epoch": 0.214248297537978,
+      "loss_partial": 0.0673828125,
+      "step": 409
+    },
+    {
+      "epoch": 0.214772132006286,
+      "grad_norm": 0.44014039635658264,
+      "learning_rate": 1.0706806282722515e-05,
+      "loss": 0.1207,
+      "step": 410
+    },
+    {
+      "epoch": 0.214772132006286,
+      "loss_partial": 0.0654296875,
+      "step": 410
+    },
+    {
+      "epoch": 0.214772132006286,
+      "loss_partial": 0.064453125,
+      "step": 410
+    },
+    {
+      "epoch": 0.21529596647459404,
+      "grad_norm": 0.13683930039405823,
+      "learning_rate": 1.0732984293193717e-05,
+      "loss": 0.1234,
+      "step": 411
+    },
+    {
+      "epoch": 0.21529596647459404,
+      "loss_partial": 0.051025390625,
+      "step": 411
+    },
+    {
+      "epoch": 0.21529596647459404,
+      "loss_partial": 0.06689453125,
+      "step": 411
+    },
+    {
+      "epoch": 0.21581980094290204,
+      "grad_norm": 0.27118706703186035,
+      "learning_rate": 1.0759162303664923e-05,
+      "loss": 0.1196,
+      "step": 412
+    },
+    {
+      "epoch": 0.21581980094290204,
+      "loss_partial": 0.061767578125,
+      "step": 412
+    },
+    {
+      "epoch": 0.21581980094290204,
+      "loss_partial": 0.05712890625,
+      "step": 412
+    },
+    {
+      "epoch": 0.21634363541121004,
+      "grad_norm": 0.17043739557266235,
+      "learning_rate": 1.0785340314136125e-05,
+      "loss": 0.1216,
+      "step": 413
+    },
+    {
+      "epoch": 0.21634363541121004,
+      "loss_partial": 0.0615234375,
+      "step": 413
+    },
+    {
+      "epoch": 0.21634363541121004,
+      "loss_partial": 0.06689453125,
+      "step": 413
+    },
+    {
+      "epoch": 0.21686746987951808,
+      "grad_norm": 0.19993828237056732,
+      "learning_rate": 1.081151832460733e-05,
+      "loss": 0.1211,
+      "step": 414
+    },
+    {
+      "epoch": 0.21686746987951808,
+      "loss_partial": 0.061279296875,
+      "step": 414
+    },
+    {
+      "epoch": 0.21686746987951808,
+      "loss_partial": 0.07275390625,
+      "step": 414
+    },
+    {
+      "epoch": 0.21739130434782608,
+      "grad_norm": 0.2037537842988968,
+      "learning_rate": 1.0837696335078536e-05,
+      "loss": 0.1229,
+      "step": 415
+    },
+    {
+      "epoch": 0.21739130434782608,
+      "loss_partial": 0.060791015625,
+      "step": 415
+    },
+    {
+      "epoch": 0.21739130434782608,
+      "loss_partial": 0.0625,
+      "step": 415
+    },
+    {
+      "epoch": 0.2179151388161341,
+      "grad_norm": 0.18145912885665894,
+      "learning_rate": 1.0863874345549739e-05,
+      "loss": 0.121,
+      "step": 416
+    },
+    {
+      "epoch": 0.2179151388161341,
+      "loss_partial": 0.059326171875,
+      "step": 416
+    },
+    {
+      "epoch": 0.2179151388161341,
+      "loss_partial": 0.0634765625,
+      "step": 416
+    },
+    {
+      "epoch": 0.21843897328444212,
+      "grad_norm": 0.14377936720848083,
+      "learning_rate": 1.0890052356020944e-05,
+      "loss": 0.1204,
+      "step": 417
+    },
+    {
+      "epoch": 0.21843897328444212,
+      "loss_partial": 0.053466796875,
+      "step": 417
+    },
+    {
+      "epoch": 0.21843897328444212,
+      "loss_partial": 0.05810546875,
+      "step": 417
+    },
+    {
+      "epoch": 0.21896280775275012,
+      "grad_norm": 0.1603909283876419,
+      "learning_rate": 1.0916230366492146e-05,
+      "loss": 0.1198,
+      "step": 418
+    },
+    {
+      "epoch": 0.21896280775275012,
+      "loss_partial": 0.06591796875,
+      "step": 418
+    },
+    {
+      "epoch": 0.21896280775275012,
+      "loss_partial": 0.06982421875,
+      "step": 418
+    },
+    {
+      "epoch": 0.21948664222105815,
+      "grad_norm": 0.3446015417575836,
+      "learning_rate": 1.0942408376963352e-05,
+      "loss": 0.1227,
+      "step": 419
+    },
+    {
+      "epoch": 0.21948664222105815,
+      "loss_partial": 0.078125,
+      "step": 419
+    },
+    {
+      "epoch": 0.21948664222105815,
+      "loss_partial": 0.06005859375,
+      "step": 419
+    },
+    {
+      "epoch": 0.22001047668936616,
+      "grad_norm": 0.34164705872535706,
+      "learning_rate": 1.0968586387434556e-05,
+      "loss": 0.122,
+      "step": 420
+    },
+    {
+      "epoch": 0.22001047668936616,
+      "loss_partial": 0.056396484375,
+      "step": 420
+    },
+    {
+      "epoch": 0.22001047668936616,
+      "loss_partial": 0.051513671875,
+      "step": 420
+    },
+    {
+      "epoch": 0.22053431115767416,
+      "grad_norm": 0.2860282063484192,
+      "learning_rate": 1.099476439790576e-05,
+      "loss": 0.1175,
+      "step": 421
+    },
+    {
+      "epoch": 0.22053431115767416,
+      "loss_partial": 0.060546875,
+      "step": 421
+    },
+    {
+      "epoch": 0.22053431115767416,
+      "loss_partial": 0.056640625,
+      "step": 421
+    },
+    {
+      "epoch": 0.2210581456259822,
+      "grad_norm": 0.2830250561237335,
+      "learning_rate": 1.1020942408376964e-05,
+      "loss": 0.1198,
+      "step": 422
+    },
+    {
+      "epoch": 0.2210581456259822,
+      "loss_partial": 0.058349609375,
+      "step": 422
+    },
+    {
+      "epoch": 0.2210581456259822,
+      "loss_partial": 0.05859375,
+      "step": 422
+    },
+    {
+      "epoch": 0.2215819800942902,
+      "grad_norm": 0.29200565814971924,
+      "learning_rate": 1.104712041884817e-05,
+      "loss": 0.1206,
+      "step": 423
+    },
+    {
+      "epoch": 0.2215819800942902,
+      "loss_partial": 0.058837890625,
+      "step": 423
+    },
+    {
+      "epoch": 0.2215819800942902,
+      "loss_partial": 0.05419921875,
+      "step": 423
+    },
+    {
+      "epoch": 0.22210581456259823,
+      "grad_norm": 0.2963371276855469,
+      "learning_rate": 1.1073298429319372e-05,
+      "loss": 0.1199,
+      "step": 424
+    },
+    {
+      "epoch": 0.22210581456259823,
+      "loss_partial": 0.061279296875,
+      "step": 424
+    },
+    {
+      "epoch": 0.22210581456259823,
+      "loss_partial": 0.0654296875,
+      "step": 424
+    },
+    {
+      "epoch": 0.22262964903090623,
+      "grad_norm": 0.2757945656776428,
+      "learning_rate": 1.1099476439790577e-05,
+      "loss": 0.1207,
+      "step": 425
+    },
+    {
+      "epoch": 0.22262964903090623,
+      "loss_partial": 0.054931640625,
+      "step": 425
+    },
+    {
+      "epoch": 0.22262964903090623,
+      "loss_partial": 0.06640625,
+      "step": 425
+    },
+    {
+      "epoch": 0.22315348349921424,
+      "grad_norm": 0.27250099182128906,
+      "learning_rate": 1.112565445026178e-05,
+      "loss": 0.1219,
+      "step": 426
+    },
+    {
+      "epoch": 0.22315348349921424,
+      "loss_partial": 0.055908203125,
+      "step": 426
+    },
+    {
+      "epoch": 0.22315348349921424,
+      "loss_partial": 0.0634765625,
+      "step": 426
+    },
+    {
+      "epoch": 0.22367731796752227,
+      "grad_norm": 0.28090643882751465,
+      "learning_rate": 1.1151832460732985e-05,
+      "loss": 0.1206,
+      "step": 427
+    },
+    {
+      "epoch": 0.22367731796752227,
+      "loss_partial": 0.05810546875,
+      "step": 427
+    },
+    {
+      "epoch": 0.22367731796752227,
+      "loss_partial": 0.05908203125,
+      "step": 427
+    },
+    {
+      "epoch": 0.22420115243583028,
+      "grad_norm": 0.47172683477401733,
+      "learning_rate": 1.117801047120419e-05,
+      "loss": 0.1194,
+      "step": 428
+    },
+    {
+      "epoch": 0.22420115243583028,
+      "loss_partial": 0.058349609375,
+      "step": 428
+    },
+    {
+      "epoch": 0.22420115243583028,
+      "loss_partial": 0.0556640625,
+      "step": 428
+    },
+    {
+      "epoch": 0.22472498690413828,
+      "grad_norm": 0.28107985854148865,
+      "learning_rate": 1.1204188481675393e-05,
+      "loss": 0.1199,
+      "step": 429
+    },
+    {
+      "epoch": 0.22472498690413828,
+      "loss_partial": 0.0498046875,
+      "step": 429
+    },
+    {
+      "epoch": 0.22472498690413828,
+      "loss_partial": 0.061279296875,
+      "step": 429
+    },
+    {
+      "epoch": 0.2252488213724463,
+      "grad_norm": 0.19957777857780457,
+      "learning_rate": 1.1230366492146599e-05,
+      "loss": 0.1215,
+      "step": 430
+    },
+    {
+      "epoch": 0.2252488213724463,
+      "loss_partial": 0.06201171875,
+      "step": 430
+    },
+    {
+      "epoch": 0.2252488213724463,
+      "loss_partial": 0.059814453125,
+      "step": 430
+    },
+    {
+      "epoch": 0.22577265584075432,
+      "grad_norm": 0.3269699513912201,
+      "learning_rate": 1.12565445026178e-05,
+      "loss": 0.124,
+      "step": 431
+    },
+    {
+      "epoch": 0.22577265584075432,
+      "loss_partial": 0.048828125,
+      "step": 431
+    },
+    {
+      "epoch": 0.22577265584075432,
+      "loss_partial": 0.06396484375,
+      "step": 431
+    },
+    {
+      "epoch": 0.22629649030906235,
+      "grad_norm": 0.5049388408660889,
+      "learning_rate": 1.1282722513089006e-05,
+      "loss": 0.1219,
+      "step": 432
+    },
+    {
+      "epoch": 0.22629649030906235,
+      "loss_partial": 0.0615234375,
+      "step": 432
+    },
+    {
+      "epoch": 0.22629649030906235,
+      "loss_partial": 0.05712890625,
+      "step": 432
+    },
+    {
+      "epoch": 0.22682032477737035,
+      "grad_norm": 0.26436886191368103,
+      "learning_rate": 1.130890052356021e-05,
+      "loss": 0.1178,
+      "step": 433
+    },
+    {
+      "epoch": 0.22682032477737035,
+      "loss_partial": 0.06298828125,
+      "step": 433
+    },
+    {
+      "epoch": 0.22682032477737035,
+      "loss_partial": 0.056396484375,
+      "step": 433
+    },
+    {
+      "epoch": 0.22734415924567836,
+      "grad_norm": 0.29416847229003906,
+      "learning_rate": 1.1335078534031414e-05,
+      "loss": 0.121,
+      "step": 434
+    },
+    {
+      "epoch": 0.22734415924567836,
+      "loss_partial": 0.05224609375,
+      "step": 434
+    },
+    {
+      "epoch": 0.22734415924567836,
+      "loss_partial": 0.06982421875,
+      "step": 434
+    },
+    {
+      "epoch": 0.2278679937139864,
+      "grad_norm": 0.4029017388820648,
+      "learning_rate": 1.1361256544502618e-05,
+      "loss": 0.1179,
+      "step": 435
+    },
+    {
+      "epoch": 0.2278679937139864,
+      "loss_partial": 0.06005859375,
+      "step": 435
+    },
+    {
+      "epoch": 0.2278679937139864,
+      "loss_partial": 0.06298828125,
+      "step": 435
+    },
+    {
+      "epoch": 0.2283918281822944,
+      "grad_norm": 0.35979095101356506,
+      "learning_rate": 1.1387434554973822e-05,
+      "loss": 0.1202,
+      "step": 436
+    },
+    {
+      "epoch": 0.2283918281822944,
+      "loss_partial": 0.052734375,
+      "step": 436
+    },
+    {
+      "epoch": 0.2283918281822944,
+      "loss_partial": 0.0625,
+      "step": 436
+    },
+    {
+      "epoch": 0.2289156626506024,
+      "grad_norm": 0.425813764333725,
+      "learning_rate": 1.1413612565445028e-05,
+      "loss": 0.1209,
+      "step": 437
+    },
+    {
+      "epoch": 0.2289156626506024,
+      "loss_partial": 0.060302734375,
+      "step": 437
+    },
+    {
+      "epoch": 0.2289156626506024,
+      "loss_partial": 0.07470703125,
+      "step": 437
+    },
+    {
+      "epoch": 0.22943949711891043,
+      "grad_norm": 0.24648378789424896,
+      "learning_rate": 1.1439790575916232e-05,
+      "loss": 0.121,
+      "step": 438
+    },
+    {
+      "epoch": 0.22943949711891043,
+      "loss_partial": 0.0634765625,
+      "step": 438
+    },
+    {
+      "epoch": 0.22943949711891043,
+      "loss_partial": 0.0654296875,
+      "step": 438
+    },
+    {
+      "epoch": 0.22996333158721843,
+      "grad_norm": 0.4888686239719391,
+      "learning_rate": 1.1465968586387436e-05,
+      "loss": 0.1245,
+      "step": 439
+    },
+    {
+      "epoch": 0.22996333158721843,
+      "loss_partial": 0.057861328125,
+      "step": 439
+    },
+    {
+      "epoch": 0.22996333158721843,
+      "loss_partial": 0.06982421875,
+      "step": 439
+    },
+    {
+      "epoch": 0.23048716605552647,
+      "grad_norm": 0.5359091758728027,
+      "learning_rate": 1.149214659685864e-05,
+      "loss": 0.1196,
+      "step": 440
+    },
+    {
+      "epoch": 0.23048716605552647,
+      "loss_partial": 0.068359375,
+      "step": 440
+    },
+    {
+      "epoch": 0.23048716605552647,
+      "loss_partial": 0.05615234375,
+      "step": 440
+    },
+    {
+      "epoch": 0.23101100052383447,
+      "grad_norm": 0.4936372935771942,
+      "learning_rate": 1.1518324607329845e-05,
+      "loss": 0.126,
+      "step": 441
+    },
+    {
+      "epoch": 0.23101100052383447,
+      "loss_partial": 0.07080078125,
+      "step": 441
+    },
+    {
+      "epoch": 0.23101100052383447,
+      "loss_partial": 0.05224609375,
+      "step": 441
+    },
+    {
+      "epoch": 0.23153483499214247,
+      "grad_norm": 0.7331179976463318,
+      "learning_rate": 1.1544502617801047e-05,
+      "loss": 0.1222,
+      "step": 442
+    },
+    {
+      "epoch": 0.23153483499214247,
+      "loss_partial": 0.05810546875,
+      "step": 442
+    },
+    {
+      "epoch": 0.23153483499214247,
+      "loss_partial": 0.053955078125,
+      "step": 442
+    },
+    {
+      "epoch": 0.2320586694604505,
+      "grad_norm": 0.275703102350235,
+      "learning_rate": 1.1570680628272253e-05,
+      "loss": 0.1154,
+      "step": 443
+    },
+    {
+      "epoch": 0.2320586694604505,
+      "loss_partial": 0.061767578125,
+      "step": 443
+    },
+    {
+      "epoch": 0.2320586694604505,
+      "loss_partial": 0.07080078125,
+      "step": 443
+    },
+    {
+      "epoch": 0.2325825039287585,
+      "grad_norm": 0.6398016214370728,
+      "learning_rate": 1.1596858638743455e-05,
+      "loss": 0.1204,
+      "step": 444
+    },
+    {
+      "epoch": 0.2325825039287585,
+      "loss_partial": 0.0625,
+      "step": 444
+    },
+    {
+      "epoch": 0.2325825039287585,
+      "loss_partial": 0.06396484375,
+      "step": 444
+    },
+    {
+      "epoch": 0.23310633839706651,
+      "grad_norm": 0.2759239077568054,
+      "learning_rate": 1.162303664921466e-05,
+      "loss": 0.1195,
+      "step": 445
+    },
+    {
+      "epoch": 0.23310633839706651,
+      "loss_partial": 0.056884765625,
+      "step": 445
+    },
+    {
+      "epoch": 0.23310633839706651,
+      "loss_partial": 0.0537109375,
+      "step": 445
+    },
+    {
+      "epoch": 0.23363017286537455,
+      "grad_norm": 0.7155484557151794,
+      "learning_rate": 1.1649214659685866e-05,
+      "loss": 0.1213,
+      "step": 446
+    },
+    {
+      "epoch": 0.23363017286537455,
+      "loss_partial": 0.06298828125,
+      "step": 446
+    },
+    {
+      "epoch": 0.23363017286537455,
+      "loss_partial": 0.06103515625,
+      "step": 446
+    },
+    {
+      "epoch": 0.23415400733368255,
+      "grad_norm": 0.3971101641654968,
+      "learning_rate": 1.1675392670157069e-05,
+      "loss": 0.1206,
+      "step": 447
+    },
+    {
+      "epoch": 0.23415400733368255,
+      "loss_partial": 0.0615234375,
+      "step": 447
+    },
+    {
+      "epoch": 0.23415400733368255,
+      "loss_partial": 0.058349609375,
+      "step": 447
+    },
+    {
+      "epoch": 0.23467784180199058,
+      "grad_norm": 0.7452484965324402,
+      "learning_rate": 1.1701570680628274e-05,
+      "loss": 0.1213,
+      "step": 448
+    },
+    {
+      "epoch": 0.23467784180199058,
+      "loss_partial": 0.05712890625,
+      "step": 448
+    },
+    {
+      "epoch": 0.23467784180199058,
+      "loss_partial": 0.0693359375,
+      "step": 448
+    },
+    {
+      "epoch": 0.2352016762702986,
+      "grad_norm": 0.5112625360488892,
+      "learning_rate": 1.1727748691099476e-05,
+      "loss": 0.1225,
+      "step": 449
+    },
+    {
+      "epoch": 0.2352016762702986,
+      "loss_partial": 0.06591796875,
+      "step": 449
+    },
+    {
+      "epoch": 0.2352016762702986,
+      "loss_partial": 0.06689453125,
+      "step": 449
+    },
+    {
+      "epoch": 0.2357255107386066,
+      "grad_norm": 0.7447317838668823,
+      "learning_rate": 1.1753926701570682e-05,
+      "loss": 0.1197,
+      "step": 450
+    },
+    {
+      "epoch": 0.2357255107386066,
+      "loss_partial": 0.06298828125,
+      "step": 450
+    },
+    {
+      "epoch": 0.2357255107386066,
+      "loss_partial": 0.057861328125,
+      "step": 450
+    },
+    {
+      "epoch": 0.23624934520691462,
+      "grad_norm": 0.5700985789299011,
+      "learning_rate": 1.1780104712041886e-05,
+      "loss": 0.1194,
+      "step": 451
+    },
+    {
+      "epoch": 0.23624934520691462,
+      "loss_partial": 0.0693359375,
+      "step": 451
+    },
+    {
+      "epoch": 0.23624934520691462,
+      "loss_partial": 0.059326171875,
+      "step": 451
+    },
+    {
+      "epoch": 0.23677317967522263,
+      "grad_norm": 0.6297487616539001,
+      "learning_rate": 1.180628272251309e-05,
+      "loss": 0.1202,
+      "step": 452
+    },
+    {
+      "epoch": 0.23677317967522263,
+      "loss_partial": 0.06396484375,
+      "step": 452
+    },
+    {
+      "epoch": 0.23677317967522263,
+      "loss_partial": 0.057861328125,
+      "step": 452
+    },
+    {
+      "epoch": 0.23729701414353063,
+      "grad_norm": 0.8351492285728455,
+      "learning_rate": 1.1832460732984294e-05,
+      "loss": 0.1207,
+      "step": 453
+    },
+    {
+      "epoch": 0.23729701414353063,
+      "loss_partial": 0.061767578125,
+      "step": 453
+    },
+    {
+      "epoch": 0.23729701414353063,
+      "loss_partial": 0.059814453125,
+      "step": 453
+    },
+    {
+      "epoch": 0.23782084861183866,
+      "grad_norm": 0.38295379281044006,
+      "learning_rate": 1.1858638743455498e-05,
+      "loss": 0.1272,
+      "step": 454
+    },
+    {
+      "epoch": 0.23782084861183866,
+      "loss_partial": 0.0625,
+      "step": 454
+    },
+    {
+      "epoch": 0.23782084861183866,
+      "loss_partial": 0.056640625,
+      "step": 454
+    },
+    {
+      "epoch": 0.23834468308014667,
+      "grad_norm": 0.39443227648735046,
+      "learning_rate": 1.1884816753926702e-05,
+      "loss": 0.1255,
+      "step": 455
+    },
+    {
+      "epoch": 0.23834468308014667,
+      "loss_partial": 0.062255859375,
+      "step": 455
+    },
+    {
+      "epoch": 0.23834468308014667,
+      "loss_partial": 0.055419921875,
+      "step": 455
+    },
+    {
+      "epoch": 0.2388685175484547,
+      "grad_norm": 0.2795475721359253,
+      "learning_rate": 1.1910994764397907e-05,
+      "loss": 0.118,
+      "step": 456
+    },
+    {
+      "epoch": 0.2388685175484547,
+      "loss_partial": 0.056396484375,
+      "step": 456
+    },
+    {
+      "epoch": 0.2388685175484547,
+      "loss_partial": 0.07861328125,
+      "step": 456
+    },
+    {
+      "epoch": 0.2393923520167627,
+      "grad_norm": 0.6722269058227539,
+      "learning_rate": 1.193717277486911e-05,
+      "loss": 0.1177,
+      "step": 457
+    },
+    {
+      "epoch": 0.2393923520167627,
+      "loss_partial": 0.058349609375,
+      "step": 457
+    },
+    {
+      "epoch": 0.2393923520167627,
+      "loss_partial": 0.06494140625,
+      "step": 457
+    },
+    {
+      "epoch": 0.2399161864850707,
+      "grad_norm": 0.6604887247085571,
+      "learning_rate": 1.1963350785340315e-05,
+      "loss": 0.124,
+      "step": 458
+    },
+    {
+      "epoch": 0.2399161864850707,
+      "loss_partial": 0.054443359375,
+      "step": 458
+    },
+    {
+      "epoch": 0.2399161864850707,
+      "loss_partial": 0.046630859375,
+      "step": 458
+    },
+    {
+      "epoch": 0.24044002095337874,
+      "grad_norm": 0.6658053398132324,
+      "learning_rate": 1.198952879581152e-05,
+      "loss": 0.1183,
+      "step": 459
+    },
+    {
+      "epoch": 0.24044002095337874,
+      "loss_partial": 0.05615234375,
+      "step": 459
+    },
+    {
+      "epoch": 0.24044002095337874,
+      "loss_partial": 0.058837890625,
+      "step": 459
+    },
+    {
+      "epoch": 0.24096385542168675,
+      "grad_norm": 1.1427894830703735,
+      "learning_rate": 1.2015706806282723e-05,
+      "loss": 0.1236,
+      "step": 460
+    },
+    {
+      "epoch": 0.24096385542168675,
+      "loss_partial": 0.058349609375,
+      "step": 460
+    },
+    {
+      "epoch": 0.24096385542168675,
+      "loss_partial": 0.059326171875,
+      "step": 460
+    },
+    {
+      "epoch": 0.24148768988999475,
+      "grad_norm": 0.6826235055923462,
+      "learning_rate": 1.2041884816753929e-05,
+      "loss": 0.1192,
+      "step": 461
+    },
+    {
+      "epoch": 0.24148768988999475,
+      "loss_partial": 0.056640625,
+      "step": 461
+    },
+    {
+      "epoch": 0.24148768988999475,
+      "loss_partial": 0.054443359375,
+      "step": 461
+    },
+    {
+      "epoch": 0.24201152435830278,
+      "grad_norm": 1.1548488140106201,
+      "learning_rate": 1.206806282722513e-05,
+      "loss": 0.1167,
+      "step": 462
+    },
+    {
+      "epoch": 0.24201152435830278,
+      "loss_partial": 0.0546875,
+      "step": 462
+    },
+    {
+      "epoch": 0.24201152435830278,
+      "loss_partial": 0.0556640625,
+      "step": 462
+    },
+    {
+      "epoch": 0.2425353588266108,
+      "grad_norm": 0.480056494474411,
+      "learning_rate": 1.2094240837696336e-05,
+      "loss": 0.1216,
+      "step": 463
+    },
+    {
+      "epoch": 0.2425353588266108,
+      "loss_partial": 0.055908203125,
+      "step": 463
+    },
+    {
+      "epoch": 0.2425353588266108,
+      "loss_partial": 0.056884765625,
+      "step": 463
+    },
+    {
+      "epoch": 0.24305919329491882,
+      "grad_norm": 1.1184483766555786,
+      "learning_rate": 1.212041884816754e-05,
+      "loss": 0.1206,
+      "step": 464
+    },
+    {
+      "epoch": 0.24305919329491882,
+      "loss_partial": 0.0634765625,
+      "step": 464
+    },
+    {
+      "epoch": 0.24305919329491882,
+      "loss_partial": 0.057861328125,
+      "step": 464
+    },
+    {
+      "epoch": 0.24358302776322682,
+      "grad_norm": 0.6215450763702393,
+      "learning_rate": 1.2146596858638744e-05,
+      "loss": 0.12,
+      "step": 465
+    },
+    {
+      "epoch": 0.24358302776322682,
+      "loss_partial": 0.05517578125,
+      "step": 465
+    },
+    {
+      "epoch": 0.24358302776322682,
+      "loss_partial": 0.0576171875,
+      "step": 465
+    },
+    {
+      "epoch": 0.24410686223153483,
+      "grad_norm": 0.8275788426399231,
+      "learning_rate": 1.2172774869109948e-05,
+      "loss": 0.1172,
+      "step": 466
+    },
+    {
+      "epoch": 0.24410686223153483,
+      "loss_partial": 0.054931640625,
+      "step": 466
+    },
+    {
+      "epoch": 0.24410686223153483,
+      "loss_partial": 0.060546875,
+      "step": 466
+    },
+    {
+      "epoch": 0.24463069669984286,
+      "grad_norm": 1.133435606956482,
+      "learning_rate": 1.2198952879581152e-05,
+      "loss": 0.1208,
+      "step": 467
+    },
+    {
+      "epoch": 0.24463069669984286,
+      "loss_partial": 0.060302734375,
+      "step": 467
+    },
+    {
+      "epoch": 0.24463069669984286,
+      "loss_partial": 0.0693359375,
+      "step": 467
+    },
+    {
+      "epoch": 0.24515453116815086,
+      "grad_norm": 1.0361342430114746,
+      "learning_rate": 1.2225130890052356e-05,
+      "loss": 0.1219,
+      "step": 468
+    },
+    {
+      "epoch": 0.24515453116815086,
+      "loss_partial": 0.06396484375,
+      "step": 468
+    },
+    {
+      "epoch": 0.24515453116815086,
+      "loss_partial": 0.0625,
+      "step": 468
+    },
+    {
+      "epoch": 0.24567836563645887,
+      "grad_norm": 0.8327014446258545,
+      "learning_rate": 1.2251308900523562e-05,
+      "loss": 0.1189,
+      "step": 469
+    },
+    {
+      "epoch": 0.24567836563645887,
+      "loss_partial": 0.0576171875,
+      "step": 469
+    },
+    {
+      "epoch": 0.24567836563645887,
+      "loss_partial": 0.0693359375,
+      "step": 469
+    },
+    {
+      "epoch": 0.2462022001047669,
+      "grad_norm": 0.5846191644668579,
+      "learning_rate": 1.2277486910994766e-05,
+      "loss": 0.1223,
+      "step": 470
+    },
+    {
+      "epoch": 0.2462022001047669,
+      "loss_partial": 0.0576171875,
+      "step": 470
+    },
+    {
+      "epoch": 0.2462022001047669,
+      "loss_partial": 0.058349609375,
+      "step": 470
+    },
+    {
+      "epoch": 0.2467260345730749,
+      "grad_norm": 0.620318591594696,
+      "learning_rate": 1.230366492146597e-05,
+      "loss": 0.118,
+      "step": 471
+    },
+    {
+      "epoch": 0.2467260345730749,
+      "loss_partial": 0.06396484375,
+      "step": 471
+    },
+    {
+      "epoch": 0.2467260345730749,
+      "loss_partial": 0.06591796875,
+      "step": 471
+    },
+    {
+      "epoch": 0.24724986904138294,
+      "grad_norm": 0.6793543696403503,
+      "learning_rate": 1.2329842931937175e-05,
+      "loss": 0.1197,
+      "step": 472
+    },
+    {
+      "epoch": 0.24724986904138294,
+      "loss_partial": 0.06298828125,
+      "step": 472
+    },
+    {
+      "epoch": 0.24724986904138294,
+      "loss_partial": 0.052978515625,
+      "step": 472
+    },
+    {
+      "epoch": 0.24777370350969094,
+      "grad_norm": 0.742216944694519,
+      "learning_rate": 1.2356020942408377e-05,
+      "loss": 0.1201,
+      "step": 473
+    },
+    {
+      "epoch": 0.24777370350969094,
+      "loss_partial": 0.0673828125,
+      "step": 473
+    },
+    {
+      "epoch": 0.24777370350969094,
+      "loss_partial": 0.05615234375,
+      "step": 473
+    },
+    {
+      "epoch": 0.24829753797799894,
+      "grad_norm": 0.6220785975456238,
+      "learning_rate": 1.2382198952879583e-05,
+      "loss": 0.122,
+      "step": 474
+    },
+    {
+      "epoch": 0.24829753797799894,
+      "loss_partial": 0.0625,
+      "step": 474
+    },
+    {
+      "epoch": 0.24829753797799894,
+      "loss_partial": 0.05615234375,
+      "step": 474
+    },
+    {
+      "epoch": 0.24882137244630698,
+      "grad_norm": 0.7873874306678772,
+      "learning_rate": 1.2408376963350785e-05,
+      "loss": 0.1181,
+      "step": 475
+    },
+    {
+      "epoch": 0.24882137244630698,
+      "loss_partial": 0.05517578125,
+      "step": 475
+    },
+    {
+      "epoch": 0.24882137244630698,
+      "loss_partial": 0.059326171875,
+      "step": 475
+    },
+    {
+      "epoch": 0.24934520691461498,
+      "grad_norm": 0.5392262935638428,
+      "learning_rate": 1.243455497382199e-05,
+      "loss": 0.1199,
+      "step": 476
+    },
+    {
+      "epoch": 0.24934520691461498,
+      "loss_partial": 0.0869140625,
+      "step": 476
+    },
+    {
+      "epoch": 0.24934520691461498,
+      "loss_partial": 0.05908203125,
+      "step": 476
+    },
+    {
+      "epoch": 0.24986904138292299,
+      "grad_norm": 0.6925792098045349,
+      "learning_rate": 1.2460732984293196e-05,
+      "loss": 0.1226,
+      "step": 477
+    },
+    {
+      "epoch": 0.24986904138292299,
+      "loss_partial": 0.06005859375,
+      "step": 477
+    },
+    {
+      "epoch": 0.24986904138292299,
+      "loss_partial": 0.058837890625,
+      "step": 477
+    },
+    {
+      "epoch": 0.250392875851231,
+      "grad_norm": 0.30498865246772766,
+      "learning_rate": 1.2486910994764399e-05,
+      "loss": 0.1172,
+      "step": 478
+    },
+    {
+      "epoch": 0.250392875851231,
+      "loss_partial": 0.06396484375,
+      "step": 478
+    },
+    {
+      "epoch": 0.250392875851231,
+      "loss_partial": 0.0693359375,
+      "step": 478
+    },
+    {
+      "epoch": 0.25091671031953905,
+      "grad_norm": 0.5977203845977783,
+      "learning_rate": 1.2513089005235604e-05,
+      "loss": 0.1196,
+      "step": 479
+    },
+    {
+      "epoch": 0.25091671031953905,
+      "loss_partial": 0.059326171875,
+      "step": 479
+    },
+    {
+      "epoch": 0.25091671031953905,
+      "loss_partial": 0.052001953125,
+      "step": 479
+    },
+    {
+      "epoch": 0.25144054478784705,
+      "grad_norm": 0.6122121810913086,
+      "learning_rate": 1.2539267015706806e-05,
+      "loss": 0.1219,
+      "step": 480
+    },
+    {
+      "epoch": 0.25144054478784705,
+      "loss_partial": 0.06396484375,
+      "step": 480
+    },
+    {
+      "epoch": 0.25144054478784705,
+      "loss_partial": 0.060302734375,
+      "step": 480
+    },
+    {
+      "epoch": 0.25196437925615506,
+      "grad_norm": 0.32883763313293457,
+      "learning_rate": 1.2565445026178012e-05,
+      "loss": 0.1163,
+      "step": 481
+    },
+    {
+      "epoch": 0.25196437925615506,
+      "loss_partial": 0.056640625,
+      "step": 481
+    },
+    {
+      "epoch": 0.25196437925615506,
+      "loss_partial": 0.0634765625,
+      "step": 481
+    },
+    {
+      "epoch": 0.25248821372446306,
+      "grad_norm": 0.8023962378501892,
+      "learning_rate": 1.2591623036649216e-05,
+      "loss": 0.1228,
+      "step": 482
+    },
+    {
+      "epoch": 0.25248821372446306,
+      "loss_partial": 0.05126953125,
+      "step": 482
+    },
+    {
+      "epoch": 0.25248821372446306,
+      "loss_partial": 0.0634765625,
+      "step": 482
+    },
+    {
+      "epoch": 0.25301204819277107,
+      "grad_norm": 0.6063456535339355,
+      "learning_rate": 1.261780104712042e-05,
+      "loss": 0.1219,
+      "step": 483
+    },
+    {
+      "epoch": 0.25301204819277107,
+      "loss_partial": 0.06689453125,
+      "step": 483
+    },
+    {
+      "epoch": 0.25301204819277107,
+      "loss_partial": 0.057861328125,
+      "step": 483
+    },
+    {
+      "epoch": 0.2535358826610791,
+      "grad_norm": 0.5791170001029968,
+      "learning_rate": 1.2643979057591624e-05,
+      "loss": 0.1194,
+      "step": 484
+    },
+    {
+      "epoch": 0.2535358826610791,
+      "loss_partial": 0.046875,
+      "step": 484
+    },
+    {
+      "epoch": 0.2535358826610791,
+      "loss_partial": 0.06005859375,
+      "step": 484
+    },
+    {
+      "epoch": 0.25405971712938713,
+      "grad_norm": 1.4599140882492065,
+      "learning_rate": 1.2670157068062828e-05,
+      "loss": 0.119,
+      "step": 485
+    },
+    {
+      "epoch": 0.25405971712938713,
+      "loss_partial": 0.06640625,
+      "step": 485
+    },
+    {
+      "epoch": 0.25405971712938713,
+      "loss_partial": 0.058349609375,
+      "step": 485
+    },
+    {
+      "epoch": 0.25458355159769513,
+      "grad_norm": 1.1946654319763184,
+      "learning_rate": 1.2696335078534032e-05,
+      "loss": 0.1173,
+      "step": 486
+    },
+    {
+      "epoch": 0.25458355159769513,
+      "loss_partial": 0.0654296875,
+      "step": 486
+    },
+    {
+      "epoch": 0.25458355159769513,
+      "loss_partial": 0.06494140625,
+      "step": 486
+    },
+    {
+      "epoch": 0.25510738606600314,
+      "grad_norm": 1.0757167339324951,
+      "learning_rate": 1.2722513089005237e-05,
+      "loss": 0.1205,
+      "step": 487
+    },
+    {
+      "epoch": 0.25510738606600314,
+      "loss_partial": 0.060791015625,
+      "step": 487
+    },
+    {
+      "epoch": 0.25510738606600314,
+      "loss_partial": 0.05810546875,
+      "step": 487
+    },
+    {
+      "epoch": 0.25563122053431114,
+      "grad_norm": 0.67076176404953,
+      "learning_rate": 1.274869109947644e-05,
+      "loss": 0.119,
+      "step": 488
+    },
+    {
+      "epoch": 0.25563122053431114,
+      "loss_partial": 0.05810546875,
+      "step": 488
+    },
+    {
+      "epoch": 0.25563122053431114,
+      "loss_partial": 0.06298828125,
+      "step": 488
+    },
+    {
+      "epoch": 0.25615505500261915,
+      "grad_norm": 1.1724108457565308,
+      "learning_rate": 1.2774869109947645e-05,
+      "loss": 0.118,
+      "step": 489
+    },
+    {
+      "epoch": 0.25615505500261915,
+      "loss_partial": 0.064453125,
+      "step": 489
+    },
+    {
+      "epoch": 0.25615505500261915,
+      "loss_partial": 0.064453125,
+      "step": 489
+    },
+    {
+      "epoch": 0.2566788894709272,
+      "grad_norm": 0.4799434244632721,
+      "learning_rate": 1.280104712041885e-05,
+      "loss": 0.1202,
+      "step": 490
+    },
+    {
+      "epoch": 0.2566788894709272,
+      "loss_partial": 0.05029296875,
+      "step": 490
+    },
+    {
+      "epoch": 0.2566788894709272,
+      "loss_partial": 0.0673828125,
+      "step": 490
+    },
+    {
+      "epoch": 0.2572027239392352,
+      "grad_norm": 0.9915763735771179,
+      "learning_rate": 1.2827225130890053e-05,
+      "loss": 0.1205,
+      "step": 491
+    },
+    {
+      "epoch": 0.2572027239392352,
+      "loss_partial": 0.05859375,
+      "step": 491
+    },
+    {
+      "epoch": 0.2572027239392352,
+      "loss_partial": 0.0576171875,
+      "step": 491
+    },
+    {
+      "epoch": 0.2577265584075432,
+      "grad_norm": 0.9803262948989868,
+      "learning_rate": 1.2853403141361259e-05,
+      "loss": 0.1182,
+      "step": 492
+    },
+    {
+      "epoch": 0.2577265584075432,
+      "loss_partial": 0.056640625,
+      "step": 492
+    },
+    {
+      "epoch": 0.2577265584075432,
+      "loss_partial": 0.052001953125,
+      "step": 492
+    },
+    {
+      "epoch": 0.2582503928758512,
+      "grad_norm": 0.4656517207622528,
+      "learning_rate": 1.287958115183246e-05,
+      "loss": 0.116,
+      "step": 493
+    },
+    {
+      "epoch": 0.2582503928758512,
+      "loss_partial": 0.06298828125,
+      "step": 493
+    },
+    {
+      "epoch": 0.2582503928758512,
+      "loss_partial": 0.06494140625,
+      "step": 493
+    },
+    {
+      "epoch": 0.2587742273441592,
+      "grad_norm": 0.7359094619750977,
+      "learning_rate": 1.2905759162303666e-05,
+      "loss": 0.1188,
+      "step": 494
+    },
+    {
+      "epoch": 0.2587742273441592,
+      "loss_partial": 0.064453125,
+      "step": 494
+    },
+    {
+      "epoch": 0.2587742273441592,
+      "loss_partial": 0.050537109375,
+      "step": 494
+    },
+    {
+      "epoch": 0.2592980618124673,
+      "grad_norm": 0.7898105978965759,
+      "learning_rate": 1.293193717277487e-05,
+      "loss": 0.1194,
+      "step": 495
+    },
+    {
+      "epoch": 0.2592980618124673,
+      "loss_partial": 0.0625,
+      "step": 495
+    },
+    {
+      "epoch": 0.2592980618124673,
+      "loss_partial": 0.055908203125,
+      "step": 495
+    },
+    {
+      "epoch": 0.2598218962807753,
+      "grad_norm": 1.489152193069458,
+      "learning_rate": 1.2958115183246074e-05,
+      "loss": 0.117,
+      "step": 496
+    },
+    {
+      "epoch": 0.2598218962807753,
+      "loss_partial": 0.05712890625,
+      "step": 496
+    },
+    {
+      "epoch": 0.2598218962807753,
+      "loss_partial": 0.07080078125,
+      "step": 496
+    },
+    {
+      "epoch": 0.2603457307490833,
+      "grad_norm": 1.0152653455734253,
+      "learning_rate": 1.2984293193717278e-05,
+      "loss": 0.1192,
+      "step": 497
+    },
+    {
+      "epoch": 0.2603457307490833,
+      "loss_partial": 0.053955078125,
+      "step": 497
+    },
+    {
+      "epoch": 0.2603457307490833,
+      "loss_partial": 0.06982421875,
+      "step": 497
+    },
+    {
+      "epoch": 0.2608695652173913,
+      "grad_norm": 0.5420125722885132,
+      "learning_rate": 1.3010471204188482e-05,
+      "loss": 0.1209,
+      "step": 498
+    },
+    {
+      "epoch": 0.2608695652173913,
+      "loss_partial": 0.05908203125,
+      "step": 498
+    },
+    {
+      "epoch": 0.2608695652173913,
+      "loss_partial": 0.0556640625,
+      "step": 498
+    },
+    {
+      "epoch": 0.2613933996856993,
+      "grad_norm": 0.7523015141487122,
+      "learning_rate": 1.3036649214659686e-05,
+      "loss": 0.1164,
+      "step": 499
+    },
+    {
+      "epoch": 0.2613933996856993,
+      "loss_partial": 0.05126953125,
+      "step": 499
+    },
+    {
+      "epoch": 0.2613933996856993,
+      "loss_partial": 0.05517578125,
+      "step": 499
+    },
+    {
+      "epoch": 0.26191723415400736,
+      "grad_norm": 0.5912072658538818,
+      "learning_rate": 1.3062827225130892e-05,
+      "loss": 0.1161,
+      "step": 500
+    },
+    {
+      "epoch": 0.26191723415400736,
+      "loss_partial": 0.0615234375,
+      "step": 500
+    },
+    {
+      "epoch": 0.26191723415400736,
+      "loss_partial": 0.061279296875,
+      "step": 500
+    },
+    {
+      "epoch": 0.26244106862231537,
+      "grad_norm": 0.6947856545448303,
+      "learning_rate": 1.3089005235602094e-05,
+      "loss": 0.1195,
+      "step": 501
+    },
+    {
+      "epoch": 0.26244106862231537,
+      "loss_partial": 0.06494140625,
+      "step": 501
+    },
+    {
+      "epoch": 0.26244106862231537,
+      "loss_partial": 0.0625,
+      "step": 501
+    },
+    {
+      "epoch": 0.26296490309062337,
+      "grad_norm": 0.5476560592651367,
+      "learning_rate": 1.31151832460733e-05,
+      "loss": 0.117,
+      "step": 502
+    },
+    {
+      "epoch": 0.26296490309062337,
+      "loss_partial": 0.060546875,
+      "step": 502
+    },
+    {
+      "epoch": 0.26296490309062337,
+      "loss_partial": 0.0546875,
+      "step": 502
+    },
+    {
+      "epoch": 0.2634887375589314,
+      "grad_norm": 1.1056994199752808,
+      "learning_rate": 1.3141361256544505e-05,
+      "loss": 0.1202,
+      "step": 503
+    },
+    {
+      "epoch": 0.2634887375589314,
+      "loss_partial": 0.0595703125,
+      "step": 503
+    },
+    {
+      "epoch": 0.2634887375589314,
+      "loss_partial": 0.0615234375,
+      "step": 503
+    },
+    {
+      "epoch": 0.2640125720272394,
+      "grad_norm": 0.7188048362731934,
+      "learning_rate": 1.3167539267015707e-05,
+      "loss": 0.1187,
+      "step": 504
+    },
+    {
+      "epoch": 0.2640125720272394,
+      "loss_partial": 0.07763671875,
+      "step": 504
+    },
+    {
+      "epoch": 0.2640125720272394,
+      "loss_partial": 0.0595703125,
+      "step": 504
+    },
+    {
+      "epoch": 0.2645364064955474,
+      "grad_norm": 1.0726604461669922,
+      "learning_rate": 1.3193717277486913e-05,
+      "loss": 0.1194,
+      "step": 505
+    },
+    {
+      "epoch": 0.2645364064955474,
+      "loss_partial": 0.0625,
+      "step": 505
+    },
+    {
+      "epoch": 0.2645364064955474,
+      "loss_partial": 0.0615234375,
+      "step": 505
+    },
+    {
+      "epoch": 0.26506024096385544,
+      "grad_norm": 1.8903950452804565,
+      "learning_rate": 1.3219895287958115e-05,
+      "loss": 0.1227,
+      "step": 506
+    },
+    {
+      "epoch": 0.26506024096385544,
+      "loss_partial": 0.056640625,
+      "step": 506
+    },
+    {
+      "epoch": 0.26506024096385544,
+      "loss_partial": 0.059814453125,
+      "step": 506
+    },
+    {
+      "epoch": 0.26558407543216345,
+      "grad_norm": 1.1368725299835205,
+      "learning_rate": 1.324607329842932e-05,
+      "loss": 0.1212,
+      "step": 507
+    },
+    {
+      "epoch": 0.26558407543216345,
+      "loss_partial": 0.0595703125,
+      "step": 507
+    },
+    {
+      "epoch": 0.26558407543216345,
+      "loss_partial": 0.060791015625,
+      "step": 507
+    },
+    {
+      "epoch": 0.26610790990047145,
+      "grad_norm": 1.2785784006118774,
+      "learning_rate": 1.3272251308900526e-05,
+      "loss": 0.1172,
+      "step": 508
+    },
+    {
+      "epoch": 0.26610790990047145,
+      "loss_partial": 0.04833984375,
+      "step": 508
+    },
+    {
+      "epoch": 0.26610790990047145,
+      "loss_partial": 0.068359375,
+      "step": 508
+    },
+    {
+      "epoch": 0.26663174436877946,
+      "grad_norm": 1.3772419691085815,
+      "learning_rate": 1.3298429319371729e-05,
+      "loss": 0.1152,
+      "step": 509
+    },
+    {
+      "epoch": 0.26663174436877946,
+      "loss_partial": 0.06201171875,
+      "step": 509
+    },
+    {
+      "epoch": 0.26663174436877946,
+      "loss_partial": 0.07080078125,
+      "step": 509
+    },
+    {
+      "epoch": 0.26715557883708746,
+      "grad_norm": 0.7208380699157715,
+      "learning_rate": 1.3324607329842934e-05,
+      "loss": 0.1185,
+      "step": 510
+    },
+    {
+      "epoch": 0.26715557883708746,
+      "loss_partial": 0.056640625,
+      "step": 510
+    },
+    {
+      "epoch": 0.26715557883708746,
+      "loss_partial": 0.058837890625,
+      "step": 510
+    },
+    {
+      "epoch": 0.2676794133053955,
+      "grad_norm": 1.5505352020263672,
+      "learning_rate": 1.3350785340314136e-05,
+      "loss": 0.118,
+      "step": 511
+    },
+    {
+      "epoch": 0.2676794133053955,
+      "loss_partial": 0.0625,
+      "step": 511
+    },
+    {
+      "epoch": 0.2676794133053955,
+      "loss_partial": 0.058837890625,
+      "step": 511
+    },
+    {
+      "epoch": 0.2682032477737035,
+      "grad_norm": 0.826796293258667,
+      "learning_rate": 1.3376963350785342e-05,
+      "loss": 0.1176,
+      "step": 512
+    },
+    {
+      "epoch": 0.2682032477737035,
+      "loss_partial": 0.053955078125,
+      "step": 512
+    },
+    {
+      "epoch": 0.2682032477737035,
+      "loss_partial": 0.0537109375,
+      "step": 512
+    },
+    {
+      "epoch": 0.26872708224201153,
+      "grad_norm": 1.6049742698669434,
+      "learning_rate": 1.3403141361256546e-05,
+      "loss": 0.1194,
+      "step": 513
+    },
+    {
+      "epoch": 0.26872708224201153,
+      "loss_partial": 0.0537109375,
+      "step": 513
+    },
+    {
+      "epoch": 0.26872708224201153,
+      "loss_partial": 0.051513671875,
+      "step": 513
+    },
+    {
+      "epoch": 0.26925091671031953,
+      "grad_norm": 1.0830178260803223,
+      "learning_rate": 1.342931937172775e-05,
+      "loss": 0.1173,
+      "step": 514
+    },
+    {
+      "epoch": 0.26925091671031953,
+      "loss_partial": 0.053955078125,
+      "step": 514
+    },
+    {
+      "epoch": 0.26925091671031953,
+      "loss_partial": 0.054931640625,
+      "step": 514
+    },
+    {
+      "epoch": 0.26977475117862754,
+      "grad_norm": 0.956086277961731,
+      "learning_rate": 1.3455497382198954e-05,
+      "loss": 0.1196,
+      "step": 515
+    },
+    {
+      "epoch": 0.26977475117862754,
+      "loss_partial": 0.0556640625,
+      "step": 515
+    },
+    {
+      "epoch": 0.26977475117862754,
+      "loss_partial": 0.06005859375,
+      "step": 515
+    },
+    {
+      "epoch": 0.27029858564693554,
+      "grad_norm": 0.8048759698867798,
+      "learning_rate": 1.3481675392670158e-05,
+      "loss": 0.1133,
+      "step": 516
+    },
+    {
+      "epoch": 0.27029858564693554,
+      "loss_partial": 0.0576171875,
+      "step": 516
+    },
+    {
+      "epoch": 0.27029858564693554,
+      "loss_partial": 0.0634765625,
+      "step": 516
+    },
+    {
+      "epoch": 0.2708224201152436,
+      "grad_norm": 1.2805330753326416,
+      "learning_rate": 1.3507853403141362e-05,
+      "loss": 0.1187,
+      "step": 517
+    },
+    {
+      "epoch": 0.2708224201152436,
+      "loss_partial": 0.076171875,
+      "step": 517
+    },
+    {
+      "epoch": 0.2708224201152436,
+      "loss_partial": 0.05810546875,
+      "step": 517
+    },
+    {
+      "epoch": 0.2713462545835516,
+      "grad_norm": 1.2382632493972778,
+      "learning_rate": 1.3534031413612567e-05,
+      "loss": 0.1218,
+      "step": 518
+    },
+    {
+      "epoch": 0.2713462545835516,
+      "loss_partial": 0.06103515625,
+      "step": 518
+    },
+    {
+      "epoch": 0.2713462545835516,
+      "loss_partial": 0.059326171875,
+      "step": 518
+    },
+    {
+      "epoch": 0.2718700890518596,
+      "grad_norm": 0.939368724822998,
+      "learning_rate": 1.356020942408377e-05,
+      "loss": 0.1214,
+      "step": 519
+    },
+    {
+      "epoch": 0.2718700890518596,
+      "loss_partial": 0.057861328125,
+      "step": 519
+    },
+    {
+      "epoch": 0.2718700890518596,
+      "loss_partial": 0.059814453125,
+      "step": 519
+    },
+    {
+      "epoch": 0.2723939235201676,
+      "grad_norm": 1.2878472805023193,
+      "learning_rate": 1.3586387434554975e-05,
+      "loss": 0.1145,
+      "step": 520
+    },
+    {
+      "epoch": 0.2723939235201676,
+      "loss_partial": 0.064453125,
+      "step": 520
+    },
+    {
+      "epoch": 0.2723939235201676,
+      "loss_partial": 0.0615234375,
+      "step": 520
+    },
+    {
+      "epoch": 0.2729177579884756,
+      "grad_norm": 0.4714389443397522,
+      "learning_rate": 1.361256544502618e-05,
+      "loss": 0.1209,
+      "step": 521
+    },
+    {
+      "epoch": 0.2729177579884756,
+      "loss_partial": 0.055419921875,
+      "step": 521
+    },
+    {
+      "epoch": 0.2729177579884756,
+      "loss_partial": 0.0546875,
+      "step": 521
+    },
+    {
+      "epoch": 0.2734415924567837,
+      "grad_norm": 0.8297640681266785,
+      "learning_rate": 1.3638743455497383e-05,
+      "loss": 0.1151,
+      "step": 522
+    },
+    {
+      "epoch": 0.2734415924567837,
+      "loss_partial": 0.058349609375,
+      "step": 522
+    },
+    {
+      "epoch": 0.2734415924567837,
+      "loss_partial": 0.060791015625,
+      "step": 522
+    },
+    {
+      "epoch": 0.2739654269250917,
+      "grad_norm": 0.7212359309196472,
+      "learning_rate": 1.3664921465968589e-05,
+      "loss": 0.1179,
+      "step": 523
+    },
+    {
+      "epoch": 0.2739654269250917,
+      "loss_partial": 0.057861328125,
+      "step": 523
+    },
+    {
+      "epoch": 0.2739654269250917,
+      "loss_partial": 0.06640625,
+      "step": 523
+    },
+    {
+      "epoch": 0.2744892613933997,
+      "grad_norm": 1.1520731449127197,
+      "learning_rate": 1.369109947643979e-05,
+      "loss": 0.1246,
+      "step": 524
+    },
+    {
+      "epoch": 0.2744892613933997,
+      "loss_partial": 0.056884765625,
+      "step": 524
+    },
+    {
+      "epoch": 0.2744892613933997,
+      "loss_partial": 0.0625,
+      "step": 524
+    },
+    {
+      "epoch": 0.2750130958617077,
+      "grad_norm": 0.7087950110435486,
+      "learning_rate": 1.3717277486910996e-05,
+      "loss": 0.1199,
+      "step": 525
+    },
+    {
+      "epoch": 0.2750130958617077,
+      "loss_partial": 0.06396484375,
+      "step": 525
+    },
+    {
+      "epoch": 0.2750130958617077,
+      "loss_partial": 0.06396484375,
+      "step": 525
+    },
+    {
+      "epoch": 0.2755369303300157,
+      "grad_norm": 0.5410304069519043,
+      "learning_rate": 1.37434554973822e-05,
+      "loss": 0.1174,
+      "step": 526
+    },
+    {
+      "epoch": 0.2755369303300157,
+      "loss_partial": 0.055908203125,
+      "step": 526
+    },
+    {
+      "epoch": 0.2755369303300157,
+      "loss_partial": 0.060791015625,
+      "step": 526
+    },
+    {
+      "epoch": 0.27606076479832375,
+      "grad_norm": 0.9374832510948181,
+      "learning_rate": 1.3769633507853404e-05,
+      "loss": 0.1188,
+      "step": 527
+    },
+    {
+      "epoch": 0.27606076479832375,
+      "loss_partial": 0.053466796875,
+      "step": 527
+    },
+    {
+      "epoch": 0.27606076479832375,
+      "loss_partial": 0.052978515625,
+      "step": 527
+    },
+    {
+      "epoch": 0.27658459926663176,
+      "grad_norm": 0.5269318222999573,
+      "learning_rate": 1.3795811518324608e-05,
+      "loss": 0.1161,
+      "step": 528
+    },
+    {
+      "epoch": 0.27658459926663176,
+      "loss_partial": 0.05908203125,
+      "step": 528
+    },
+    {
+      "epoch": 0.27658459926663176,
+      "loss_partial": 0.05615234375,
+      "step": 528
+    },
+    {
+      "epoch": 0.27710843373493976,
+      "grad_norm": 0.9833793640136719,
+      "learning_rate": 1.3821989528795812e-05,
+      "loss": 0.1151,
+      "step": 529
+    },
+    {
+      "epoch": 0.27710843373493976,
+      "loss_partial": 0.057861328125,
+      "step": 529
+    },
+    {
+      "epoch": 0.27710843373493976,
+      "loss_partial": 0.048583984375,
+      "step": 529
+    },
+    {
+      "epoch": 0.27763226820324777,
+      "grad_norm": 0.9383687376976013,
+      "learning_rate": 1.3848167539267016e-05,
+      "loss": 0.1186,
+      "step": 530
+    },
+    {
+      "epoch": 0.27763226820324777,
+      "loss_partial": 0.06396484375,
+      "step": 530
+    },
+    {
+      "epoch": 0.27763226820324777,
+      "loss_partial": 0.06201171875,
+      "step": 530
+    },
+    {
+      "epoch": 0.27815610267155577,
+      "grad_norm": 0.6647336483001709,
+      "learning_rate": 1.3874345549738222e-05,
+      "loss": 0.1185,
+      "step": 531
+    },
+    {
+      "epoch": 0.27815610267155577,
+      "loss_partial": 0.050048828125,
+      "step": 531
+    },
+    {
+      "epoch": 0.27815610267155577,
+      "loss_partial": 0.0556640625,
+      "step": 531
+    },
+    {
+      "epoch": 0.2786799371398638,
+      "grad_norm": 1.3723640441894531,
+      "learning_rate": 1.3900523560209424e-05,
+      "loss": 0.1176,
+      "step": 532
+    },
+    {
+      "epoch": 0.2786799371398638,
+      "loss_partial": 0.06103515625,
+      "step": 532
+    },
+    {
+      "epoch": 0.2786799371398638,
+      "loss_partial": 0.064453125,
+      "step": 532
+    },
+    {
+      "epoch": 0.27920377160817184,
+      "grad_norm": 1.3675214052200317,
+      "learning_rate": 1.392670157068063e-05,
+      "loss": 0.116,
+      "step": 533
+    },
+    {
+      "epoch": 0.27920377160817184,
+      "loss_partial": 0.056396484375,
+      "step": 533
+    },
+    {
+      "epoch": 0.27920377160817184,
+      "loss_partial": 0.07177734375,
+      "step": 533
+    },
+    {
+      "epoch": 0.27972760607647984,
+      "grad_norm": 0.6791559457778931,
+      "learning_rate": 1.3952879581151832e-05,
+      "loss": 0.1171,
+      "step": 534
+    },
+    {
+      "epoch": 0.27972760607647984,
+      "loss_partial": 0.060302734375,
+      "step": 534
+    },
+    {
+      "epoch": 0.27972760607647984,
+      "loss_partial": 0.06591796875,
+      "step": 534
+    },
+    {
+      "epoch": 0.28025144054478784,
+      "grad_norm": 1.7702178955078125,
+      "learning_rate": 1.3979057591623037e-05,
+      "loss": 0.117,
+      "step": 535
+    },
+    {
+      "epoch": 0.28025144054478784,
+      "loss_partial": 0.057373046875,
+      "step": 535
+    },
+    {
+      "epoch": 0.28025144054478784,
+      "loss_partial": 0.050048828125,
+      "step": 535
+    },
+    {
+      "epoch": 0.28077527501309585,
+      "grad_norm": 1.818332314491272,
+      "learning_rate": 1.4005235602094243e-05,
+      "loss": 0.1155,
+      "step": 536
+    },
+    {
+      "epoch": 0.28077527501309585,
+      "loss_partial": 0.055419921875,
+      "step": 536
+    },
+    {
+      "epoch": 0.28077527501309585,
+      "loss_partial": 0.058349609375,
+      "step": 536
+    },
+    {
+      "epoch": 0.28129910948140385,
+      "grad_norm": 0.5375508069992065,
+      "learning_rate": 1.4031413612565445e-05,
+      "loss": 0.1164,
+      "step": 537
+    },
+    {
+      "epoch": 0.28129910948140385,
+      "loss_partial": 0.05712890625,
+      "step": 537
+    },
+    {
+      "epoch": 0.28129910948140385,
+      "loss_partial": 0.056884765625,
+      "step": 537
+    },
+    {
+      "epoch": 0.2818229439497119,
+      "grad_norm": 1.635233759880066,
+      "learning_rate": 1.405759162303665e-05,
+      "loss": 0.1158,
+      "step": 538
+    },
+    {
+      "epoch": 0.2818229439497119,
+      "loss_partial": 0.06640625,
+      "step": 538
+    },
+    {
+      "epoch": 0.2818229439497119,
+      "loss_partial": 0.048583984375,
+      "step": 538
+    },
+    {
+      "epoch": 0.2823467784180199,
+      "grad_norm": 1.13425612449646,
+      "learning_rate": 1.4083769633507855e-05,
+      "loss": 0.1187,
+      "step": 539
+    },
+    {
+      "epoch": 0.2823467784180199,
+      "loss_partial": 0.060302734375,
+      "step": 539
+    },
+    {
+      "epoch": 0.2823467784180199,
+      "loss_partial": 0.0673828125,
+      "step": 539
+    },
+    {
+      "epoch": 0.2828706128863279,
+      "grad_norm": 1.1286367177963257,
+      "learning_rate": 1.4109947643979059e-05,
+      "loss": 0.1164,
+      "step": 540
+    },
+    {
+      "epoch": 0.2828706128863279,
+      "loss_partial": 0.06201171875,
+      "step": 540
+    },
+    {
+      "epoch": 0.2828706128863279,
+      "loss_partial": 0.056640625,
+      "step": 540
+    },
+    {
+      "epoch": 0.2833944473546359,
+      "grad_norm": 1.9721400737762451,
+      "learning_rate": 1.4136125654450264e-05,
+      "loss": 0.1159,
+      "step": 541
+    },
+    {
+      "epoch": 0.2833944473546359,
+      "loss_partial": 0.05517578125,
+      "step": 541
+    },
+    {
+      "epoch": 0.2833944473546359,
+      "loss_partial": 0.052490234375,
+      "step": 541
+    },
+    {
+      "epoch": 0.28391828182294393,
+      "grad_norm": 0.4379724860191345,
+      "learning_rate": 1.4162303664921466e-05,
+      "loss": 0.1179,
+      "step": 542
+    },
+    {
+      "epoch": 0.28391828182294393,
+      "loss_partial": 0.05908203125,
+      "step": 542
+    },
+    {
+      "epoch": 0.28391828182294393,
+      "loss_partial": 0.057861328125,
+      "step": 542
+    },
+    {
+      "epoch": 0.284442116291252,
+      "grad_norm": 1.8200002908706665,
+      "learning_rate": 1.4188481675392672e-05,
+      "loss": 0.1131,
+      "step": 543
+    },
+    {
+      "epoch": 0.284442116291252,
+      "loss_partial": 0.0546875,
+      "step": 543
+    },
+    {
+      "epoch": 0.284442116291252,
+      "loss_partial": 0.052001953125,
+      "step": 543
+    },
+    {
+      "epoch": 0.28496595075956,
+      "grad_norm": 0.9815565347671509,
+      "learning_rate": 1.4214659685863876e-05,
+      "loss": 0.1135,
+      "step": 544
+    },
+    {
+      "epoch": 0.28496595075956,
+      "loss_partial": 0.0576171875,
+      "step": 544
+    },
+    {
+      "epoch": 0.28496595075956,
+      "loss_partial": 0.057373046875,
+      "step": 544
+    },
+    {
+      "epoch": 0.285489785227868,
+      "grad_norm": 2.1194496154785156,
+      "learning_rate": 1.424083769633508e-05,
+      "loss": 0.1192,
+      "step": 545
+    },
+    {
+      "epoch": 0.285489785227868,
+      "loss_partial": 0.0556640625,
+      "step": 545
+    },
+    {
+      "epoch": 0.285489785227868,
+      "loss_partial": 0.053466796875,
+      "step": 545
+    },
+    {
+      "epoch": 0.286013619696176,
+      "grad_norm": 1.6489022970199585,
+      "learning_rate": 1.4267015706806284e-05,
+      "loss": 0.1201,
+      "step": 546
+    },
+    {
+      "epoch": 0.286013619696176,
+      "loss_partial": 0.051513671875,
+      "step": 546
+    },
+    {
+      "epoch": 0.286013619696176,
+      "loss_partial": 0.0634765625,
+      "step": 546
+    },
+    {
+      "epoch": 0.286537454164484,
+      "grad_norm": 0.95640629529953,
+      "learning_rate": 1.4293193717277488e-05,
+      "loss": 0.118,
+      "step": 547
+    },
+    {
+      "epoch": 0.286537454164484,
+      "loss_partial": 0.0673828125,
+      "step": 547
+    },
+    {
+      "epoch": 0.286537454164484,
+      "loss_partial": 0.06787109375,
+      "step": 547
+    },
+    {
+      "epoch": 0.287061288632792,
+      "grad_norm": 1.4802289009094238,
+      "learning_rate": 1.4319371727748692e-05,
+      "loss": 0.1196,
+      "step": 548
+    },
+    {
+      "epoch": 0.287061288632792,
+      "loss_partial": 0.048095703125,
+      "step": 548
+    },
+    {
+      "epoch": 0.287061288632792,
+      "loss_partial": 0.060302734375,
+      "step": 548
+    },
+    {
+      "epoch": 0.28758512310110007,
+      "grad_norm": 0.7960492968559265,
+      "learning_rate": 1.4345549738219897e-05,
+      "loss": 0.1199,
+      "step": 549
+    },
+    {
+      "epoch": 0.28758512310110007,
+      "loss_partial": 0.05419921875,
+      "step": 549
+    },
+    {
+      "epoch": 0.28758512310110007,
+      "loss_partial": 0.06396484375,
+      "step": 549
+    },
+    {
+      "epoch": 0.2881089575694081,
+      "grad_norm": 1.168573021888733,
+      "learning_rate": 1.43717277486911e-05,
+      "loss": 0.1132,
+      "step": 550
+    },
+    {
+      "epoch": 0.2881089575694081,
+      "loss_partial": 0.06640625,
+      "step": 550
+    },
+    {
+      "epoch": 0.2881089575694081,
+      "loss_partial": 0.059326171875,
+      "step": 550
+    },
+    {
+      "epoch": 0.2886327920377161,
+      "grad_norm": 0.6939902901649475,
+      "learning_rate": 1.4397905759162305e-05,
+      "loss": 0.115,
+      "step": 551
+    },
+    {
+      "epoch": 0.2886327920377161,
+      "loss_partial": 0.053466796875,
+      "step": 551
+    },
+    {
+      "epoch": 0.2886327920377161,
+      "loss_partial": 0.05029296875,
+      "step": 551
+    },
+    {
+      "epoch": 0.2891566265060241,
+      "grad_norm": 1.4399957656860352,
+      "learning_rate": 1.442408376963351e-05,
+      "loss": 0.1214,
+      "step": 552
+    },
+    {
+      "epoch": 0.2891566265060241,
+      "loss_partial": 0.060791015625,
+      "step": 552
+    },
+    {
+      "epoch": 0.2891566265060241,
+      "loss_partial": 0.0595703125,
+      "step": 552
+    },
+    {
+      "epoch": 0.2896804609743321,
+      "grad_norm": 1.1592351198196411,
+      "learning_rate": 1.4450261780104713e-05,
+      "loss": 0.1186,
+      "step": 553
+    },
+    {
+      "epoch": 0.2896804609743321,
+      "loss_partial": 0.052001953125,
+      "step": 553
+    },
+    {
+      "epoch": 0.2896804609743321,
+      "loss_partial": 0.050537109375,
+      "step": 553
+    },
+    {
+      "epoch": 0.29020429544264015,
+      "grad_norm": 1.0691250562667847,
+      "learning_rate": 1.4476439790575919e-05,
+      "loss": 0.1181,
+      "step": 554
+    },
+    {
+      "epoch": 0.29020429544264015,
+      "loss_partial": 0.057861328125,
+      "step": 554
+    },
+    {
+      "epoch": 0.29020429544264015,
+      "loss_partial": 0.05810546875,
+      "step": 554
+    },
+    {
+      "epoch": 0.29072812991094815,
+      "grad_norm": 1.2405943870544434,
+      "learning_rate": 1.450261780104712e-05,
+      "loss": 0.1139,
+      "step": 555
+    },
+    {
+      "epoch": 0.29072812991094815,
+      "loss_partial": 0.055908203125,
+      "step": 555
+    },
+    {
+      "epoch": 0.29072812991094815,
+      "loss_partial": 0.058349609375,
+      "step": 555
+    },
+    {
+      "epoch": 0.29125196437925616,
+      "grad_norm": 1.3418548107147217,
+      "learning_rate": 1.4528795811518326e-05,
+      "loss": 0.1164,
+      "step": 556
+    },
+    {
+      "epoch": 0.29125196437925616,
+      "loss_partial": 0.0712890625,
+      "step": 556
+    },
+    {
+      "epoch": 0.29125196437925616,
+      "loss_partial": 0.060302734375,
+      "step": 556
+    },
+    {
+      "epoch": 0.29177579884756416,
+      "grad_norm": 1.4245002269744873,
+      "learning_rate": 1.455497382198953e-05,
+      "loss": 0.1195,
+      "step": 557
+    },
+    {
+      "epoch": 0.29177579884756416,
+      "loss_partial": 0.0576171875,
+      "step": 557
+    },
+    {
+      "epoch": 0.29177579884756416,
+      "loss_partial": 0.0615234375,
+      "step": 557
+    },
+    {
+      "epoch": 0.29229963331587216,
+      "grad_norm": 0.7489694952964783,
+      "learning_rate": 1.4581151832460734e-05,
+      "loss": 0.1176,
+      "step": 558
+    },
+    {
+      "epoch": 0.29229963331587216,
+      "loss_partial": 0.056396484375,
+      "step": 558
+    },
+    {
+      "epoch": 0.29229963331587216,
+      "loss_partial": 0.06201171875,
+      "step": 558
+    },
+    {
+      "epoch": 0.2928234677841802,
+      "grad_norm": 1.3087351322174072,
+      "learning_rate": 1.4607329842931938e-05,
+      "loss": 0.1172,
+      "step": 559
+    },
+    {
+      "epoch": 0.2928234677841802,
+      "loss_partial": 0.054443359375,
+      "step": 559
+    },
+    {
+      "epoch": 0.2928234677841802,
+      "loss_partial": 0.06005859375,
+      "step": 559
+    },
+    {
+      "epoch": 0.29334730225248823,
+      "grad_norm": 0.8546871542930603,
+      "learning_rate": 1.4633507853403142e-05,
+      "loss": 0.1165,
+      "step": 560
+    },
+    {
+      "epoch": 0.29334730225248823,
+      "loss_partial": 0.06201171875,
+      "step": 560
+    },
+    {
+      "epoch": 0.29334730225248823,
+      "loss_partial": 0.04931640625,
+      "step": 560
+    },
+    {
+      "epoch": 0.29387113672079623,
+      "grad_norm": 0.8943086266517639,
+      "learning_rate": 1.4659685863874346e-05,
+      "loss": 0.1115,
+      "step": 561
+    },
+    {
+      "epoch": 0.29387113672079623,
+      "loss_partial": 0.060791015625,
+      "step": 561
+    },
+    {
+      "epoch": 0.29387113672079623,
+      "loss_partial": 0.052001953125,
+      "step": 561
+    },
+    {
+      "epoch": 0.29439497118910424,
+      "grad_norm": 0.9010061621665955,
+      "learning_rate": 1.4685863874345552e-05,
+      "loss": 0.1173,
+      "step": 562
+    },
+    {
+      "epoch": 0.29439497118910424,
+      "loss_partial": 0.05615234375,
+      "step": 562
+    },
+    {
+      "epoch": 0.29439497118910424,
+      "loss_partial": 0.048828125,
+      "step": 562
+    },
+    {
+      "epoch": 0.29491880565741224,
+      "grad_norm": 0.4797144830226898,
+      "learning_rate": 1.4712041884816754e-05,
+      "loss": 0.1144,
+      "step": 563
+    },
+    {
+      "epoch": 0.29491880565741224,
+      "loss_partial": 0.053955078125,
+      "step": 563
+    },
+    {
+      "epoch": 0.29491880565741224,
+      "loss_partial": 0.050048828125,
+      "step": 563
+    },
+    {
+      "epoch": 0.29544264012572025,
+      "grad_norm": 0.8783281445503235,
+      "learning_rate": 1.473821989528796e-05,
+      "loss": 0.1155,
+      "step": 564
+    },
+    {
+      "epoch": 0.29544264012572025,
+      "loss_partial": 0.052978515625,
+      "step": 564
+    },
+    {
+      "epoch": 0.29544264012572025,
+      "loss_partial": 0.06396484375,
+      "step": 564
+    },
+    {
+      "epoch": 0.2959664745940283,
+      "grad_norm": 0.5652477145195007,
+      "learning_rate": 1.4764397905759162e-05,
+      "loss": 0.1155,
+      "step": 565
+    },
+    {
+      "epoch": 0.2959664745940283,
+      "loss_partial": 0.055419921875,
+      "step": 565
+    },
+    {
+      "epoch": 0.2959664745940283,
+      "loss_partial": 0.068359375,
+      "step": 565
+    },
+    {
+      "epoch": 0.2964903090623363,
+      "grad_norm": 1.1131712198257446,
+      "learning_rate": 1.4790575916230367e-05,
+      "loss": 0.116,
+      "step": 566
+    },
+    {
+      "epoch": 0.2964903090623363,
+      "loss_partial": 0.0576171875,
+      "step": 566
+    },
+    {
+      "epoch": 0.2964903090623363,
+      "loss_partial": 0.061279296875,
+      "step": 566
+    },
+    {
+      "epoch": 0.2970141435306443,
+      "grad_norm": 0.9266590476036072,
+      "learning_rate": 1.4816753926701573e-05,
+      "loss": 0.1159,
+      "step": 567
+    },
+    {
+      "epoch": 0.2970141435306443,
+      "loss_partial": 0.05615234375,
+      "step": 567
+    },
+    {
+      "epoch": 0.2970141435306443,
+      "loss_partial": 0.056640625,
+      "step": 567
+    },
+    {
+      "epoch": 0.2975379779989523,
+      "grad_norm": 0.7287365794181824,
+      "learning_rate": 1.4842931937172775e-05,
+      "loss": 0.114,
+      "step": 568
+    },
+    {
+      "epoch": 0.2975379779989523,
+      "loss_partial": 0.056640625,
+      "step": 568
+    },
+    {
+      "epoch": 0.2975379779989523,
+      "loss_partial": 0.064453125,
+      "step": 568
+    },
+    {
+      "epoch": 0.2980618124672603,
+      "grad_norm": 0.858201801776886,
+      "learning_rate": 1.486910994764398e-05,
+      "loss": 0.1164,
+      "step": 569
+    },
+    {
+      "epoch": 0.2980618124672603,
+      "loss_partial": 0.061767578125,
+      "step": 569
+    },
+    {
+      "epoch": 0.2980618124672603,
+      "loss_partial": 0.0556640625,
+      "step": 569
+    },
+    {
+      "epoch": 0.2985856469355684,
+      "grad_norm": 1.1244839429855347,
+      "learning_rate": 1.4895287958115185e-05,
+      "loss": 0.1178,
+      "step": 570
+    },
+    {
+      "epoch": 0.2985856469355684,
+      "loss_partial": 0.07080078125,
+      "step": 570
+    },
+    {
+      "epoch": 0.2985856469355684,
+      "loss_partial": 0.06201171875,
+      "step": 570
+    },
+    {
+      "epoch": 0.2991094814038764,
+      "grad_norm": 0.7514625191688538,
+      "learning_rate": 1.4921465968586389e-05,
+      "loss": 0.1141,
+      "step": 571
+    },
+    {
+      "epoch": 0.2991094814038764,
+      "loss_partial": 0.06689453125,
+      "step": 571
+    },
+    {
+      "epoch": 0.2991094814038764,
+      "loss_partial": 0.0654296875,
+      "step": 571
+    },
+    {
+      "epoch": 0.2996333158721844,
+      "grad_norm": 1.0957118272781372,
+      "learning_rate": 1.4947643979057593e-05,
+      "loss": 0.1172,
+      "step": 572
+    },
+    {
+      "epoch": 0.2996333158721844,
+      "loss_partial": 0.06982421875,
+      "step": 572
+    },
+    {
+      "epoch": 0.2996333158721844,
+      "loss_partial": 0.05126953125,
+      "step": 572
+    },
+    {
+      "epoch": 0.3001571503404924,
+      "grad_norm": 1.2538069486618042,
+      "learning_rate": 1.4973821989528796e-05,
+      "loss": 0.1192,
+      "step": 573
+    },
+    {
+      "epoch": 0.3001571503404924,
+      "loss_partial": 0.0673828125,
+      "step": 573
+    },
+    {
+      "epoch": 0.3001571503404924,
+      "loss_partial": 0.0498046875,
+      "step": 573
+    },
+    {
+      "epoch": 0.3006809848088004,
+      "grad_norm": 1.8168107271194458,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.1197,
+      "step": 574
+    },
+    {
+      "epoch": 0.3006809848088004,
+      "loss_partial": 0.0556640625,
+      "step": 574
+    },
+    {
+      "epoch": 0.3006809848088004,
+      "loss_partial": 0.0595703125,
+      "step": 574
+    },
+    {
+      "epoch": 0.30120481927710846,
+      "grad_norm": 1.398218035697937,
+      "learning_rate": 1.5026178010471206e-05,
+      "loss": 0.1195,
+      "step": 575
+    },
+    {
+      "epoch": 0.30120481927710846,
+      "loss_partial": 0.059814453125,
+      "step": 575
+    },
+    {
+      "epoch": 0.30120481927710846,
+      "loss_partial": 0.0576171875,
+      "step": 575
+    },
+    {
+      "epoch": 0.30172865374541646,
+      "grad_norm": 1.3904602527618408,
+      "learning_rate": 1.505235602094241e-05,
+      "loss": 0.1162,
+      "step": 576
+    },
+    {
+      "epoch": 0.30172865374541646,
+      "loss_partial": 0.0595703125,
+      "step": 576
+    },
+    {
+      "epoch": 0.30172865374541646,
+      "loss_partial": 0.04833984375,
+      "step": 576
+    },
+    {
+      "epoch": 0.30225248821372447,
+      "grad_norm": 2.2827043533325195,
+      "learning_rate": 1.5078534031413614e-05,
+      "loss": 0.1117,
+      "step": 577
+    },
+    {
+      "epoch": 0.30225248821372447,
+      "loss_partial": 0.05224609375,
+      "step": 577
+    },
+    {
+      "epoch": 0.30225248821372447,
+      "loss_partial": 0.055419921875,
+      "step": 577
+    },
+    {
+      "epoch": 0.3027763226820325,
+      "grad_norm": 0.9616119265556335,
+      "learning_rate": 1.5104712041884818e-05,
+      "loss": 0.1165,
+      "step": 578
+    },
+    {
+      "epoch": 0.3027763226820325,
+      "loss_partial": 0.05224609375,
+      "step": 578
+    },
+    {
+      "epoch": 0.3027763226820325,
+      "loss_partial": 0.052978515625,
+      "step": 578
+    },
+    {
+      "epoch": 0.3033001571503405,
+      "grad_norm": 1.7989139556884766,
+      "learning_rate": 1.5130890052356022e-05,
+      "loss": 0.1147,
+      "step": 579
+    },
+    {
+      "epoch": 0.3033001571503405,
+      "loss_partial": 0.05859375,
+      "step": 579
+    },
+    {
+      "epoch": 0.3033001571503405,
+      "loss_partial": 0.05615234375,
+      "step": 579
+    },
+    {
+      "epoch": 0.3038239916186485,
+      "grad_norm": 2.0364534854888916,
+      "learning_rate": 1.5157068062827227e-05,
+      "loss": 0.1169,
+      "step": 580
+    },
+    {
+      "epoch": 0.3038239916186485,
+      "loss_partial": 0.05908203125,
+      "step": 580
+    },
+    {
+      "epoch": 0.3038239916186485,
+      "loss_partial": 0.058349609375,
+      "step": 580
+    },
+    {
+      "epoch": 0.30434782608695654,
+      "grad_norm": 0.6847519874572754,
+      "learning_rate": 1.518324607329843e-05,
+      "loss": 0.1146,
+      "step": 581
+    },
+    {
+      "epoch": 0.30434782608695654,
+      "loss_partial": 0.07177734375,
+      "step": 581
+    },
+    {
+      "epoch": 0.30434782608695654,
+      "loss_partial": 0.060302734375,
+      "step": 581
+    },
+    {
+      "epoch": 0.30487166055526455,
+      "grad_norm": 1.7854543924331665,
+      "learning_rate": 1.5209424083769635e-05,
+      "loss": 0.1203,
+      "step": 582
+    },
+    {
+      "epoch": 0.30487166055526455,
+      "loss_partial": 0.052734375,
+      "step": 582
+    },
+    {
+      "epoch": 0.30487166055526455,
+      "loss_partial": 0.064453125,
+      "step": 582
+    },
+    {
+      "epoch": 0.30539549502357255,
+      "grad_norm": 1.6083542108535767,
+      "learning_rate": 1.523560209424084e-05,
+      "loss": 0.1197,
+      "step": 583
+    },
+    {
+      "epoch": 0.30539549502357255,
+      "loss_partial": 0.060302734375,
+      "step": 583
+    },
+    {
+      "epoch": 0.30539549502357255,
+      "loss_partial": 0.06298828125,
+      "step": 583
+    },
+    {
+      "epoch": 0.30591932949188055,
+      "grad_norm": 1.2472540140151978,
+      "learning_rate": 1.526178010471204e-05,
+      "loss": 0.1174,
+      "step": 584
+    },
+    {
+      "epoch": 0.30591932949188055,
+      "loss_partial": 0.052978515625,
+      "step": 584
+    },
+    {
+      "epoch": 0.30591932949188055,
+      "loss_partial": 0.0654296875,
+      "step": 584
+    },
+    {
+      "epoch": 0.30644316396018856,
+      "grad_norm": 2.3406026363372803,
+      "learning_rate": 1.528795811518325e-05,
+      "loss": 0.1166,
+      "step": 585
+    },
+    {
+      "epoch": 0.30644316396018856,
+      "loss_partial": 0.048095703125,
+      "step": 585
+    },
+    {
+      "epoch": 0.30644316396018856,
+      "loss_partial": 0.0625,
+      "step": 585
+    },
+    {
+      "epoch": 0.3069669984284966,
+      "grad_norm": 1.4858561754226685,
+      "learning_rate": 1.531413612565445e-05,
+      "loss": 0.118,
+      "step": 586
+    },
+    {
+      "epoch": 0.3069669984284966,
+      "loss_partial": 0.052978515625,
+      "step": 586
+    },
+    {
+      "epoch": 0.3069669984284966,
+      "loss_partial": 0.06396484375,
+      "step": 586
+    },
+    {
+      "epoch": 0.3074908328968046,
+      "grad_norm": 1.4010021686553955,
+      "learning_rate": 1.5340314136125656e-05,
+      "loss": 0.1154,
+      "step": 587
+    },
+    {
+      "epoch": 0.3074908328968046,
+      "loss_partial": 0.0498046875,
+      "step": 587
+    },
+    {
+      "epoch": 0.3074908328968046,
+      "loss_partial": 0.0732421875,
+      "step": 587
+    },
+    {
+      "epoch": 0.3080146673651126,
+      "grad_norm": 1.3916840553283691,
+      "learning_rate": 1.536649214659686e-05,
+      "loss": 0.1173,
+      "step": 588
+    },
+    {
+      "epoch": 0.3080146673651126,
+      "loss_partial": 0.052734375,
+      "step": 588
+    },
+    {
+      "epoch": 0.3080146673651126,
+      "loss_partial": 0.060546875,
+      "step": 588
+    },
+    {
+      "epoch": 0.30853850183342063,
+      "grad_norm": 1.283528447151184,
+      "learning_rate": 1.5392670157068064e-05,
+      "loss": 0.1214,
+      "step": 589
+    },
+    {
+      "epoch": 0.30853850183342063,
+      "loss_partial": 0.06494140625,
+      "step": 589
+    },
+    {
+      "epoch": 0.30853850183342063,
+      "loss_partial": 0.0615234375,
+      "step": 589
+    },
+    {
+      "epoch": 0.30906233630172864,
+      "grad_norm": 1.7020272016525269,
+      "learning_rate": 1.5418848167539268e-05,
+      "loss": 0.1223,
+      "step": 590
+    },
+    {
+      "epoch": 0.30906233630172864,
+      "loss_partial": 0.055908203125,
+      "step": 590
+    },
+    {
+      "epoch": 0.30906233630172864,
+      "loss_partial": 0.05810546875,
+      "step": 590
+    },
+    {
+      "epoch": 0.3095861707700367,
+      "grad_norm": 1.2503600120544434,
+      "learning_rate": 1.5445026178010472e-05,
+      "loss": 0.1147,
+      "step": 591
+    },
+    {
+      "epoch": 0.3095861707700367,
+      "loss_partial": 0.05908203125,
+      "step": 591
+    },
+    {
+      "epoch": 0.3095861707700367,
+      "loss_partial": 0.049560546875,
+      "step": 591
+    },
+    {
+      "epoch": 0.3101100052383447,
+      "grad_norm": 2.0286426544189453,
+      "learning_rate": 1.5471204188481676e-05,
+      "loss": 0.1199,
+      "step": 592
+    },
+    {
+      "epoch": 0.3101100052383447,
+      "loss_partial": 0.06884765625,
+      "step": 592
+    },
+    {
+      "epoch": 0.3101100052383447,
+      "loss_partial": 0.05224609375,
+      "step": 592
+    },
+    {
+      "epoch": 0.3106338397066527,
+      "grad_norm": 1.1532933712005615,
+      "learning_rate": 1.5497382198952883e-05,
+      "loss": 0.1163,
+      "step": 593
+    },
+    {
+      "epoch": 0.3106338397066527,
+      "loss_partial": 0.05615234375,
+      "step": 593
+    },
+    {
+      "epoch": 0.3106338397066527,
+      "loss_partial": 0.07275390625,
+      "step": 593
+    },
+    {
+      "epoch": 0.3111576741749607,
+      "grad_norm": 1.7483525276184082,
+      "learning_rate": 1.5523560209424084e-05,
+      "loss": 0.1196,
+      "step": 594
+    },
+    {
+      "epoch": 0.3111576741749607,
+      "loss_partial": 0.06103515625,
+      "step": 594
+    },
+    {
+      "epoch": 0.3111576741749607,
+      "loss_partial": 0.0537109375,
+      "step": 594
+    },
+    {
+      "epoch": 0.3116815086432687,
+      "grad_norm": 1.7140519618988037,
+      "learning_rate": 1.554973821989529e-05,
+      "loss": 0.1128,
+      "step": 595
+    },
+    {
+      "epoch": 0.3116815086432687,
+      "loss_partial": 0.057861328125,
+      "step": 595
+    },
+    {
+      "epoch": 0.3116815086432687,
+      "loss_partial": 0.053466796875,
+      "step": 595
+    },
+    {
+      "epoch": 0.3122053431115767,
+      "grad_norm": 0.849955141544342,
+      "learning_rate": 1.5575916230366492e-05,
+      "loss": 0.1182,
+      "step": 596
+    },
+    {
+      "epoch": 0.3122053431115767,
+      "loss_partial": 0.0625,
+      "step": 596
+    },
+    {
+      "epoch": 0.3122053431115767,
+      "loss_partial": 0.05322265625,
+      "step": 596
+    },
+    {
+      "epoch": 0.3127291775798848,
+      "grad_norm": 1.3978350162506104,
+      "learning_rate": 1.56020942408377e-05,
+      "loss": 0.1111,
+      "step": 597
+    },
+    {
+      "epoch": 0.3127291775798848,
+      "loss_partial": 0.054931640625,
+      "step": 597
+    },
+    {
+      "epoch": 0.3127291775798848,
+      "loss_partial": 0.061279296875,
+      "step": 597
+    },
+    {
+      "epoch": 0.3132530120481928,
+      "grad_norm": 1.3528066873550415,
+      "learning_rate": 1.5628272251308903e-05,
+      "loss": 0.1167,
+      "step": 598
+    },
+    {
+      "epoch": 0.3132530120481928,
+      "loss_partial": 0.0556640625,
+      "step": 598
+    },
+    {
+      "epoch": 0.3132530120481928,
+      "loss_partial": 0.068359375,
+      "step": 598
+    },
+    {
+      "epoch": 0.3137768465165008,
+      "grad_norm": 1.0091869831085205,
+      "learning_rate": 1.5654450261780107e-05,
+      "loss": 0.1168,
+      "step": 599
+    },
+    {
+      "epoch": 0.3137768465165008,
+      "loss_partial": 0.07177734375,
+      "step": 599
+    },
+    {
+      "epoch": 0.3137768465165008,
+      "loss_partial": 0.0595703125,
+      "step": 599
+    },
+    {
+      "epoch": 0.3143006809848088,
+      "grad_norm": 0.9173312187194824,
+      "learning_rate": 1.568062827225131e-05,
+      "loss": 0.1232,
+      "step": 600
+    },
+    {
+      "epoch": 0.3143006809848088,
+      "loss_partial": 0.059814453125,
+      "step": 600
+    },
+    {
+      "epoch": 0.3143006809848088,
+      "loss_partial": 0.052978515625,
+      "step": 600
+    },
+    {
+      "epoch": 0.3148245154531168,
+      "grad_norm": 0.9427313804626465,
+      "learning_rate": 1.5706806282722515e-05,
+      "loss": 0.1175,
+      "step": 601
+    },
+    {
+      "epoch": 0.3148245154531168,
+      "loss_partial": 0.05224609375,
+      "step": 601
+    },
+    {
+      "epoch": 0.3148245154531168,
+      "loss_partial": 0.06689453125,
+      "step": 601
+    },
+    {
+      "epoch": 0.31534834992142485,
+      "grad_norm": 1.1972391605377197,
+      "learning_rate": 1.573298429319372e-05,
+      "loss": 0.117,
+      "step": 602
+    },
+    {
+      "epoch": 0.31534834992142485,
+      "loss_partial": 0.06103515625,
+      "step": 602
+    },
+    {
+      "epoch": 0.31534834992142485,
+      "loss_partial": 0.062255859375,
+      "step": 602
+    },
+    {
+      "epoch": 0.31587218438973286,
+      "grad_norm": 1.4883958101272583,
+      "learning_rate": 1.5759162303664923e-05,
+      "loss": 0.1205,
+      "step": 603
+    },
+    {
+      "epoch": 0.31587218438973286,
+      "loss_partial": 0.06298828125,
+      "step": 603
+    },
+    {
+      "epoch": 0.31587218438973286,
+      "loss_partial": 0.060546875,
+      "step": 603
+    },
+    {
+      "epoch": 0.31639601885804086,
+      "grad_norm": 1.123306393623352,
+      "learning_rate": 1.5785340314136126e-05,
+      "loss": 0.1164,
+      "step": 604
+    },
+    {
+      "epoch": 0.31639601885804086,
+      "loss_partial": 0.0537109375,
+      "step": 604
+    },
+    {
+      "epoch": 0.31639601885804086,
+      "loss_partial": 0.06884765625,
+      "step": 604
+    },
+    {
+      "epoch": 0.31691985332634887,
+      "grad_norm": 1.1174871921539307,
+      "learning_rate": 1.581151832460733e-05,
+      "loss": 0.1178,
+      "step": 605
+    },
+    {
+      "epoch": 0.31691985332634887,
+      "loss_partial": 0.04736328125,
+      "step": 605
+    },
+    {
+      "epoch": 0.31691985332634887,
+      "loss_partial": 0.061279296875,
+      "step": 605
+    },
+    {
+      "epoch": 0.31744368779465687,
+      "grad_norm": 1.3643866777420044,
+      "learning_rate": 1.5837696335078538e-05,
+      "loss": 0.1162,
+      "step": 606
+    },
+    {
+      "epoch": 0.31744368779465687,
+      "loss_partial": 0.051513671875,
+      "step": 606
+    },
+    {
+      "epoch": 0.31744368779465687,
+      "loss_partial": 0.06884765625,
+      "step": 606
+    },
+    {
+      "epoch": 0.31796752226296493,
+      "grad_norm": 0.924239993095398,
+      "learning_rate": 1.5863874345549738e-05,
+      "loss": 0.1132,
+      "step": 607
+    },
+    {
+      "epoch": 0.31796752226296493,
+      "loss_partial": 0.0654296875,
+      "step": 607
+    },
+    {
+      "epoch": 0.31796752226296493,
+      "loss_partial": 0.059814453125,
+      "step": 607
+    },
+    {
+      "epoch": 0.31849135673127293,
+      "grad_norm": 1.7670350074768066,
+      "learning_rate": 1.5890052356020946e-05,
+      "loss": 0.1201,
+      "step": 608
+    },
+    {
+      "epoch": 0.31849135673127293,
+      "loss_partial": 0.0498046875,
+      "step": 608
+    },
+    {
+      "epoch": 0.31849135673127293,
+      "loss_partial": 0.05810546875,
+      "step": 608
+    },
+    {
+      "epoch": 0.31901519119958094,
+      "grad_norm": 1.4204803705215454,
+      "learning_rate": 1.5916230366492146e-05,
+      "loss": 0.1162,
+      "step": 609
+    },
+    {
+      "epoch": 0.31901519119958094,
+      "loss_partial": 0.05908203125,
+      "step": 609
+    },
+    {
+      "epoch": 0.31901519119958094,
+      "loss_partial": 0.05712890625,
+      "step": 609
+    },
+    {
+      "epoch": 0.31953902566788894,
+      "grad_norm": 1.0448371171951294,
+      "learning_rate": 1.5942408376963353e-05,
+      "loss": 0.1205,
+      "step": 610
+    },
+    {
+      "epoch": 0.31953902566788894,
+      "loss_partial": 0.06298828125,
+      "step": 610
+    },
+    {
+      "epoch": 0.31953902566788894,
+      "loss_partial": 0.0517578125,
+      "step": 610
+    },
+    {
+      "epoch": 0.32006286013619695,
+      "grad_norm": 1.7896711826324463,
+      "learning_rate": 1.5968586387434557e-05,
+      "loss": 0.1155,
+      "step": 611
+    },
+    {
+      "epoch": 0.32006286013619695,
+      "loss_partial": 0.064453125,
+      "step": 611
+    },
+    {
+      "epoch": 0.32006286013619695,
+      "loss_partial": 0.058837890625,
+      "step": 611
+    },
+    {
+      "epoch": 0.32058669460450495,
+      "grad_norm": 1.4659695625305176,
+      "learning_rate": 1.599476439790576e-05,
+      "loss": 0.1187,
+      "step": 612
+    },
+    {
+      "epoch": 0.32058669460450495,
+      "loss_partial": 0.056884765625,
+      "step": 612
+    },
+    {
+      "epoch": 0.32058669460450495,
+      "loss_partial": 0.059326171875,
+      "step": 612
+    },
+    {
+      "epoch": 0.321110529072813,
+      "grad_norm": 1.0563347339630127,
+      "learning_rate": 1.6020942408376965e-05,
+      "loss": 0.118,
+      "step": 613
+    },
+    {
+      "epoch": 0.321110529072813,
+      "loss_partial": 0.052490234375,
+      "step": 613
+    },
+    {
+      "epoch": 0.321110529072813,
+      "loss_partial": 0.053466796875,
+      "step": 613
+    },
+    {
+      "epoch": 0.321634363541121,
+      "grad_norm": 1.0164827108383179,
+      "learning_rate": 1.604712041884817e-05,
+      "loss": 0.1176,
+      "step": 614
+    },
+    {
+      "epoch": 0.321634363541121,
+      "loss_partial": 0.06640625,
+      "step": 614
+    },
+    {
+      "epoch": 0.321634363541121,
+      "loss_partial": 0.049560546875,
+      "step": 614
+    },
+    {
+      "epoch": 0.322158198009429,
+      "grad_norm": 1.1465169191360474,
+      "learning_rate": 1.6073298429319373e-05,
+      "loss": 0.1133,
+      "step": 615
+    },
+    {
+      "epoch": 0.322158198009429,
+      "loss_partial": 0.0537109375,
+      "step": 615
+    },
+    {
+      "epoch": 0.322158198009429,
+      "loss_partial": 0.06005859375,
+      "step": 615
+    },
+    {
+      "epoch": 0.322682032477737,
+      "grad_norm": 1.6628837585449219,
+      "learning_rate": 1.6099476439790577e-05,
+      "loss": 0.1187,
+      "step": 616
+    },
+    {
+      "epoch": 0.322682032477737,
+      "loss_partial": 0.053955078125,
+      "step": 616
+    },
+    {
+      "epoch": 0.322682032477737,
+      "loss_partial": 0.0537109375,
+      "step": 616
+    },
+    {
+      "epoch": 0.32320586694604503,
+      "grad_norm": 1.5963950157165527,
+      "learning_rate": 1.612565445026178e-05,
+      "loss": 0.1161,
+      "step": 617
+    },
+    {
+      "epoch": 0.32320586694604503,
+      "loss_partial": 0.0546875,
+      "step": 617
+    },
+    {
+      "epoch": 0.32320586694604503,
+      "loss_partial": 0.057373046875,
+      "step": 617
+    },
+    {
+      "epoch": 0.3237297014143531,
+      "grad_norm": 1.3660812377929688,
+      "learning_rate": 1.6151832460732985e-05,
+      "loss": 0.116,
+      "step": 618
+    },
+    {
+      "epoch": 0.3237297014143531,
+      "loss_partial": 0.0703125,
+      "step": 618
+    },
+    {
+      "epoch": 0.3237297014143531,
+      "loss_partial": 0.06396484375,
+      "step": 618
+    },
+    {
+      "epoch": 0.3242535358826611,
+      "grad_norm": 1.6790350675582886,
+      "learning_rate": 1.6178010471204192e-05,
+      "loss": 0.1188,
+      "step": 619
+    },
+    {
+      "epoch": 0.3242535358826611,
+      "loss_partial": 0.05859375,
+      "step": 619
+    },
+    {
+      "epoch": 0.3242535358826611,
+      "loss_partial": 0.07080078125,
+      "step": 619
+    },
+    {
+      "epoch": 0.3247773703509691,
+      "grad_norm": 2.828580617904663,
+      "learning_rate": 1.6204188481675393e-05,
+      "loss": 0.1192,
+      "step": 620
+    },
+    {
+      "epoch": 0.3247773703509691,
+      "loss_partial": 0.07275390625,
+      "step": 620
+    },
+    {
+      "epoch": 0.3247773703509691,
+      "loss_partial": 0.0556640625,
+      "step": 620
+    },
+    {
+      "epoch": 0.3253012048192771,
+      "grad_norm": 1.5233465433120728,
+      "learning_rate": 1.62303664921466e-05,
+      "loss": 0.122,
+      "step": 621
+    },
+    {
+      "epoch": 0.3253012048192771,
+      "loss_partial": 0.06103515625,
+      "step": 621
+    },
+    {
+      "epoch": 0.3253012048192771,
+      "loss_partial": 0.0576171875,
+      "step": 621
+    },
+    {
+      "epoch": 0.3258250392875851,
+      "grad_norm": 2.6138358116149902,
+      "learning_rate": 1.62565445026178e-05,
+      "loss": 0.1197,
+      "step": 622
+    },
+    {
+      "epoch": 0.3258250392875851,
+      "loss_partial": 0.06982421875,
+      "step": 622
+    },
+    {
+      "epoch": 0.3258250392875851,
+      "loss_partial": 0.061767578125,
+      "step": 622
+    },
+    {
+      "epoch": 0.32634887375589317,
+      "grad_norm": 1.549716830253601,
+      "learning_rate": 1.6282722513089008e-05,
+      "loss": 0.1179,
+      "step": 623
+    },
+    {
+      "epoch": 0.32634887375589317,
+      "loss_partial": 0.054931640625,
+      "step": 623
+    },
+    {
+      "epoch": 0.32634887375589317,
+      "loss_partial": 0.05712890625,
+      "step": 623
+    },
+    {
+      "epoch": 0.32687270822420117,
+      "grad_norm": 2.3961052894592285,
+      "learning_rate": 1.630890052356021e-05,
+      "loss": 0.116,
+      "step": 624
+    },
+    {
+      "epoch": 0.32687270822420117,
+      "loss_partial": 0.060791015625,
+      "step": 624
+    },
+    {
+      "epoch": 0.32687270822420117,
+      "loss_partial": 0.06689453125,
+      "step": 624
+    },
+    {
+      "epoch": 0.3273965426925092,
+      "grad_norm": 2.018993854522705,
+      "learning_rate": 1.6335078534031416e-05,
+      "loss": 0.1167,
+      "step": 625
+    },
+    {
+      "epoch": 0.3273965426925092,
+      "loss_partial": 0.054931640625,
+      "step": 625
+    },
+    {
+      "epoch": 0.3273965426925092,
+      "loss_partial": 0.0595703125,
+      "step": 625
+    },
+    {
+      "epoch": 0.3279203771608172,
+      "grad_norm": 1.1115331649780273,
+      "learning_rate": 1.636125654450262e-05,
+      "loss": 0.1176,
+      "step": 626
+    },
+    {
+      "epoch": 0.3279203771608172,
+      "loss_partial": 0.05908203125,
+      "step": 626
+    },
+    {
+      "epoch": 0.3279203771608172,
+      "loss_partial": 0.0673828125,
+      "step": 626
+    },
+    {
+      "epoch": 0.3284442116291252,
+      "grad_norm": 1.6887903213500977,
+      "learning_rate": 1.6387434554973823e-05,
+      "loss": 0.1166,
+      "step": 627
+    },
+    {
+      "epoch": 0.3284442116291252,
+      "loss_partial": 0.05078125,
+      "step": 627
+    },
+    {
+      "epoch": 0.3284442116291252,
+      "loss_partial": 0.060302734375,
+      "step": 627
+    },
+    {
+      "epoch": 0.3289680460974332,
+      "grad_norm": 1.7336153984069824,
+      "learning_rate": 1.6413612565445027e-05,
+      "loss": 0.1173,
+      "step": 628
+    },
+    {
+      "epoch": 0.3289680460974332,
+      "loss_partial": 0.05517578125,
+      "step": 628
+    },
+    {
+      "epoch": 0.3289680460974332,
+      "loss_partial": 0.0517578125,
+      "step": 628
+    },
+    {
+      "epoch": 0.32949188056574125,
+      "grad_norm": 1.1957073211669922,
+      "learning_rate": 1.643979057591623e-05,
+      "loss": 0.1149,
+      "step": 629
+    },
+    {
+      "epoch": 0.32949188056574125,
+      "loss_partial": 0.0634765625,
+      "step": 629
+    },
+    {
+      "epoch": 0.32949188056574125,
+      "loss_partial": 0.06689453125,
+      "step": 629
+    },
+    {
+      "epoch": 0.33001571503404925,
+      "grad_norm": 1.3983234167099,
+      "learning_rate": 1.6465968586387435e-05,
+      "loss": 0.1157,
+      "step": 630
+    },
+    {
+      "epoch": 0.33001571503404925,
+      "loss_partial": 0.050048828125,
+      "step": 630
+    },
+    {
+      "epoch": 0.33001571503404925,
+      "loss_partial": 0.05712890625,
+      "step": 630
+    },
+    {
+      "epoch": 0.33053954950235725,
+      "grad_norm": 0.7299147844314575,
+      "learning_rate": 1.649214659685864e-05,
+      "loss": 0.1152,
+      "step": 631
+    },
+    {
+      "epoch": 0.33053954950235725,
+      "loss_partial": 0.059326171875,
+      "step": 631
+    },
+    {
+      "epoch": 0.33053954950235725,
+      "loss_partial": 0.07568359375,
+      "step": 631
+    },
+    {
+      "epoch": 0.33106338397066526,
+      "grad_norm": 1.6629928350448608,
+      "learning_rate": 1.6518324607329846e-05,
+      "loss": 0.1166,
+      "step": 632
+    },
+    {
+      "epoch": 0.33106338397066526,
+      "loss_partial": 0.0615234375,
+      "step": 632
+    },
+    {
+      "epoch": 0.33106338397066526,
+      "loss_partial": 0.06005859375,
+      "step": 632
+    },
+    {
+      "epoch": 0.33158721843897326,
+      "grad_norm": 1.5369395017623901,
+      "learning_rate": 1.6544502617801047e-05,
+      "loss": 0.1185,
+      "step": 633
+    },
+    {
+      "epoch": 0.33158721843897326,
+      "loss_partial": 0.0673828125,
+      "step": 633
+    },
+    {
+      "epoch": 0.33158721843897326,
+      "loss_partial": 0.056640625,
+      "step": 633
+    },
+    {
+      "epoch": 0.3321110529072813,
+      "grad_norm": 0.9346281290054321,
+      "learning_rate": 1.6570680628272254e-05,
+      "loss": 0.1176,
+      "step": 634
+    },
+    {
+      "epoch": 0.3321110529072813,
+      "loss_partial": 0.062255859375,
+      "step": 634
+    },
+    {
+      "epoch": 0.3321110529072813,
+      "loss_partial": 0.06591796875,
+      "step": 634
+    },
+    {
+      "epoch": 0.3326348873755893,
+      "grad_norm": 1.6333281993865967,
+      "learning_rate": 1.6596858638743455e-05,
+      "loss": 0.1169,
+      "step": 635
+    },
+    {
+      "epoch": 0.3326348873755893,
+      "loss_partial": 0.0625,
+      "step": 635
+    },
+    {
+      "epoch": 0.3326348873755893,
+      "loss_partial": 0.05615234375,
+      "step": 635
+    },
+    {
+      "epoch": 0.33315872184389733,
+      "grad_norm": 1.0381044149398804,
+      "learning_rate": 1.6623036649214662e-05,
+      "loss": 0.1194,
+      "step": 636
+    },
+    {
+      "epoch": 0.33315872184389733,
+      "loss_partial": 0.05712890625,
+      "step": 636
+    },
+    {
+      "epoch": 0.33315872184389733,
+      "loss_partial": 0.057861328125,
+      "step": 636
+    },
+    {
+      "epoch": 0.33368255631220534,
+      "grad_norm": 1.615931510925293,
+      "learning_rate": 1.6649214659685866e-05,
+      "loss": 0.1203,
+      "step": 637
+    },
+    {
+      "epoch": 0.33368255631220534,
+      "loss_partial": 0.05859375,
+      "step": 637
+    },
+    {
+      "epoch": 0.33368255631220534,
+      "loss_partial": 0.05615234375,
+      "step": 637
+    },
+    {
+      "epoch": 0.33420639078051334,
+      "grad_norm": 2.6961495876312256,
+      "learning_rate": 1.667539267015707e-05,
+      "loss": 0.1157,
+      "step": 638
+    },
+    {
+      "epoch": 0.33420639078051334,
+      "loss_partial": 0.0625,
+      "step": 638
+    },
+    {
+      "epoch": 0.33420639078051334,
+      "loss_partial": 0.0517578125,
+      "step": 638
+    },
+    {
+      "epoch": 0.33473022524882134,
+      "grad_norm": 1.9027595520019531,
+      "learning_rate": 1.6701570680628274e-05,
+      "loss": 0.1173,
+      "step": 639
+    },
+    {
+      "epoch": 0.33473022524882134,
+      "loss_partial": 0.0625,
+      "step": 639
+    },
+    {
+      "epoch": 0.33473022524882134,
+      "loss_partial": 0.052734375,
+      "step": 639
+    },
+    {
+      "epoch": 0.3352540597171294,
+      "grad_norm": 1.2416445016860962,
+      "learning_rate": 1.6727748691099478e-05,
+      "loss": 0.1153,
+      "step": 640
+    },
+    {
+      "epoch": 0.3352540597171294,
+      "loss_partial": 0.06591796875,
+      "step": 640
+    },
+    {
+      "epoch": 0.3352540597171294,
+      "loss_partial": 0.05615234375,
+      "step": 640
+    },
+    {
+      "epoch": 0.3357778941854374,
+      "grad_norm": 2.318073034286499,
+      "learning_rate": 1.675392670157068e-05,
+      "loss": 0.1137,
+      "step": 641
+    },
+    {
+      "epoch": 0.3357778941854374,
+      "loss_partial": 0.058349609375,
+      "step": 641
+    },
+    {
+      "epoch": 0.3357778941854374,
+      "loss_partial": 0.056640625,
+      "step": 641
+    },
+    {
+      "epoch": 0.3363017286537454,
+      "grad_norm": 1.147790789604187,
+      "learning_rate": 1.6780104712041886e-05,
+      "loss": 0.1152,
+      "step": 642
+    },
+    {
+      "epoch": 0.3363017286537454,
+      "loss_partial": 0.05712890625,
+      "step": 642
+    },
+    {
+      "epoch": 0.3363017286537454,
+      "loss_partial": 0.0634765625,
+      "step": 642
+    },
+    {
+      "epoch": 0.3368255631220534,
+      "grad_norm": 1.6573280096054077,
+      "learning_rate": 1.680628272251309e-05,
+      "loss": 0.1188,
+      "step": 643
+    },
+    {
+      "epoch": 0.3368255631220534,
+      "loss_partial": 0.0595703125,
+      "step": 643
+    },
+    {
+      "epoch": 0.3368255631220534,
+      "loss_partial": 0.060302734375,
+      "step": 643
+    },
+    {
+      "epoch": 0.3373493975903614,
+      "grad_norm": 1.0488643646240234,
+      "learning_rate": 1.6832460732984293e-05,
+      "loss": 0.114,
+      "step": 644
+    },
+    {
+      "epoch": 0.3373493975903614,
+      "loss_partial": 0.056396484375,
+      "step": 644
+    },
+    {
+      "epoch": 0.3373493975903614,
+      "loss_partial": 0.06005859375,
+      "step": 644
+    },
+    {
+      "epoch": 0.3378732320586695,
+      "grad_norm": 1.1434345245361328,
+      "learning_rate": 1.6858638743455497e-05,
+      "loss": 0.1143,
+      "step": 645
+    },
+    {
+      "epoch": 0.3378732320586695,
+      "loss_partial": 0.060791015625,
+      "step": 645
+    },
+    {
+      "epoch": 0.3378732320586695,
+      "loss_partial": 0.0537109375,
+      "step": 645
+    },
+    {
+      "epoch": 0.3383970665269775,
+      "grad_norm": 1.301935076713562,
+      "learning_rate": 1.68848167539267e-05,
+      "loss": 0.1139,
+      "step": 646
+    },
+    {
+      "epoch": 0.3383970665269775,
+      "loss_partial": 0.05810546875,
+      "step": 646
+    },
+    {
+      "epoch": 0.3383970665269775,
+      "loss_partial": 0.044189453125,
+      "step": 646
+    },
+    {
+      "epoch": 0.3389209009952855,
+      "grad_norm": 1.1992590427398682,
+      "learning_rate": 1.691099476439791e-05,
+      "loss": 0.1157,
+      "step": 647
+    },
+    {
+      "epoch": 0.3389209009952855,
+      "loss_partial": 0.0595703125,
+      "step": 647
+    },
+    {
+      "epoch": 0.3389209009952855,
+      "loss_partial": 0.050537109375,
+      "step": 647
+    },
+    {
+      "epoch": 0.3394447354635935,
+      "grad_norm": 1.1970548629760742,
+      "learning_rate": 1.693717277486911e-05,
+      "loss": 0.1171,
+      "step": 648
+    },
+    {
+      "epoch": 0.3394447354635935,
+      "loss_partial": 0.07177734375,
+      "step": 648
+    },
+    {
+      "epoch": 0.3394447354635935,
+      "loss_partial": 0.0576171875,
+      "step": 648
+    },
+    {
+      "epoch": 0.3399685699319015,
+      "grad_norm": 1.4279611110687256,
+      "learning_rate": 1.6963350785340316e-05,
+      "loss": 0.1175,
+      "step": 649
+    },
+    {
+      "epoch": 0.3399685699319015,
+      "loss_partial": 0.052734375,
+      "step": 649
+    },
+    {
+      "epoch": 0.3399685699319015,
+      "loss_partial": 0.06689453125,
+      "step": 649
+    },
+    {
+      "epoch": 0.34049240440020956,
+      "grad_norm": 1.2866655588150024,
+      "learning_rate": 1.698952879581152e-05,
+      "loss": 0.1192,
+      "step": 650
+    },
+    {
+      "epoch": 0.34049240440020956,
+      "loss_partial": 0.06884765625,
+      "step": 650
+    },
+    {
+      "epoch": 0.34049240440020956,
+      "loss_partial": 0.051025390625,
+      "step": 650
+    },
+    {
+      "epoch": 0.34101623886851756,
+      "grad_norm": 1.1179622411727905,
+      "learning_rate": 1.7015706806282724e-05,
+      "loss": 0.1142,
+      "step": 651
+    },
+    {
+      "epoch": 0.34101623886851756,
+      "loss_partial": 0.055419921875,
+      "step": 651
+    },
+    {
+      "epoch": 0.34101623886851756,
+      "loss_partial": 0.05322265625,
+      "step": 651
+    },
+    {
+      "epoch": 0.34154007333682557,
+      "grad_norm": 1.5898587703704834,
+      "learning_rate": 1.7041884816753928e-05,
+      "loss": 0.115,
+      "step": 652
+    },
+    {
+      "epoch": 0.34154007333682557,
+      "loss_partial": 0.06689453125,
+      "step": 652
+    },
+    {
+      "epoch": 0.34154007333682557,
+      "loss_partial": 0.06982421875,
+      "step": 652
+    },
+    {
+      "epoch": 0.34206390780513357,
+      "grad_norm": 1.230438232421875,
+      "learning_rate": 1.7068062827225132e-05,
+      "loss": 0.1181,
+      "step": 653
+    },
+    {
+      "epoch": 0.34206390780513357,
+      "loss_partial": 0.06396484375,
+      "step": 653
+    },
+    {
+      "epoch": 0.34206390780513357,
+      "loss_partial": 0.05615234375,
+      "step": 653
+    },
+    {
+      "epoch": 0.3425877422734416,
+      "grad_norm": 0.8835859298706055,
+      "learning_rate": 1.7094240837696336e-05,
+      "loss": 0.1129,
+      "step": 654
+    },
+    {
+      "epoch": 0.3425877422734416,
+      "loss_partial": 0.055419921875,
+      "step": 654
+    },
+    {
+      "epoch": 0.3425877422734416,
+      "loss_partial": 0.051513671875,
+      "step": 654
+    },
+    {
+      "epoch": 0.3431115767417496,
+      "grad_norm": 1.2033016681671143,
+      "learning_rate": 1.712041884816754e-05,
+      "loss": 0.1137,
+      "step": 655
+    },
+    {
+      "epoch": 0.3431115767417496,
+      "loss_partial": 0.06005859375,
+      "step": 655
+    },
+    {
+      "epoch": 0.3431115767417496,
+      "loss_partial": 0.06005859375,
+      "step": 655
+    },
+    {
+      "epoch": 0.34363541121005764,
+      "grad_norm": 0.8993368744850159,
+      "learning_rate": 1.7146596858638744e-05,
+      "loss": 0.1135,
+      "step": 656
+    },
+    {
+      "epoch": 0.34363541121005764,
+      "loss_partial": 0.049560546875,
+      "step": 656
+    },
+    {
+      "epoch": 0.34363541121005764,
+      "loss_partial": 0.06201171875,
+      "step": 656
+    },
+    {
+      "epoch": 0.34415924567836564,
+      "grad_norm": 1.203160285949707,
+      "learning_rate": 1.7172774869109948e-05,
+      "loss": 0.1176,
+      "step": 657
+    },
+    {
+      "epoch": 0.34415924567836564,
+      "loss_partial": 0.0615234375,
+      "step": 657
+    },
+    {
+      "epoch": 0.34415924567836564,
+      "loss_partial": 0.053955078125,
+      "step": 657
+    },
+    {
+      "epoch": 0.34468308014667365,
+      "grad_norm": 1.3121893405914307,
+      "learning_rate": 1.719895287958115e-05,
+      "loss": 0.1162,
+      "step": 658
+    },
+    {
+      "epoch": 0.34468308014667365,
+      "loss_partial": 0.052490234375,
+      "step": 658
+    },
+    {
+      "epoch": 0.34468308014667365,
+      "loss_partial": 0.06787109375,
+      "step": 658
+    },
+    {
+      "epoch": 0.34520691461498165,
+      "grad_norm": 1.3986907005310059,
+      "learning_rate": 1.7225130890052356e-05,
+      "loss": 0.1151,
+      "step": 659
+    },
+    {
+      "epoch": 0.34520691461498165,
+      "loss_partial": 0.055908203125,
+      "step": 659
+    },
+    {
+      "epoch": 0.34520691461498165,
+      "loss_partial": 0.05859375,
+      "step": 659
+    },
+    {
+      "epoch": 0.34573074908328966,
+      "grad_norm": 0.7627560496330261,
+      "learning_rate": 1.7251308900523563e-05,
+      "loss": 0.1174,
+      "step": 660
+    },
+    {
+      "epoch": 0.34573074908328966,
+      "loss_partial": 0.05029296875,
+      "step": 660
+    },
+    {
+      "epoch": 0.34573074908328966,
+      "loss_partial": 0.05908203125,
+      "step": 660
+    },
+    {
+      "epoch": 0.3462545835515977,
+      "grad_norm": 1.1189239025115967,
+      "learning_rate": 1.7277486910994767e-05,
+      "loss": 0.115,
+      "step": 661
+    },
+    {
+      "epoch": 0.3462545835515977,
+      "loss_partial": 0.0546875,
+      "step": 661
+    },
+    {
+      "epoch": 0.3462545835515977,
+      "loss_partial": 0.05908203125,
+      "step": 661
+    },
+    {
+      "epoch": 0.3467784180199057,
+      "grad_norm": 0.9747763276100159,
+      "learning_rate": 1.730366492146597e-05,
+      "loss": 0.116,
+      "step": 662
+    },
+    {
+      "epoch": 0.3467784180199057,
+      "loss_partial": 0.052734375,
+      "step": 662
+    },
+    {
+      "epoch": 0.3467784180199057,
+      "loss_partial": 0.056396484375,
+      "step": 662
+    },
+    {
+      "epoch": 0.3473022524882137,
+      "grad_norm": 0.8820493817329407,
+      "learning_rate": 1.7329842931937175e-05,
+      "loss": 0.1153,
+      "step": 663
+    },
+    {
+      "epoch": 0.3473022524882137,
+      "loss_partial": 0.058349609375,
+      "step": 663
+    },
+    {
+      "epoch": 0.3473022524882137,
+      "loss_partial": 0.05078125,
+      "step": 663
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "grad_norm": 1.2384777069091797,
+      "learning_rate": 1.735602094240838e-05,
+      "loss": 0.1184,
+      "step": 664
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "loss_partial": 0.059326171875,
+      "step": 664
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "loss_partial": 0.054443359375,
+      "step": 664
+    },
+    {
+      "epoch": 0.34834992142482973,
+      "grad_norm": 1.2581779956817627,
+      "learning_rate": 1.7382198952879583e-05,
+      "loss": 0.1197,
+      "step": 665
+    },
+    {
+      "epoch": 0.34834992142482973,
+      "loss_partial": 0.06103515625,
+      "step": 665
+    },
+    {
+      "epoch": 0.34834992142482973,
+      "loss_partial": 0.049072265625,
+      "step": 665
+    },
+    {
+      "epoch": 0.3488737558931378,
+      "grad_norm": 1.3455864191055298,
+      "learning_rate": 1.7408376963350786e-05,
+      "loss": 0.1144,
+      "step": 666
+    },
+    {
+      "epoch": 0.3488737558931378,
+      "loss_partial": 0.054931640625,
+      "step": 666
+    },
+    {
+      "epoch": 0.3488737558931378,
+      "loss_partial": 0.05517578125,
+      "step": 666
+    },
+    {
+      "epoch": 0.3493975903614458,
+      "grad_norm": 1.056415319442749,
+      "learning_rate": 1.743455497382199e-05,
+      "loss": 0.1166,
+      "step": 667
+    },
+    {
+      "epoch": 0.3493975903614458,
+      "loss_partial": 0.047607421875,
+      "step": 667
+    },
+    {
+      "epoch": 0.3493975903614458,
+      "loss_partial": 0.05908203125,
+      "step": 667
+    },
+    {
+      "epoch": 0.3499214248297538,
+      "grad_norm": 1.963565707206726,
+      "learning_rate": 1.7460732984293198e-05,
+      "loss": 0.117,
+      "step": 668
+    },
+    {
+      "epoch": 0.3499214248297538,
+      "loss_partial": 0.06201171875,
+      "step": 668
+    },
+    {
+      "epoch": 0.3499214248297538,
+      "loss_partial": 0.0634765625,
+      "step": 668
+    },
+    {
+      "epoch": 0.3504452592980618,
+      "grad_norm": 2.3525350093841553,
+      "learning_rate": 1.7486910994764398e-05,
+      "loss": 0.1168,
+      "step": 669
+    },
+    {
+      "epoch": 0.3504452592980618,
+      "loss_partial": 0.0546875,
+      "step": 669
+    },
+    {
+      "epoch": 0.3504452592980618,
+      "loss_partial": 0.061279296875,
+      "step": 669
+    },
+    {
+      "epoch": 0.3509690937663698,
+      "grad_norm": 1.921631932258606,
+      "learning_rate": 1.7513089005235606e-05,
+      "loss": 0.1184,
+      "step": 670
+    },
+    {
+      "epoch": 0.3509690937663698,
+      "loss_partial": 0.06201171875,
+      "step": 670
+    },
+    {
+      "epoch": 0.3509690937663698,
+      "loss_partial": 0.05078125,
+      "step": 670
+    },
+    {
+      "epoch": 0.3514929282346778,
+      "grad_norm": 0.9081473350524902,
+      "learning_rate": 1.7539267015706806e-05,
+      "loss": 0.1174,
+      "step": 671
+    },
+    {
+      "epoch": 0.3514929282346778,
+      "loss_partial": 0.054931640625,
+      "step": 671
+    },
+    {
+      "epoch": 0.3514929282346778,
+      "loss_partial": 0.06103515625,
+      "step": 671
+    },
+    {
+      "epoch": 0.3520167627029859,
+      "grad_norm": 2.3186025619506836,
+      "learning_rate": 1.7565445026178013e-05,
+      "loss": 0.1156,
+      "step": 672
+    },
+    {
+      "epoch": 0.3520167627029859,
+      "loss_partial": 0.057373046875,
+      "step": 672
+    },
+    {
+      "epoch": 0.3520167627029859,
+      "loss_partial": 0.056884765625,
+      "step": 672
+    },
+    {
+      "epoch": 0.3525405971712939,
+      "grad_norm": 2.1421101093292236,
+      "learning_rate": 1.7591623036649217e-05,
+      "loss": 0.1151,
+      "step": 673
+    },
+    {
+      "epoch": 0.3525405971712939,
+      "loss_partial": 0.049560546875,
+      "step": 673
+    },
+    {
+      "epoch": 0.3525405971712939,
+      "loss_partial": 0.04931640625,
+      "step": 673
+    },
+    {
+      "epoch": 0.3530644316396019,
+      "grad_norm": 1.6448079347610474,
+      "learning_rate": 1.761780104712042e-05,
+      "loss": 0.1136,
+      "step": 674
+    },
+    {
+      "epoch": 0.3530644316396019,
+      "loss_partial": 0.0634765625,
+      "step": 674
+    },
+    {
+      "epoch": 0.3530644316396019,
+      "loss_partial": 0.048828125,
+      "step": 674
+    },
+    {
+      "epoch": 0.3535882661079099,
+      "grad_norm": 2.1362850666046143,
+      "learning_rate": 1.7643979057591625e-05,
+      "loss": 0.1164,
+      "step": 675
+    },
+    {
+      "epoch": 0.3535882661079099,
+      "loss_partial": 0.06103515625,
+      "step": 675
+    },
+    {
+      "epoch": 0.3535882661079099,
+      "loss_partial": 0.0595703125,
+      "step": 675
+    },
+    {
+      "epoch": 0.3541121005762179,
+      "grad_norm": 2.5235116481781006,
+      "learning_rate": 1.767015706806283e-05,
+      "loss": 0.1154,
+      "step": 676
+    },
+    {
+      "epoch": 0.3541121005762179,
+      "loss_partial": 0.05859375,
+      "step": 676
+    },
+    {
+      "epoch": 0.3541121005762179,
+      "loss_partial": 0.050048828125,
+      "step": 676
+    },
+    {
+      "epoch": 0.35463593504452595,
+      "grad_norm": 1.2432148456573486,
+      "learning_rate": 1.7696335078534033e-05,
+      "loss": 0.1125,
+      "step": 677
+    },
+    {
+      "epoch": 0.35463593504452595,
+      "loss_partial": 0.061279296875,
+      "step": 677
+    },
+    {
+      "epoch": 0.35463593504452595,
+      "loss_partial": 0.06201171875,
+      "step": 677
+    },
+    {
+      "epoch": 0.35515976951283396,
+      "grad_norm": 1.8355249166488647,
+      "learning_rate": 1.7722513089005237e-05,
+      "loss": 0.1209,
+      "step": 678
+    },
+    {
+      "epoch": 0.35515976951283396,
+      "loss_partial": 0.0595703125,
+      "step": 678
+    },
+    {
+      "epoch": 0.35515976951283396,
+      "loss_partial": 0.058837890625,
+      "step": 678
+    },
+    {
+      "epoch": 0.35568360398114196,
+      "grad_norm": 1.9406406879425049,
+      "learning_rate": 1.774869109947644e-05,
+      "loss": 0.1167,
+      "step": 679
+    },
+    {
+      "epoch": 0.35568360398114196,
+      "loss_partial": 0.06689453125,
+      "step": 679
+    },
+    {
+      "epoch": 0.35568360398114196,
+      "loss_partial": 0.056640625,
+      "step": 679
+    },
+    {
+      "epoch": 0.35620743844944996,
+      "grad_norm": 1.5537017583847046,
+      "learning_rate": 1.7774869109947645e-05,
+      "loss": 0.1155,
+      "step": 680
+    },
+    {
+      "epoch": 0.35620743844944996,
+      "loss_partial": 0.05615234375,
+      "step": 680
+    },
+    {
+      "epoch": 0.35620743844944996,
+      "loss_partial": 0.06591796875,
+      "step": 680
+    },
+    {
+      "epoch": 0.35673127291775797,
+      "grad_norm": 1.3787569999694824,
+      "learning_rate": 1.7801047120418852e-05,
+      "loss": 0.1191,
+      "step": 681
+    },
+    {
+      "epoch": 0.35673127291775797,
+      "loss_partial": 0.0634765625,
+      "step": 681
+    },
+    {
+      "epoch": 0.35673127291775797,
+      "loss_partial": 0.06005859375,
+      "step": 681
+    },
+    {
+      "epoch": 0.35725510738606603,
+      "grad_norm": 1.9430887699127197,
+      "learning_rate": 1.7827225130890053e-05,
+      "loss": 0.1174,
+      "step": 682
+    },
+    {
+      "epoch": 0.35725510738606603,
+      "loss_partial": 0.060302734375,
+      "step": 682
+    },
+    {
+      "epoch": 0.35725510738606603,
+      "loss_partial": 0.0654296875,
+      "step": 682
+    },
+    {
+      "epoch": 0.35777894185437403,
+      "grad_norm": 1.0107072591781616,
+      "learning_rate": 1.785340314136126e-05,
+      "loss": 0.1137,
+      "step": 683
+    },
+    {
+      "epoch": 0.35777894185437403,
+      "loss_partial": 0.05908203125,
+      "step": 683
+    },
+    {
+      "epoch": 0.35777894185437403,
+      "loss_partial": 0.052978515625,
+      "step": 683
+    },
+    {
+      "epoch": 0.35830277632268204,
+      "grad_norm": 1.5686736106872559,
+      "learning_rate": 1.787958115183246e-05,
+      "loss": 0.1207,
+      "step": 684
+    },
+    {
+      "epoch": 0.35830277632268204,
+      "loss_partial": 0.05419921875,
+      "step": 684
+    },
+    {
+      "epoch": 0.35830277632268204,
+      "loss_partial": 0.05029296875,
+      "step": 684
+    },
+    {
+      "epoch": 0.35882661079099004,
+      "grad_norm": 1.5140116214752197,
+      "learning_rate": 1.7905759162303668e-05,
+      "loss": 0.112,
+      "step": 685
+    },
+    {
+      "epoch": 0.35882661079099004,
+      "loss_partial": 0.056884765625,
+      "step": 685
+    },
+    {
+      "epoch": 0.35882661079099004,
+      "loss_partial": 0.055908203125,
+      "step": 685
+    },
+    {
+      "epoch": 0.35935044525929805,
+      "grad_norm": 1.3362791538238525,
+      "learning_rate": 1.793193717277487e-05,
+      "loss": 0.112,
+      "step": 686
+    },
+    {
+      "epoch": 0.35935044525929805,
+      "loss_partial": 0.056396484375,
+      "step": 686
+    },
+    {
+      "epoch": 0.35935044525929805,
+      "loss_partial": 0.060546875,
+      "step": 686
+    },
+    {
+      "epoch": 0.35987427972760605,
+      "grad_norm": 1.191327452659607,
+      "learning_rate": 1.7958115183246076e-05,
+      "loss": 0.1192,
+      "step": 687
+    },
+    {
+      "epoch": 0.35987427972760605,
+      "loss_partial": 0.05419921875,
+      "step": 687
+    },
+    {
+      "epoch": 0.35987427972760605,
+      "loss_partial": 0.06298828125,
+      "step": 687
+    },
+    {
+      "epoch": 0.3603981141959141,
+      "grad_norm": 1.6991393566131592,
+      "learning_rate": 1.798429319371728e-05,
+      "loss": 0.12,
+      "step": 688
+    },
+    {
+      "epoch": 0.3603981141959141,
+      "loss_partial": 0.0556640625,
+      "step": 688
+    },
+    {
+      "epoch": 0.3603981141959141,
+      "loss_partial": 0.06298828125,
+      "step": 688
+    },
+    {
+      "epoch": 0.3609219486642221,
+      "grad_norm": 0.9647210240364075,
+      "learning_rate": 1.8010471204188483e-05,
+      "loss": 0.1165,
+      "step": 689
+    },
+    {
+      "epoch": 0.3609219486642221,
+      "loss_partial": 0.05810546875,
+      "step": 689
+    },
+    {
+      "epoch": 0.3609219486642221,
+      "loss_partial": 0.048828125,
+      "step": 689
+    },
+    {
+      "epoch": 0.3614457831325301,
+      "grad_norm": 1.3127620220184326,
+      "learning_rate": 1.8036649214659687e-05,
+      "loss": 0.1118,
+      "step": 690
+    },
+    {
+      "epoch": 0.3614457831325301,
+      "loss_partial": 0.0546875,
+      "step": 690
+    },
+    {
+      "epoch": 0.3614457831325301,
+      "loss_partial": 0.052001953125,
+      "step": 690
+    },
+    {
+      "epoch": 0.3619696176008381,
+      "grad_norm": 2.011439085006714,
+      "learning_rate": 1.806282722513089e-05,
+      "loss": 0.1156,
+      "step": 691
+    },
+    {
+      "epoch": 0.3619696176008381,
+      "loss_partial": 0.049560546875,
+      "step": 691
+    },
+    {
+      "epoch": 0.3619696176008381,
+      "loss_partial": 0.047119140625,
+      "step": 691
+    },
+    {
+      "epoch": 0.3624934520691461,
+      "grad_norm": 2.038118362426758,
+      "learning_rate": 1.8089005235602095e-05,
+      "loss": 0.1161,
+      "step": 692
+    },
+    {
+      "epoch": 0.3624934520691461,
+      "loss_partial": 0.06640625,
+      "step": 692
+    },
+    {
+      "epoch": 0.3624934520691461,
+      "loss_partial": 0.064453125,
+      "step": 692
+    },
+    {
+      "epoch": 0.3630172865374542,
+      "grad_norm": 1.4187815189361572,
+      "learning_rate": 1.81151832460733e-05,
+      "loss": 0.1152,
+      "step": 693
+    },
+    {
+      "epoch": 0.3630172865374542,
+      "loss_partial": 0.053955078125,
+      "step": 693
+    },
+    {
+      "epoch": 0.3630172865374542,
+      "loss_partial": 0.0576171875,
+      "step": 693
+    },
+    {
+      "epoch": 0.3635411210057622,
+      "grad_norm": 1.360778570175171,
+      "learning_rate": 1.8141361256544506e-05,
+      "loss": 0.1147,
+      "step": 694
+    },
+    {
+      "epoch": 0.3635411210057622,
+      "loss_partial": 0.050048828125,
+      "step": 694
+    },
+    {
+      "epoch": 0.3635411210057622,
+      "loss_partial": 0.058349609375,
+      "step": 694
+    },
+    {
+      "epoch": 0.3640649554740702,
+      "grad_norm": 2.976351499557495,
+      "learning_rate": 1.8167539267015707e-05,
+      "loss": 0.1159,
+      "step": 695
+    },
+    {
+      "epoch": 0.3640649554740702,
+      "loss_partial": 0.06494140625,
+      "step": 695
+    },
+    {
+      "epoch": 0.3640649554740702,
+      "loss_partial": 0.064453125,
+      "step": 695
+    },
+    {
+      "epoch": 0.3645887899423782,
+      "grad_norm": 3.030999183654785,
+      "learning_rate": 1.8193717277486914e-05,
+      "loss": 0.1184,
+      "step": 696
+    },
+    {
+      "epoch": 0.3645887899423782,
+      "loss_partial": 0.05224609375,
+      "step": 696
+    },
+    {
+      "epoch": 0.3645887899423782,
+      "loss_partial": 0.05419921875,
+      "step": 696
+    },
+    {
+      "epoch": 0.3651126244106862,
+      "grad_norm": 1.7023307085037231,
+      "learning_rate": 1.8219895287958115e-05,
+      "loss": 0.1096,
+      "step": 697
+    },
+    {
+      "epoch": 0.3651126244106862,
+      "loss_partial": 0.056640625,
+      "step": 697
+    },
+    {
+      "epoch": 0.3651126244106862,
+      "loss_partial": 0.05126953125,
+      "step": 697
+    },
+    {
+      "epoch": 0.36563645887899426,
+      "grad_norm": 1.5823994874954224,
+      "learning_rate": 1.8246073298429322e-05,
+      "loss": 0.1151,
+      "step": 698
+    },
+    {
+      "epoch": 0.36563645887899426,
+      "loss_partial": 0.05908203125,
+      "step": 698
+    },
+    {
+      "epoch": 0.36563645887899426,
+      "loss_partial": 0.046630859375,
+      "step": 698
+    },
+    {
+      "epoch": 0.36616029334730227,
+      "grad_norm": 2.606627941131592,
+      "learning_rate": 1.8272251308900526e-05,
+      "loss": 0.1143,
+      "step": 699
+    },
+    {
+      "epoch": 0.36616029334730227,
+      "loss_partial": 0.055908203125,
+      "step": 699
+    },
+    {
+      "epoch": 0.36616029334730227,
+      "loss_partial": 0.060546875,
+      "step": 699
+    },
+    {
+      "epoch": 0.3666841278156103,
+      "grad_norm": 2.831486701965332,
+      "learning_rate": 1.829842931937173e-05,
+      "loss": 0.1156,
+      "step": 700
+    },
+    {
+      "epoch": 0.3666841278156103,
+      "loss_partial": 0.06005859375,
+      "step": 700
+    },
+    {
+      "epoch": 0.3666841278156103,
+      "loss_partial": 0.05126953125,
+      "step": 700
+    },
+    {
+      "epoch": 0.3672079622839183,
+      "grad_norm": 1.3430317640304565,
+      "learning_rate": 1.8324607329842934e-05,
+      "loss": 0.1177,
+      "step": 701
+    },
+    {
+      "epoch": 0.3672079622839183,
+      "loss_partial": 0.058837890625,
+      "step": 701
+    },
+    {
+      "epoch": 0.3672079622839183,
+      "loss_partial": 0.05517578125,
+      "step": 701
+    },
+    {
+      "epoch": 0.3677317967522263,
+      "grad_norm": 3.211317300796509,
+      "learning_rate": 1.8350785340314138e-05,
+      "loss": 0.1176,
+      "step": 702
+    },
+    {
+      "epoch": 0.3677317967522263,
+      "loss_partial": 0.057861328125,
+      "step": 702
+    },
+    {
+      "epoch": 0.3677317967522263,
+      "loss_partial": 0.060791015625,
+      "step": 702
+    },
+    {
+      "epoch": 0.3682556312205343,
+      "grad_norm": 2.1536428928375244,
+      "learning_rate": 1.837696335078534e-05,
+      "loss": 0.1128,
+      "step": 703
+    },
+    {
+      "epoch": 0.3682556312205343,
+      "loss_partial": 0.06005859375,
+      "step": 703
+    },
+    {
+      "epoch": 0.3682556312205343,
+      "loss_partial": 0.055908203125,
+      "step": 703
+    },
+    {
+      "epoch": 0.36877946568884235,
+      "grad_norm": 1.4852138757705688,
+      "learning_rate": 1.8403141361256546e-05,
+      "loss": 0.116,
+      "step": 704
+    },
+    {
+      "epoch": 0.36877946568884235,
+      "loss_partial": 0.053955078125,
+      "step": 704
+    },
+    {
+      "epoch": 0.36877946568884235,
+      "loss_partial": 0.061279296875,
+      "step": 704
+    },
+    {
+      "epoch": 0.36930330015715035,
+      "grad_norm": 2.2470703125,
+      "learning_rate": 1.842931937172775e-05,
+      "loss": 0.1114,
+      "step": 705
+    },
+    {
+      "epoch": 0.36930330015715035,
+      "loss_partial": 0.053466796875,
+      "step": 705
+    },
+    {
+      "epoch": 0.36930330015715035,
+      "loss_partial": 0.059814453125,
+      "step": 705
+    },
+    {
+      "epoch": 0.36982713462545835,
+      "grad_norm": 1.073407769203186,
+      "learning_rate": 1.8455497382198953e-05,
+      "loss": 0.1144,
+      "step": 706
+    },
+    {
+      "epoch": 0.36982713462545835,
+      "loss_partial": 0.062255859375,
+      "step": 706
+    },
+    {
+      "epoch": 0.36982713462545835,
+      "loss_partial": 0.06103515625,
+      "step": 706
+    },
+    {
+      "epoch": 0.37035096909376636,
+      "grad_norm": 1.9461934566497803,
+      "learning_rate": 1.8481675392670157e-05,
+      "loss": 0.1183,
+      "step": 707
+    },
+    {
+      "epoch": 0.37035096909376636,
+      "loss_partial": 0.056396484375,
+      "step": 707
+    },
+    {
+      "epoch": 0.37035096909376636,
+      "loss_partial": 0.05615234375,
+      "step": 707
+    },
+    {
+      "epoch": 0.37087480356207436,
+      "grad_norm": 1.759114146232605,
+      "learning_rate": 1.850785340314136e-05,
+      "loss": 0.1082,
+      "step": 708
+    },
+    {
+      "epoch": 0.37087480356207436,
+      "loss_partial": 0.053955078125,
+      "step": 708
+    },
+    {
+      "epoch": 0.37087480356207436,
+      "loss_partial": 0.06494140625,
+      "step": 708
+    },
+    {
+      "epoch": 0.3713986380303824,
+      "grad_norm": 0.7778896689414978,
+      "learning_rate": 1.853403141361257e-05,
+      "loss": 0.1176,
+      "step": 709
+    },
+    {
+      "epoch": 0.3713986380303824,
+      "loss_partial": 0.06884765625,
+      "step": 709
+    },
+    {
+      "epoch": 0.3713986380303824,
+      "loss_partial": 0.05224609375,
+      "step": 709
+    },
+    {
+      "epoch": 0.3719224724986904,
+      "grad_norm": 1.6541705131530762,
+      "learning_rate": 1.856020942408377e-05,
+      "loss": 0.1152,
+      "step": 710
+    },
+    {
+      "epoch": 0.3719224724986904,
+      "loss_partial": 0.057861328125,
+      "step": 710
+    },
+    {
+      "epoch": 0.3719224724986904,
+      "loss_partial": 0.059814453125,
+      "step": 710
+    },
+    {
+      "epoch": 0.37244630696699843,
+      "grad_norm": 1.2571650743484497,
+      "learning_rate": 1.8586387434554976e-05,
+      "loss": 0.1148,
+      "step": 711
+    },
+    {
+      "epoch": 0.37244630696699843,
+      "loss_partial": 0.05810546875,
+      "step": 711
+    },
+    {
+      "epoch": 0.37244630696699843,
+      "loss_partial": 0.045654296875,
+      "step": 711
+    },
+    {
+      "epoch": 0.37297014143530643,
+      "grad_norm": 1.1161404848098755,
+      "learning_rate": 1.861256544502618e-05,
+      "loss": 0.1175,
+      "step": 712
+    },
+    {
+      "epoch": 0.37297014143530643,
+      "loss_partial": 0.055908203125,
+      "step": 712
+    },
+    {
+      "epoch": 0.37297014143530643,
+      "loss_partial": 0.055908203125,
+      "step": 712
+    },
+    {
+      "epoch": 0.37349397590361444,
+      "grad_norm": 1.802951693534851,
+      "learning_rate": 1.8638743455497384e-05,
+      "loss": 0.1129,
+      "step": 713
+    },
+    {
+      "epoch": 0.37349397590361444,
+      "loss_partial": 0.0625,
+      "step": 713
+    },
+    {
+      "epoch": 0.37349397590361444,
+      "loss_partial": 0.059814453125,
+      "step": 713
+    },
+    {
+      "epoch": 0.3740178103719225,
+      "grad_norm": 1.4033969640731812,
+      "learning_rate": 1.8664921465968588e-05,
+      "loss": 0.1178,
+      "step": 714
+    },
+    {
+      "epoch": 0.3740178103719225,
+      "loss_partial": 0.053955078125,
+      "step": 714
+    },
+    {
+      "epoch": 0.3740178103719225,
+      "loss_partial": 0.054443359375,
+      "step": 714
+    },
+    {
+      "epoch": 0.3745416448402305,
+      "grad_norm": 0.9478284120559692,
+      "learning_rate": 1.8691099476439792e-05,
+      "loss": 0.1193,
+      "step": 715
+    },
+    {
+      "epoch": 0.3745416448402305,
+      "loss_partial": 0.054931640625,
+      "step": 715
+    },
+    {
+      "epoch": 0.3745416448402305,
+      "loss_partial": 0.06640625,
+      "step": 715
+    },
+    {
+      "epoch": 0.3750654793085385,
+      "grad_norm": 1.63286554813385,
+      "learning_rate": 1.8717277486910996e-05,
+      "loss": 0.1129,
+      "step": 716
+    },
+    {
+      "epoch": 0.3750654793085385,
+      "loss_partial": 0.059326171875,
+      "step": 716
+    },
+    {
+      "epoch": 0.3750654793085385,
+      "loss_partial": 0.0625,
+      "step": 716
+    },
+    {
+      "epoch": 0.3755893137768465,
+      "grad_norm": 1.4723081588745117,
+      "learning_rate": 1.87434554973822e-05,
+      "loss": 0.116,
+      "step": 717
+    },
+    {
+      "epoch": 0.3755893137768465,
+      "loss_partial": 0.05517578125,
+      "step": 717
+    },
+    {
+      "epoch": 0.3755893137768465,
+      "loss_partial": 0.056884765625,
+      "step": 717
+    },
+    {
+      "epoch": 0.3761131482451545,
+      "grad_norm": 1.2525311708450317,
+      "learning_rate": 1.8769633507853404e-05,
+      "loss": 0.1154,
+      "step": 718
+    },
+    {
+      "epoch": 0.3761131482451545,
+      "loss_partial": 0.058349609375,
+      "step": 718
+    },
+    {
+      "epoch": 0.3761131482451545,
+      "loss_partial": 0.060791015625,
+      "step": 718
+    },
+    {
+      "epoch": 0.3766369827134625,
+      "grad_norm": 1.9110609292984009,
+      "learning_rate": 1.8795811518324608e-05,
+      "loss": 0.1136,
+      "step": 719
+    },
+    {
+      "epoch": 0.3766369827134625,
+      "loss_partial": 0.06640625,
+      "step": 719
+    },
+    {
+      "epoch": 0.3766369827134625,
+      "loss_partial": 0.052490234375,
+      "step": 719
+    },
+    {
+      "epoch": 0.3771608171817706,
+      "grad_norm": 1.9771496057510376,
+      "learning_rate": 1.882198952879581e-05,
+      "loss": 0.1129,
+      "step": 720
+    },
+    {
+      "epoch": 0.3771608171817706,
+      "loss_partial": 0.05029296875,
+      "step": 720
+    },
+    {
+      "epoch": 0.3771608171817706,
+      "loss_partial": 0.05224609375,
+      "step": 720
+    },
+    {
+      "epoch": 0.3776846516500786,
+      "grad_norm": 1.240181565284729,
+      "learning_rate": 1.8848167539267016e-05,
+      "loss": 0.117,
+      "step": 721
+    },
+    {
+      "epoch": 0.3776846516500786,
+      "loss_partial": 0.0615234375,
+      "step": 721
+    },
+    {
+      "epoch": 0.3776846516500786,
+      "loss_partial": 0.057861328125,
+      "step": 721
+    },
+    {
+      "epoch": 0.3782084861183866,
+      "grad_norm": 1.4072682857513428,
+      "learning_rate": 1.8874345549738223e-05,
+      "loss": 0.1182,
+      "step": 722
+    },
+    {
+      "epoch": 0.3782084861183866,
+      "loss_partial": 0.06787109375,
+      "step": 722
+    },
+    {
+      "epoch": 0.3782084861183866,
+      "loss_partial": 0.05029296875,
+      "step": 722
+    },
+    {
+      "epoch": 0.3787323205866946,
+      "grad_norm": 1.1417090892791748,
+      "learning_rate": 1.8900523560209423e-05,
+      "loss": 0.113,
+      "step": 723
+    },
+    {
+      "epoch": 0.3787323205866946,
+      "loss_partial": 0.050537109375,
+      "step": 723
+    },
+    {
+      "epoch": 0.3787323205866946,
+      "loss_partial": 0.0625,
+      "step": 723
+    },
+    {
+      "epoch": 0.3792561550550026,
+      "grad_norm": 1.3047901391983032,
+      "learning_rate": 1.892670157068063e-05,
+      "loss": 0.1127,
+      "step": 724
+    },
+    {
+      "epoch": 0.3792561550550026,
+      "loss_partial": 0.0634765625,
+      "step": 724
+    },
+    {
+      "epoch": 0.3792561550550026,
+      "loss_partial": 0.05126953125,
+      "step": 724
+    },
+    {
+      "epoch": 0.37977998952331066,
+      "grad_norm": 1.5438144207000732,
+      "learning_rate": 1.895287958115183e-05,
+      "loss": 0.116,
+      "step": 725
+    },
+    {
+      "epoch": 0.37977998952331066,
+      "loss_partial": 0.0673828125,
+      "step": 725
+    },
+    {
+      "epoch": 0.37977998952331066,
+      "loss_partial": 0.055419921875,
+      "step": 725
+    },
+    {
+      "epoch": 0.38030382399161866,
+      "grad_norm": 1.0432322025299072,
+      "learning_rate": 1.897905759162304e-05,
+      "loss": 0.1139,
+      "step": 726
+    },
+    {
+      "epoch": 0.38030382399161866,
+      "loss_partial": 0.052978515625,
+      "step": 726
+    },
+    {
+      "epoch": 0.38030382399161866,
+      "loss_partial": 0.05078125,
+      "step": 726
+    },
+    {
+      "epoch": 0.38082765845992667,
+      "grad_norm": 0.8491865396499634,
+      "learning_rate": 1.9005235602094243e-05,
+      "loss": 0.114,
+      "step": 727
+    },
+    {
+      "epoch": 0.38082765845992667,
+      "loss_partial": 0.059326171875,
+      "step": 727
+    },
+    {
+      "epoch": 0.38082765845992667,
+      "loss_partial": 0.0625,
+      "step": 727
+    },
+    {
+      "epoch": 0.38135149292823467,
+      "grad_norm": 1.7017719745635986,
+      "learning_rate": 1.9031413612565446e-05,
+      "loss": 0.1161,
+      "step": 728
+    },
+    {
+      "epoch": 0.38135149292823467,
+      "loss_partial": 0.050537109375,
+      "step": 728
+    },
+    {
+      "epoch": 0.38135149292823467,
+      "loss_partial": 0.0517578125,
+      "step": 728
+    },
+    {
+      "epoch": 0.3818753273965427,
+      "grad_norm": 1.9081051349639893,
+      "learning_rate": 1.905759162303665e-05,
+      "loss": 0.1123,
+      "step": 729
+    },
+    {
+      "epoch": 0.3818753273965427,
+      "loss_partial": 0.050537109375,
+      "step": 729
+    },
+    {
+      "epoch": 0.3818753273965427,
+      "loss_partial": 0.06494140625,
+      "step": 729
+    },
+    {
+      "epoch": 0.38239916186485073,
+      "grad_norm": 2.1978437900543213,
+      "learning_rate": 1.9083769633507854e-05,
+      "loss": 0.114,
+      "step": 730
+    },
+    {
+      "epoch": 0.38239916186485073,
+      "loss_partial": 0.05322265625,
+      "step": 730
+    },
+    {
+      "epoch": 0.38239916186485073,
+      "loss_partial": 0.05029296875,
+      "step": 730
+    },
+    {
+      "epoch": 0.38292299633315874,
+      "grad_norm": 2.2926220893859863,
+      "learning_rate": 1.9109947643979058e-05,
+      "loss": 0.114,
+      "step": 731
+    },
+    {
+      "epoch": 0.38292299633315874,
+      "loss_partial": 0.05322265625,
+      "step": 731
+    },
+    {
+      "epoch": 0.38292299633315874,
+      "loss_partial": 0.0517578125,
+      "step": 731
+    },
+    {
+      "epoch": 0.38344683080146674,
+      "grad_norm": 1.5374516248703003,
+      "learning_rate": 1.9136125654450265e-05,
+      "loss": 0.1176,
+      "step": 732
+    },
+    {
+      "epoch": 0.38344683080146674,
+      "loss_partial": 0.0576171875,
+      "step": 732
+    },
+    {
+      "epoch": 0.38344683080146674,
+      "loss_partial": 0.06298828125,
+      "step": 732
+    },
+    {
+      "epoch": 0.38397066526977475,
+      "grad_norm": 1.4940398931503296,
+      "learning_rate": 1.9162303664921466e-05,
+      "loss": 0.1171,
+      "step": 733
+    },
+    {
+      "epoch": 0.38397066526977475,
+      "loss_partial": 0.05078125,
+      "step": 733
+    },
+    {
+      "epoch": 0.38397066526977475,
+      "loss_partial": 0.06201171875,
+      "step": 733
+    },
+    {
+      "epoch": 0.38449449973808275,
+      "grad_norm": 1.0516942739486694,
+      "learning_rate": 1.9188481675392673e-05,
+      "loss": 0.1137,
+      "step": 734
+    },
+    {
+      "epoch": 0.38449449973808275,
+      "loss_partial": 0.0537109375,
+      "step": 734
+    },
+    {
+      "epoch": 0.38449449973808275,
+      "loss_partial": 0.054931640625,
+      "step": 734
+    },
+    {
+      "epoch": 0.38501833420639076,
+      "grad_norm": 1.1087830066680908,
+      "learning_rate": 1.9214659685863877e-05,
+      "loss": 0.1123,
+      "step": 735
+    },
+    {
+      "epoch": 0.38501833420639076,
+      "loss_partial": 0.060546875,
+      "step": 735
+    },
+    {
+      "epoch": 0.38501833420639076,
+      "loss_partial": 0.0615234375,
+      "step": 735
+    },
+    {
+      "epoch": 0.3855421686746988,
+      "grad_norm": 1.3794641494750977,
+      "learning_rate": 1.924083769633508e-05,
+      "loss": 0.1178,
+      "step": 736
+    },
+    {
+      "epoch": 0.3855421686746988,
+      "loss_partial": 0.05517578125,
+      "step": 736
+    },
+    {
+      "epoch": 0.3855421686746988,
+      "loss_partial": 0.05419921875,
+      "step": 736
+    },
+    {
+      "epoch": 0.3860660031430068,
+      "grad_norm": 1.002105474472046,
+      "learning_rate": 1.9267015706806285e-05,
+      "loss": 0.1146,
+      "step": 737
+    },
+    {
+      "epoch": 0.3860660031430068,
+      "loss_partial": 0.06005859375,
+      "step": 737
+    },
+    {
+      "epoch": 0.3860660031430068,
+      "loss_partial": 0.064453125,
+      "step": 737
+    },
+    {
+      "epoch": 0.3865898376113148,
+      "grad_norm": 1.388738751411438,
+      "learning_rate": 1.929319371727749e-05,
+      "loss": 0.1176,
+      "step": 738
+    },
+    {
+      "epoch": 0.3865898376113148,
+      "loss_partial": 0.06103515625,
+      "step": 738
+    },
+    {
+      "epoch": 0.3865898376113148,
+      "loss_partial": 0.05517578125,
+      "step": 738
+    },
+    {
+      "epoch": 0.38711367207962283,
+      "grad_norm": 1.492421269416809,
+      "learning_rate": 1.9319371727748693e-05,
+      "loss": 0.1111,
+      "step": 739
+    },
+    {
+      "epoch": 0.38711367207962283,
+      "loss_partial": 0.055419921875,
+      "step": 739
+    },
+    {
+      "epoch": 0.38711367207962283,
+      "loss_partial": 0.057373046875,
+      "step": 739
+    },
+    {
+      "epoch": 0.38763750654793083,
+      "grad_norm": 1.9023735523223877,
+      "learning_rate": 1.9345549738219897e-05,
+      "loss": 0.113,
+      "step": 740
+    },
+    {
+      "epoch": 0.38763750654793083,
+      "loss_partial": 0.058837890625,
+      "step": 740
+    },
+    {
+      "epoch": 0.38763750654793083,
+      "loss_partial": 0.056640625,
+      "step": 740
+    },
+    {
+      "epoch": 0.3881613410162389,
+      "grad_norm": 1.796362280845642,
+      "learning_rate": 1.93717277486911e-05,
+      "loss": 0.1163,
+      "step": 741
+    },
+    {
+      "epoch": 0.3881613410162389,
+      "loss_partial": 0.0517578125,
+      "step": 741
+    },
+    {
+      "epoch": 0.3881613410162389,
+      "loss_partial": 0.05029296875,
+      "step": 741
+    },
+    {
+      "epoch": 0.3886851754845469,
+      "grad_norm": 1.5628836154937744,
+      "learning_rate": 1.9397905759162305e-05,
+      "loss": 0.1115,
+      "step": 742
+    },
+    {
+      "epoch": 0.3886851754845469,
+      "loss_partial": 0.068359375,
+      "step": 742
+    },
+    {
+      "epoch": 0.3886851754845469,
+      "loss_partial": 0.0576171875,
+      "step": 742
+    },
+    {
+      "epoch": 0.3892090099528549,
+      "grad_norm": 2.4459116458892822,
+      "learning_rate": 1.9424083769633512e-05,
+      "loss": 0.1144,
+      "step": 743
+    },
+    {
+      "epoch": 0.3892090099528549,
+      "loss_partial": 0.06005859375,
+      "step": 743
+    },
+    {
+      "epoch": 0.3892090099528549,
+      "loss_partial": 0.055419921875,
+      "step": 743
+    },
+    {
+      "epoch": 0.3897328444211629,
+      "grad_norm": 1.2941135168075562,
+      "learning_rate": 1.9450261780104713e-05,
+      "loss": 0.1117,
+      "step": 744
+    },
+    {
+      "epoch": 0.3897328444211629,
+      "loss_partial": 0.060546875,
+      "step": 744
+    },
+    {
+      "epoch": 0.3897328444211629,
+      "loss_partial": 0.05517578125,
+      "step": 744
+    },
+    {
+      "epoch": 0.3902566788894709,
+      "grad_norm": 1.7356597185134888,
+      "learning_rate": 1.947643979057592e-05,
+      "loss": 0.1119,
+      "step": 745
+    },
+    {
+      "epoch": 0.3902566788894709,
+      "loss_partial": 0.054931640625,
+      "step": 745
+    },
+    {
+      "epoch": 0.3902566788894709,
+      "loss_partial": 0.051513671875,
+      "step": 745
+    },
+    {
+      "epoch": 0.39078051335777897,
+      "grad_norm": 2.5171666145324707,
+      "learning_rate": 1.950261780104712e-05,
+      "loss": 0.1115,
+      "step": 746
+    },
+    {
+      "epoch": 0.39078051335777897,
+      "loss_partial": 0.060546875,
+      "step": 746
+    },
+    {
+      "epoch": 0.39078051335777897,
+      "loss_partial": 0.055419921875,
+      "step": 746
+    },
+    {
+      "epoch": 0.391304347826087,
+      "grad_norm": 2.7877321243286133,
+      "learning_rate": 1.9528795811518328e-05,
+      "loss": 0.1139,
+      "step": 747
+    },
+    {
+      "epoch": 0.391304347826087,
+      "loss_partial": 0.053466796875,
+      "step": 747
+    },
+    {
+      "epoch": 0.391304347826087,
+      "loss_partial": 0.0546875,
+      "step": 747
+    },
+    {
+      "epoch": 0.391828182294395,
+      "grad_norm": 3.060483694076538,
+      "learning_rate": 1.955497382198953e-05,
+      "loss": 0.1147,
+      "step": 748
+    },
+    {
+      "epoch": 0.391828182294395,
+      "loss_partial": 0.06298828125,
+      "step": 748
+    },
+    {
+      "epoch": 0.391828182294395,
+      "loss_partial": 0.061767578125,
+      "step": 748
+    },
+    {
+      "epoch": 0.392352016762703,
+      "grad_norm": 2.4794883728027344,
+      "learning_rate": 1.9581151832460736e-05,
+      "loss": 0.1146,
+      "step": 749
+    },
+    {
+      "epoch": 0.392352016762703,
+      "loss_partial": 0.054931640625,
+      "step": 749
+    },
+    {
+      "epoch": 0.392352016762703,
+      "loss_partial": 0.06298828125,
+      "step": 749
+    },
+    {
+      "epoch": 0.392875851231011,
+      "grad_norm": 2.1987626552581787,
+      "learning_rate": 1.960732984293194e-05,
+      "loss": 0.1154,
+      "step": 750
+    },
+    {
+      "epoch": 0.392875851231011,
+      "loss_partial": 0.0517578125,
+      "step": 750
+    },
+    {
+      "epoch": 0.392875851231011,
+      "loss_partial": 0.04931640625,
+      "step": 750
+    },
+    {
+      "epoch": 0.393399685699319,
+      "grad_norm": 2.164353847503662,
+      "learning_rate": 1.9633507853403143e-05,
+      "loss": 0.1142,
+      "step": 751
+    },
+    {
+      "epoch": 0.393399685699319,
+      "loss_partial": 0.0537109375,
+      "step": 751
+    },
+    {
+      "epoch": 0.393399685699319,
+      "loss_partial": 0.0556640625,
+      "step": 751
+    },
+    {
+      "epoch": 0.39392352016762705,
+      "grad_norm": 1.6272916793823242,
+      "learning_rate": 1.9659685863874347e-05,
+      "loss": 0.112,
+      "step": 752
+    },
+    {
+      "epoch": 0.39392352016762705,
+      "loss_partial": 0.06640625,
+      "step": 752
+    },
+    {
+      "epoch": 0.39392352016762705,
+      "loss_partial": 0.05908203125,
+      "step": 752
+    },
+    {
+      "epoch": 0.39444735463593505,
+      "grad_norm": 1.7089800834655762,
+      "learning_rate": 1.968586387434555e-05,
+      "loss": 0.1153,
+      "step": 753
+    },
+    {
+      "epoch": 0.39444735463593505,
+      "loss_partial": 0.05029296875,
+      "step": 753
+    },
+    {
+      "epoch": 0.39444735463593505,
+      "loss_partial": 0.06201171875,
+      "step": 753
+    },
+    {
+      "epoch": 0.39497118910424306,
+      "grad_norm": 1.9634721279144287,
+      "learning_rate": 1.9712041884816755e-05,
+      "loss": 0.1191,
+      "step": 754
+    },
+    {
+      "epoch": 0.39497118910424306,
+      "loss_partial": 0.0625,
+      "step": 754
+    },
+    {
+      "epoch": 0.39497118910424306,
+      "loss_partial": 0.0625,
+      "step": 754
+    },
+    {
+      "epoch": 0.39549502357255106,
+      "grad_norm": 1.280761480331421,
+      "learning_rate": 1.973821989528796e-05,
+      "loss": 0.1119,
+      "step": 755
+    },
+    {
+      "epoch": 0.39549502357255106,
+      "loss_partial": 0.0478515625,
+      "step": 755
+    },
+    {
+      "epoch": 0.39549502357255106,
+      "loss_partial": 0.06396484375,
+      "step": 755
+    },
+    {
+      "epoch": 0.39601885804085907,
+      "grad_norm": 1.7637323141098022,
+      "learning_rate": 1.9764397905759163e-05,
+      "loss": 0.1197,
+      "step": 756
+    },
+    {
+      "epoch": 0.39601885804085907,
+      "loss_partial": 0.05419921875,
+      "step": 756
+    },
+    {
+      "epoch": 0.39601885804085907,
+      "loss_partial": 0.056884765625,
+      "step": 756
+    },
+    {
+      "epoch": 0.3965426925091671,
+      "grad_norm": 2.1437506675720215,
+      "learning_rate": 1.9790575916230367e-05,
+      "loss": 0.1132,
+      "step": 757
+    },
+    {
+      "epoch": 0.3965426925091671,
+      "loss_partial": 0.057373046875,
+      "step": 757
+    },
+    {
+      "epoch": 0.3965426925091671,
+      "loss_partial": 0.06298828125,
+      "step": 757
+    },
+    {
+      "epoch": 0.39706652697747513,
+      "grad_norm": 3.003790855407715,
+      "learning_rate": 1.9816753926701574e-05,
+      "loss": 0.1174,
+      "step": 758
+    },
+    {
+      "epoch": 0.39706652697747513,
+      "loss_partial": 0.06494140625,
+      "step": 758
+    },
+    {
+      "epoch": 0.39706652697747513,
+      "loss_partial": 0.06591796875,
+      "step": 758
+    },
+    {
+      "epoch": 0.39759036144578314,
+      "grad_norm": 1.5441396236419678,
+      "learning_rate": 1.9842931937172775e-05,
+      "loss": 0.1173,
+      "step": 759
+    },
+    {
+      "epoch": 0.39759036144578314,
+      "loss_partial": 0.053466796875,
+      "step": 759
+    },
+    {
+      "epoch": 0.39759036144578314,
+      "loss_partial": 0.056396484375,
+      "step": 759
+    },
+    {
+      "epoch": 0.39811419591409114,
+      "grad_norm": 1.5172356367111206,
+      "learning_rate": 1.9869109947643982e-05,
+      "loss": 0.1165,
+      "step": 760
+    },
+    {
+      "epoch": 0.39811419591409114,
+      "loss_partial": 0.05419921875,
+      "step": 760
+    },
+    {
+      "epoch": 0.39811419591409114,
+      "loss_partial": 0.0478515625,
+      "step": 760
+    },
+    {
+      "epoch": 0.39863803038239914,
+      "grad_norm": 1.5124863386154175,
+      "learning_rate": 1.9895287958115186e-05,
+      "loss": 0.1126,
+      "step": 761
+    },
+    {
+      "epoch": 0.39863803038239914,
+      "loss_partial": 0.058837890625,
+      "step": 761
+    },
+    {
+      "epoch": 0.39863803038239914,
+      "loss_partial": 0.054931640625,
+      "step": 761
+    },
+    {
+      "epoch": 0.3991618648507072,
+      "grad_norm": 2.0935513973236084,
+      "learning_rate": 1.992146596858639e-05,
+      "loss": 0.1175,
+      "step": 762
+    },
+    {
+      "epoch": 0.3991618648507072,
+      "loss_partial": 0.04931640625,
+      "step": 762
+    },
+    {
+      "epoch": 0.3991618648507072,
+      "loss_partial": 0.050048828125,
+      "step": 762
+    },
+    {
+      "epoch": 0.3996856993190152,
+      "grad_norm": 1.637799859046936,
+      "learning_rate": 1.9947643979057594e-05,
+      "loss": 0.1119,
+      "step": 763
+    },
+    {
+      "epoch": 0.3996856993190152,
+      "loss_partial": 0.045654296875,
+      "step": 763
+    },
+    {
+      "epoch": 0.3996856993190152,
+      "loss_partial": 0.06787109375,
+      "step": 763
+    },
+    {
+      "epoch": 0.4002095337873232,
+      "grad_norm": 1.3277908563613892,
+      "learning_rate": 1.9973821989528798e-05,
+      "loss": 0.1129,
+      "step": 764
+    },
+    {
+      "epoch": 0.4002095337873232,
+      "loss_partial": 0.049560546875,
+      "step": 764
+    },
+    {
+      "epoch": 0.4002095337873232,
+      "loss_partial": 0.055419921875,
+      "step": 764
+    },
+    {
+      "epoch": 0.4007333682556312,
+      "grad_norm": 1.4582921266555786,
+      "learning_rate": 2e-05,
+      "loss": 0.1139,
+      "step": 765
+    },
+    {
+      "epoch": 0.4007333682556312,
+      "loss_partial": 0.06005859375,
+      "step": 765
+    },
+    {
+      "epoch": 0.4007333682556312,
+      "loss_partial": 0.049560546875,
+      "step": 765
+    },
+    {
+      "epoch": 0.4012572027239392,
+      "grad_norm": 2.342686176300049,
+      "learning_rate": 1.999999895503092e-05,
+      "loss": 0.1136,
+      "step": 766
+    },
+    {
+      "epoch": 0.4012572027239392,
+      "loss_partial": 0.055908203125,
+      "step": 766
+    },
+    {
+      "epoch": 0.4012572027239392,
+      "loss_partial": 0.05126953125,
+      "step": 766
+    },
+    {
+      "epoch": 0.4017810371922472,
+      "grad_norm": 2.555981159210205,
+      "learning_rate": 1.9999995820123898e-05,
+      "loss": 0.1168,
+      "step": 767
+    },
+    {
+      "epoch": 0.4017810371922472,
+      "loss_partial": 0.04833984375,
+      "step": 767
+    },
+    {
+      "epoch": 0.4017810371922472,
+      "loss_partial": 0.050537109375,
+      "step": 767
+    },
+    {
+      "epoch": 0.4023048716605553,
+      "grad_norm": 1.9616669416427612,
+      "learning_rate": 1.9999990595279584e-05,
+      "loss": 0.1136,
+      "step": 768
+    },
+    {
+      "epoch": 0.4023048716605553,
+      "loss_partial": 0.050537109375,
+      "step": 768
+    },
+    {
+      "epoch": 0.4023048716605553,
+      "loss_partial": 0.0673828125,
+      "step": 768
+    },
+    {
+      "epoch": 0.4028287061288633,
+      "grad_norm": 2.07094407081604,
+      "learning_rate": 1.9999983280499075e-05,
+      "loss": 0.118,
+      "step": 769
+    },
+    {
+      "epoch": 0.4028287061288633,
+      "loss_partial": 0.048583984375,
+      "step": 769
+    },
+    {
+      "epoch": 0.4028287061288633,
+      "loss_partial": 0.056640625,
+      "step": 769
+    },
+    {
+      "epoch": 0.4033525405971713,
+      "grad_norm": 1.4839967489242554,
+      "learning_rate": 1.99999738757839e-05,
+      "loss": 0.1126,
+      "step": 770
+    },
+    {
+      "epoch": 0.4033525405971713,
+      "loss_partial": 0.0654296875,
+      "step": 770
+    },
+    {
+      "epoch": 0.4033525405971713,
+      "loss_partial": 0.06298828125,
+      "step": 770
+    },
+    {
+      "epoch": 0.4038763750654793,
+      "grad_norm": 1.5439138412475586,
+      "learning_rate": 1.9999962381136018e-05,
+      "loss": 0.1176,
+      "step": 771
+    },
+    {
+      "epoch": 0.4038763750654793,
+      "loss_partial": 0.05322265625,
+      "step": 771
+    },
+    {
+      "epoch": 0.4038763750654793,
+      "loss_partial": 0.05712890625,
+      "step": 771
+    },
+    {
+      "epoch": 0.4044002095337873,
+      "grad_norm": 1.8508546352386475,
+      "learning_rate": 1.999994879655784e-05,
+      "loss": 0.1146,
+      "step": 772
+    },
+    {
+      "epoch": 0.4044002095337873,
+      "loss_partial": 0.05908203125,
+      "step": 772
+    },
+    {
+      "epoch": 0.4044002095337873,
+      "loss_partial": 0.06298828125,
+      "step": 772
+    },
+    {
+      "epoch": 0.40492404400209536,
+      "grad_norm": 2.6969680786132812,
+      "learning_rate": 1.99999331220522e-05,
+      "loss": 0.1133,
+      "step": 773
+    },
+    {
+      "epoch": 0.40492404400209536,
+      "loss_partial": 0.052490234375,
+      "step": 773
+    },
+    {
+      "epoch": 0.40492404400209536,
+      "loss_partial": 0.052490234375,
+      "step": 773
+    },
+    {
+      "epoch": 0.40544787847040337,
+      "grad_norm": 2.3233108520507812,
+      "learning_rate": 1.9999915357622378e-05,
+      "loss": 0.1177,
+      "step": 774
+    },
+    {
+      "epoch": 0.40544787847040337,
+      "loss_partial": 0.0634765625,
+      "step": 774
+    },
+    {
+      "epoch": 0.40544787847040337,
+      "loss_partial": 0.06103515625,
+      "step": 774
+    },
+    {
+      "epoch": 0.40597171293871137,
+      "grad_norm": 1.6282570362091064,
+      "learning_rate": 1.999989550327208e-05,
+      "loss": 0.1128,
+      "step": 775
+    },
+    {
+      "epoch": 0.40597171293871137,
+      "loss_partial": 0.06689453125,
+      "step": 775
+    },
+    {
+      "epoch": 0.40597171293871137,
+      "loss_partial": 0.051025390625,
+      "step": 775
+    },
+    {
+      "epoch": 0.4064955474070194,
+      "grad_norm": 1.6509393453598022,
+      "learning_rate": 1.9999873559005463e-05,
+      "loss": 0.1203,
+      "step": 776
+    },
+    {
+      "epoch": 0.4064955474070194,
+      "loss_partial": 0.05224609375,
+      "step": 776
+    },
+    {
+      "epoch": 0.4064955474070194,
+      "loss_partial": 0.05517578125,
+      "step": 776
+    },
+    {
+      "epoch": 0.4070193818753274,
+      "grad_norm": 1.9719444513320923,
+      "learning_rate": 1.9999849524827108e-05,
+      "loss": 0.1166,
+      "step": 777
+    },
+    {
+      "epoch": 0.4070193818753274,
+      "loss_partial": 0.0517578125,
+      "step": 777
+    },
+    {
+      "epoch": 0.4070193818753274,
+      "loss_partial": 0.056396484375,
+      "step": 777
+    },
+    {
+      "epoch": 0.4075432163436354,
+      "grad_norm": 1.6004106998443604,
+      "learning_rate": 1.999982340074204e-05,
+      "loss": 0.117,
+      "step": 778
+    },
+    {
+      "epoch": 0.4075432163436354,
+      "loss_partial": 0.053955078125,
+      "step": 778
+    },
+    {
+      "epoch": 0.4075432163436354,
+      "loss_partial": 0.058837890625,
+      "step": 778
+    },
+    {
+      "epoch": 0.40806705081194344,
+      "grad_norm": 1.9237130880355835,
+      "learning_rate": 1.9999795186755716e-05,
+      "loss": 0.1158,
+      "step": 779
+    },
+    {
+      "epoch": 0.40806705081194344,
+      "loss_partial": 0.05078125,
+      "step": 779
+    },
+    {
+      "epoch": 0.40806705081194344,
+      "loss_partial": 0.06201171875,
+      "step": 779
+    },
+    {
+      "epoch": 0.40859088528025145,
+      "grad_norm": 2.634411334991455,
+      "learning_rate": 1.9999764882874038e-05,
+      "loss": 0.117,
+      "step": 780
+    },
+    {
+      "epoch": 0.40859088528025145,
+      "loss_partial": 0.0576171875,
+      "step": 780
+    },
+    {
+      "epoch": 0.40859088528025145,
+      "loss_partial": 0.05810546875,
+      "step": 780
+    },
+    {
+      "epoch": 0.40911471974855945,
+      "grad_norm": 1.931278944015503,
+      "learning_rate": 1.9999732489103334e-05,
+      "loss": 0.1122,
+      "step": 781
+    },
+    {
+      "epoch": 0.40911471974855945,
+      "loss_partial": 0.0478515625,
+      "step": 781
+    },
+    {
+      "epoch": 0.40911471974855945,
+      "loss_partial": 0.051513671875,
+      "step": 781
+    },
+    {
+      "epoch": 0.40963855421686746,
+      "grad_norm": 1.797959566116333,
+      "learning_rate": 1.9999698005450378e-05,
+      "loss": 0.1147,
+      "step": 782
+    },
+    {
+      "epoch": 0.40963855421686746,
+      "loss_partial": 0.056640625,
+      "step": 782
+    },
+    {
+      "epoch": 0.40963855421686746,
+      "loss_partial": 0.060302734375,
+      "step": 782
+    },
+    {
+      "epoch": 0.41016238868517546,
+      "grad_norm": 1.595902919769287,
+      "learning_rate": 1.9999661431922375e-05,
+      "loss": 0.1116,
+      "step": 783
+    },
+    {
+      "epoch": 0.41016238868517546,
+      "loss_partial": 0.051025390625,
+      "step": 783
+    },
+    {
+      "epoch": 0.41016238868517546,
+      "loss_partial": 0.060791015625,
+      "step": 783
+    },
+    {
+      "epoch": 0.4106862231534835,
+      "grad_norm": 1.7588170766830444,
+      "learning_rate": 1.9999622768526967e-05,
+      "loss": 0.1115,
+      "step": 784
+    },
+    {
+      "epoch": 0.4106862231534835,
+      "loss_partial": 0.06103515625,
+      "step": 784
+    },
+    {
+      "epoch": 0.4106862231534835,
+      "loss_partial": 0.057861328125,
+      "step": 784
+    },
+    {
+      "epoch": 0.4112100576217915,
+      "grad_norm": 2.9138529300689697,
+      "learning_rate": 1.9999582015272236e-05,
+      "loss": 0.1166,
+      "step": 785
+    },
+    {
+      "epoch": 0.4112100576217915,
+      "loss_partial": 0.059326171875,
+      "step": 785
+    },
+    {
+      "epoch": 0.4112100576217915,
+      "loss_partial": 0.056640625,
+      "step": 785
+    },
+    {
+      "epoch": 0.41173389209009953,
+      "grad_norm": 1.663454532623291,
+      "learning_rate": 1.99995391721667e-05,
+      "loss": 0.113,
+      "step": 786
+    },
+    {
+      "epoch": 0.41173389209009953,
+      "loss_partial": 0.056396484375,
+      "step": 786
+    },
+    {
+      "epoch": 0.41173389209009953,
+      "loss_partial": 0.06396484375,
+      "step": 786
+    },
+    {
+      "epoch": 0.41225772655840753,
+      "grad_norm": 1.790297269821167,
+      "learning_rate": 1.9999494239219313e-05,
+      "loss": 0.1141,
+      "step": 787
+    },
+    {
+      "epoch": 0.41225772655840753,
+      "loss_partial": 0.0576171875,
+      "step": 787
+    },
+    {
+      "epoch": 0.41225772655840753,
+      "loss_partial": 0.052001953125,
+      "step": 787
+    },
+    {
+      "epoch": 0.41278156102671554,
+      "grad_norm": 1.3720468282699585,
+      "learning_rate": 1.9999447216439465e-05,
+      "loss": 0.1114,
+      "step": 788
+    },
+    {
+      "epoch": 0.41278156102671554,
+      "loss_partial": 0.051513671875,
+      "step": 788
+    },
+    {
+      "epoch": 0.41278156102671554,
+      "loss_partial": 0.078125,
+      "step": 788
+    },
+    {
+      "epoch": 0.4133053954950236,
+      "grad_norm": 1.4562180042266846,
+      "learning_rate": 1.9999398103836984e-05,
+      "loss": 0.1171,
+      "step": 789
+    },
+    {
+      "epoch": 0.4133053954950236,
+      "loss_partial": 0.0693359375,
+      "step": 789
+    },
+    {
+      "epoch": 0.4133053954950236,
+      "loss_partial": 0.0634765625,
+      "step": 789
+    },
+    {
+      "epoch": 0.4138292299633316,
+      "grad_norm": 1.252482533454895,
+      "learning_rate": 1.999934690142213e-05,
+      "loss": 0.113,
+      "step": 790
+    },
+    {
+      "epoch": 0.4138292299633316,
+      "loss_partial": 0.06640625,
+      "step": 790
+    },
+    {
+      "epoch": 0.4138292299633316,
+      "loss_partial": 0.05419921875,
+      "step": 790
+    },
+    {
+      "epoch": 0.4143530644316396,
+      "grad_norm": 1.7458158731460571,
+      "learning_rate": 1.9999293609205612e-05,
+      "loss": 0.1146,
+      "step": 791
+    },
+    {
+      "epoch": 0.4143530644316396,
+      "loss_partial": 0.048828125,
+      "step": 791
+    },
+    {
+      "epoch": 0.4143530644316396,
+      "loss_partial": 0.059814453125,
+      "step": 791
+    },
+    {
+      "epoch": 0.4148768988999476,
+      "grad_norm": 1.9501049518585205,
+      "learning_rate": 1.9999238227198564e-05,
+      "loss": 0.1123,
+      "step": 792
+    },
+    {
+      "epoch": 0.4148768988999476,
+      "loss_partial": 0.052490234375,
+      "step": 792
+    },
+    {
+      "epoch": 0.4148768988999476,
+      "loss_partial": 0.0654296875,
+      "step": 792
+    },
+    {
+      "epoch": 0.4154007333682556,
+      "grad_norm": 1.989865779876709,
+      "learning_rate": 1.9999180755412556e-05,
+      "loss": 0.1158,
+      "step": 793
+    },
+    {
+      "epoch": 0.4154007333682556,
+      "loss_partial": 0.06689453125,
+      "step": 793
+    },
+    {
+      "epoch": 0.4154007333682556,
+      "loss_partial": 0.05908203125,
+      "step": 793
+    },
+    {
+      "epoch": 0.4159245678365636,
+      "grad_norm": 1.3630294799804688,
+      "learning_rate": 1.9999121193859604e-05,
+      "loss": 0.1192,
+      "step": 794
+    },
+    {
+      "epoch": 0.4159245678365636,
+      "loss_partial": 0.050048828125,
+      "step": 794
+    },
+    {
+      "epoch": 0.4159245678365636,
+      "loss_partial": 0.06298828125,
+      "step": 794
+    },
+    {
+      "epoch": 0.4164484023048717,
+      "grad_norm": 1.552640676498413,
+      "learning_rate": 1.999905954255216e-05,
+      "loss": 0.1173,
+      "step": 795
+    },
+    {
+      "epoch": 0.4164484023048717,
+      "loss_partial": 0.056640625,
+      "step": 795
+    },
+    {
+      "epoch": 0.4164484023048717,
+      "loss_partial": 0.04833984375,
+      "step": 795
+    },
+    {
+      "epoch": 0.4169722367731797,
+      "grad_norm": 1.79812753200531,
+      "learning_rate": 1.99989958015031e-05,
+      "loss": 0.1117,
+      "step": 796
+    },
+    {
+      "epoch": 0.4169722367731797,
+      "loss_partial": 0.0634765625,
+      "step": 796
+    },
+    {
+      "epoch": 0.4169722367731797,
+      "loss_partial": 0.047607421875,
+      "step": 796
+    },
+    {
+      "epoch": 0.4174960712414877,
+      "grad_norm": 1.1519502401351929,
+      "learning_rate": 1.999892997072575e-05,
+      "loss": 0.113,
+      "step": 797
+    },
+    {
+      "epoch": 0.4174960712414877,
+      "loss_partial": 0.053955078125,
+      "step": 797
+    },
+    {
+      "epoch": 0.4174960712414877,
+      "loss_partial": 0.060791015625,
+      "step": 797
+    },
+    {
+      "epoch": 0.4180199057097957,
+      "grad_norm": 2.2884621620178223,
+      "learning_rate": 1.9998862050233864e-05,
+      "loss": 0.1114,
+      "step": 798
+    },
+    {
+      "epoch": 0.4180199057097957,
+      "loss_partial": 0.0673828125,
+      "step": 798
+    },
+    {
+      "epoch": 0.4180199057097957,
+      "loss_partial": 0.06640625,
+      "step": 798
+    },
+    {
+      "epoch": 0.4185437401781037,
+      "grad_norm": 2.7330431938171387,
+      "learning_rate": 1.9998792040041643e-05,
+      "loss": 0.1159,
+      "step": 799
+    },
+    {
+      "epoch": 0.4185437401781037,
+      "loss_partial": 0.06298828125,
+      "step": 799
+    },
+    {
+      "epoch": 0.4185437401781037,
+      "loss_partial": 0.0830078125,
+      "step": 799
+    },
+    {
+      "epoch": 0.41906757464641176,
+      "grad_norm": 1.4423155784606934,
+      "learning_rate": 1.999871994016372e-05,
+      "loss": 0.1159,
+      "step": 800
+    },
+    {
+      "epoch": 0.41906757464641176,
+      "loss_partial": 0.06884765625,
+      "step": 800
+    },
+    {
+      "epoch": 0.41906757464641176,
+      "loss_partial": 0.057373046875,
+      "step": 800
+    },
+    {
+      "epoch": 0.41959140911471976,
+      "grad_norm": 1.9763787984848022,
+      "learning_rate": 1.9998645750615155e-05,
+      "loss": 0.1175,
+      "step": 801
+    },
+    {
+      "epoch": 0.41959140911471976,
+      "loss_partial": 0.05322265625,
+      "step": 801
+    },
+    {
+      "epoch": 0.41959140911471976,
+      "loss_partial": 0.052001953125,
+      "step": 801
+    },
+    {
+      "epoch": 0.42011524358302776,
+      "grad_norm": 2.010662317276001,
+      "learning_rate": 1.999856947141146e-05,
+      "loss": 0.1177,
+      "step": 802
+    },
+    {
+      "epoch": 0.42011524358302776,
+      "loss_partial": 0.058349609375,
+      "step": 802
+    },
+    {
+      "epoch": 0.42011524358302776,
+      "loss_partial": 0.06591796875,
+      "step": 802
+    },
+    {
+      "epoch": 0.42063907805133577,
+      "grad_norm": 1.620468258857727,
+      "learning_rate": 1.9998491102568575e-05,
+      "loss": 0.1114,
+      "step": 803
+    },
+    {
+      "epoch": 0.42063907805133577,
+      "loss_partial": 0.06298828125,
+      "step": 803
+    },
+    {
+      "epoch": 0.42063907805133577,
+      "loss_partial": 0.05859375,
+      "step": 803
+    },
+    {
+      "epoch": 0.4211629125196438,
+      "grad_norm": 1.371483325958252,
+      "learning_rate": 1.999841064410288e-05,
+      "loss": 0.1171,
+      "step": 804
+    },
+    {
+      "epoch": 0.4211629125196438,
+      "loss_partial": 0.056884765625,
+      "step": 804
+    },
+    {
+      "epoch": 0.4211629125196438,
+      "loss_partial": 0.05859375,
+      "step": 804
+    },
+    {
+      "epoch": 0.42168674698795183,
+      "grad_norm": 1.952289342880249,
+      "learning_rate": 1.9998328096031186e-05,
+      "loss": 0.1151,
+      "step": 805
+    },
+    {
+      "epoch": 0.42168674698795183,
+      "loss_partial": 0.053955078125,
+      "step": 805
+    },
+    {
+      "epoch": 0.42168674698795183,
+      "loss_partial": 0.053955078125,
+      "step": 805
+    },
+    {
+      "epoch": 0.42221058145625984,
+      "grad_norm": 2.1545932292938232,
+      "learning_rate": 1.999824345837075e-05,
+      "loss": 0.1159,
+      "step": 806
+    },
+    {
+      "epoch": 0.42221058145625984,
+      "loss_partial": 0.054931640625,
+      "step": 806
+    },
+    {
+      "epoch": 0.42221058145625984,
+      "loss_partial": 0.061279296875,
+      "step": 806
+    },
+    {
+      "epoch": 0.42273441592456784,
+      "grad_norm": 2.1476857662200928,
+      "learning_rate": 1.9998156731139254e-05,
+      "loss": 0.1165,
+      "step": 807
+    },
+    {
+      "epoch": 0.42273441592456784,
+      "loss_partial": 0.05712890625,
+      "step": 807
+    },
+    {
+      "epoch": 0.42273441592456784,
+      "loss_partial": 0.055908203125,
+      "step": 807
+    },
+    {
+      "epoch": 0.42325825039287585,
+      "grad_norm": 2.028904438018799,
+      "learning_rate": 1.9998067914354832e-05,
+      "loss": 0.1162,
+      "step": 808
+    },
+    {
+      "epoch": 0.42325825039287585,
+      "loss_partial": 0.055908203125,
+      "step": 808
+    },
+    {
+      "epoch": 0.42325825039287585,
+      "loss_partial": 0.06591796875,
+      "step": 808
+    },
+    {
+      "epoch": 0.42378208486118385,
+      "grad_norm": 2.1423215866088867,
+      "learning_rate": 1.999797700803604e-05,
+      "loss": 0.1139,
+      "step": 809
+    },
+    {
+      "epoch": 0.42378208486118385,
+      "loss_partial": 0.0625,
+      "step": 809
+    },
+    {
+      "epoch": 0.42378208486118385,
+      "loss_partial": 0.05712890625,
+      "step": 809
+    },
+    {
+      "epoch": 0.42430591932949185,
+      "grad_norm": 2.267958164215088,
+      "learning_rate": 1.999788401220188e-05,
+      "loss": 0.1168,
+      "step": 810
+    },
+    {
+      "epoch": 0.42430591932949185,
+      "loss_partial": 0.06591796875,
+      "step": 810
+    },
+    {
+      "epoch": 0.42430591932949185,
+      "loss_partial": 0.057861328125,
+      "step": 810
+    },
+    {
+      "epoch": 0.4248297537977999,
+      "grad_norm": 1.9203699827194214,
+      "learning_rate": 1.9997788926871787e-05,
+      "loss": 0.113,
+      "step": 811
+    },
+    {
+      "epoch": 0.4248297537977999,
+      "loss_partial": 0.04541015625,
+      "step": 811
+    },
+    {
+      "epoch": 0.4248297537977999,
+      "loss_partial": 0.06103515625,
+      "step": 811
+    },
+    {
+      "epoch": 0.4253535882661079,
+      "grad_norm": 1.9822297096252441,
+      "learning_rate": 1.9997691752065635e-05,
+      "loss": 0.1125,
+      "step": 812
+    },
+    {
+      "epoch": 0.4253535882661079,
+      "loss_partial": 0.059814453125,
+      "step": 812
+    },
+    {
+      "epoch": 0.4253535882661079,
+      "loss_partial": 0.056884765625,
+      "step": 812
+    },
+    {
+      "epoch": 0.4258774227344159,
+      "grad_norm": 1.8722342252731323,
+      "learning_rate": 1.9997592487803724e-05,
+      "loss": 0.1162,
+      "step": 813
+    },
+    {
+      "epoch": 0.4258774227344159,
+      "loss_partial": 0.06298828125,
+      "step": 813
+    },
+    {
+      "epoch": 0.4258774227344159,
+      "loss_partial": 0.056396484375,
+      "step": 813
+    },
+    {
+      "epoch": 0.4264012572027239,
+      "grad_norm": 1.7499629259109497,
+      "learning_rate": 1.999749113410681e-05,
+      "loss": 0.1104,
+      "step": 814
+    },
+    {
+      "epoch": 0.4264012572027239,
+      "loss_partial": 0.05126953125,
+      "step": 814
+    },
+    {
+      "epoch": 0.4264012572027239,
+      "loss_partial": 0.05029296875,
+      "step": 814
+    },
+    {
+      "epoch": 0.42692509167103193,
+      "grad_norm": 3.0445024967193604,
+      "learning_rate": 1.9997387690996072e-05,
+      "loss": 0.1124,
+      "step": 815
+    },
+    {
+      "epoch": 0.42692509167103193,
+      "loss_partial": 0.0546875,
+      "step": 815
+    },
+    {
+      "epoch": 0.42692509167103193,
+      "loss_partial": 0.057373046875,
+      "step": 815
+    },
+    {
+      "epoch": 0.42744892613934,
+      "grad_norm": 2.9422316551208496,
+      "learning_rate": 1.999728215849313e-05,
+      "loss": 0.1101,
+      "step": 816
+    },
+    {
+      "epoch": 0.42744892613934,
+      "loss_partial": 0.047607421875,
+      "step": 816
+    },
+    {
+      "epoch": 0.42744892613934,
+      "loss_partial": 0.060546875,
+      "step": 816
+    },
+    {
+      "epoch": 0.427972760607648,
+      "grad_norm": 1.8268113136291504,
+      "learning_rate": 1.9997174536620035e-05,
+      "loss": 0.1121,
+      "step": 817
+    },
+    {
+      "epoch": 0.427972760607648,
+      "loss_partial": 0.0478515625,
+      "step": 817
+    },
+    {
+      "epoch": 0.427972760607648,
+      "loss_partial": 0.058837890625,
+      "step": 817
+    },
+    {
+      "epoch": 0.428496595075956,
+      "grad_norm": 3.0715243816375732,
+      "learning_rate": 1.9997064825399282e-05,
+      "loss": 0.1127,
+      "step": 818
+    },
+    {
+      "epoch": 0.428496595075956,
+      "loss_partial": 0.060302734375,
+      "step": 818
+    },
+    {
+      "epoch": 0.428496595075956,
+      "loss_partial": 0.056396484375,
+      "step": 818
+    },
+    {
+      "epoch": 0.429020429544264,
+      "grad_norm": 2.182481288909912,
+      "learning_rate": 1.9996953024853803e-05,
+      "loss": 0.115,
+      "step": 819
+    },
+    {
+      "epoch": 0.429020429544264,
+      "loss_partial": 0.05517578125,
+      "step": 819
+    },
+    {
+      "epoch": 0.429020429544264,
+      "loss_partial": 0.049560546875,
+      "step": 819
+    },
+    {
+      "epoch": 0.429544264012572,
+      "grad_norm": 2.4219372272491455,
+      "learning_rate": 1.9996839135006963e-05,
+      "loss": 0.1151,
+      "step": 820
+    },
+    {
+      "epoch": 0.429544264012572,
+      "loss_partial": 0.0703125,
+      "step": 820
+    },
+    {
+      "epoch": 0.429544264012572,
+      "loss_partial": 0.058349609375,
+      "step": 820
+    },
+    {
+      "epoch": 0.43006809848088007,
+      "grad_norm": 2.1910781860351562,
+      "learning_rate": 1.999672315588256e-05,
+      "loss": 0.114,
+      "step": 821
+    },
+    {
+      "epoch": 0.43006809848088007,
+      "loss_partial": 0.08251953125,
+      "step": 821
+    },
+    {
+      "epoch": 0.43006809848088007,
+      "loss_partial": 0.05322265625,
+      "step": 821
+    },
+    {
+      "epoch": 0.43059193294918807,
+      "grad_norm": 3.0139007568359375,
+      "learning_rate": 1.9996605087504834e-05,
+      "loss": 0.1192,
+      "step": 822
+    },
+    {
+      "epoch": 0.43059193294918807,
+      "loss_partial": 0.061767578125,
+      "step": 822
+    },
+    {
+      "epoch": 0.43059193294918807,
+      "loss_partial": 0.055908203125,
+      "step": 822
+    },
+    {
+      "epoch": 0.4311157674174961,
+      "grad_norm": 1.58257257938385,
+      "learning_rate": 1.9996484929898465e-05,
+      "loss": 0.112,
+      "step": 823
+    },
+    {
+      "epoch": 0.4311157674174961,
+      "loss_partial": 0.05712890625,
+      "step": 823
+    },
+    {
+      "epoch": 0.4311157674174961,
+      "loss_partial": 0.06689453125,
+      "step": 823
+    },
+    {
+      "epoch": 0.4316396018858041,
+      "grad_norm": 2.342312812805176,
+      "learning_rate": 1.9996362683088558e-05,
+      "loss": 0.114,
+      "step": 824
+    },
+    {
+      "epoch": 0.4316396018858041,
+      "loss_partial": 0.05322265625,
+      "step": 824
+    },
+    {
+      "epoch": 0.4316396018858041,
+      "loss_partial": 0.059326171875,
+      "step": 824
+    },
+    {
+      "epoch": 0.4321634363541121,
+      "grad_norm": 2.057569742202759,
+      "learning_rate": 1.999623834710067e-05,
+      "loss": 0.1156,
+      "step": 825
+    },
+    {
+      "epoch": 0.4321634363541121,
+      "loss_partial": 0.064453125,
+      "step": 825
+    },
+    {
+      "epoch": 0.4321634363541121,
+      "loss_partial": 0.0546875,
+      "step": 825
+    },
+    {
+      "epoch": 0.4326872708224201,
+      "grad_norm": 2.3218977451324463,
+      "learning_rate": 1.999611192196078e-05,
+      "loss": 0.1198,
+      "step": 826
+    },
+    {
+      "epoch": 0.4326872708224201,
+      "loss_partial": 0.056640625,
+      "step": 826
+    },
+    {
+      "epoch": 0.4326872708224201,
+      "loss_partial": 0.060546875,
+      "step": 826
+    },
+    {
+      "epoch": 0.43321110529072815,
+      "grad_norm": 2.0732767581939697,
+      "learning_rate": 1.999598340769531e-05,
+      "loss": 0.1147,
+      "step": 827
+    },
+    {
+      "epoch": 0.43321110529072815,
+      "loss_partial": 0.056884765625,
+      "step": 827
+    },
+    {
+      "epoch": 0.43321110529072815,
+      "loss_partial": 0.061767578125,
+      "step": 827
+    },
+    {
+      "epoch": 0.43373493975903615,
+      "grad_norm": 1.34905207157135,
+      "learning_rate": 1.9995852804331128e-05,
+      "loss": 0.1164,
+      "step": 828
+    },
+    {
+      "epoch": 0.43373493975903615,
+      "loss_partial": 0.058837890625,
+      "step": 828
+    },
+    {
+      "epoch": 0.43373493975903615,
+      "loss_partial": 0.061279296875,
+      "step": 828
+    },
+    {
+      "epoch": 0.43425877422734416,
+      "grad_norm": 1.2299573421478271,
+      "learning_rate": 1.9995720111895516e-05,
+      "loss": 0.1142,
+      "step": 829
+    },
+    {
+      "epoch": 0.43425877422734416,
+      "loss_partial": 0.061767578125,
+      "step": 829
+    },
+    {
+      "epoch": 0.43425877422734416,
+      "loss_partial": 0.0498046875,
+      "step": 829
+    },
+    {
+      "epoch": 0.43478260869565216,
+      "grad_norm": 1.79463791847229,
+      "learning_rate": 1.9995585330416217e-05,
+      "loss": 0.1116,
+      "step": 830
+    },
+    {
+      "epoch": 0.43478260869565216,
+      "loss_partial": 0.05859375,
+      "step": 830
+    },
+    {
+      "epoch": 0.43478260869565216,
+      "loss_partial": 0.05322265625,
+      "step": 830
+    },
+    {
+      "epoch": 0.43530644316396017,
+      "grad_norm": 1.4542195796966553,
+      "learning_rate": 1.999544845992139e-05,
+      "loss": 0.1166,
+      "step": 831
+    },
+    {
+      "epoch": 0.43530644316396017,
+      "loss_partial": 0.0498046875,
+      "step": 831
+    },
+    {
+      "epoch": 0.43530644316396017,
+      "loss_partial": 0.05908203125,
+      "step": 831
+    },
+    {
+      "epoch": 0.4358302776322682,
+      "grad_norm": 1.4859585762023926,
+      "learning_rate": 1.9995309500439645e-05,
+      "loss": 0.1128,
+      "step": 832
+    },
+    {
+      "epoch": 0.4358302776322682,
+      "loss_partial": 0.05224609375,
+      "step": 832
+    },
+    {
+      "epoch": 0.4358302776322682,
+      "loss_partial": 0.0615234375,
+      "step": 832
+    },
+    {
+      "epoch": 0.43635411210057623,
+      "grad_norm": 1.3908532857894897,
+      "learning_rate": 1.999516845200003e-05,
+      "loss": 0.1192,
+      "step": 833
+    },
+    {
+      "epoch": 0.43635411210057623,
+      "loss_partial": 0.0625,
+      "step": 833
+    },
+    {
+      "epoch": 0.43635411210057623,
+      "loss_partial": 0.0556640625,
+      "step": 833
+    },
+    {
+      "epoch": 0.43687794656888423,
+      "grad_norm": 1.0729831457138062,
+      "learning_rate": 1.999502531463201e-05,
+      "loss": 0.1138,
+      "step": 834
+    },
+    {
+      "epoch": 0.43687794656888423,
+      "loss_partial": 0.064453125,
+      "step": 834
+    },
+    {
+      "epoch": 0.43687794656888423,
+      "loss_partial": 0.0537109375,
+      "step": 834
+    },
+    {
+      "epoch": 0.43740178103719224,
+      "grad_norm": 1.7257969379425049,
+      "learning_rate": 1.999488008836551e-05,
+      "loss": 0.1156,
+      "step": 835
+    },
+    {
+      "epoch": 0.43740178103719224,
+      "loss_partial": 0.05419921875,
+      "step": 835
+    },
+    {
+      "epoch": 0.43740178103719224,
+      "loss_partial": 0.048583984375,
+      "step": 835
+    },
+    {
+      "epoch": 0.43792561550550024,
+      "grad_norm": 1.2270234823226929,
+      "learning_rate": 1.999473277323088e-05,
+      "loss": 0.1076,
+      "step": 836
+    },
+    {
+      "epoch": 0.43792561550550024,
+      "loss_partial": 0.058349609375,
+      "step": 836
+    },
+    {
+      "epoch": 0.43792561550550024,
+      "loss_partial": 0.05908203125,
+      "step": 836
+    },
+    {
+      "epoch": 0.4384494499738083,
+      "grad_norm": 2.1113717555999756,
+      "learning_rate": 1.99945833692589e-05,
+      "loss": 0.1122,
+      "step": 837
+    },
+    {
+      "epoch": 0.4384494499738083,
+      "loss_partial": 0.052490234375,
+      "step": 837
+    },
+    {
+      "epoch": 0.4384494499738083,
+      "loss_partial": 0.0615234375,
+      "step": 837
+    },
+    {
+      "epoch": 0.4389732844421163,
+      "grad_norm": 1.3499785661697388,
+      "learning_rate": 1.99944318764808e-05,
+      "loss": 0.1137,
+      "step": 838
+    },
+    {
+      "epoch": 0.4389732844421163,
+      "loss_partial": 0.05029296875,
+      "step": 838
+    },
+    {
+      "epoch": 0.4389732844421163,
+      "loss_partial": 0.051513671875,
+      "step": 838
+    },
+    {
+      "epoch": 0.4394971189104243,
+      "grad_norm": 1.4272596836090088,
+      "learning_rate": 1.999427829492825e-05,
+      "loss": 0.1084,
+      "step": 839
+    },
+    {
+      "epoch": 0.4394971189104243,
+      "loss_partial": 0.06005859375,
+      "step": 839
+    },
+    {
+      "epoch": 0.4394971189104243,
+      "loss_partial": 0.048583984375,
+      "step": 839
+    },
+    {
+      "epoch": 0.4400209533787323,
+      "grad_norm": 1.4691433906555176,
+      "learning_rate": 1.9994122624633333e-05,
+      "loss": 0.1136,
+      "step": 840
+    },
+    {
+      "epoch": 0.4400209533787323,
+      "loss_partial": 0.053955078125,
+      "step": 840
+    },
+    {
+      "epoch": 0.4400209533787323,
+      "loss_partial": 0.045166015625,
+      "step": 840
+    },
+    {
+      "epoch": 0.4405447878470403,
+      "grad_norm": 1.235945701599121,
+      "learning_rate": 1.999396486562859e-05,
+      "loss": 0.113,
+      "step": 841
+    },
+    {
+      "epoch": 0.4405447878470403,
+      "loss_partial": 0.054931640625,
+      "step": 841
+    },
+    {
+      "epoch": 0.4405447878470403,
+      "loss_partial": 0.051025390625,
+      "step": 841
+    },
+    {
+      "epoch": 0.4410686223153483,
+      "grad_norm": 1.3102614879608154,
+      "learning_rate": 1.9993805017946994e-05,
+      "loss": 0.112,
+      "step": 842
+    },
+    {
+      "epoch": 0.4410686223153483,
+      "loss_partial": 0.064453125,
+      "step": 842
+    },
+    {
+      "epoch": 0.4410686223153483,
+      "loss_partial": 0.06640625,
+      "step": 842
+    },
+    {
+      "epoch": 0.4415924567836564,
+      "grad_norm": 1.7917448282241821,
+      "learning_rate": 1.9993643081621943e-05,
+      "loss": 0.1163,
+      "step": 843
+    },
+    {
+      "epoch": 0.4415924567836564,
+      "loss_partial": 0.0673828125,
+      "step": 843
+    },
+    {
+      "epoch": 0.4415924567836564,
+      "loss_partial": 0.06103515625,
+      "step": 843
+    },
+    {
+      "epoch": 0.4421162912519644,
+      "grad_norm": 2.107288360595703,
+      "learning_rate": 1.9993479056687294e-05,
+      "loss": 0.1162,
+      "step": 844
+    },
+    {
+      "epoch": 0.4421162912519644,
+      "loss_partial": 0.056640625,
+      "step": 844
+    },
+    {
+      "epoch": 0.4421162912519644,
+      "loss_partial": 0.0634765625,
+      "step": 844
+    },
+    {
+      "epoch": 0.4426401257202724,
+      "grad_norm": 2.679953098297119,
+      "learning_rate": 1.9993312943177316e-05,
+      "loss": 0.1152,
+      "step": 845
+    },
+    {
+      "epoch": 0.4426401257202724,
+      "loss_partial": 0.048095703125,
+      "step": 845
+    },
+    {
+      "epoch": 0.4426401257202724,
+      "loss_partial": 0.059326171875,
+      "step": 845
+    },
+    {
+      "epoch": 0.4431639601885804,
+      "grad_norm": 1.8857934474945068,
+      "learning_rate": 1.9993144741126728e-05,
+      "loss": 0.113,
+      "step": 846
+    },
+    {
+      "epoch": 0.4431639601885804,
+      "loss_partial": 0.049560546875,
+      "step": 846
+    },
+    {
+      "epoch": 0.4431639601885804,
+      "loss_partial": 0.0654296875,
+      "step": 846
+    },
+    {
+      "epoch": 0.4436877946568884,
+      "grad_norm": 1.4373725652694702,
+      "learning_rate": 1.9992974450570684e-05,
+      "loss": 0.1179,
+      "step": 847
+    },
+    {
+      "epoch": 0.4436877946568884,
+      "loss_partial": 0.05810546875,
+      "step": 847
+    },
+    {
+      "epoch": 0.4436877946568884,
+      "loss_partial": 0.048828125,
+      "step": 847
+    },
+    {
+      "epoch": 0.44421162912519646,
+      "grad_norm": 1.674368977546692,
+      "learning_rate": 1.999280207154478e-05,
+      "loss": 0.1153,
+      "step": 848
+    },
+    {
+      "epoch": 0.44421162912519646,
+      "loss_partial": 0.05859375,
+      "step": 848
+    },
+    {
+      "epoch": 0.44421162912519646,
+      "loss_partial": 0.05078125,
+      "step": 848
+    },
+    {
+      "epoch": 0.44473546359350447,
+      "grad_norm": 2.071584701538086,
+      "learning_rate": 1.9992627604085032e-05,
+      "loss": 0.1117,
+      "step": 849
+    },
+    {
+      "epoch": 0.44473546359350447,
+      "loss_partial": 0.04736328125,
+      "step": 849
+    },
+    {
+      "epoch": 0.44473546359350447,
+      "loss_partial": 0.058837890625,
+      "step": 849
+    },
+    {
+      "epoch": 0.44525929806181247,
+      "grad_norm": 1.5561100244522095,
+      "learning_rate": 1.9992451048227913e-05,
+      "loss": 0.1143,
+      "step": 850
+    },
+    {
+      "epoch": 0.44525929806181247,
+      "loss_partial": 0.058349609375,
+      "step": 850
+    },
+    {
+      "epoch": 0.44525929806181247,
+      "loss_partial": 0.053466796875,
+      "step": 850
+    },
+    {
+      "epoch": 0.4457831325301205,
+      "grad_norm": 1.922202706336975,
+      "learning_rate": 1.999227240401031e-05,
+      "loss": 0.112,
+      "step": 851
+    },
+    {
+      "epoch": 0.4457831325301205,
+      "loss_partial": 0.05517578125,
+      "step": 851
+    },
+    {
+      "epoch": 0.4457831325301205,
+      "loss_partial": 0.049560546875,
+      "step": 851
+    },
+    {
+      "epoch": 0.4463069669984285,
+      "grad_norm": 1.6812705993652344,
+      "learning_rate": 1.999209167146957e-05,
+      "loss": 0.1146,
+      "step": 852
+    },
+    {
+      "epoch": 0.4463069669984285,
+      "loss_partial": 0.056396484375,
+      "step": 852
+    },
+    {
+      "epoch": 0.4463069669984285,
+      "loss_partial": 0.0498046875,
+      "step": 852
+    },
+    {
+      "epoch": 0.44683080146673654,
+      "grad_norm": 1.0490691661834717,
+      "learning_rate": 1.9991908850643455e-05,
+      "loss": 0.1143,
+      "step": 853
+    },
+    {
+      "epoch": 0.44683080146673654,
+      "loss_partial": 0.056884765625,
+      "step": 853
+    },
+    {
+      "epoch": 0.44683080146673654,
+      "loss_partial": 0.05615234375,
+      "step": 853
+    },
+    {
+      "epoch": 0.44735463593504454,
+      "grad_norm": 1.6563372611999512,
+      "learning_rate": 1.999172394157018e-05,
+      "loss": 0.1147,
+      "step": 854
+    },
+    {
+      "epoch": 0.44735463593504454,
+      "loss_partial": 0.05078125,
+      "step": 854
+    },
+    {
+      "epoch": 0.44735463593504454,
+      "loss_partial": 0.056396484375,
+      "step": 854
+    },
+    {
+      "epoch": 0.44787847040335255,
+      "grad_norm": 1.7329884767532349,
+      "learning_rate": 1.999153694428839e-05,
+      "loss": 0.1152,
+      "step": 855
+    },
+    {
+      "epoch": 0.44787847040335255,
+      "loss_partial": 0.05908203125,
+      "step": 855
+    },
+    {
+      "epoch": 0.44787847040335255,
+      "loss_partial": 0.05810546875,
+      "step": 855
+    },
+    {
+      "epoch": 0.44840230487166055,
+      "grad_norm": 1.9675676822662354,
+      "learning_rate": 1.9991347858837163e-05,
+      "loss": 0.1122,
+      "step": 856
+    },
+    {
+      "epoch": 0.44840230487166055,
+      "loss_partial": 0.064453125,
+      "step": 856
+    },
+    {
+      "epoch": 0.44840230487166055,
+      "loss_partial": 0.06591796875,
+      "step": 856
+    },
+    {
+      "epoch": 0.44892613933996856,
+      "grad_norm": 1.5867431163787842,
+      "learning_rate": 1.999115668525602e-05,
+      "loss": 0.1162,
+      "step": 857
+    },
+    {
+      "epoch": 0.44892613933996856,
+      "loss_partial": 0.056640625,
+      "step": 857
+    },
+    {
+      "epoch": 0.44892613933996856,
+      "loss_partial": 0.058349609375,
+      "step": 857
+    },
+    {
+      "epoch": 0.44944997380827656,
+      "grad_norm": 1.9565672874450684,
+      "learning_rate": 1.999096342358491e-05,
+      "loss": 0.1096,
+      "step": 858
+    },
+    {
+      "epoch": 0.44944997380827656,
+      "loss_partial": 0.06005859375,
+      "step": 858
+    },
+    {
+      "epoch": 0.44944997380827656,
+      "loss_partial": 0.05517578125,
+      "step": 858
+    },
+    {
+      "epoch": 0.4499738082765846,
+      "grad_norm": 2.1413049697875977,
+      "learning_rate": 1.999076807386423e-05,
+      "loss": 0.1093,
+      "step": 859
+    },
+    {
+      "epoch": 0.4499738082765846,
+      "loss_partial": 0.0537109375,
+      "step": 859
+    },
+    {
+      "epoch": 0.4499738082765846,
+      "loss_partial": 0.06591796875,
+      "step": 859
+    },
+    {
+      "epoch": 0.4504976427448926,
+      "grad_norm": 2.3876802921295166,
+      "learning_rate": 1.99905706361348e-05,
+      "loss": 0.1185,
+      "step": 860
+    },
+    {
+      "epoch": 0.4504976427448926,
+      "loss_partial": 0.057373046875,
+      "step": 860
+    },
+    {
+      "epoch": 0.4504976427448926,
+      "loss_partial": 0.0546875,
+      "step": 860
+    },
+    {
+      "epoch": 0.45102147721320063,
+      "grad_norm": 2.396432399749756,
+      "learning_rate": 1.999037111043789e-05,
+      "loss": 0.1119,
+      "step": 861
+    },
+    {
+      "epoch": 0.45102147721320063,
+      "loss_partial": 0.057861328125,
+      "step": 861
+    },
+    {
+      "epoch": 0.45102147721320063,
+      "loss_partial": 0.05859375,
+      "step": 861
+    },
+    {
+      "epoch": 0.45154531168150863,
+      "grad_norm": 2.4295897483825684,
+      "learning_rate": 1.9990169496815194e-05,
+      "loss": 0.113,
+      "step": 862
+    },
+    {
+      "epoch": 0.45154531168150863,
+      "loss_partial": 0.049072265625,
+      "step": 862
+    },
+    {
+      "epoch": 0.45154531168150863,
+      "loss_partial": 0.056884765625,
+      "step": 862
+    },
+    {
+      "epoch": 0.45206914614981664,
+      "grad_norm": 1.9039413928985596,
+      "learning_rate": 1.998996579530885e-05,
+      "loss": 0.1145,
+      "step": 863
+    },
+    {
+      "epoch": 0.45206914614981664,
+      "loss_partial": 0.058837890625,
+      "step": 863
+    },
+    {
+      "epoch": 0.45206914614981664,
+      "loss_partial": 0.054931640625,
+      "step": 863
+    },
+    {
+      "epoch": 0.4525929806181247,
+      "grad_norm": 2.129211187362671,
+      "learning_rate": 1.9989760005961435e-05,
+      "loss": 0.1129,
+      "step": 864
+    },
+    {
+      "epoch": 0.4525929806181247,
+      "loss_partial": 0.056640625,
+      "step": 864
+    },
+    {
+      "epoch": 0.4525929806181247,
+      "loss_partial": 0.0546875,
+      "step": 864
+    },
+    {
+      "epoch": 0.4531168150864327,
+      "grad_norm": 3.329042911529541,
+      "learning_rate": 1.9989552128815952e-05,
+      "loss": 0.1143,
+      "step": 865
+    },
+    {
+      "epoch": 0.4531168150864327,
+      "loss_partial": 0.053466796875,
+      "step": 865
+    },
+    {
+      "epoch": 0.4531168150864327,
+      "loss_partial": 0.054443359375,
+      "step": 865
+    },
+    {
+      "epoch": 0.4536406495547407,
+      "grad_norm": 2.2092936038970947,
+      "learning_rate": 1.9989342163915844e-05,
+      "loss": 0.1148,
+      "step": 866
+    },
+    {
+      "epoch": 0.4536406495547407,
+      "loss_partial": 0.0556640625,
+      "step": 866
+    },
+    {
+      "epoch": 0.4536406495547407,
+      "loss_partial": 0.052490234375,
+      "step": 866
+    },
+    {
+      "epoch": 0.4541644840230487,
+      "grad_norm": 1.9364285469055176,
+      "learning_rate": 1.9989130111305e-05,
+      "loss": 0.1125,
+      "step": 867
+    },
+    {
+      "epoch": 0.4541644840230487,
+      "loss_partial": 0.057861328125,
+      "step": 867
+    },
+    {
+      "epoch": 0.4541644840230487,
+      "loss_partial": 0.062255859375,
+      "step": 867
+    },
+    {
+      "epoch": 0.4546883184913567,
+      "grad_norm": 1.382449746131897,
+      "learning_rate": 1.9988915971027734e-05,
+      "loss": 0.114,
+      "step": 868
+    },
+    {
+      "epoch": 0.4546883184913567,
+      "loss_partial": 0.060546875,
+      "step": 868
+    },
+    {
+      "epoch": 0.4546883184913567,
+      "loss_partial": 0.06591796875,
+      "step": 868
+    },
+    {
+      "epoch": 0.4552121529596648,
+      "grad_norm": 1.3695353269577026,
+      "learning_rate": 1.99886997431288e-05,
+      "loss": 0.1155,
+      "step": 869
+    },
+    {
+      "epoch": 0.4552121529596648,
+      "loss_partial": 0.06494140625,
+      "step": 869
+    },
+    {
+      "epoch": 0.4552121529596648,
+      "loss_partial": 0.06982421875,
+      "step": 869
+    },
+    {
+      "epoch": 0.4557359874279728,
+      "grad_norm": 1.4382096529006958,
+      "learning_rate": 1.9988481427653384e-05,
+      "loss": 0.1153,
+      "step": 870
+    },
+    {
+      "epoch": 0.4557359874279728,
+      "loss_partial": 0.054443359375,
+      "step": 870
+    },
+    {
+      "epoch": 0.4557359874279728,
+      "loss_partial": 0.0634765625,
+      "step": 870
+    },
+    {
+      "epoch": 0.4562598218962808,
+      "grad_norm": 1.372860074043274,
+      "learning_rate": 1.998826102464712e-05,
+      "loss": 0.1104,
+      "step": 871
+    },
+    {
+      "epoch": 0.4562598218962808,
+      "loss_partial": 0.064453125,
+      "step": 871
+    },
+    {
+      "epoch": 0.4562598218962808,
+      "loss_partial": 0.0556640625,
+      "step": 871
+    },
+    {
+      "epoch": 0.4567836563645888,
+      "grad_norm": 2.263254404067993,
+      "learning_rate": 1.9988038534156066e-05,
+      "loss": 0.1117,
+      "step": 872
+    },
+    {
+      "epoch": 0.4567836563645888,
+      "loss_partial": 0.05419921875,
+      "step": 872
+    },
+    {
+      "epoch": 0.4567836563645888,
+      "loss_partial": 0.0615234375,
+      "step": 872
+    },
+    {
+      "epoch": 0.4573074908328968,
+      "grad_norm": 1.3320266008377075,
+      "learning_rate": 1.9987813956226723e-05,
+      "loss": 0.1151,
+      "step": 873
+    },
+    {
+      "epoch": 0.4573074908328968,
+      "loss_partial": 0.061767578125,
+      "step": 873
+    },
+    {
+      "epoch": 0.4573074908328968,
+      "loss_partial": 0.0537109375,
+      "step": 873
+    },
+    {
+      "epoch": 0.4578313253012048,
+      "grad_norm": 2.4667773246765137,
+      "learning_rate": 1.9987587290906025e-05,
+      "loss": 0.1118,
+      "step": 874
+    },
+    {
+      "epoch": 0.4578313253012048,
+      "loss_partial": 0.05126953125,
+      "step": 874
+    },
+    {
+      "epoch": 0.4578313253012048,
+      "loss_partial": 0.052490234375,
+      "step": 874
+    },
+    {
+      "epoch": 0.45835515976951285,
+      "grad_norm": 1.3248450756072998,
+      "learning_rate": 1.9987358538241346e-05,
+      "loss": 0.1123,
+      "step": 875
+    },
+    {
+      "epoch": 0.45835515976951285,
+      "loss_partial": 0.05908203125,
+      "step": 875
+    },
+    {
+      "epoch": 0.45835515976951285,
+      "loss_partial": 0.046875,
+      "step": 875
+    },
+    {
+      "epoch": 0.45887899423782086,
+      "grad_norm": 1.323224663734436,
+      "learning_rate": 1.998712769828049e-05,
+      "loss": 0.1115,
+      "step": 876
+    },
+    {
+      "epoch": 0.45887899423782086,
+      "loss_partial": 0.05712890625,
+      "step": 876
+    },
+    {
+      "epoch": 0.45887899423782086,
+      "loss_partial": 0.06201171875,
+      "step": 876
+    },
+    {
+      "epoch": 0.45940282870612886,
+      "grad_norm": 1.4431065320968628,
+      "learning_rate": 1.9986894771071707e-05,
+      "loss": 0.1136,
+      "step": 877
+    },
+    {
+      "epoch": 0.45940282870612886,
+      "loss_partial": 0.057373046875,
+      "step": 877
+    },
+    {
+      "epoch": 0.45940282870612886,
+      "loss_partial": 0.06591796875,
+      "step": 877
+    },
+    {
+      "epoch": 0.45992666317443687,
+      "grad_norm": 1.5910017490386963,
+      "learning_rate": 1.9986659756663674e-05,
+      "loss": 0.1155,
+      "step": 878
+    },
+    {
+      "epoch": 0.45992666317443687,
+      "loss_partial": 0.068359375,
+      "step": 878
+    },
+    {
+      "epoch": 0.45992666317443687,
+      "loss_partial": 0.068359375,
+      "step": 878
+    },
+    {
+      "epoch": 0.46045049764274487,
+      "grad_norm": 2.0188562870025635,
+      "learning_rate": 1.9986422655105503e-05,
+      "loss": 0.1187,
+      "step": 879
+    },
+    {
+      "epoch": 0.46045049764274487,
+      "loss_partial": 0.06298828125,
+      "step": 879
+    },
+    {
+      "epoch": 0.46045049764274487,
+      "loss_partial": 0.05908203125,
+      "step": 879
+    },
+    {
+      "epoch": 0.46097433211105293,
+      "grad_norm": 1.909438133239746,
+      "learning_rate": 1.9986183466446755e-05,
+      "loss": 0.1157,
+      "step": 880
+    },
+    {
+      "epoch": 0.46097433211105293,
+      "loss_partial": 0.0576171875,
+      "step": 880
+    },
+    {
+      "epoch": 0.46097433211105293,
+      "loss_partial": 0.049072265625,
+      "step": 880
+    },
+    {
+      "epoch": 0.46149816657936094,
+      "grad_norm": 1.9288233518600464,
+      "learning_rate": 1.9985942190737413e-05,
+      "loss": 0.1117,
+      "step": 881
+    },
+    {
+      "epoch": 0.46149816657936094,
+      "loss_partial": 0.04931640625,
+      "step": 881
+    },
+    {
+      "epoch": 0.46149816657936094,
+      "loss_partial": 0.057373046875,
+      "step": 881
+    },
+    {
+      "epoch": 0.46202200104766894,
+      "grad_norm": 1.255985140800476,
+      "learning_rate": 1.9985698828027905e-05,
+      "loss": 0.1093,
+      "step": 882
+    },
+    {
+      "epoch": 0.46202200104766894,
+      "loss_partial": 0.057373046875,
+      "step": 882
+    },
+    {
+      "epoch": 0.46202200104766894,
+      "loss_partial": 0.048095703125,
+      "step": 882
+    },
+    {
+      "epoch": 0.46254583551597694,
+      "grad_norm": 1.764944076538086,
+      "learning_rate": 1.9985453378369094e-05,
+      "loss": 0.1114,
+      "step": 883
+    },
+    {
+      "epoch": 0.46254583551597694,
+      "loss_partial": 0.04833984375,
+      "step": 883
+    },
+    {
+      "epoch": 0.46254583551597694,
+      "loss_partial": 0.05419921875,
+      "step": 883
+    },
+    {
+      "epoch": 0.46306966998428495,
+      "grad_norm": 1.8941699266433716,
+      "learning_rate": 1.9985205841812273e-05,
+      "loss": 0.1133,
+      "step": 884
+    },
+    {
+      "epoch": 0.46306966998428495,
+      "loss_partial": 0.057861328125,
+      "step": 884
+    },
+    {
+      "epoch": 0.46306966998428495,
+      "loss_partial": 0.0634765625,
+      "step": 884
+    },
+    {
+      "epoch": 0.463593504452593,
+      "grad_norm": 2.5118582248687744,
+      "learning_rate": 1.9984956218409176e-05,
+      "loss": 0.1165,
+      "step": 885
+    },
+    {
+      "epoch": 0.463593504452593,
+      "loss_partial": 0.072265625,
+      "step": 885
+    },
+    {
+      "epoch": 0.463593504452593,
+      "loss_partial": 0.0546875,
+      "step": 885
+    },
+    {
+      "epoch": 0.464117338920901,
+      "grad_norm": 2.604283094406128,
+      "learning_rate": 1.998470450821198e-05,
+      "loss": 0.1172,
+      "step": 886
+    },
+    {
+      "epoch": 0.464117338920901,
+      "loss_partial": 0.052734375,
+      "step": 886
+    },
+    {
+      "epoch": 0.464117338920901,
+      "loss_partial": 0.05224609375,
+      "step": 886
+    },
+    {
+      "epoch": 0.464641173389209,
+      "grad_norm": 1.8031281232833862,
+      "learning_rate": 1.998445071127328e-05,
+      "loss": 0.1161,
+      "step": 887
+    },
+    {
+      "epoch": 0.464641173389209,
+      "loss_partial": 0.05029296875,
+      "step": 887
+    },
+    {
+      "epoch": 0.464641173389209,
+      "loss_partial": 0.0654296875,
+      "step": 887
+    },
+    {
+      "epoch": 0.465165007857517,
+      "grad_norm": 1.4913355112075806,
+      "learning_rate": 1.9984194827646126e-05,
+      "loss": 0.117,
+      "step": 888
+    },
+    {
+      "epoch": 0.465165007857517,
+      "loss_partial": 0.05419921875,
+      "step": 888
+    },
+    {
+      "epoch": 0.465165007857517,
+      "loss_partial": 0.048583984375,
+      "step": 888
+    },
+    {
+      "epoch": 0.465688842325825,
+      "grad_norm": 2.4904491901397705,
+      "learning_rate": 1.9983936857383993e-05,
+      "loss": 0.1105,
+      "step": 889
+    },
+    {
+      "epoch": 0.465688842325825,
+      "loss_partial": 0.052978515625,
+      "step": 889
+    },
+    {
+      "epoch": 0.465688842325825,
+      "loss_partial": 0.0634765625,
+      "step": 889
+    },
+    {
+      "epoch": 0.46621267679413303,
+      "grad_norm": 2.313584327697754,
+      "learning_rate": 1.9983676800540794e-05,
+      "loss": 0.1124,
+      "step": 890
+    },
+    {
+      "epoch": 0.46621267679413303,
+      "loss_partial": 0.051513671875,
+      "step": 890
+    },
+    {
+      "epoch": 0.46621267679413303,
+      "loss_partial": 0.05029296875,
+      "step": 890
+    },
+    {
+      "epoch": 0.4667365112624411,
+      "grad_norm": 1.2049787044525146,
+      "learning_rate": 1.9983414657170882e-05,
+      "loss": 0.1114,
+      "step": 891
+    },
+    {
+      "epoch": 0.4667365112624411,
+      "loss_partial": 0.06005859375,
+      "step": 891
+    },
+    {
+      "epoch": 0.4667365112624411,
+      "loss_partial": 0.060302734375,
+      "step": 891
+    },
+    {
+      "epoch": 0.4672603457307491,
+      "grad_norm": 2.3323700428009033,
+      "learning_rate": 1.9983150427329042e-05,
+      "loss": 0.1117,
+      "step": 892
+    },
+    {
+      "epoch": 0.4672603457307491,
+      "loss_partial": 0.055419921875,
+      "step": 892
+    },
+    {
+      "epoch": 0.4672603457307491,
+      "loss_partial": 0.053466796875,
+      "step": 892
+    },
+    {
+      "epoch": 0.4677841801990571,
+      "grad_norm": 1.7091115713119507,
+      "learning_rate": 1.9982884111070496e-05,
+      "loss": 0.1111,
+      "step": 893
+    },
+    {
+      "epoch": 0.4677841801990571,
+      "loss_partial": 0.06298828125,
+      "step": 893
+    },
+    {
+      "epoch": 0.4677841801990571,
+      "loss_partial": 0.057373046875,
+      "step": 893
+    },
+    {
+      "epoch": 0.4683080146673651,
+      "grad_norm": 1.530231237411499,
+      "learning_rate": 1.9982615708450905e-05,
+      "loss": 0.1138,
+      "step": 894
+    },
+    {
+      "epoch": 0.4683080146673651,
+      "loss_partial": 0.0595703125,
+      "step": 894
+    },
+    {
+      "epoch": 0.4683080146673651,
+      "loss_partial": 0.04541015625,
+      "step": 894
+    },
+    {
+      "epoch": 0.4688318491356731,
+      "grad_norm": 1.3548346757888794,
+      "learning_rate": 1.998234521952636e-05,
+      "loss": 0.1108,
+      "step": 895
+    },
+    {
+      "epoch": 0.4688318491356731,
+      "loss_partial": 0.047607421875,
+      "step": 895
+    },
+    {
+      "epoch": 0.4688318491356731,
+      "loss_partial": 0.054443359375,
+      "step": 895
+    },
+    {
+      "epoch": 0.46935568360398117,
+      "grad_norm": 1.5346709489822388,
+      "learning_rate": 1.9982072644353396e-05,
+      "loss": 0.1133,
+      "step": 896
+    },
+    {
+      "epoch": 0.46935568360398117,
+      "loss_partial": 0.057861328125,
+      "step": 896
+    },
+    {
+      "epoch": 0.46935568360398117,
+      "loss_partial": 0.07080078125,
+      "step": 896
+    },
+    {
+      "epoch": 0.46987951807228917,
+      "grad_norm": 1.2529239654541016,
+      "learning_rate": 1.9981797982988973e-05,
+      "loss": 0.1126,
+      "step": 897
+    },
+    {
+      "epoch": 0.46987951807228917,
+      "loss_partial": 0.060546875,
+      "step": 897
+    },
+    {
+      "epoch": 0.46987951807228917,
+      "loss_partial": 0.052978515625,
+      "step": 897
+    },
+    {
+      "epoch": 0.4704033525405972,
+      "grad_norm": 1.9998096227645874,
+      "learning_rate": 1.99815212354905e-05,
+      "loss": 0.1129,
+      "step": 898
+    },
+    {
+      "epoch": 0.4704033525405972,
+      "loss_partial": 0.062255859375,
+      "step": 898
+    },
+    {
+      "epoch": 0.4704033525405972,
+      "loss_partial": 0.0546875,
+      "step": 898
+    },
+    {
+      "epoch": 0.4709271870089052,
+      "grad_norm": 2.095952033996582,
+      "learning_rate": 1.9981242401915808e-05,
+      "loss": 0.1107,
+      "step": 899
+    },
+    {
+      "epoch": 0.4709271870089052,
+      "loss_partial": 0.06103515625,
+      "step": 899
+    },
+    {
+      "epoch": 0.4709271870089052,
+      "loss_partial": 0.0517578125,
+      "step": 899
+    },
+    {
+      "epoch": 0.4714510214772132,
+      "grad_norm": 2.781738042831421,
+      "learning_rate": 1.998096148232318e-05,
+      "loss": 0.1165,
+      "step": 900
+    },
+    {
+      "epoch": 0.4714510214772132,
+      "loss_partial": 0.05712890625,
+      "step": 900
+    },
+    {
+      "epoch": 0.4714510214772132,
+      "loss_partial": 0.04931640625,
+      "step": 900
+    },
+    {
+      "epoch": 0.4719748559455212,
+      "grad_norm": 2.9680304527282715,
+      "learning_rate": 1.9980678476771322e-05,
+      "loss": 0.1132,
+      "step": 901
+    },
+    {
+      "epoch": 0.4719748559455212,
+      "loss_partial": 0.056884765625,
+      "step": 901
+    },
+    {
+      "epoch": 0.4719748559455212,
+      "loss_partial": 0.052001953125,
+      "step": 901
+    },
+    {
+      "epoch": 0.47249869041382925,
+      "grad_norm": 2.1780803203582764,
+      "learning_rate": 1.998039338531938e-05,
+      "loss": 0.114,
+      "step": 902
+    },
+    {
+      "epoch": 0.47249869041382925,
+      "loss_partial": 0.05908203125,
+      "step": 902
+    },
+    {
+      "epoch": 0.47249869041382925,
+      "loss_partial": 0.05419921875,
+      "step": 902
+    },
+    {
+      "epoch": 0.47302252488213725,
+      "grad_norm": 2.081393003463745,
+      "learning_rate": 1.998010620802694e-05,
+      "loss": 0.1117,
+      "step": 903
+    },
+    {
+      "epoch": 0.47302252488213725,
+      "loss_partial": 0.0576171875,
+      "step": 903
+    },
+    {
+      "epoch": 0.47302252488213725,
+      "loss_partial": 0.059326171875,
+      "step": 903
+    },
+    {
+      "epoch": 0.47354635935044526,
+      "grad_norm": 2.4369356632232666,
+      "learning_rate": 1.9979816944954017e-05,
+      "loss": 0.1113,
+      "step": 904
+    },
+    {
+      "epoch": 0.47354635935044526,
+      "loss_partial": 0.0595703125,
+      "step": 904
+    },
+    {
+      "epoch": 0.47354635935044526,
+      "loss_partial": 0.052001953125,
+      "step": 904
+    },
+    {
+      "epoch": 0.47407019381875326,
+      "grad_norm": 1.6765094995498657,
+      "learning_rate": 1.9979525596161062e-05,
+      "loss": 0.1136,
+      "step": 905
+    },
+    {
+      "epoch": 0.47407019381875326,
+      "loss_partial": 0.07763671875,
+      "step": 905
+    },
+    {
+      "epoch": 0.47407019381875326,
+      "loss_partial": 0.049560546875,
+      "step": 905
+    },
+    {
+      "epoch": 0.47459402828706126,
+      "grad_norm": 2.020612955093384,
+      "learning_rate": 1.9979232161708972e-05,
+      "loss": 0.1172,
+      "step": 906
+    },
+    {
+      "epoch": 0.47459402828706126,
+      "loss_partial": 0.06689453125,
+      "step": 906
+    },
+    {
+      "epoch": 0.47459402828706126,
+      "loss_partial": 0.056396484375,
+      "step": 906
+    },
+    {
+      "epoch": 0.4751178627553693,
+      "grad_norm": 2.474872589111328,
+      "learning_rate": 1.997893664165907e-05,
+      "loss": 0.1123,
+      "step": 907
+    },
+    {
+      "epoch": 0.4751178627553693,
+      "loss_partial": 0.060302734375,
+      "step": 907
+    },
+    {
+      "epoch": 0.4751178627553693,
+      "loss_partial": 0.048095703125,
+      "step": 907
+    },
+    {
+      "epoch": 0.47564169722367733,
+      "grad_norm": 2.2355546951293945,
+      "learning_rate": 1.9978639036073116e-05,
+      "loss": 0.1087,
+      "step": 908
+    },
+    {
+      "epoch": 0.47564169722367733,
+      "loss_partial": 0.056884765625,
+      "step": 908
+    },
+    {
+      "epoch": 0.47564169722367733,
+      "loss_partial": 0.05859375,
+      "step": 908
+    },
+    {
+      "epoch": 0.47616553169198533,
+      "grad_norm": 2.3834004402160645,
+      "learning_rate": 1.9978339345013313e-05,
+      "loss": 0.1155,
+      "step": 909
+    },
+    {
+      "epoch": 0.47616553169198533,
+      "loss_partial": 0.062255859375,
+      "step": 909
+    },
+    {
+      "epoch": 0.47616553169198533,
+      "loss_partial": 0.059326171875,
+      "step": 909
+    },
+    {
+      "epoch": 0.47668936616029334,
+      "grad_norm": 3.254819869995117,
+      "learning_rate": 1.997803756854229e-05,
+      "loss": 0.1134,
+      "step": 910
+    },
+    {
+      "epoch": 0.47668936616029334,
+      "loss_partial": 0.058349609375,
+      "step": 910
+    },
+    {
+      "epoch": 0.47668936616029334,
+      "loss_partial": 0.06591796875,
+      "step": 910
+    },
+    {
+      "epoch": 0.47721320062860134,
+      "grad_norm": 1.7093243598937988,
+      "learning_rate": 1.9977733706723116e-05,
+      "loss": 0.1139,
+      "step": 911
+    },
+    {
+      "epoch": 0.47721320062860134,
+      "loss_partial": 0.055908203125,
+      "step": 911
+    },
+    {
+      "epoch": 0.47721320062860134,
+      "loss_partial": 0.059326171875,
+      "step": 911
+    },
+    {
+      "epoch": 0.4777370350969094,
+      "grad_norm": 2.985297203063965,
+      "learning_rate": 1.9977427759619296e-05,
+      "loss": 0.1111,
+      "step": 912
+    },
+    {
+      "epoch": 0.4777370350969094,
+      "loss_partial": 0.06396484375,
+      "step": 912
+    },
+    {
+      "epoch": 0.4777370350969094,
+      "loss_partial": 0.0537109375,
+      "step": 912
+    },
+    {
+      "epoch": 0.4782608695652174,
+      "grad_norm": 2.5239133834838867,
+      "learning_rate": 1.9977119727294775e-05,
+      "loss": 0.1133,
+      "step": 913
+    },
+    {
+      "epoch": 0.4782608695652174,
+      "loss_partial": 0.05517578125,
+      "step": 913
+    },
+    {
+      "epoch": 0.4782608695652174,
+      "loss_partial": 0.0615234375,
+      "step": 913
+    },
+    {
+      "epoch": 0.4787847040335254,
+      "grad_norm": 1.913838505744934,
+      "learning_rate": 1.9976809609813926e-05,
+      "loss": 0.1179,
+      "step": 914
+    },
+    {
+      "epoch": 0.4787847040335254,
+      "loss_partial": 0.057861328125,
+      "step": 914
+    },
+    {
+      "epoch": 0.4787847040335254,
+      "loss_partial": 0.052490234375,
+      "step": 914
+    },
+    {
+      "epoch": 0.4793085385018334,
+      "grad_norm": 1.7744715213775635,
+      "learning_rate": 1.9976497407241565e-05,
+      "loss": 0.1116,
+      "step": 915
+    },
+    {
+      "epoch": 0.4793085385018334,
+      "loss_partial": 0.05126953125,
+      "step": 915
+    },
+    {
+      "epoch": 0.4793085385018334,
+      "loss_partial": 0.052001953125,
+      "step": 915
+    },
+    {
+      "epoch": 0.4798323729701414,
+      "grad_norm": 1.44873046875,
+      "learning_rate": 1.997618311964294e-05,
+      "loss": 0.1106,
+      "step": 916
+    },
+    {
+      "epoch": 0.4798323729701414,
+      "loss_partial": 0.059326171875,
+      "step": 916
+    },
+    {
+      "epoch": 0.4798323729701414,
+      "loss_partial": 0.06298828125,
+      "step": 916
+    },
+    {
+      "epoch": 0.4803562074384494,
+      "grad_norm": 1.9395575523376465,
+      "learning_rate": 1.9975866747083734e-05,
+      "loss": 0.1148,
+      "step": 917
+    },
+    {
+      "epoch": 0.4803562074384494,
+      "loss_partial": 0.0517578125,
+      "step": 917
+    },
+    {
+      "epoch": 0.4803562074384494,
+      "loss_partial": 0.04736328125,
+      "step": 917
+    },
+    {
+      "epoch": 0.4808800419067575,
+      "grad_norm": 1.9249712228775024,
+      "learning_rate": 1.9975548289630065e-05,
+      "loss": 0.1148,
+      "step": 918
+    },
+    {
+      "epoch": 0.4808800419067575,
+      "loss_partial": 0.056640625,
+      "step": 918
+    },
+    {
+      "epoch": 0.4808800419067575,
+      "loss_partial": 0.060546875,
+      "step": 918
+    },
+    {
+      "epoch": 0.4814038763750655,
+      "grad_norm": 0.9786761999130249,
+      "learning_rate": 1.9975227747348488e-05,
+      "loss": 0.1137,
+      "step": 919
+    },
+    {
+      "epoch": 0.4814038763750655,
+      "loss_partial": 0.05859375,
+      "step": 919
+    },
+    {
+      "epoch": 0.4814038763750655,
+      "loss_partial": 0.061767578125,
+      "step": 919
+    },
+    {
+      "epoch": 0.4819277108433735,
+      "grad_norm": 1.5055928230285645,
+      "learning_rate": 1.9974905120306e-05,
+      "loss": 0.1127,
+      "step": 920
+    },
+    {
+      "epoch": 0.4819277108433735,
+      "loss_partial": 0.0478515625,
+      "step": 920
+    },
+    {
+      "epoch": 0.4819277108433735,
+      "loss_partial": 0.04931640625,
+      "step": 920
+    },
+    {
+      "epoch": 0.4824515453116815,
+      "grad_norm": 1.8890553712844849,
+      "learning_rate": 1.9974580408570024e-05,
+      "loss": 0.1112,
+      "step": 921
+    },
+    {
+      "epoch": 0.4824515453116815,
+      "loss_partial": 0.05810546875,
+      "step": 921
+    },
+    {
+      "epoch": 0.4824515453116815,
+      "loss_partial": 0.05419921875,
+      "step": 921
+    },
+    {
+      "epoch": 0.4829753797799895,
+      "grad_norm": 1.5315805673599243,
+      "learning_rate": 1.9974253612208425e-05,
+      "loss": 0.1106,
+      "step": 922
+    },
+    {
+      "epoch": 0.4829753797799895,
+      "loss_partial": 0.05078125,
+      "step": 922
+    },
+    {
+      "epoch": 0.4829753797799895,
+      "loss_partial": 0.0537109375,
+      "step": 922
+    },
+    {
+      "epoch": 0.48349921424829756,
+      "grad_norm": 1.892657995223999,
+      "learning_rate": 1.9973924731289498e-05,
+      "loss": 0.111,
+      "step": 923
+    },
+    {
+      "epoch": 0.48349921424829756,
+      "loss_partial": 0.059326171875,
+      "step": 923
+    },
+    {
+      "epoch": 0.48349921424829756,
+      "loss_partial": 0.0693359375,
+      "step": 923
+    },
+    {
+      "epoch": 0.48402304871660556,
+      "grad_norm": 1.9845294952392578,
+      "learning_rate": 1.997359376588198e-05,
+      "loss": 0.1147,
+      "step": 924
+    },
+    {
+      "epoch": 0.48402304871660556,
+      "loss_partial": 0.049072265625,
+      "step": 924
+    },
+    {
+      "epoch": 0.48402304871660556,
+      "loss_partial": 0.05224609375,
+      "step": 924
+    },
+    {
+      "epoch": 0.48454688318491357,
+      "grad_norm": 1.31414794921875,
+      "learning_rate": 1.997326071605504e-05,
+      "loss": 0.1086,
+      "step": 925
+    },
+    {
+      "epoch": 0.48454688318491357,
+      "loss_partial": 0.06298828125,
+      "step": 925
+    },
+    {
+      "epoch": 0.48454688318491357,
+      "loss_partial": 0.05078125,
+      "step": 925
+    },
+    {
+      "epoch": 0.4850707176532216,
+      "grad_norm": 1.613341212272644,
+      "learning_rate": 1.9972925581878284e-05,
+      "loss": 0.1099,
+      "step": 926
+    },
+    {
+      "epoch": 0.4850707176532216,
+      "loss_partial": 0.05419921875,
+      "step": 926
+    },
+    {
+      "epoch": 0.4850707176532216,
+      "loss_partial": 0.052490234375,
+      "step": 926
+    },
+    {
+      "epoch": 0.4855945521215296,
+      "grad_norm": 2.347219705581665,
+      "learning_rate": 1.997258836342175e-05,
+      "loss": 0.1142,
+      "step": 927
+    },
+    {
+      "epoch": 0.4855945521215296,
+      "loss_partial": 0.05810546875,
+      "step": 927
+    },
+    {
+      "epoch": 0.4855945521215296,
+      "loss_partial": 0.0634765625,
+      "step": 927
+    },
+    {
+      "epoch": 0.48611838658983764,
+      "grad_norm": 1.8564143180847168,
+      "learning_rate": 1.9972249060755917e-05,
+      "loss": 0.1104,
+      "step": 928
+    },
+    {
+      "epoch": 0.48611838658983764,
+      "loss_partial": 0.059326171875,
+      "step": 928
+    },
+    {
+      "epoch": 0.48611838658983764,
+      "loss_partial": 0.056640625,
+      "step": 928
+    },
+    {
+      "epoch": 0.48664222105814564,
+      "grad_norm": 1.5099155902862549,
+      "learning_rate": 1.9971907673951695e-05,
+      "loss": 0.116,
+      "step": 929
+    },
+    {
+      "epoch": 0.48664222105814564,
+      "loss_partial": 0.048095703125,
+      "step": 929
+    },
+    {
+      "epoch": 0.48664222105814564,
+      "loss_partial": 0.048828125,
+      "step": 929
+    },
+    {
+      "epoch": 0.48716605552645365,
+      "grad_norm": 2.388002634048462,
+      "learning_rate": 1.9971564203080437e-05,
+      "loss": 0.1116,
+      "step": 930
+    },
+    {
+      "epoch": 0.48716605552645365,
+      "loss_partial": 0.0458984375,
+      "step": 930
+    },
+    {
+      "epoch": 0.48716605552645365,
+      "loss_partial": 0.05029296875,
+      "step": 930
+    },
+    {
+      "epoch": 0.48768988999476165,
+      "grad_norm": 2.024631977081299,
+      "learning_rate": 1.9971218648213923e-05,
+      "loss": 0.1123,
+      "step": 931
+    },
+    {
+      "epoch": 0.48768988999476165,
+      "loss_partial": 0.0517578125,
+      "step": 931
+    },
+    {
+      "epoch": 0.48768988999476165,
+      "loss_partial": 0.06298828125,
+      "step": 931
+    },
+    {
+      "epoch": 0.48821372446306965,
+      "grad_norm": 1.8788487911224365,
+      "learning_rate": 1.997087100942437e-05,
+      "loss": 0.114,
+      "step": 932
+    },
+    {
+      "epoch": 0.48821372446306965,
+      "loss_partial": 0.05126953125,
+      "step": 932
+    },
+    {
+      "epoch": 0.48821372446306965,
+      "loss_partial": 0.05517578125,
+      "step": 932
+    },
+    {
+      "epoch": 0.48873755893137766,
+      "grad_norm": 1.5528393983840942,
+      "learning_rate": 1.9970521286784433e-05,
+      "loss": 0.1106,
+      "step": 933
+    },
+    {
+      "epoch": 0.48873755893137766,
+      "loss_partial": 0.059326171875,
+      "step": 933
+    },
+    {
+      "epoch": 0.48873755893137766,
+      "loss_partial": 0.05419921875,
+      "step": 933
+    },
+    {
+      "epoch": 0.4892613933996857,
+      "grad_norm": 1.700645923614502,
+      "learning_rate": 1.9970169480367207e-05,
+      "loss": 0.1096,
+      "step": 934
+    },
+    {
+      "epoch": 0.4892613933996857,
+      "loss_partial": 0.0537109375,
+      "step": 934
+    },
+    {
+      "epoch": 0.4892613933996857,
+      "loss_partial": 0.057373046875,
+      "step": 934
+    },
+    {
+      "epoch": 0.4897852278679937,
+      "grad_norm": 1.7822870016098022,
+      "learning_rate": 1.996981559024621e-05,
+      "loss": 0.1114,
+      "step": 935
+    },
+    {
+      "epoch": 0.4897852278679937,
+      "loss_partial": 0.06298828125,
+      "step": 935
+    },
+    {
+      "epoch": 0.4897852278679937,
+      "loss_partial": 0.050537109375,
+      "step": 935
+    },
+    {
+      "epoch": 0.4903090623363017,
+      "grad_norm": 2.7795419692993164,
+      "learning_rate": 1.9969459616495404e-05,
+      "loss": 0.1123,
+      "step": 936
+    },
+    {
+      "epoch": 0.4903090623363017,
+      "loss_partial": 0.06103515625,
+      "step": 936
+    },
+    {
+      "epoch": 0.4903090623363017,
+      "loss_partial": 0.048583984375,
+      "step": 936
+    },
+    {
+      "epoch": 0.49083289680460973,
+      "grad_norm": 2.2132887840270996,
+      "learning_rate": 1.996910155918919e-05,
+      "loss": 0.1136,
+      "step": 937
+    },
+    {
+      "epoch": 0.49083289680460973,
+      "loss_partial": 0.04638671875,
+      "step": 937
+    },
+    {
+      "epoch": 0.49083289680460973,
+      "loss_partial": 0.058837890625,
+      "step": 937
+    },
+    {
+      "epoch": 0.49135673127291773,
+      "grad_norm": 2.174363136291504,
+      "learning_rate": 1.9968741418402397e-05,
+      "loss": 0.1103,
+      "step": 938
+    },
+    {
+      "epoch": 0.49135673127291773,
+      "loss_partial": 0.048828125,
+      "step": 938
+    },
+    {
+      "epoch": 0.49135673127291773,
+      "loss_partial": 0.056640625,
+      "step": 938
+    },
+    {
+      "epoch": 0.4918805657412258,
+      "grad_norm": 1.3841314315795898,
+      "learning_rate": 1.9968379194210295e-05,
+      "loss": 0.1105,
+      "step": 939
+    },
+    {
+      "epoch": 0.4918805657412258,
+      "loss_partial": 0.056396484375,
+      "step": 939
+    },
+    {
+      "epoch": 0.4918805657412258,
+      "loss_partial": 0.0537109375,
+      "step": 939
+    },
+    {
+      "epoch": 0.4924044002095338,
+      "grad_norm": 2.0050344467163086,
+      "learning_rate": 1.996801488668858e-05,
+      "loss": 0.1139,
+      "step": 940
+    },
+    {
+      "epoch": 0.4924044002095338,
+      "loss_partial": 0.0595703125,
+      "step": 940
+    },
+    {
+      "epoch": 0.4924044002095338,
+      "loss_partial": 0.0576171875,
+      "step": 940
+    },
+    {
+      "epoch": 0.4929282346778418,
+      "grad_norm": 2.4418437480926514,
+      "learning_rate": 1.9967648495913396e-05,
+      "loss": 0.1147,
+      "step": 941
+    },
+    {
+      "epoch": 0.4929282346778418,
+      "loss_partial": 0.058837890625,
+      "step": 941
+    },
+    {
+      "epoch": 0.4929282346778418,
+      "loss_partial": 0.053466796875,
+      "step": 941
+    },
+    {
+      "epoch": 0.4934520691461498,
+      "grad_norm": 2.391571283340454,
+      "learning_rate": 1.9967280021961314e-05,
+      "loss": 0.1155,
+      "step": 942
+    },
+    {
+      "epoch": 0.4934520691461498,
+      "loss_partial": 0.06201171875,
+      "step": 942
+    },
+    {
+      "epoch": 0.4934520691461498,
+      "loss_partial": 0.056884765625,
+      "step": 942
+    },
+    {
+      "epoch": 0.4939759036144578,
+      "grad_norm": 1.8850547075271606,
+      "learning_rate": 1.9966909464909342e-05,
+      "loss": 0.1161,
+      "step": 943
+    },
+    {
+      "epoch": 0.4939759036144578,
+      "loss_partial": 0.05029296875,
+      "step": 943
+    },
+    {
+      "epoch": 0.4939759036144578,
+      "loss_partial": 0.052734375,
+      "step": 943
+    },
+    {
+      "epoch": 0.49449973808276587,
+      "grad_norm": 2.6463451385498047,
+      "learning_rate": 1.996653682483493e-05,
+      "loss": 0.1132,
+      "step": 944
+    },
+    {
+      "epoch": 0.49449973808276587,
+      "loss_partial": 0.056884765625,
+      "step": 944
+    },
+    {
+      "epoch": 0.49449973808276587,
+      "loss_partial": 0.059814453125,
+      "step": 944
+    },
+    {
+      "epoch": 0.4950235725510739,
+      "grad_norm": 1.3778513669967651,
+      "learning_rate": 1.996616210181595e-05,
+      "loss": 0.112,
+      "step": 945
+    },
+    {
+      "epoch": 0.4950235725510739,
+      "loss_partial": 0.05322265625,
+      "step": 945
+    },
+    {
+      "epoch": 0.4950235725510739,
+      "loss_partial": 0.0712890625,
+      "step": 945
+    },
+    {
+      "epoch": 0.4955474070193819,
+      "grad_norm": 1.9448747634887695,
+      "learning_rate": 1.9965785295930716e-05,
+      "loss": 0.1147,
+      "step": 946
+    },
+    {
+      "epoch": 0.4955474070193819,
+      "loss_partial": 0.056640625,
+      "step": 946
+    },
+    {
+      "epoch": 0.4955474070193819,
+      "loss_partial": 0.058349609375,
+      "step": 946
+    },
+    {
+      "epoch": 0.4960712414876899,
+      "grad_norm": 1.8210512399673462,
+      "learning_rate": 1.9965406407257988e-05,
+      "loss": 0.1174,
+      "step": 947
+    },
+    {
+      "epoch": 0.4960712414876899,
+      "loss_partial": 0.05908203125,
+      "step": 947
+    },
+    {
+      "epoch": 0.4960712414876899,
+      "loss_partial": 0.045654296875,
+      "step": 947
+    },
+    {
+      "epoch": 0.4965950759559979,
+      "grad_norm": 1.5933568477630615,
+      "learning_rate": 1.9965025435876942e-05,
+      "loss": 0.1099,
+      "step": 948
+    },
+    {
+      "epoch": 0.4965950759559979,
+      "loss_partial": 0.058349609375,
+      "step": 948
+    },
+    {
+      "epoch": 0.4965950759559979,
+      "loss_partial": 0.059814453125,
+      "step": 948
+    },
+    {
+      "epoch": 0.4971189104243059,
+      "grad_norm": 2.1950008869171143,
+      "learning_rate": 1.99646423818672e-05,
+      "loss": 0.1123,
+      "step": 949
+    },
+    {
+      "epoch": 0.4971189104243059,
+      "loss_partial": 0.05322265625,
+      "step": 949
+    },
+    {
+      "epoch": 0.4971189104243059,
+      "loss_partial": 0.058837890625,
+      "step": 949
+    },
+    {
+      "epoch": 0.49764274489261395,
+      "grad_norm": 1.8338720798492432,
+      "learning_rate": 1.9964257245308823e-05,
+      "loss": 0.1166,
+      "step": 950
+    },
+    {
+      "epoch": 0.49764274489261395,
+      "loss_partial": 0.05322265625,
+      "step": 950
+    },
+    {
+      "epoch": 0.49764274489261395,
+      "loss_partial": 0.0595703125,
+      "step": 950
+    },
+    {
+      "epoch": 0.49816657936092196,
+      "grad_norm": 1.9034348726272583,
+      "learning_rate": 1.9963870026282296e-05,
+      "loss": 0.1143,
+      "step": 951
+    },
+    {
+      "epoch": 0.49816657936092196,
+      "loss_partial": 0.05712890625,
+      "step": 951
+    },
+    {
+      "epoch": 0.49816657936092196,
+      "loss_partial": 0.0556640625,
+      "step": 951
+    },
+    {
+      "epoch": 0.49869041382922996,
+      "grad_norm": 1.2723523378372192,
+      "learning_rate": 1.996348072486855e-05,
+      "loss": 0.1163,
+      "step": 952
+    },
+    {
+      "epoch": 0.49869041382922996,
+      "loss_partial": 0.0654296875,
+      "step": 952
+    },
+    {
+      "epoch": 0.49869041382922996,
+      "loss_partial": 0.046875,
+      "step": 952
+    },
+    {
+      "epoch": 0.49921424829753797,
+      "grad_norm": 1.5545631647109985,
+      "learning_rate": 1.9963089341148945e-05,
+      "loss": 0.1121,
+      "step": 953
+    },
+    {
+      "epoch": 0.49921424829753797,
+      "loss_partial": 0.049560546875,
+      "step": 953
+    },
+    {
+      "epoch": 0.49921424829753797,
+      "loss_partial": 0.055419921875,
+      "step": 953
+    },
+    {
+      "epoch": 0.49973808276584597,
+      "grad_norm": 1.3837624788284302,
+      "learning_rate": 1.9962695875205277e-05,
+      "loss": 0.1127,
+      "step": 954
+    },
+    {
+      "epoch": 0.49973808276584597,
+      "loss_partial": 0.05419921875,
+      "step": 954
+    },
+    {
+      "epoch": 0.49973808276584597,
+      "loss_partial": 0.054931640625,
+      "step": 954
+    },
+    {
+      "epoch": 0.500261917234154,
+      "grad_norm": 1.8790316581726074,
+      "learning_rate": 1.9962300327119778e-05,
+      "loss": 0.1128,
+      "step": 955
+    },
+    {
+      "epoch": 0.500261917234154,
+      "loss_partial": 0.057861328125,
+      "step": 955
+    },
+    {
+      "epoch": 0.500261917234154,
+      "loss_partial": 0.055419921875,
+      "step": 955
+    },
+    {
+      "epoch": 0.500785751702462,
+      "grad_norm": 1.393887996673584,
+      "learning_rate": 1.9961902696975116e-05,
+      "loss": 0.1151,
+      "step": 956
+    },
+    {
+      "epoch": 0.500785751702462,
+      "loss_partial": 0.055908203125,
+      "step": 956
+    },
+    {
+      "epoch": 0.500785751702462,
+      "loss_partial": 0.051513671875,
+      "step": 956
+    },
+    {
+      "epoch": 0.50130958617077,
+      "grad_norm": 1.0824428796768188,
+      "learning_rate": 1.9961502984854394e-05,
+      "loss": 0.1116,
+      "step": 957
+    },
+    {
+      "epoch": 0.50130958617077,
+      "loss_partial": 0.048828125,
+      "step": 957
+    },
+    {
+      "epoch": 0.50130958617077,
+      "loss_partial": 0.051513671875,
+      "step": 957
+    },
+    {
+      "epoch": 0.5018334206390781,
+      "grad_norm": 1.1963006258010864,
+      "learning_rate": 1.9961101190841145e-05,
+      "loss": 0.1128,
+      "step": 958
+    },
+    {
+      "epoch": 0.5018334206390781,
+      "loss_partial": 0.056640625,
+      "step": 958
+    },
+    {
+      "epoch": 0.5018334206390781,
+      "loss_partial": 0.04931640625,
+      "step": 958
+    },
+    {
+      "epoch": 0.502357255107386,
+      "grad_norm": 1.7325907945632935,
+      "learning_rate": 1.9960697315019348e-05,
+      "loss": 0.1147,
+      "step": 959
+    },
+    {
+      "epoch": 0.502357255107386,
+      "loss_partial": 0.0615234375,
+      "step": 959
+    },
+    {
+      "epoch": 0.502357255107386,
+      "loss_partial": 0.05859375,
+      "step": 959
+    },
+    {
+      "epoch": 0.5028810895756941,
+      "grad_norm": 1.590585708618164,
+      "learning_rate": 1.9960291357473408e-05,
+      "loss": 0.112,
+      "step": 960
+    },
+    {
+      "epoch": 0.5028810895756941,
+      "loss_partial": 0.04638671875,
+      "step": 960
+    },
+    {
+      "epoch": 0.5028810895756941,
+      "loss_partial": 0.060302734375,
+      "step": 960
+    },
+    {
+      "epoch": 0.5034049240440021,
+      "grad_norm": 1.5999681949615479,
+      "learning_rate": 1.9959883318288163e-05,
+      "loss": 0.1134,
+      "step": 961
+    },
+    {
+      "epoch": 0.5034049240440021,
+      "loss_partial": 0.054443359375,
+      "step": 961
+    },
+    {
+      "epoch": 0.5034049240440021,
+      "loss_partial": 0.06298828125,
+      "step": 961
+    },
+    {
+      "epoch": 0.5039287585123101,
+      "grad_norm": 1.6574335098266602,
+      "learning_rate": 1.9959473197548897e-05,
+      "loss": 0.1141,
+      "step": 962
+    },
+    {
+      "epoch": 0.5039287585123101,
+      "loss_partial": 0.0615234375,
+      "step": 962
+    },
+    {
+      "epoch": 0.5039287585123101,
+      "loss_partial": 0.058349609375,
+      "step": 962
+    },
+    {
+      "epoch": 0.5044525929806182,
+      "grad_norm": 2.1104519367218018,
+      "learning_rate": 1.995906099534132e-05,
+      "loss": 0.1152,
+      "step": 963
+    },
+    {
+      "epoch": 0.5044525929806182,
+      "loss_partial": 0.0498046875,
+      "step": 963
+    },
+    {
+      "epoch": 0.5044525929806182,
+      "loss_partial": 0.057861328125,
+      "step": 963
+    },
+    {
+      "epoch": 0.5049764274489261,
+      "grad_norm": 2.21069598197937,
+      "learning_rate": 1.995864671175158e-05,
+      "loss": 0.1098,
+      "step": 964
+    },
+    {
+      "epoch": 0.5049764274489261,
+      "loss_partial": 0.050048828125,
+      "step": 964
+    },
+    {
+      "epoch": 0.5049764274489261,
+      "loss_partial": 0.068359375,
+      "step": 964
+    },
+    {
+      "epoch": 0.5055002619172342,
+      "grad_norm": 1.5053514242172241,
+      "learning_rate": 1.995823034686626e-05,
+      "loss": 0.1155,
+      "step": 965
+    },
+    {
+      "epoch": 0.5055002619172342,
+      "loss_partial": 0.054931640625,
+      "step": 965
+    },
+    {
+      "epoch": 0.5055002619172342,
+      "loss_partial": 0.059326171875,
+      "step": 965
+    },
+    {
+      "epoch": 0.5060240963855421,
+      "grad_norm": 1.6727110147476196,
+      "learning_rate": 1.9957811900772377e-05,
+      "loss": 0.1142,
+      "step": 966
+    },
+    {
+      "epoch": 0.5060240963855421,
+      "loss_partial": 0.055908203125,
+      "step": 966
+    },
+    {
+      "epoch": 0.5060240963855421,
+      "loss_partial": 0.0732421875,
+      "step": 966
+    },
+    {
+      "epoch": 0.5065479308538502,
+      "grad_norm": 2.2249226570129395,
+      "learning_rate": 1.9957391373557385e-05,
+      "loss": 0.1145,
+      "step": 967
+    },
+    {
+      "epoch": 0.5065479308538502,
+      "loss_partial": 0.0625,
+      "step": 967
+    },
+    {
+      "epoch": 0.5065479308538502,
+      "loss_partial": 0.0537109375,
+      "step": 967
+    },
+    {
+      "epoch": 0.5070717653221583,
+      "grad_norm": 1.581803321838379,
+      "learning_rate": 1.9956968765309167e-05,
+      "loss": 0.1125,
+      "step": 968
+    },
+    {
+      "epoch": 0.5070717653221583,
+      "loss_partial": 0.05322265625,
+      "step": 968
+    },
+    {
+      "epoch": 0.5070717653221583,
+      "loss_partial": 0.056396484375,
+      "step": 968
+    },
+    {
+      "epoch": 0.5075955997904662,
+      "grad_norm": 1.7447553873062134,
+      "learning_rate": 1.9956544076116052e-05,
+      "loss": 0.1133,
+      "step": 969
+    },
+    {
+      "epoch": 0.5075955997904662,
+      "loss_partial": 0.06201171875,
+      "step": 969
+    },
+    {
+      "epoch": 0.5075955997904662,
+      "loss_partial": 0.055419921875,
+      "step": 969
+    },
+    {
+      "epoch": 0.5081194342587743,
+      "grad_norm": 2.22318696975708,
+      "learning_rate": 1.9956117306066794e-05,
+      "loss": 0.1158,
+      "step": 970
+    },
+    {
+      "epoch": 0.5081194342587743,
+      "loss_partial": 0.05419921875,
+      "step": 970
+    },
+    {
+      "epoch": 0.5081194342587743,
+      "loss_partial": 0.054931640625,
+      "step": 970
+    },
+    {
+      "epoch": 0.5086432687270822,
+      "grad_norm": 3.188004493713379,
+      "learning_rate": 1.9955688455250587e-05,
+      "loss": 0.1177,
+      "step": 971
+    },
+    {
+      "epoch": 0.5086432687270822,
+      "loss_partial": 0.0693359375,
+      "step": 971
+    },
+    {
+      "epoch": 0.5086432687270822,
+      "loss_partial": 0.051513671875,
+      "step": 971
+    },
+    {
+      "epoch": 0.5091671031953903,
+      "grad_norm": 2.4117815494537354,
+      "learning_rate": 1.9955257523757054e-05,
+      "loss": 0.1136,
+      "step": 972
+    },
+    {
+      "epoch": 0.5091671031953903,
+      "loss_partial": 0.057373046875,
+      "step": 972
+    },
+    {
+      "epoch": 0.5091671031953903,
+      "loss_partial": 0.05517578125,
+      "step": 972
+    },
+    {
+      "epoch": 0.5096909376636982,
+      "grad_norm": 2.1021652221679688,
+      "learning_rate": 1.9954824511676262e-05,
+      "loss": 0.1103,
+      "step": 973
+    },
+    {
+      "epoch": 0.5096909376636982,
+      "loss_partial": 0.051025390625,
+      "step": 973
+    },
+    {
+      "epoch": 0.5096909376636982,
+      "loss_partial": 0.06982421875,
+      "step": 973
+    },
+    {
+      "epoch": 0.5102147721320063,
+      "grad_norm": 3.151110887527466,
+      "learning_rate": 1.9954389419098705e-05,
+      "loss": 0.1094,
+      "step": 974
+    },
+    {
+      "epoch": 0.5102147721320063,
+      "loss_partial": 0.059326171875,
+      "step": 974
+    },
+    {
+      "epoch": 0.5102147721320063,
+      "loss_partial": 0.053955078125,
+      "step": 974
+    },
+    {
+      "epoch": 0.5107386066003143,
+      "grad_norm": 2.0050485134124756,
+      "learning_rate": 1.9953952246115317e-05,
+      "loss": 0.1115,
+      "step": 975
+    },
+    {
+      "epoch": 0.5107386066003143,
+      "loss_partial": 0.0576171875,
+      "step": 975
+    },
+    {
+      "epoch": 0.5107386066003143,
+      "loss_partial": 0.046142578125,
+      "step": 975
+    },
+    {
+      "epoch": 0.5112624410686223,
+      "grad_norm": 2.8537325859069824,
+      "learning_rate": 1.9953512992817462e-05,
+      "loss": 0.1133,
+      "step": 976
+    },
+    {
+      "epoch": 0.5112624410686223,
+      "loss_partial": 0.056884765625,
+      "step": 976
+    },
+    {
+      "epoch": 0.5112624410686223,
+      "loss_partial": 0.0576171875,
+      "step": 976
+    },
+    {
+      "epoch": 0.5117862755369303,
+      "grad_norm": 2.133848190307617,
+      "learning_rate": 1.9953071659296942e-05,
+      "loss": 0.1141,
+      "step": 977
+    },
+    {
+      "epoch": 0.5117862755369303,
+      "loss_partial": 0.05712890625,
+      "step": 977
+    },
+    {
+      "epoch": 0.5117862755369303,
+      "loss_partial": 0.05126953125,
+      "step": 977
+    },
+    {
+      "epoch": 0.5123101100052383,
+      "grad_norm": 1.5941627025604248,
+      "learning_rate": 1.995262824564599e-05,
+      "loss": 0.114,
+      "step": 978
+    },
+    {
+      "epoch": 0.5123101100052383,
+      "loss_partial": 0.064453125,
+      "step": 978
+    },
+    {
+      "epoch": 0.5123101100052383,
+      "loss_partial": 0.053955078125,
+      "step": 978
+    },
+    {
+      "epoch": 0.5128339444735464,
+      "grad_norm": 1.946256160736084,
+      "learning_rate": 1.9952182751957286e-05,
+      "loss": 0.1128,
+      "step": 979
+    },
+    {
+      "epoch": 0.5128339444735464,
+      "loss_partial": 0.06640625,
+      "step": 979
+    },
+    {
+      "epoch": 0.5128339444735464,
+      "loss_partial": 0.0537109375,
+      "step": 979
+    },
+    {
+      "epoch": 0.5133577789418544,
+      "grad_norm": 1.5830131769180298,
+      "learning_rate": 1.995173517832393e-05,
+      "loss": 0.1187,
+      "step": 980
+    },
+    {
+      "epoch": 0.5133577789418544,
+      "loss_partial": 0.051513671875,
+      "step": 980
+    },
+    {
+      "epoch": 0.5133577789418544,
+      "loss_partial": 0.05615234375,
+      "step": 980
+    },
+    {
+      "epoch": 0.5138816134101624,
+      "grad_norm": 1.6746649742126465,
+      "learning_rate": 1.9951285524839458e-05,
+      "loss": 0.1115,
+      "step": 981
+    },
+    {
+      "epoch": 0.5138816134101624,
+      "loss_partial": 0.05322265625,
+      "step": 981
+    },
+    {
+      "epoch": 0.5138816134101624,
+      "loss_partial": 0.0576171875,
+      "step": 981
+    },
+    {
+      "epoch": 0.5144054478784704,
+      "grad_norm": 1.983910322189331,
+      "learning_rate": 1.9950833791597845e-05,
+      "loss": 0.112,
+      "step": 982
+    },
+    {
+      "epoch": 0.5144054478784704,
+      "loss_partial": 0.057373046875,
+      "step": 982
+    },
+    {
+      "epoch": 0.5144054478784704,
+      "loss_partial": 0.0546875,
+      "step": 982
+    },
+    {
+      "epoch": 0.5149292823467784,
+      "grad_norm": 2.464210271835327,
+      "learning_rate": 1.995037997869351e-05,
+      "loss": 0.1101,
+      "step": 983
+    },
+    {
+      "epoch": 0.5149292823467784,
+      "loss_partial": 0.05029296875,
+      "step": 983
+    },
+    {
+      "epoch": 0.5149292823467784,
+      "loss_partial": 0.056396484375,
+      "step": 983
+    },
+    {
+      "epoch": 0.5154531168150864,
+      "grad_norm": 1.9574872255325317,
+      "learning_rate": 1.9949924086221285e-05,
+      "loss": 0.1106,
+      "step": 984
+    },
+    {
+      "epoch": 0.5154531168150864,
+      "loss_partial": 0.0634765625,
+      "step": 984
+    },
+    {
+      "epoch": 0.5154531168150864,
+      "loss_partial": 0.04931640625,
+      "step": 984
+    },
+    {
+      "epoch": 0.5159769512833945,
+      "grad_norm": 2.6570253372192383,
+      "learning_rate": 1.9949466114276457e-05,
+      "loss": 0.1131,
+      "step": 985
+    },
+    {
+      "epoch": 0.5159769512833945,
+      "loss_partial": 0.0576171875,
+      "step": 985
+    },
+    {
+      "epoch": 0.5159769512833945,
+      "loss_partial": 0.052490234375,
+      "step": 985
+    },
+    {
+      "epoch": 0.5165007857517024,
+      "grad_norm": 2.036383628845215,
+      "learning_rate": 1.9949006062954736e-05,
+      "loss": 0.1142,
+      "step": 986
+    },
+    {
+      "epoch": 0.5165007857517024,
+      "loss_partial": 0.04638671875,
+      "step": 986
+    },
+    {
+      "epoch": 0.5165007857517024,
+      "loss_partial": 0.0498046875,
+      "step": 986
+    },
+    {
+      "epoch": 0.5170246202200105,
+      "grad_norm": 3.121446132659912,
+      "learning_rate": 1.9948543932352274e-05,
+      "loss": 0.1094,
+      "step": 987
+    },
+    {
+      "epoch": 0.5170246202200105,
+      "loss_partial": 0.06640625,
+      "step": 987
+    },
+    {
+      "epoch": 0.5170246202200105,
+      "loss_partial": 0.05322265625,
+      "step": 987
+    },
+    {
+      "epoch": 0.5175484546883184,
+      "grad_norm": 3.12756609916687,
+      "learning_rate": 1.9948079722565646e-05,
+      "loss": 0.113,
+      "step": 988
+    },
+    {
+      "epoch": 0.5175484546883184,
+      "loss_partial": 0.060791015625,
+      "step": 988
+    },
+    {
+      "epoch": 0.5175484546883184,
+      "loss_partial": 0.044189453125,
+      "step": 988
+    },
+    {
+      "epoch": 0.5180722891566265,
+      "grad_norm": 3.000704765319824,
+      "learning_rate": 1.9947613433691874e-05,
+      "loss": 0.1125,
+      "step": 989
+    },
+    {
+      "epoch": 0.5180722891566265,
+      "loss_partial": 0.05517578125,
+      "step": 989
+    },
+    {
+      "epoch": 0.5180722891566265,
+      "loss_partial": 0.060302734375,
+      "step": 989
+    },
+    {
+      "epoch": 0.5185961236249346,
+      "grad_norm": 1.5470890998840332,
+      "learning_rate": 1.994714506582841e-05,
+      "loss": 0.1127,
+      "step": 990
+    },
+    {
+      "epoch": 0.5185961236249346,
+      "loss_partial": 0.06640625,
+      "step": 990
+    },
+    {
+      "epoch": 0.5185961236249346,
+      "loss_partial": 0.06201171875,
+      "step": 990
+    },
+    {
+      "epoch": 0.5191199580932425,
+      "grad_norm": 1.616701364517212,
+      "learning_rate": 1.9946674619073137e-05,
+      "loss": 0.1151,
+      "step": 991
+    },
+    {
+      "epoch": 0.5191199580932425,
+      "loss_partial": 0.06005859375,
+      "step": 991
+    },
+    {
+      "epoch": 0.5191199580932425,
+      "loss_partial": 0.05859375,
+      "step": 991
+    },
+    {
+      "epoch": 0.5196437925615506,
+      "grad_norm": 1.931157112121582,
+      "learning_rate": 1.9946202093524378e-05,
+      "loss": 0.1147,
+      "step": 992
+    },
+    {
+      "epoch": 0.5196437925615506,
+      "loss_partial": 0.057373046875,
+      "step": 992
+    },
+    {
+      "epoch": 0.5196437925615506,
+      "loss_partial": 0.068359375,
+      "step": 992
+    },
+    {
+      "epoch": 0.5201676270298585,
+      "grad_norm": 1.6979904174804688,
+      "learning_rate": 1.9945727489280886e-05,
+      "loss": 0.1098,
+      "step": 993
+    },
+    {
+      "epoch": 0.5201676270298585,
+      "loss_partial": 0.060302734375,
+      "step": 993
+    },
+    {
+      "epoch": 0.5201676270298585,
+      "loss_partial": 0.048583984375,
+      "step": 993
+    },
+    {
+      "epoch": 0.5206914614981666,
+      "grad_norm": 1.5608294010162354,
+      "learning_rate": 1.994525080644185e-05,
+      "loss": 0.1136,
+      "step": 994
+    },
+    {
+      "epoch": 0.5206914614981666,
+      "loss_partial": 0.06982421875,
+      "step": 994
+    },
+    {
+      "epoch": 0.5206914614981666,
+      "loss_partial": 0.05419921875,
+      "step": 994
+    },
+    {
+      "epoch": 0.5212152959664746,
+      "grad_norm": 2.1038730144500732,
+      "learning_rate": 1.99447720451069e-05,
+      "loss": 0.11,
+      "step": 995
+    },
+    {
+      "epoch": 0.5212152959664746,
+      "loss_partial": 0.053466796875,
+      "step": 995
+    },
+    {
+      "epoch": 0.5212152959664746,
+      "loss_partial": 0.050537109375,
+      "step": 995
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "grad_norm": 1.932984709739685,
+      "learning_rate": 1.9944291205376084e-05,
+      "loss": 0.1101,
+      "step": 996
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "loss_partial": 0.05029296875,
+      "step": 996
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "loss_partial": 0.051513671875,
+      "step": 996
+    },
+    {
+      "epoch": 0.5222629649030907,
+      "grad_norm": 2.050504446029663,
+      "learning_rate": 1.9943808287349902e-05,
+      "loss": 0.114,
+      "step": 997
+    },
+    {
+      "epoch": 0.5222629649030907,
+      "loss_partial": 0.057861328125,
+      "step": 997
+    },
+    {
+      "epoch": 0.5222629649030907,
+      "loss_partial": 0.05908203125,
+      "step": 997
+    },
+    {
+      "epoch": 0.5227867993713986,
+      "grad_norm": 2.3910646438598633,
+      "learning_rate": 1.994332329112928e-05,
+      "loss": 0.1169,
+      "step": 998
+    },
+    {
+      "epoch": 0.5227867993713986,
+      "loss_partial": 0.06396484375,
+      "step": 998
+    },
+    {
+      "epoch": 0.5227867993713986,
+      "loss_partial": 0.056640625,
+      "step": 998
+    },
+    {
+      "epoch": 0.5233106338397067,
+      "grad_norm": 2.3323707580566406,
+      "learning_rate": 1.9942836216815576e-05,
+      "loss": 0.1116,
+      "step": 999
+    },
+    {
+      "epoch": 0.5233106338397067,
+      "loss_partial": 0.06103515625,
+      "step": 999
+    },
+    {
+      "epoch": 0.5233106338397067,
+      "loss_partial": 0.054931640625,
+      "step": 999
+    },
+    {
+      "epoch": 0.5238344683080147,
+      "grad_norm": 2.0052144527435303,
+      "learning_rate": 1.9942347064510587e-05,
+      "loss": 0.1143,
+      "step": 1000
+    },
+    {
+      "epoch": 0.5238344683080147,
+      "loss_partial": 0.0576171875,
+      "step": 1000
+    },
+    {
+      "epoch": 0.5238344683080147,
+      "loss_partial": 0.0634765625,
+      "step": 1000
+    },
+    {
+      "epoch": 0.5243583027763227,
+      "grad_norm": 1.565406084060669,
+      "learning_rate": 1.9941855834316543e-05,
+      "loss": 0.1136,
+      "step": 1001
+    },
+    {
+      "epoch": 0.5243583027763227,
+      "loss_partial": 0.06591796875,
+      "step": 1001
+    },
+    {
+      "epoch": 0.5243583027763227,
+      "loss_partial": 0.054931640625,
+      "step": 1001
+    },
+    {
+      "epoch": 0.5248821372446307,
+      "grad_norm": 2.3559951782226562,
+      "learning_rate": 1.994136252633611e-05,
+      "loss": 0.1118,
+      "step": 1002
+    },
+    {
+      "epoch": 0.5248821372446307,
+      "loss_partial": 0.059326171875,
+      "step": 1002
+    },
+    {
+      "epoch": 0.5248821372446307,
+      "loss_partial": 0.055419921875,
+      "step": 1002
+    },
+    {
+      "epoch": 0.5254059717129387,
+      "grad_norm": 1.2506792545318604,
+      "learning_rate": 1.9940867140672383e-05,
+      "loss": 0.1138,
+      "step": 1003
+    },
+    {
+      "epoch": 0.5254059717129387,
+      "loss_partial": 0.055419921875,
+      "step": 1003
+    },
+    {
+      "epoch": 0.5254059717129387,
+      "loss_partial": 0.044921875,
+      "step": 1003
+    },
+    {
+      "epoch": 0.5259298061812467,
+      "grad_norm": 1.8791555166244507,
+      "learning_rate": 1.9940369677428896e-05,
+      "loss": 0.1119,
+      "step": 1004
+    },
+    {
+      "epoch": 0.5259298061812467,
+      "loss_partial": 0.06201171875,
+      "step": 1004
+    },
+    {
+      "epoch": 0.5259298061812467,
+      "loss_partial": 0.056396484375,
+      "step": 1004
+    },
+    {
+      "epoch": 0.5264536406495547,
+      "grad_norm": 1.359347939491272,
+      "learning_rate": 1.9939870136709617e-05,
+      "loss": 0.1105,
+      "step": 1005
+    },
+    {
+      "epoch": 0.5264536406495547,
+      "loss_partial": 0.06787109375,
+      "step": 1005
+    },
+    {
+      "epoch": 0.5264536406495547,
+      "loss_partial": 0.05029296875,
+      "step": 1005
+    },
+    {
+      "epoch": 0.5269774751178627,
+      "grad_norm": 2.204306125640869,
+      "learning_rate": 1.9939368518618944e-05,
+      "loss": 0.1117,
+      "step": 1006
+    },
+    {
+      "epoch": 0.5269774751178627,
+      "loss_partial": 0.06103515625,
+      "step": 1006
+    },
+    {
+      "epoch": 0.5269774751178627,
+      "loss_partial": 0.05078125,
+      "step": 1006
+    },
+    {
+      "epoch": 0.5275013095861708,
+      "grad_norm": 1.4527775049209595,
+      "learning_rate": 1.9938864823261714e-05,
+      "loss": 0.1069,
+      "step": 1007
+    },
+    {
+      "epoch": 0.5275013095861708,
+      "loss_partial": 0.055908203125,
+      "step": 1007
+    },
+    {
+      "epoch": 0.5275013095861708,
+      "loss_partial": 0.0634765625,
+      "step": 1007
+    },
+    {
+      "epoch": 0.5280251440544788,
+      "grad_norm": 1.4286500215530396,
+      "learning_rate": 1.99383590507432e-05,
+      "loss": 0.1118,
+      "step": 1008
+    },
+    {
+      "epoch": 0.5280251440544788,
+      "loss_partial": 0.05419921875,
+      "step": 1008
+    },
+    {
+      "epoch": 0.5280251440544788,
+      "loss_partial": 0.050537109375,
+      "step": 1008
+    },
+    {
+      "epoch": 0.5285489785227868,
+      "grad_norm": 1.6138265132904053,
+      "learning_rate": 1.99378512011691e-05,
+      "loss": 0.1175,
+      "step": 1009
+    },
+    {
+      "epoch": 0.5285489785227868,
+      "loss_partial": 0.05078125,
+      "step": 1009
+    },
+    {
+      "epoch": 0.5285489785227868,
+      "loss_partial": 0.064453125,
+      "step": 1009
+    },
+    {
+      "epoch": 0.5290728129910948,
+      "grad_norm": 1.94916570186615,
+      "learning_rate": 1.9937341274645546e-05,
+      "loss": 0.1116,
+      "step": 1010
+    },
+    {
+      "epoch": 0.5290728129910948,
+      "loss_partial": 0.06005859375,
+      "step": 1010
+    },
+    {
+      "epoch": 0.5290728129910948,
+      "loss_partial": 0.05419921875,
+      "step": 1010
+    },
+    {
+      "epoch": 0.5295966474594028,
+      "grad_norm": 2.0233113765716553,
+      "learning_rate": 1.9936829271279125e-05,
+      "loss": 0.1141,
+      "step": 1011
+    },
+    {
+      "epoch": 0.5295966474594028,
+      "loss_partial": 0.054443359375,
+      "step": 1011
+    },
+    {
+      "epoch": 0.5295966474594028,
+      "loss_partial": 0.06494140625,
+      "step": 1011
+    },
+    {
+      "epoch": 0.5301204819277109,
+      "grad_norm": 1.0953514575958252,
+      "learning_rate": 1.993631519117683e-05,
+      "loss": 0.1122,
+      "step": 1012
+    },
+    {
+      "epoch": 0.5301204819277109,
+      "loss_partial": 0.061279296875,
+      "step": 1012
+    },
+    {
+      "epoch": 0.5301204819277109,
+      "loss_partial": 0.05419921875,
+      "step": 1012
+    },
+    {
+      "epoch": 0.5306443163960188,
+      "grad_norm": 2.4114773273468018,
+      "learning_rate": 1.9935799034446103e-05,
+      "loss": 0.1146,
+      "step": 1013
+    },
+    {
+      "epoch": 0.5306443163960188,
+      "loss_partial": 0.05517578125,
+      "step": 1013
+    },
+    {
+      "epoch": 0.5306443163960188,
+      "loss_partial": 0.055908203125,
+      "step": 1013
+    },
+    {
+      "epoch": 0.5311681508643269,
+      "grad_norm": 1.5720771551132202,
+      "learning_rate": 1.993528080119482e-05,
+      "loss": 0.1145,
+      "step": 1014
+    },
+    {
+      "epoch": 0.5311681508643269,
+      "loss_partial": 0.051513671875,
+      "step": 1014
+    },
+    {
+      "epoch": 0.5311681508643269,
+      "loss_partial": 0.055419921875,
+      "step": 1014
+    },
+    {
+      "epoch": 0.5316919853326348,
+      "grad_norm": 1.6050612926483154,
+      "learning_rate": 1.993476049153129e-05,
+      "loss": 0.1153,
+      "step": 1015
+    },
+    {
+      "epoch": 0.5316919853326348,
+      "loss_partial": 0.055419921875,
+      "step": 1015
+    },
+    {
+      "epoch": 0.5316919853326348,
+      "loss_partial": 0.06103515625,
+      "step": 1015
+    },
+    {
+      "epoch": 0.5322158198009429,
+      "grad_norm": 1.7404890060424805,
+      "learning_rate": 1.9934238105564246e-05,
+      "loss": 0.1129,
+      "step": 1016
+    },
+    {
+      "epoch": 0.5322158198009429,
+      "loss_partial": 0.05126953125,
+      "step": 1016
+    },
+    {
+      "epoch": 0.5322158198009429,
+      "loss_partial": 0.06103515625,
+      "step": 1016
+    },
+    {
+      "epoch": 0.532739654269251,
+      "grad_norm": 2.03824520111084,
+      "learning_rate": 1.9933713643402877e-05,
+      "loss": 0.1094,
+      "step": 1017
+    },
+    {
+      "epoch": 0.532739654269251,
+      "loss_partial": 0.053466796875,
+      "step": 1017
+    },
+    {
+      "epoch": 0.532739654269251,
+      "loss_partial": 0.055908203125,
+      "step": 1017
+    },
+    {
+      "epoch": 0.5332634887375589,
+      "grad_norm": 2.743818521499634,
+      "learning_rate": 1.993318710515678e-05,
+      "loss": 0.1124,
+      "step": 1018
+    },
+    {
+      "epoch": 0.5332634887375589,
+      "loss_partial": 0.0537109375,
+      "step": 1018
+    },
+    {
+      "epoch": 0.5332634887375589,
+      "loss_partial": 0.057861328125,
+      "step": 1018
+    },
+    {
+      "epoch": 0.533787323205867,
+      "grad_norm": 2.208622455596924,
+      "learning_rate": 1.9932658490936e-05,
+      "loss": 0.1106,
+      "step": 1019
+    },
+    {
+      "epoch": 0.533787323205867,
+      "loss_partial": 0.047119140625,
+      "step": 1019
+    },
+    {
+      "epoch": 0.533787323205867,
+      "loss_partial": 0.047607421875,
+      "step": 1019
+    },
+    {
+      "epoch": 0.5343111576741749,
+      "grad_norm": 1.6403371095657349,
+      "learning_rate": 1.9932127800851026e-05,
+      "loss": 0.1136,
+      "step": 1020
+    },
+    {
+      "epoch": 0.5343111576741749,
+      "loss_partial": 0.052978515625,
+      "step": 1020
+    },
+    {
+      "epoch": 0.5343111576741749,
+      "loss_partial": 0.06591796875,
+      "step": 1020
+    },
+    {
+      "epoch": 0.534834992142483,
+      "grad_norm": 2.3111467361450195,
+      "learning_rate": 1.9931595035012754e-05,
+      "loss": 0.1127,
+      "step": 1021
+    },
+    {
+      "epoch": 0.534834992142483,
+      "loss_partial": 0.056396484375,
+      "step": 1021
+    },
+    {
+      "epoch": 0.534834992142483,
+      "loss_partial": 0.05859375,
+      "step": 1021
+    },
+    {
+      "epoch": 0.535358826610791,
+      "grad_norm": 2.1001293659210205,
+      "learning_rate": 1.9931060193532534e-05,
+      "loss": 0.1178,
+      "step": 1022
+    },
+    {
+      "epoch": 0.535358826610791,
+      "loss_partial": 0.059814453125,
+      "step": 1022
+    },
+    {
+      "epoch": 0.535358826610791,
+      "loss_partial": 0.052490234375,
+      "step": 1022
+    },
+    {
+      "epoch": 0.535882661079099,
+      "grad_norm": 1.8964799642562866,
+      "learning_rate": 1.993052327652215e-05,
+      "loss": 0.116,
+      "step": 1023
+    },
+    {
+      "epoch": 0.535882661079099,
+      "loss_partial": 0.05419921875,
+      "step": 1023
+    },
+    {
+      "epoch": 0.535882661079099,
+      "loss_partial": 0.0703125,
+      "step": 1023
+    },
+    {
+      "epoch": 0.536406495547407,
+      "grad_norm": 2.1285784244537354,
+      "learning_rate": 1.9929984284093807e-05,
+      "loss": 0.1133,
+      "step": 1024
+    },
+    {
+      "epoch": 0.536406495547407,
+      "loss_partial": 0.054443359375,
+      "step": 1024
+    },
+    {
+      "epoch": 0.536406495547407,
+      "loss_partial": 0.049560546875,
+      "step": 1024
+    },
+    {
+      "epoch": 0.536930330015715,
+      "grad_norm": 1.744103193283081,
+      "learning_rate": 1.9929443216360154e-05,
+      "loss": 0.1122,
+      "step": 1025
+    },
+    {
+      "epoch": 0.536930330015715,
+      "loss_partial": 0.059326171875,
+      "step": 1025
+    },
+    {
+      "epoch": 0.536930330015715,
+      "loss_partial": 0.048583984375,
+      "step": 1025
+    },
+    {
+      "epoch": 0.5374541644840231,
+      "grad_norm": 1.6106587648391724,
+      "learning_rate": 1.992890007343427e-05,
+      "loss": 0.1071,
+      "step": 1026
+    },
+    {
+      "epoch": 0.5374541644840231,
+      "loss_partial": 0.0537109375,
+      "step": 1026
+    },
+    {
+      "epoch": 0.5374541644840231,
+      "loss_partial": 0.052978515625,
+      "step": 1026
+    },
+    {
+      "epoch": 0.5379779989523311,
+      "grad_norm": 1.4077717065811157,
+      "learning_rate": 1.9928354855429668e-05,
+      "loss": 0.1097,
+      "step": 1027
+    },
+    {
+      "epoch": 0.5379779989523311,
+      "loss_partial": 0.06689453125,
+      "step": 1027
+    },
+    {
+      "epoch": 0.5379779989523311,
+      "loss_partial": 0.0478515625,
+      "step": 1027
+    },
+    {
+      "epoch": 0.5385018334206391,
+      "grad_norm": 1.8502495288848877,
+      "learning_rate": 1.99278075624603e-05,
+      "loss": 0.1144,
+      "step": 1028
+    },
+    {
+      "epoch": 0.5385018334206391,
+      "loss_partial": 0.055908203125,
+      "step": 1028
+    },
+    {
+      "epoch": 0.5385018334206391,
+      "loss_partial": 0.06494140625,
+      "step": 1028
+    },
+    {
+      "epoch": 0.5390256678889471,
+      "grad_norm": 2.1362202167510986,
+      "learning_rate": 1.992725819464054e-05,
+      "loss": 0.1158,
+      "step": 1029
+    },
+    {
+      "epoch": 0.5390256678889471,
+      "loss_partial": 0.046630859375,
+      "step": 1029
+    },
+    {
+      "epoch": 0.5390256678889471,
+      "loss_partial": 0.053466796875,
+      "step": 1029
+    },
+    {
+      "epoch": 0.5395495023572551,
+      "grad_norm": 1.2089097499847412,
+      "learning_rate": 1.992670675208521e-05,
+      "loss": 0.112,
+      "step": 1030
+    },
+    {
+      "epoch": 0.5395495023572551,
+      "loss_partial": 0.0634765625,
+      "step": 1030
+    },
+    {
+      "epoch": 0.5395495023572551,
+      "loss_partial": 0.05419921875,
+      "step": 1030
+    },
+    {
+      "epoch": 0.5400733368255631,
+      "grad_norm": 2.4934439659118652,
+      "learning_rate": 1.992615323490955e-05,
+      "loss": 0.1119,
+      "step": 1031
+    },
+    {
+      "epoch": 0.5400733368255631,
+      "loss_partial": 0.051513671875,
+      "step": 1031
+    },
+    {
+      "epoch": 0.5400733368255631,
+      "loss_partial": 0.059814453125,
+      "step": 1031
+    },
+    {
+      "epoch": 0.5405971712938711,
+      "grad_norm": 2.2117295265197754,
+      "learning_rate": 1.992559764322924e-05,
+      "loss": 0.1114,
+      "step": 1032
+    },
+    {
+      "epoch": 0.5405971712938711,
+      "loss_partial": 0.053466796875,
+      "step": 1032
+    },
+    {
+      "epoch": 0.5405971712938711,
+      "loss_partial": 0.061767578125,
+      "step": 1032
+    },
+    {
+      "epoch": 0.5411210057621791,
+      "grad_norm": 2.7633697986602783,
+      "learning_rate": 1.992503997716041e-05,
+      "loss": 0.112,
+      "step": 1033
+    },
+    {
+      "epoch": 0.5411210057621791,
+      "loss_partial": 0.05615234375,
+      "step": 1033
+    },
+    {
+      "epoch": 0.5411210057621791,
+      "loss_partial": 0.057373046875,
+      "step": 1033
+    },
+    {
+      "epoch": 0.5416448402304872,
+      "grad_norm": 1.5322961807250977,
+      "learning_rate": 1.9924480236819598e-05,
+      "loss": 0.11,
+      "step": 1034
+    },
+    {
+      "epoch": 0.5416448402304872,
+      "loss_partial": 0.05419921875,
+      "step": 1034
+    },
+    {
+      "epoch": 0.5416448402304872,
+      "loss_partial": 0.055908203125,
+      "step": 1034
+    },
+    {
+      "epoch": 0.5421686746987951,
+      "grad_norm": 2.275934934616089,
+      "learning_rate": 1.9923918422323783e-05,
+      "loss": 0.1138,
+      "step": 1035
+    },
+    {
+      "epoch": 0.5421686746987951,
+      "loss_partial": 0.064453125,
+      "step": 1035
+    },
+    {
+      "epoch": 0.5421686746987951,
+      "loss_partial": 0.0654296875,
+      "step": 1035
+    },
+    {
+      "epoch": 0.5426925091671032,
+      "grad_norm": 1.815812110900879,
+      "learning_rate": 1.9923354533790386e-05,
+      "loss": 0.1125,
+      "step": 1036
+    },
+    {
+      "epoch": 0.5426925091671032,
+      "loss_partial": 0.050048828125,
+      "step": 1036
+    },
+    {
+      "epoch": 0.5426925091671032,
+      "loss_partial": 0.06103515625,
+      "step": 1036
+    },
+    {
+      "epoch": 0.5432163436354112,
+      "grad_norm": 2.183906316757202,
+      "learning_rate": 1.992278857133726e-05,
+      "loss": 0.1126,
+      "step": 1037
+    },
+    {
+      "epoch": 0.5432163436354112,
+      "loss_partial": 0.0556640625,
+      "step": 1037
+    },
+    {
+      "epoch": 0.5432163436354112,
+      "loss_partial": 0.06640625,
+      "step": 1037
+    },
+    {
+      "epoch": 0.5437401781037192,
+      "grad_norm": 1.293330192565918,
+      "learning_rate": 1.992222053508268e-05,
+      "loss": 0.114,
+      "step": 1038
+    },
+    {
+      "epoch": 0.5437401781037192,
+      "loss_partial": 0.051513671875,
+      "step": 1038
+    },
+    {
+      "epoch": 0.5437401781037192,
+      "loss_partial": 0.05419921875,
+      "step": 1038
+    },
+    {
+      "epoch": 0.5442640125720273,
+      "grad_norm": 1.770542860031128,
+      "learning_rate": 1.9921650425145368e-05,
+      "loss": 0.1097,
+      "step": 1039
+    },
+    {
+      "epoch": 0.5442640125720273,
+      "loss_partial": 0.048828125,
+      "step": 1039
+    },
+    {
+      "epoch": 0.5442640125720273,
+      "loss_partial": 0.05419921875,
+      "step": 1039
+    },
+    {
+      "epoch": 0.5447878470403352,
+      "grad_norm": 2.0630059242248535,
+      "learning_rate": 1.9921078241644467e-05,
+      "loss": 0.1097,
+      "step": 1040
+    },
+    {
+      "epoch": 0.5447878470403352,
+      "loss_partial": 0.06640625,
+      "step": 1040
+    },
+    {
+      "epoch": 0.5447878470403352,
+      "loss_partial": 0.045166015625,
+      "step": 1040
+    },
+    {
+      "epoch": 0.5453116815086433,
+      "grad_norm": 1.6954290866851807,
+      "learning_rate": 1.9920503984699567e-05,
+      "loss": 0.1123,
+      "step": 1041
+    },
+    {
+      "epoch": 0.5453116815086433,
+      "loss_partial": 0.059326171875,
+      "step": 1041
+    },
+    {
+      "epoch": 0.5453116815086433,
+      "loss_partial": 0.051025390625,
+      "step": 1041
+    },
+    {
+      "epoch": 0.5458355159769512,
+      "grad_norm": 1.7919913530349731,
+      "learning_rate": 1.991992765443068e-05,
+      "loss": 0.1135,
+      "step": 1042
+    },
+    {
+      "epoch": 0.5458355159769512,
+      "loss_partial": 0.0537109375,
+      "step": 1042
+    },
+    {
+      "epoch": 0.5458355159769512,
+      "loss_partial": 0.0595703125,
+      "step": 1042
+    },
+    {
+      "epoch": 0.5463593504452593,
+      "grad_norm": 1.9916304349899292,
+      "learning_rate": 1.9919349250958256e-05,
+      "loss": 0.1121,
+      "step": 1043
+    },
+    {
+      "epoch": 0.5463593504452593,
+      "loss_partial": 0.053955078125,
+      "step": 1043
+    },
+    {
+      "epoch": 0.5463593504452593,
+      "loss_partial": 0.049072265625,
+      "step": 1043
+    },
+    {
+      "epoch": 0.5468831849135674,
+      "grad_norm": 1.7330372333526611,
+      "learning_rate": 1.9918768774403174e-05,
+      "loss": 0.1116,
+      "step": 1044
+    },
+    {
+      "epoch": 0.5468831849135674,
+      "loss_partial": 0.051513671875,
+      "step": 1044
+    },
+    {
+      "epoch": 0.5468831849135674,
+      "loss_partial": 0.056884765625,
+      "step": 1044
+    },
+    {
+      "epoch": 0.5474070193818753,
+      "grad_norm": 1.8701972961425781,
+      "learning_rate": 1.9918186224886756e-05,
+      "loss": 0.1091,
+      "step": 1045
+    },
+    {
+      "epoch": 0.5474070193818753,
+      "loss_partial": 0.0537109375,
+      "step": 1045
+    },
+    {
+      "epoch": 0.5474070193818753,
+      "loss_partial": 0.0517578125,
+      "step": 1045
+    },
+    {
+      "epoch": 0.5479308538501834,
+      "grad_norm": 1.7141889333724976,
+      "learning_rate": 1.991760160253075e-05,
+      "loss": 0.112,
+      "step": 1046
+    },
+    {
+      "epoch": 0.5479308538501834,
+      "loss_partial": 0.06982421875,
+      "step": 1046
+    },
+    {
+      "epoch": 0.5479308538501834,
+      "loss_partial": 0.052490234375,
+      "step": 1046
+    },
+    {
+      "epoch": 0.5484546883184913,
+      "grad_norm": 1.241444706916809,
+      "learning_rate": 1.991701490745734e-05,
+      "loss": 0.1129,
+      "step": 1047
+    },
+    {
+      "epoch": 0.5484546883184913,
+      "loss_partial": 0.050048828125,
+      "step": 1047
+    },
+    {
+      "epoch": 0.5484546883184913,
+      "loss_partial": 0.056884765625,
+      "step": 1047
+    },
+    {
+      "epoch": 0.5489785227867994,
+      "grad_norm": 2.371305227279663,
+      "learning_rate": 1.9916426139789136e-05,
+      "loss": 0.1122,
+      "step": 1048
+    },
+    {
+      "epoch": 0.5489785227867994,
+      "loss_partial": 0.0576171875,
+      "step": 1048
+    },
+    {
+      "epoch": 0.5489785227867994,
+      "loss_partial": 0.061767578125,
+      "step": 1048
+    },
+    {
+      "epoch": 0.5495023572551074,
+      "grad_norm": 2.268573760986328,
+      "learning_rate": 1.9915835299649197e-05,
+      "loss": 0.1107,
+      "step": 1049
+    },
+    {
+      "epoch": 0.5495023572551074,
+      "loss_partial": 0.043701171875,
+      "step": 1049
+    },
+    {
+      "epoch": 0.5495023572551074,
+      "loss_partial": 0.051025390625,
+      "step": 1049
+    },
+    {
+      "epoch": 0.5500261917234154,
+      "grad_norm": 1.9345781803131104,
+      "learning_rate": 1.991524238716099e-05,
+      "loss": 0.1133,
+      "step": 1050
+    },
+    {
+      "epoch": 0.5500261917234154,
+      "loss_partial": 0.0537109375,
+      "step": 1050
+    },
+    {
+      "epoch": 0.5500261917234154,
+      "loss_partial": 0.052978515625,
+      "step": 1050
+    },
+    {
+      "epoch": 0.5505500261917234,
+      "grad_norm": 2.269200563430786,
+      "learning_rate": 1.9914647402448443e-05,
+      "loss": 0.1136,
+      "step": 1051
+    },
+    {
+      "epoch": 0.5505500261917234,
+      "loss_partial": 0.056640625,
+      "step": 1051
+    },
+    {
+      "epoch": 0.5505500261917234,
+      "loss_partial": 0.052490234375,
+      "step": 1051
+    },
+    {
+      "epoch": 0.5510738606600314,
+      "grad_norm": 2.1687521934509277,
+      "learning_rate": 1.9914050345635897e-05,
+      "loss": 0.1102,
+      "step": 1052
+    },
+    {
+      "epoch": 0.5510738606600314,
+      "loss_partial": 0.050048828125,
+      "step": 1052
+    },
+    {
+      "epoch": 0.5510738606600314,
+      "loss_partial": 0.06298828125,
+      "step": 1052
+    },
+    {
+      "epoch": 0.5515976951283394,
+      "grad_norm": 1.7193535566329956,
+      "learning_rate": 1.9913451216848133e-05,
+      "loss": 0.1146,
+      "step": 1053
+    },
+    {
+      "epoch": 0.5515976951283394,
+      "loss_partial": 0.06103515625,
+      "step": 1053
+    },
+    {
+      "epoch": 0.5515976951283394,
+      "loss_partial": 0.0537109375,
+      "step": 1053
+    },
+    {
+      "epoch": 0.5521215295966475,
+      "grad_norm": 2.62015438079834,
+      "learning_rate": 1.991285001621037e-05,
+      "loss": 0.1137,
+      "step": 1054
+    },
+    {
+      "epoch": 0.5521215295966475,
+      "loss_partial": 0.0693359375,
+      "step": 1054
+    },
+    {
+      "epoch": 0.5521215295966475,
+      "loss_partial": 0.044189453125,
+      "step": 1054
+    },
+    {
+      "epoch": 0.5526453640649555,
+      "grad_norm": 2.171715021133423,
+      "learning_rate": 1.9912246743848252e-05,
+      "loss": 0.1109,
+      "step": 1055
+    },
+    {
+      "epoch": 0.5526453640649555,
+      "loss_partial": 0.060791015625,
+      "step": 1055
+    },
+    {
+      "epoch": 0.5526453640649555,
+      "loss_partial": 0.044189453125,
+      "step": 1055
+    },
+    {
+      "epoch": 0.5531691985332635,
+      "grad_norm": 2.007647752761841,
+      "learning_rate": 1.9911641399887862e-05,
+      "loss": 0.114,
+      "step": 1056
+    },
+    {
+      "epoch": 0.5531691985332635,
+      "loss_partial": 0.05859375,
+      "step": 1056
+    },
+    {
+      "epoch": 0.5531691985332635,
+      "loss_partial": 0.0595703125,
+      "step": 1056
+    },
+    {
+      "epoch": 0.5536930330015715,
+      "grad_norm": 2.31152606010437,
+      "learning_rate": 1.991103398445571e-05,
+      "loss": 0.1109,
+      "step": 1057
+    },
+    {
+      "epoch": 0.5536930330015715,
+      "loss_partial": 0.05810546875,
+      "step": 1057
+    },
+    {
+      "epoch": 0.5536930330015715,
+      "loss_partial": 0.055419921875,
+      "step": 1057
+    },
+    {
+      "epoch": 0.5542168674698795,
+      "grad_norm": 1.5686171054840088,
+      "learning_rate": 1.9910424497678746e-05,
+      "loss": 0.1108,
+      "step": 1058
+    },
+    {
+      "epoch": 0.5542168674698795,
+      "loss_partial": 0.061767578125,
+      "step": 1058
+    },
+    {
+      "epoch": 0.5542168674698795,
+      "loss_partial": 0.06201171875,
+      "step": 1058
+    },
+    {
+      "epoch": 0.5547407019381876,
+      "grad_norm": 2.231790542602539,
+      "learning_rate": 1.990981293968434e-05,
+      "loss": 0.1124,
+      "step": 1059
+    },
+    {
+      "epoch": 0.5547407019381876,
+      "loss_partial": 0.06494140625,
+      "step": 1059
+    },
+    {
+      "epoch": 0.5547407019381876,
+      "loss_partial": 0.0576171875,
+      "step": 1059
+    },
+    {
+      "epoch": 0.5552645364064955,
+      "grad_norm": 2.3046514987945557,
+      "learning_rate": 1.9909199310600317e-05,
+      "loss": 0.1115,
+      "step": 1060
+    },
+    {
+      "epoch": 0.5552645364064955,
+      "loss_partial": 0.05224609375,
+      "step": 1060
+    },
+    {
+      "epoch": 0.5552645364064955,
+      "loss_partial": 0.060302734375,
+      "step": 1060
+    },
+    {
+      "epoch": 0.5557883708748036,
+      "grad_norm": 2.6659066677093506,
+      "learning_rate": 1.990858361055491e-05,
+      "loss": 0.111,
+      "step": 1061
+    },
+    {
+      "epoch": 0.5557883708748036,
+      "loss_partial": 0.056884765625,
+      "step": 1061
+    },
+    {
+      "epoch": 0.5557883708748036,
+      "loss_partial": 0.054931640625,
+      "step": 1061
+    },
+    {
+      "epoch": 0.5563122053431115,
+      "grad_norm": 2.04217267036438,
+      "learning_rate": 1.9907965839676802e-05,
+      "loss": 0.1128,
+      "step": 1062
+    },
+    {
+      "epoch": 0.5563122053431115,
+      "loss_partial": 0.06787109375,
+      "step": 1062
+    },
+    {
+      "epoch": 0.5563122053431115,
+      "loss_partial": 0.05908203125,
+      "step": 1062
+    },
+    {
+      "epoch": 0.5568360398114196,
+      "grad_norm": 1.5101131200790405,
+      "learning_rate": 1.9907345998095105e-05,
+      "loss": 0.1133,
+      "step": 1063
+    },
+    {
+      "epoch": 0.5568360398114196,
+      "loss_partial": 0.059814453125,
+      "step": 1063
+    },
+    {
+      "epoch": 0.5568360398114196,
+      "loss_partial": 0.054443359375,
+      "step": 1063
+    },
+    {
+      "epoch": 0.5573598742797276,
+      "grad_norm": 1.609043836593628,
+      "learning_rate": 1.9906724085939356e-05,
+      "loss": 0.1161,
+      "step": 1064
+    },
+    {
+      "epoch": 0.5573598742797276,
+      "loss_partial": 0.059326171875,
+      "step": 1064
+    },
+    {
+      "epoch": 0.5573598742797276,
+      "loss_partial": 0.0498046875,
+      "step": 1064
+    },
+    {
+      "epoch": 0.5578837087480356,
+      "grad_norm": 1.3095420598983765,
+      "learning_rate": 1.9906100103339536e-05,
+      "loss": 0.1115,
+      "step": 1065
+    },
+    {
+      "epoch": 0.5578837087480356,
+      "loss_partial": 0.05712890625,
+      "step": 1065
+    },
+    {
+      "epoch": 0.5578837087480356,
+      "loss_partial": 0.05908203125,
+      "step": 1065
+    },
+    {
+      "epoch": 0.5584075432163437,
+      "grad_norm": 1.462333083152771,
+      "learning_rate": 1.9905474050426053e-05,
+      "loss": 0.1111,
+      "step": 1066
+    },
+    {
+      "epoch": 0.5584075432163437,
+      "loss_partial": 0.060791015625,
+      "step": 1066
+    },
+    {
+      "epoch": 0.5584075432163437,
+      "loss_partial": 0.052001953125,
+      "step": 1066
+    },
+    {
+      "epoch": 0.5589313776846516,
+      "grad_norm": 1.5706316232681274,
+      "learning_rate": 1.9904845927329742e-05,
+      "loss": 0.1089,
+      "step": 1067
+    },
+    {
+      "epoch": 0.5589313776846516,
+      "loss_partial": 0.048583984375,
+      "step": 1067
+    },
+    {
+      "epoch": 0.5589313776846516,
+      "loss_partial": 0.050537109375,
+      "step": 1067
+    },
+    {
+      "epoch": 0.5594552121529597,
+      "grad_norm": 2.015862464904785,
+      "learning_rate": 1.990421573418189e-05,
+      "loss": 0.113,
+      "step": 1068
+    },
+    {
+      "epoch": 0.5594552121529597,
+      "loss_partial": 0.0498046875,
+      "step": 1068
+    },
+    {
+      "epoch": 0.5594552121529597,
+      "loss_partial": 0.057373046875,
+      "step": 1068
+    },
+    {
+      "epoch": 0.5599790466212676,
+      "grad_norm": 2.255983352661133,
+      "learning_rate": 1.990358347111419e-05,
+      "loss": 0.1127,
+      "step": 1069
+    },
+    {
+      "epoch": 0.5599790466212676,
+      "loss_partial": 0.062255859375,
+      "step": 1069
+    },
+    {
+      "epoch": 0.5599790466212676,
+      "loss_partial": 0.05419921875,
+      "step": 1069
+    },
+    {
+      "epoch": 0.5605028810895757,
+      "grad_norm": 1.4277225732803345,
+      "learning_rate": 1.9902949138258785e-05,
+      "loss": 0.1145,
+      "step": 1070
+    },
+    {
+      "epoch": 0.5605028810895757,
+      "loss_partial": 0.04833984375,
+      "step": 1070
+    },
+    {
+      "epoch": 0.5605028810895757,
+      "loss_partial": 0.0537109375,
+      "step": 1070
+    },
+    {
+      "epoch": 0.5610267155578837,
+      "grad_norm": 1.5006954669952393,
+      "learning_rate": 1.990231273574825e-05,
+      "loss": 0.1093,
+      "step": 1071
+    },
+    {
+      "epoch": 0.5610267155578837,
+      "loss_partial": 0.05126953125,
+      "step": 1071
+    },
+    {
+      "epoch": 0.5610267155578837,
+      "loss_partial": 0.054931640625,
+      "step": 1071
+    },
+    {
+      "epoch": 0.5615505500261917,
+      "grad_norm": 1.4388582706451416,
+      "learning_rate": 1.9901674263715592e-05,
+      "loss": 0.1078,
+      "step": 1072
+    },
+    {
+      "epoch": 0.5615505500261917,
+      "loss_partial": 0.054931640625,
+      "step": 1072
+    },
+    {
+      "epoch": 0.5615505500261917,
+      "loss_partial": 0.047607421875,
+      "step": 1072
+    },
+    {
+      "epoch": 0.5620743844944998,
+      "grad_norm": 1.7269781827926636,
+      "learning_rate": 1.990103372229424e-05,
+      "loss": 0.1133,
+      "step": 1073
+    },
+    {
+      "epoch": 0.5620743844944998,
+      "loss_partial": 0.06396484375,
+      "step": 1073
+    },
+    {
+      "epoch": 0.5620743844944998,
+      "loss_partial": 0.05322265625,
+      "step": 1073
+    },
+    {
+      "epoch": 0.5625982189628077,
+      "grad_norm": 1.6790763139724731,
+      "learning_rate": 1.9900391111618064e-05,
+      "loss": 0.1125,
+      "step": 1074
+    },
+    {
+      "epoch": 0.5625982189628077,
+      "loss_partial": 0.049560546875,
+      "step": 1074
+    },
+    {
+      "epoch": 0.5625982189628077,
+      "loss_partial": 0.057373046875,
+      "step": 1074
+    },
+    {
+      "epoch": 0.5631220534311158,
+      "grad_norm": 1.426659345626831,
+      "learning_rate": 1.989974643182137e-05,
+      "loss": 0.1138,
+      "step": 1075
+    },
+    {
+      "epoch": 0.5631220534311158,
+      "loss_partial": 0.054443359375,
+      "step": 1075
+    },
+    {
+      "epoch": 0.5631220534311158,
+      "loss_partial": 0.058349609375,
+      "step": 1075
+    },
+    {
+      "epoch": 0.5636458878994238,
+      "grad_norm": 1.4121755361557007,
+      "learning_rate": 1.9899099683038886e-05,
+      "loss": 0.1134,
+      "step": 1076
+    },
+    {
+      "epoch": 0.5636458878994238,
+      "loss_partial": 0.06201171875,
+      "step": 1076
+    },
+    {
+      "epoch": 0.5636458878994238,
+      "loss_partial": 0.06103515625,
+      "step": 1076
+    },
+    {
+      "epoch": 0.5641697223677318,
+      "grad_norm": 1.3419263362884521,
+      "learning_rate": 1.9898450865405786e-05,
+      "loss": 0.1087,
+      "step": 1077
+    },
+    {
+      "epoch": 0.5641697223677318,
+      "loss_partial": 0.054443359375,
+      "step": 1077
+    },
+    {
+      "epoch": 0.5641697223677318,
+      "loss_partial": 0.05712890625,
+      "step": 1077
+    },
+    {
+      "epoch": 0.5646935568360398,
+      "grad_norm": 1.3715968132019043,
+      "learning_rate": 1.9897799979057667e-05,
+      "loss": 0.1122,
+      "step": 1078
+    },
+    {
+      "epoch": 0.5646935568360398,
+      "loss_partial": 0.06298828125,
+      "step": 1078
+    },
+    {
+      "epoch": 0.5646935568360398,
+      "loss_partial": 0.0537109375,
+      "step": 1078
+    },
+    {
+      "epoch": 0.5652173913043478,
+      "grad_norm": 1.488145112991333,
+      "learning_rate": 1.9897147024130557e-05,
+      "loss": 0.1091,
+      "step": 1079
+    },
+    {
+      "epoch": 0.5652173913043478,
+      "loss_partial": 0.057861328125,
+      "step": 1079
+    },
+    {
+      "epoch": 0.5652173913043478,
+      "loss_partial": 0.0634765625,
+      "step": 1079
+    },
+    {
+      "epoch": 0.5657412257726558,
+      "grad_norm": 2.124333143234253,
+      "learning_rate": 1.989649200076092e-05,
+      "loss": 0.1107,
+      "step": 1080
+    },
+    {
+      "epoch": 0.5657412257726558,
+      "loss_partial": 0.05810546875,
+      "step": 1080
+    },
+    {
+      "epoch": 0.5657412257726558,
+      "loss_partial": 0.06591796875,
+      "step": 1080
+    },
+    {
+      "epoch": 0.5662650602409639,
+      "grad_norm": 1.9060728549957275,
+      "learning_rate": 1.9895834909085652e-05,
+      "loss": 0.11,
+      "step": 1081
+    },
+    {
+      "epoch": 0.5662650602409639,
+      "loss_partial": 0.062255859375,
+      "step": 1081
+    },
+    {
+      "epoch": 0.5662650602409639,
+      "loss_partial": 0.05712890625,
+      "step": 1081
+    },
+    {
+      "epoch": 0.5667888947092719,
+      "grad_norm": 1.860827088356018,
+      "learning_rate": 1.9895175749242086e-05,
+      "loss": 0.1125,
+      "step": 1082
+    },
+    {
+      "epoch": 0.5667888947092719,
+      "loss_partial": 0.048095703125,
+      "step": 1082
+    },
+    {
+      "epoch": 0.5667888947092719,
+      "loss_partial": 0.06005859375,
+      "step": 1082
+    },
+    {
+      "epoch": 0.5673127291775799,
+      "grad_norm": 2.6013004779815674,
+      "learning_rate": 1.9894514521367975e-05,
+      "loss": 0.1093,
+      "step": 1083
+    },
+    {
+      "epoch": 0.5673127291775799,
+      "loss_partial": 0.06103515625,
+      "step": 1083
+    },
+    {
+      "epoch": 0.5673127291775799,
+      "loss_partial": 0.044189453125,
+      "step": 1083
+    },
+    {
+      "epoch": 0.5678365636458879,
+      "grad_norm": 2.4738779067993164,
+      "learning_rate": 1.9893851225601515e-05,
+      "loss": 0.1113,
+      "step": 1084
+    },
+    {
+      "epoch": 0.5678365636458879,
+      "loss_partial": 0.05322265625,
+      "step": 1084
+    },
+    {
+      "epoch": 0.5678365636458879,
+      "loss_partial": 0.0634765625,
+      "step": 1084
+    },
+    {
+      "epoch": 0.5683603981141959,
+      "grad_norm": 2.730713129043579,
+      "learning_rate": 1.989318586208133e-05,
+      "loss": 0.114,
+      "step": 1085
+    },
+    {
+      "epoch": 0.5683603981141959,
+      "loss_partial": 0.056640625,
+      "step": 1085
+    },
+    {
+      "epoch": 0.5683603981141959,
+      "loss_partial": 0.053955078125,
+      "step": 1085
+    },
+    {
+      "epoch": 0.568884232582504,
+      "grad_norm": 3.383035659790039,
+      "learning_rate": 1.989251843094648e-05,
+      "loss": 0.1092,
+      "step": 1086
+    },
+    {
+      "epoch": 0.568884232582504,
+      "loss_partial": 0.0615234375,
+      "step": 1086
+    },
+    {
+      "epoch": 0.568884232582504,
+      "loss_partial": 0.046142578125,
+      "step": 1086
+    },
+    {
+      "epoch": 0.5694080670508119,
+      "grad_norm": 2.4168975353240967,
+      "learning_rate": 1.989184893233645e-05,
+      "loss": 0.11,
+      "step": 1087
+    },
+    {
+      "epoch": 0.5694080670508119,
+      "loss_partial": 0.05029296875,
+      "step": 1087
+    },
+    {
+      "epoch": 0.5694080670508119,
+      "loss_partial": 0.05517578125,
+      "step": 1087
+    },
+    {
+      "epoch": 0.56993190151912,
+      "grad_norm": 2.1941514015197754,
+      "learning_rate": 1.9891177366391163e-05,
+      "loss": 0.1117,
+      "step": 1088
+    },
+    {
+      "epoch": 0.56993190151912,
+      "loss_partial": 0.06640625,
+      "step": 1088
+    },
+    {
+      "epoch": 0.56993190151912,
+      "loss_partial": 0.0576171875,
+      "step": 1088
+    },
+    {
+      "epoch": 0.5704557359874279,
+      "grad_norm": 1.773014783859253,
+      "learning_rate": 1.989050373325097e-05,
+      "loss": 0.1127,
+      "step": 1089
+    },
+    {
+      "epoch": 0.5704557359874279,
+      "loss_partial": 0.044921875,
+      "step": 1089
+    },
+    {
+      "epoch": 0.5704557359874279,
+      "loss_partial": 0.04931640625,
+      "step": 1089
+    },
+    {
+      "epoch": 0.570979570455736,
+      "grad_norm": 2.337623357772827,
+      "learning_rate": 1.988982803305666e-05,
+      "loss": 0.1156,
+      "step": 1090
+    },
+    {
+      "epoch": 0.570979570455736,
+      "loss_partial": 0.05029296875,
+      "step": 1090
+    },
+    {
+      "epoch": 0.570979570455736,
+      "loss_partial": 0.0732421875,
+      "step": 1090
+    },
+    {
+      "epoch": 0.5715034049240441,
+      "grad_norm": 3.079332113265991,
+      "learning_rate": 1.9889150265949446e-05,
+      "loss": 0.1133,
+      "step": 1091
+    },
+    {
+      "epoch": 0.5715034049240441,
+      "loss_partial": 0.06298828125,
+      "step": 1091
+    },
+    {
+      "epoch": 0.5715034049240441,
+      "loss_partial": 0.04931640625,
+      "step": 1091
+    },
+    {
+      "epoch": 0.572027239392352,
+      "grad_norm": 2.35335636138916,
+      "learning_rate": 1.988847043207098e-05,
+      "loss": 0.112,
+      "step": 1092
+    },
+    {
+      "epoch": 0.572027239392352,
+      "loss_partial": 0.0517578125,
+      "step": 1092
+    },
+    {
+      "epoch": 0.572027239392352,
+      "loss_partial": 0.06298828125,
+      "step": 1092
+    },
+    {
+      "epoch": 0.5725510738606601,
+      "grad_norm": 2.4805853366851807,
+      "learning_rate": 1.9887788531563343e-05,
+      "loss": 0.1126,
+      "step": 1093
+    },
+    {
+      "epoch": 0.5725510738606601,
+      "loss_partial": 0.057373046875,
+      "step": 1093
+    },
+    {
+      "epoch": 0.5725510738606601,
+      "loss_partial": 0.058837890625,
+      "step": 1093
+    },
+    {
+      "epoch": 0.573074908328968,
+      "grad_norm": 2.3628199100494385,
+      "learning_rate": 1.9887104564569048e-05,
+      "loss": 0.1133,
+      "step": 1094
+    },
+    {
+      "epoch": 0.573074908328968,
+      "loss_partial": 0.05419921875,
+      "step": 1094
+    },
+    {
+      "epoch": 0.573074908328968,
+      "loss_partial": 0.048828125,
+      "step": 1094
+    },
+    {
+      "epoch": 0.5735987427972761,
+      "grad_norm": 2.1923165321350098,
+      "learning_rate": 1.9886418531231034e-05,
+      "loss": 0.1079,
+      "step": 1095
+    },
+    {
+      "epoch": 0.5735987427972761,
+      "loss_partial": 0.052978515625,
+      "step": 1095
+    },
+    {
+      "epoch": 0.5735987427972761,
+      "loss_partial": 0.05908203125,
+      "step": 1095
+    },
+    {
+      "epoch": 0.574122577265584,
+      "grad_norm": 1.7969064712524414,
+      "learning_rate": 1.9885730431692687e-05,
+      "loss": 0.1122,
+      "step": 1096
+    },
+    {
+      "epoch": 0.574122577265584,
+      "loss_partial": 0.05224609375,
+      "step": 1096
+    },
+    {
+      "epoch": 0.574122577265584,
+      "loss_partial": 0.059326171875,
+      "step": 1096
+    },
+    {
+      "epoch": 0.5746464117338921,
+      "grad_norm": 2.380741596221924,
+      "learning_rate": 1.988504026609781e-05,
+      "loss": 0.1151,
+      "step": 1097
+    },
+    {
+      "epoch": 0.5746464117338921,
+      "loss_partial": 0.06396484375,
+      "step": 1097
+    },
+    {
+      "epoch": 0.5746464117338921,
+      "loss_partial": 0.0576171875,
+      "step": 1097
+    },
+    {
+      "epoch": 0.5751702462022001,
+      "grad_norm": 1.9079923629760742,
+      "learning_rate": 1.9884348034590647e-05,
+      "loss": 0.1109,
+      "step": 1098
+    },
+    {
+      "epoch": 0.5751702462022001,
+      "loss_partial": 0.05126953125,
+      "step": 1098
+    },
+    {
+      "epoch": 0.5751702462022001,
+      "loss_partial": 0.05810546875,
+      "step": 1098
+    },
+    {
+      "epoch": 0.5756940806705081,
+      "grad_norm": 2.06380295753479,
+      "learning_rate": 1.9883653737315864e-05,
+      "loss": 0.1106,
+      "step": 1099
+    },
+    {
+      "epoch": 0.5756940806705081,
+      "loss_partial": 0.06005859375,
+      "step": 1099
+    },
+    {
+      "epoch": 0.5756940806705081,
+      "loss_partial": 0.052978515625,
+      "step": 1099
+    },
+    {
+      "epoch": 0.5762179151388162,
+      "grad_norm": 1.9307208061218262,
+      "learning_rate": 1.988295737441857e-05,
+      "loss": 0.1103,
+      "step": 1100
+    },
+    {
+      "epoch": 0.5762179151388162,
+      "loss_partial": 0.05419921875,
+      "step": 1100
+    },
+    {
+      "epoch": 0.5762179151388162,
+      "loss_partial": 0.05908203125,
+      "step": 1100
+    },
+    {
+      "epoch": 0.5767417496071241,
+      "grad_norm": 2.4275336265563965,
+      "learning_rate": 1.9882258946044298e-05,
+      "loss": 0.1121,
+      "step": 1101
+    },
+    {
+      "epoch": 0.5767417496071241,
+      "loss_partial": 0.06494140625,
+      "step": 1101
+    },
+    {
+      "epoch": 0.5767417496071241,
+      "loss_partial": 0.05224609375,
+      "step": 1101
+    },
+    {
+      "epoch": 0.5772655840754322,
+      "grad_norm": 2.4421324729919434,
+      "learning_rate": 1.988155845233902e-05,
+      "loss": 0.1145,
+      "step": 1102
+    },
+    {
+      "epoch": 0.5772655840754322,
+      "loss_partial": 0.06640625,
+      "step": 1102
+    },
+    {
+      "epoch": 0.5772655840754322,
+      "loss_partial": 0.051513671875,
+      "step": 1102
+    },
+    {
+      "epoch": 0.5777894185437402,
+      "grad_norm": 1.4680297374725342,
+      "learning_rate": 1.9880855893449128e-05,
+      "loss": 0.1132,
+      "step": 1103
+    },
+    {
+      "epoch": 0.5777894185437402,
+      "loss_partial": 0.051025390625,
+      "step": 1103
+    },
+    {
+      "epoch": 0.5777894185437402,
+      "loss_partial": 0.06591796875,
+      "step": 1103
+    },
+    {
+      "epoch": 0.5783132530120482,
+      "grad_norm": 2.1772994995117188,
+      "learning_rate": 1.9880151269521458e-05,
+      "loss": 0.1139,
+      "step": 1104
+    },
+    {
+      "epoch": 0.5783132530120482,
+      "loss_partial": 0.055419921875,
+      "step": 1104
+    },
+    {
+      "epoch": 0.5783132530120482,
+      "loss_partial": 0.047607421875,
+      "step": 1104
+    },
+    {
+      "epoch": 0.5788370874803562,
+      "grad_norm": 2.5384042263031006,
+      "learning_rate": 1.9879444580703266e-05,
+      "loss": 0.1097,
+      "step": 1105
+    },
+    {
+      "epoch": 0.5788370874803562,
+      "loss_partial": 0.0556640625,
+      "step": 1105
+    },
+    {
+      "epoch": 0.5788370874803562,
+      "loss_partial": 0.05029296875,
+      "step": 1105
+    },
+    {
+      "epoch": 0.5793609219486642,
+      "grad_norm": 2.6641314029693604,
+      "learning_rate": 1.987873582714225e-05,
+      "loss": 0.1084,
+      "step": 1106
+    },
+    {
+      "epoch": 0.5793609219486642,
+      "loss_partial": 0.049072265625,
+      "step": 1106
+    },
+    {
+      "epoch": 0.5793609219486642,
+      "loss_partial": 0.0595703125,
+      "step": 1106
+    },
+    {
+      "epoch": 0.5798847564169722,
+      "grad_norm": 2.2866764068603516,
+      "learning_rate": 1.9878025008986535e-05,
+      "loss": 0.1134,
+      "step": 1107
+    },
+    {
+      "epoch": 0.5798847564169722,
+      "loss_partial": 0.061767578125,
+      "step": 1107
+    },
+    {
+      "epoch": 0.5798847564169722,
+      "loss_partial": 0.0537109375,
+      "step": 1107
+    },
+    {
+      "epoch": 0.5804085908852803,
+      "grad_norm": 2.7046594619750977,
+      "learning_rate": 1.987731212638468e-05,
+      "loss": 0.1131,
+      "step": 1108
+    },
+    {
+      "epoch": 0.5804085908852803,
+      "loss_partial": 0.061279296875,
+      "step": 1108
+    },
+    {
+      "epoch": 0.5804085908852803,
+      "loss_partial": 0.052490234375,
+      "step": 1108
+    },
+    {
+      "epoch": 0.5809324253535882,
+      "grad_norm": 1.6873124837875366,
+      "learning_rate": 1.987659717948567e-05,
+      "loss": 0.11,
+      "step": 1109
+    },
+    {
+      "epoch": 0.5809324253535882,
+      "loss_partial": 0.053466796875,
+      "step": 1109
+    },
+    {
+      "epoch": 0.5809324253535882,
+      "loss_partial": 0.06103515625,
+      "step": 1109
+    },
+    {
+      "epoch": 0.5814562598218963,
+      "grad_norm": 2.228196859359741,
+      "learning_rate": 1.987588016843892e-05,
+      "loss": 0.1131,
+      "step": 1110
+    },
+    {
+      "epoch": 0.5814562598218963,
+      "loss_partial": 0.048828125,
+      "step": 1110
+    },
+    {
+      "epoch": 0.5814562598218963,
+      "loss_partial": 0.0625,
+      "step": 1110
+    },
+    {
+      "epoch": 0.5819800942902043,
+      "grad_norm": 2.802778720855713,
+      "learning_rate": 1.987516109339429e-05,
+      "loss": 0.1117,
+      "step": 1111
+    },
+    {
+      "epoch": 0.5819800942902043,
+      "loss_partial": 0.05322265625,
+      "step": 1111
+    },
+    {
+      "epoch": 0.5819800942902043,
+      "loss_partial": 0.05322265625,
+      "step": 1111
+    },
+    {
+      "epoch": 0.5825039287585123,
+      "grad_norm": 2.2583975791931152,
+      "learning_rate": 1.9874439954502052e-05,
+      "loss": 0.1107,
+      "step": 1112
+    },
+    {
+      "epoch": 0.5825039287585123,
+      "loss_partial": 0.0546875,
+      "step": 1112
+    },
+    {
+      "epoch": 0.5825039287585123,
+      "loss_partial": 0.052978515625,
+      "step": 1112
+    },
+    {
+      "epoch": 0.5830277632268204,
+      "grad_norm": 2.6579902172088623,
+      "learning_rate": 1.9873716751912932e-05,
+      "loss": 0.1138,
+      "step": 1113
+    },
+    {
+      "epoch": 0.5830277632268204,
+      "loss_partial": 0.047607421875,
+      "step": 1113
+    },
+    {
+      "epoch": 0.5830277632268204,
+      "loss_partial": 0.0537109375,
+      "step": 1113
+    },
+    {
+      "epoch": 0.5835515976951283,
+      "grad_norm": 2.1153881549835205,
+      "learning_rate": 1.9872991485778064e-05,
+      "loss": 0.1111,
+      "step": 1114
+    },
+    {
+      "epoch": 0.5835515976951283,
+      "loss_partial": 0.057373046875,
+      "step": 1114
+    },
+    {
+      "epoch": 0.5835515976951283,
+      "loss_partial": 0.06787109375,
+      "step": 1114
+    },
+    {
+      "epoch": 0.5840754321634364,
+      "grad_norm": 1.7338578701019287,
+      "learning_rate": 1.9872264156249028e-05,
+      "loss": 0.1116,
+      "step": 1115
+    },
+    {
+      "epoch": 0.5840754321634364,
+      "loss_partial": 0.05126953125,
+      "step": 1115
+    },
+    {
+      "epoch": 0.5840754321634364,
+      "loss_partial": 0.0546875,
+      "step": 1115
+    },
+    {
+      "epoch": 0.5845992666317443,
+      "grad_norm": 2.5150744915008545,
+      "learning_rate": 1.9871534763477834e-05,
+      "loss": 0.1115,
+      "step": 1116
+    },
+    {
+      "epoch": 0.5845992666317443,
+      "loss_partial": 0.0546875,
+      "step": 1116
+    },
+    {
+      "epoch": 0.5845992666317443,
+      "loss_partial": 0.064453125,
+      "step": 1116
+    },
+    {
+      "epoch": 0.5851231011000524,
+      "grad_norm": 2.012875556945801,
+      "learning_rate": 1.9870803307616916e-05,
+      "loss": 0.1084,
+      "step": 1117
+    },
+    {
+      "epoch": 0.5851231011000524,
+      "loss_partial": 0.0537109375,
+      "step": 1117
+    },
+    {
+      "epoch": 0.5851231011000524,
+      "loss_partial": 0.050537109375,
+      "step": 1117
+    },
+    {
+      "epoch": 0.5856469355683604,
+      "grad_norm": 1.5451334714889526,
+      "learning_rate": 1.9870069788819148e-05,
+      "loss": 0.1156,
+      "step": 1118
+    },
+    {
+      "epoch": 0.5856469355683604,
+      "loss_partial": 0.054443359375,
+      "step": 1118
+    },
+    {
+      "epoch": 0.5856469355683604,
+      "loss_partial": 0.0458984375,
+      "step": 1118
+    },
+    {
+      "epoch": 0.5861707700366684,
+      "grad_norm": 2.12459397315979,
+      "learning_rate": 1.9869334207237828e-05,
+      "loss": 0.1144,
+      "step": 1119
+    },
+    {
+      "epoch": 0.5861707700366684,
+      "loss_partial": 0.052978515625,
+      "step": 1119
+    },
+    {
+      "epoch": 0.5861707700366684,
+      "loss_partial": 0.050048828125,
+      "step": 1119
+    },
+    {
+      "epoch": 0.5866946045049765,
+      "grad_norm": 1.979750633239746,
+      "learning_rate": 1.9868596563026687e-05,
+      "loss": 0.1134,
+      "step": 1120
+    },
+    {
+      "epoch": 0.5866946045049765,
+      "loss_partial": 0.0498046875,
+      "step": 1120
+    },
+    {
+      "epoch": 0.5866946045049765,
+      "loss_partial": 0.06591796875,
+      "step": 1120
+    },
+    {
+      "epoch": 0.5872184389732844,
+      "grad_norm": 1.960740566253662,
+      "learning_rate": 1.986785685633989e-05,
+      "loss": 0.1161,
+      "step": 1121
+    },
+    {
+      "epoch": 0.5872184389732844,
+      "loss_partial": 0.05859375,
+      "step": 1121
+    },
+    {
+      "epoch": 0.5872184389732844,
+      "loss_partial": 0.050537109375,
+      "step": 1121
+    },
+    {
+      "epoch": 0.5877422734415925,
+      "grad_norm": 1.8577007055282593,
+      "learning_rate": 1.9867115087332034e-05,
+      "loss": 0.1158,
+      "step": 1122
+    },
+    {
+      "epoch": 0.5877422734415925,
+      "loss_partial": 0.058349609375,
+      "step": 1122
+    },
+    {
+      "epoch": 0.5877422734415925,
+      "loss_partial": 0.061767578125,
+      "step": 1122
+    },
+    {
+      "epoch": 0.5882661079099005,
+      "grad_norm": 1.933915138244629,
+      "learning_rate": 1.986637125615814e-05,
+      "loss": 0.1127,
+      "step": 1123
+    },
+    {
+      "epoch": 0.5882661079099005,
+      "loss_partial": 0.052734375,
+      "step": 1123
+    },
+    {
+      "epoch": 0.5882661079099005,
+      "loss_partial": 0.052978515625,
+      "step": 1123
+    },
+    {
+      "epoch": 0.5887899423782085,
+      "grad_norm": 1.845471739768982,
+      "learning_rate": 1.9865625362973666e-05,
+      "loss": 0.1083,
+      "step": 1124
+    },
+    {
+      "epoch": 0.5887899423782085,
+      "loss_partial": 0.051025390625,
+      "step": 1124
+    },
+    {
+      "epoch": 0.5887899423782085,
+      "loss_partial": 0.05517578125,
+      "step": 1124
+    },
+    {
+      "epoch": 0.5893137768465165,
+      "grad_norm": 1.3821734189987183,
+      "learning_rate": 1.9864877407934493e-05,
+      "loss": 0.1093,
+      "step": 1125
+    },
+    {
+      "epoch": 0.5893137768465165,
+      "loss_partial": 0.058349609375,
+      "step": 1125
+    },
+    {
+      "epoch": 0.5893137768465165,
+      "loss_partial": 0.064453125,
+      "step": 1125
+    },
+    {
+      "epoch": 0.5898376113148245,
+      "grad_norm": 1.2540818452835083,
+      "learning_rate": 1.986412739119695e-05,
+      "loss": 0.1112,
+      "step": 1126
+    },
+    {
+      "epoch": 0.5898376113148245,
+      "loss_partial": 0.06103515625,
+      "step": 1126
+    },
+    {
+      "epoch": 0.5898376113148245,
+      "loss_partial": 0.05517578125,
+      "step": 1126
+    },
+    {
+      "epoch": 0.5903614457831325,
+      "grad_norm": 1.774559497833252,
+      "learning_rate": 1.9863375312917778e-05,
+      "loss": 0.1108,
+      "step": 1127
+    },
+    {
+      "epoch": 0.5903614457831325,
+      "loss_partial": 0.061279296875,
+      "step": 1127
+    },
+    {
+      "epoch": 0.5903614457831325,
+      "loss_partial": 0.06640625,
+      "step": 1127
+    },
+    {
+      "epoch": 0.5908852802514405,
+      "grad_norm": 1.7104103565216064,
+      "learning_rate": 1.9862621173254158e-05,
+      "loss": 0.1185,
+      "step": 1128
+    },
+    {
+      "epoch": 0.5908852802514405,
+      "loss_partial": 0.048583984375,
+      "step": 1128
+    },
+    {
+      "epoch": 0.5908852802514405,
+      "loss_partial": 0.0693359375,
+      "step": 1128
+    },
+    {
+      "epoch": 0.5914091147197486,
+      "grad_norm": 1.637734293937683,
+      "learning_rate": 1.9861864972363705e-05,
+      "loss": 0.1145,
+      "step": 1129
+    },
+    {
+      "epoch": 0.5914091147197486,
+      "loss_partial": 0.050048828125,
+      "step": 1129
+    },
+    {
+      "epoch": 0.5914091147197486,
+      "loss_partial": 0.053466796875,
+      "step": 1129
+    },
+    {
+      "epoch": 0.5919329491880566,
+      "grad_norm": 2.0833301544189453,
+      "learning_rate": 1.986110671040445e-05,
+      "loss": 0.1106,
+      "step": 1130
+    },
+    {
+      "epoch": 0.5919329491880566,
+      "loss_partial": 0.06201171875,
+      "step": 1130
+    },
+    {
+      "epoch": 0.5919329491880566,
+      "loss_partial": 0.056884765625,
+      "step": 1130
+    },
+    {
+      "epoch": 0.5924567836563646,
+      "grad_norm": 2.5516743659973145,
+      "learning_rate": 1.986034638753488e-05,
+      "loss": 0.1125,
+      "step": 1131
+    },
+    {
+      "epoch": 0.5924567836563646,
+      "loss_partial": 0.053955078125,
+      "step": 1131
+    },
+    {
+      "epoch": 0.5924567836563646,
+      "loss_partial": 0.06103515625,
+      "step": 1131
+    },
+    {
+      "epoch": 0.5929806181246726,
+      "grad_norm": 1.5582430362701416,
+      "learning_rate": 1.9859584003913883e-05,
+      "loss": 0.1102,
+      "step": 1132
+    },
+    {
+      "epoch": 0.5929806181246726,
+      "loss_partial": 0.060302734375,
+      "step": 1132
+    },
+    {
+      "epoch": 0.5929806181246726,
+      "loss_partial": 0.05517578125,
+      "step": 1132
+    },
+    {
+      "epoch": 0.5935044525929806,
+      "grad_norm": 2.2994425296783447,
+      "learning_rate": 1.9858819559700804e-05,
+      "loss": 0.1143,
+      "step": 1133
+    },
+    {
+      "epoch": 0.5935044525929806,
+      "loss_partial": 0.050048828125,
+      "step": 1133
+    },
+    {
+      "epoch": 0.5935044525929806,
+      "loss_partial": 0.0537109375,
+      "step": 1133
+    },
+    {
+      "epoch": 0.5940282870612886,
+      "grad_norm": 1.8871008157730103,
+      "learning_rate": 1.98580530550554e-05,
+      "loss": 0.11,
+      "step": 1134
+    },
+    {
+      "epoch": 0.5940282870612886,
+      "loss_partial": 0.05908203125,
+      "step": 1134
+    },
+    {
+      "epoch": 0.5940282870612886,
+      "loss_partial": 0.05859375,
+      "step": 1134
+    },
+    {
+      "epoch": 0.5945521215295967,
+      "grad_norm": 1.9845936298370361,
+      "learning_rate": 1.9857284490137868e-05,
+      "loss": 0.1096,
+      "step": 1135
+    },
+    {
+      "epoch": 0.5945521215295967,
+      "loss_partial": 0.050537109375,
+      "step": 1135
+    },
+    {
+      "epoch": 0.5945521215295967,
+      "loss_partial": 0.051025390625,
+      "step": 1135
+    },
+    {
+      "epoch": 0.5950759559979046,
+      "grad_norm": 2.259840965270996,
+      "learning_rate": 1.9856513865108835e-05,
+      "loss": 0.1072,
+      "step": 1136
+    },
+    {
+      "epoch": 0.5950759559979046,
+      "loss_partial": 0.0625,
+      "step": 1136
+    },
+    {
+      "epoch": 0.5950759559979046,
+      "loss_partial": 0.05322265625,
+      "step": 1136
+    },
+    {
+      "epoch": 0.5955997904662127,
+      "grad_norm": 1.3072491884231567,
+      "learning_rate": 1.9855741180129348e-05,
+      "loss": 0.1138,
+      "step": 1137
+    },
+    {
+      "epoch": 0.5955997904662127,
+      "loss_partial": 0.06005859375,
+      "step": 1137
+    },
+    {
+      "epoch": 0.5955997904662127,
+      "loss_partial": 0.059326171875,
+      "step": 1137
+    },
+    {
+      "epoch": 0.5961236249345206,
+      "grad_norm": 2.2081234455108643,
+      "learning_rate": 1.9854966435360907e-05,
+      "loss": 0.1126,
+      "step": 1138
+    },
+    {
+      "epoch": 0.5961236249345206,
+      "loss_partial": 0.055908203125,
+      "step": 1138
+    },
+    {
+      "epoch": 0.5961236249345206,
+      "loss_partial": 0.05078125,
+      "step": 1138
+    },
+    {
+      "epoch": 0.5966474594028287,
+      "grad_norm": 2.250922918319702,
+      "learning_rate": 1.985418963096542e-05,
+      "loss": 0.1107,
+      "step": 1139
+    },
+    {
+      "epoch": 0.5966474594028287,
+      "loss_partial": 0.056396484375,
+      "step": 1139
+    },
+    {
+      "epoch": 0.5966474594028287,
+      "loss_partial": 0.056396484375,
+      "step": 1139
+    },
+    {
+      "epoch": 0.5971712938711368,
+      "grad_norm": 1.9821327924728394,
+      "learning_rate": 1.985341076710524e-05,
+      "loss": 0.1096,
+      "step": 1140
+    },
+    {
+      "epoch": 0.5971712938711368,
+      "loss_partial": 0.054931640625,
+      "step": 1140
+    },
+    {
+      "epoch": 0.5971712938711368,
+      "loss_partial": 0.0498046875,
+      "step": 1140
+    },
+    {
+      "epoch": 0.5976951283394447,
+      "grad_norm": 2.0749971866607666,
+      "learning_rate": 1.9852629843943137e-05,
+      "loss": 0.109,
+      "step": 1141
+    },
+    {
+      "epoch": 0.5976951283394447,
+      "loss_partial": 0.055908203125,
+      "step": 1141
+    },
+    {
+      "epoch": 0.5976951283394447,
+      "loss_partial": 0.0537109375,
+      "step": 1141
+    },
+    {
+      "epoch": 0.5982189628077528,
+      "grad_norm": 2.0314812660217285,
+      "learning_rate": 1.9851846861642325e-05,
+      "loss": 0.1078,
+      "step": 1142
+    },
+    {
+      "epoch": 0.5982189628077528,
+      "loss_partial": 0.056884765625,
+      "step": 1142
+    },
+    {
+      "epoch": 0.5982189628077528,
+      "loss_partial": 0.053466796875,
+      "step": 1142
+    },
+    {
+      "epoch": 0.5987427972760607,
+      "grad_norm": 2.044513463973999,
+      "learning_rate": 1.9851061820366443e-05,
+      "loss": 0.109,
+      "step": 1143
+    },
+    {
+      "epoch": 0.5987427972760607,
+      "loss_partial": 0.05810546875,
+      "step": 1143
+    },
+    {
+      "epoch": 0.5987427972760607,
+      "loss_partial": 0.048828125,
+      "step": 1143
+    },
+    {
+      "epoch": 0.5992666317443688,
+      "grad_norm": 1.9740843772888184,
+      "learning_rate": 1.9850274720279553e-05,
+      "loss": 0.1131,
+      "step": 1144
+    },
+    {
+      "epoch": 0.5992666317443688,
+      "loss_partial": 0.057861328125,
+      "step": 1144
+    },
+    {
+      "epoch": 0.5992666317443688,
+      "loss_partial": 0.060791015625,
+      "step": 1144
+    },
+    {
+      "epoch": 0.5997904662126768,
+      "grad_norm": 2.558269500732422,
+      "learning_rate": 1.984948556154616e-05,
+      "loss": 0.1098,
+      "step": 1145
+    },
+    {
+      "epoch": 0.5997904662126768,
+      "loss_partial": 0.0634765625,
+      "step": 1145
+    },
+    {
+      "epoch": 0.5997904662126768,
+      "loss_partial": 0.059814453125,
+      "step": 1145
+    },
+    {
+      "epoch": 0.6003143006809848,
+      "grad_norm": 3.134458065032959,
+      "learning_rate": 1.9848694344331198e-05,
+      "loss": 0.1147,
+      "step": 1146
+    },
+    {
+      "epoch": 0.6003143006809848,
+      "loss_partial": 0.0625,
+      "step": 1146
+    },
+    {
+      "epoch": 0.6003143006809848,
+      "loss_partial": 0.05908203125,
+      "step": 1146
+    },
+    {
+      "epoch": 0.6008381351492929,
+      "grad_norm": 3.454909563064575,
+      "learning_rate": 1.9847901068800014e-05,
+      "loss": 0.1174,
+      "step": 1147
+    },
+    {
+      "epoch": 0.6008381351492929,
+      "loss_partial": 0.06201171875,
+      "step": 1147
+    },
+    {
+      "epoch": 0.6008381351492929,
+      "loss_partial": 0.05615234375,
+      "step": 1147
+    },
+    {
+      "epoch": 0.6013619696176008,
+      "grad_norm": 3.6273255348205566,
+      "learning_rate": 1.9847105735118406e-05,
+      "loss": 0.1155,
+      "step": 1148
+    },
+    {
+      "epoch": 0.6013619696176008,
+      "loss_partial": 0.053955078125,
+      "step": 1148
+    },
+    {
+      "epoch": 0.6013619696176008,
+      "loss_partial": 0.0556640625,
+      "step": 1148
+    },
+    {
+      "epoch": 0.6018858040859089,
+      "grad_norm": 2.34250545501709,
+      "learning_rate": 1.984630834345259e-05,
+      "loss": 0.1088,
+      "step": 1149
+    },
+    {
+      "epoch": 0.6018858040859089,
+      "loss_partial": 0.058837890625,
+      "step": 1149
+    },
+    {
+      "epoch": 0.6018858040859089,
+      "loss_partial": 0.05712890625,
+      "step": 1149
+    },
+    {
+      "epoch": 0.6024096385542169,
+      "grad_norm": 2.5193495750427246,
+      "learning_rate": 1.984550889396922e-05,
+      "loss": 0.1151,
+      "step": 1150
+    },
+    {
+      "epoch": 0.6024096385542169,
+      "loss_partial": 0.053955078125,
+      "step": 1150
+    },
+    {
+      "epoch": 0.6024096385542169,
+      "loss_partial": 0.049072265625,
+      "step": 1150
+    },
+    {
+      "epoch": 0.6029334730225249,
+      "grad_norm": 1.7890493869781494,
+      "learning_rate": 1.9844707386835375e-05,
+      "loss": 0.1154,
+      "step": 1151
+    },
+    {
+      "epoch": 0.6029334730225249,
+      "loss_partial": 0.05224609375,
+      "step": 1151
+    },
+    {
+      "epoch": 0.6029334730225249,
+      "loss_partial": 0.057861328125,
+      "step": 1151
+    },
+    {
+      "epoch": 0.6034573074908329,
+      "grad_norm": 2.679840087890625,
+      "learning_rate": 1.9843903822218565e-05,
+      "loss": 0.1109,
+      "step": 1152
+    },
+    {
+      "epoch": 0.6034573074908329,
+      "loss_partial": 0.06591796875,
+      "step": 1152
+    },
+    {
+      "epoch": 0.6034573074908329,
+      "loss_partial": 0.0595703125,
+      "step": 1152
+    },
+    {
+      "epoch": 0.6039811419591409,
+      "grad_norm": 2.2769131660461426,
+      "learning_rate": 1.9843098200286725e-05,
+      "loss": 0.1149,
+      "step": 1153
+    },
+    {
+      "epoch": 0.6039811419591409,
+      "loss_partial": 0.049072265625,
+      "step": 1153
+    },
+    {
+      "epoch": 0.6039811419591409,
+      "loss_partial": 0.0615234375,
+      "step": 1153
+    },
+    {
+      "epoch": 0.6045049764274489,
+      "grad_norm": 1.9441603422164917,
+      "learning_rate": 1.9842290521208233e-05,
+      "loss": 0.1076,
+      "step": 1154
+    },
+    {
+      "epoch": 0.6045049764274489,
+      "loss_partial": 0.055908203125,
+      "step": 1154
+    },
+    {
+      "epoch": 0.6045049764274489,
+      "loss_partial": 0.054443359375,
+      "step": 1154
+    },
+    {
+      "epoch": 0.6050288108957569,
+      "grad_norm": 1.6082491874694824,
+      "learning_rate": 1.984148078515188e-05,
+      "loss": 0.1122,
+      "step": 1155
+    },
+    {
+      "epoch": 0.6050288108957569,
+      "loss_partial": 0.05078125,
+      "step": 1155
+    },
+    {
+      "epoch": 0.6050288108957569,
+      "loss_partial": 0.050048828125,
+      "step": 1155
+    },
+    {
+      "epoch": 0.605552645364065,
+      "grad_norm": 1.6452765464782715,
+      "learning_rate": 1.9840668992286908e-05,
+      "loss": 0.1071,
+      "step": 1156
+    },
+    {
+      "epoch": 0.605552645364065,
+      "loss_partial": 0.053955078125,
+      "step": 1156
+    },
+    {
+      "epoch": 0.605552645364065,
+      "loss_partial": 0.05712890625,
+      "step": 1156
+    },
+    {
+      "epoch": 0.606076479832373,
+      "grad_norm": 2.4672820568084717,
+      "learning_rate": 1.983985514278296e-05,
+      "loss": 0.1114,
+      "step": 1157
+    },
+    {
+      "epoch": 0.606076479832373,
+      "loss_partial": 0.06591796875,
+      "step": 1157
+    },
+    {
+      "epoch": 0.606076479832373,
+      "loss_partial": 0.0517578125,
+      "step": 1157
+    },
+    {
+      "epoch": 0.606600314300681,
+      "grad_norm": 1.8816066980361938,
+      "learning_rate": 1.9839039236810142e-05,
+      "loss": 0.1113,
+      "step": 1158
+    },
+    {
+      "epoch": 0.606600314300681,
+      "loss_partial": 0.05810546875,
+      "step": 1158
+    },
+    {
+      "epoch": 0.606600314300681,
+      "loss_partial": 0.064453125,
+      "step": 1158
+    },
+    {
+      "epoch": 0.607124148768989,
+      "grad_norm": 1.7380952835083008,
+      "learning_rate": 1.9838221274538964e-05,
+      "loss": 0.1154,
+      "step": 1159
+    },
+    {
+      "epoch": 0.607124148768989,
+      "loss_partial": 0.053466796875,
+      "step": 1159
+    },
+    {
+      "epoch": 0.607124148768989,
+      "loss_partial": 0.0576171875,
+      "step": 1159
+    },
+    {
+      "epoch": 0.607647983237297,
+      "grad_norm": 1.8192715644836426,
+      "learning_rate": 1.983740125614038e-05,
+      "loss": 0.1145,
+      "step": 1160
+    },
+    {
+      "epoch": 0.607647983237297,
+      "loss_partial": 0.051513671875,
+      "step": 1160
+    },
+    {
+      "epoch": 0.607647983237297,
+      "loss_partial": 0.056396484375,
+      "step": 1160
+    },
+    {
+      "epoch": 0.608171817705605,
+      "grad_norm": 1.865319848060608,
+      "learning_rate": 1.9836579181785764e-05,
+      "loss": 0.1111,
+      "step": 1161
+    },
+    {
+      "epoch": 0.608171817705605,
+      "loss_partial": 0.0458984375,
+      "step": 1161
+    },
+    {
+      "epoch": 0.608171817705605,
+      "loss_partial": 0.064453125,
+      "step": 1161
+    },
+    {
+      "epoch": 0.6086956521739131,
+      "grad_norm": 1.8613146543502808,
+      "learning_rate": 1.9835755051646927e-05,
+      "loss": 0.1127,
+      "step": 1162
+    },
+    {
+      "epoch": 0.6086956521739131,
+      "loss_partial": 0.0634765625,
+      "step": 1162
+    },
+    {
+      "epoch": 0.6086956521739131,
+      "loss_partial": 0.0556640625,
+      "step": 1162
+    },
+    {
+      "epoch": 0.609219486642221,
+      "grad_norm": 1.727420687675476,
+      "learning_rate": 1.9834928865896105e-05,
+      "loss": 0.11,
+      "step": 1163
+    },
+    {
+      "epoch": 0.609219486642221,
+      "loss_partial": 0.0625,
+      "step": 1163
+    },
+    {
+      "epoch": 0.609219486642221,
+      "loss_partial": 0.053955078125,
+      "step": 1163
+    },
+    {
+      "epoch": 0.6097433211105291,
+      "grad_norm": 1.6994582414627075,
+      "learning_rate": 1.983410062470597e-05,
+      "loss": 0.1106,
+      "step": 1164
+    },
+    {
+      "epoch": 0.6097433211105291,
+      "loss_partial": 0.05224609375,
+      "step": 1164
+    },
+    {
+      "epoch": 0.6097433211105291,
+      "loss_partial": 0.05224609375,
+      "step": 1164
+    },
+    {
+      "epoch": 0.610267155578837,
+      "grad_norm": 1.4161678552627563,
+      "learning_rate": 1.9833270328249614e-05,
+      "loss": 0.1146,
+      "step": 1165
+    },
+    {
+      "epoch": 0.610267155578837,
+      "loss_partial": 0.056884765625,
+      "step": 1165
+    },
+    {
+      "epoch": 0.610267155578837,
+      "loss_partial": 0.064453125,
+      "step": 1165
+    },
+    {
+      "epoch": 0.6107909900471451,
+      "grad_norm": 1.9230269193649292,
+      "learning_rate": 1.9832437976700567e-05,
+      "loss": 0.1122,
+      "step": 1166
+    },
+    {
+      "epoch": 0.6107909900471451,
+      "loss_partial": 0.064453125,
+      "step": 1166
+    },
+    {
+      "epoch": 0.6107909900471451,
+      "loss_partial": 0.057373046875,
+      "step": 1166
+    },
+    {
+      "epoch": 0.6113148245154532,
+      "grad_norm": 1.9662905931472778,
+      "learning_rate": 1.983160357023279e-05,
+      "loss": 0.111,
+      "step": 1167
+    },
+    {
+      "epoch": 0.6113148245154532,
+      "loss_partial": 0.051513671875,
+      "step": 1167
+    },
+    {
+      "epoch": 0.6113148245154532,
+      "loss_partial": 0.05419921875,
+      "step": 1167
+    },
+    {
+      "epoch": 0.6118386589837611,
+      "grad_norm": 2.0443413257598877,
+      "learning_rate": 1.9830767109020658e-05,
+      "loss": 0.1083,
+      "step": 1168
+    },
+    {
+      "epoch": 0.6118386589837611,
+      "loss_partial": 0.0595703125,
+      "step": 1168
+    },
+    {
+      "epoch": 0.6118386589837611,
+      "loss_partial": 0.06103515625,
+      "step": 1168
+    },
+    {
+      "epoch": 0.6123624934520692,
+      "grad_norm": 1.7708910703659058,
+      "learning_rate": 1.982992859323899e-05,
+      "loss": 0.1111,
+      "step": 1169
+    },
+    {
+      "epoch": 0.6123624934520692,
+      "loss_partial": 0.0576171875,
+      "step": 1169
+    },
+    {
+      "epoch": 0.6123624934520692,
+      "loss_partial": 0.056640625,
+      "step": 1169
+    },
+    {
+      "epoch": 0.6128863279203771,
+      "grad_norm": 1.5575945377349854,
+      "learning_rate": 1.982908802306304e-05,
+      "loss": 0.1115,
+      "step": 1170
+    },
+    {
+      "epoch": 0.6128863279203771,
+      "loss_partial": 0.05078125,
+      "step": 1170
+    },
+    {
+      "epoch": 0.6128863279203771,
+      "loss_partial": 0.051513671875,
+      "step": 1170
+    },
+    {
+      "epoch": 0.6134101623886852,
+      "grad_norm": 1.5587291717529297,
+      "learning_rate": 1.9828245398668468e-05,
+      "loss": 0.1138,
+      "step": 1171
+    },
+    {
+      "epoch": 0.6134101623886852,
+      "loss_partial": 0.053466796875,
+      "step": 1171
+    },
+    {
+      "epoch": 0.6134101623886852,
+      "loss_partial": 0.05029296875,
+      "step": 1171
+    },
+    {
+      "epoch": 0.6139339968569932,
+      "grad_norm": 1.899531364440918,
+      "learning_rate": 1.9827400720231386e-05,
+      "loss": 0.1102,
+      "step": 1172
+    },
+    {
+      "epoch": 0.6139339968569932,
+      "loss_partial": 0.052978515625,
+      "step": 1172
+    },
+    {
+      "epoch": 0.6139339968569932,
+      "loss_partial": 0.056884765625,
+      "step": 1172
+    },
+    {
+      "epoch": 0.6144578313253012,
+      "grad_norm": 2.3804807662963867,
+      "learning_rate": 1.9826553987928325e-05,
+      "loss": 0.1148,
+      "step": 1173
+    },
+    {
+      "epoch": 0.6144578313253012,
+      "loss_partial": 0.061279296875,
+      "step": 1173
+    },
+    {
+      "epoch": 0.6144578313253012,
+      "loss_partial": 0.04931640625,
+      "step": 1173
+    },
+    {
+      "epoch": 0.6149816657936092,
+      "grad_norm": 1.4005321264266968,
+      "learning_rate": 1.9825705201936242e-05,
+      "loss": 0.1121,
+      "step": 1174
+    },
+    {
+      "epoch": 0.6149816657936092,
+      "loss_partial": 0.052001953125,
+      "step": 1174
+    },
+    {
+      "epoch": 0.6149816657936092,
+      "loss_partial": 0.052978515625,
+      "step": 1174
+    },
+    {
+      "epoch": 0.6155055002619172,
+      "grad_norm": 1.925033450126648,
+      "learning_rate": 1.9824854362432538e-05,
+      "loss": 0.112,
+      "step": 1175
+    },
+    {
+      "epoch": 0.6155055002619172,
+      "loss_partial": 0.057373046875,
+      "step": 1175
+    },
+    {
+      "epoch": 0.6155055002619172,
+      "loss_partial": 0.06494140625,
+      "step": 1175
+    },
+    {
+      "epoch": 0.6160293347302253,
+      "grad_norm": 1.5526385307312012,
+      "learning_rate": 1.9824001469595023e-05,
+      "loss": 0.1109,
+      "step": 1176
+    },
+    {
+      "epoch": 0.6160293347302253,
+      "loss_partial": 0.057373046875,
+      "step": 1176
+    },
+    {
+      "epoch": 0.6160293347302253,
+      "loss_partial": 0.0634765625,
+      "step": 1176
+    },
+    {
+      "epoch": 0.6165531691985333,
+      "grad_norm": 2.6413938999176025,
+      "learning_rate": 1.9823146523601954e-05,
+      "loss": 0.115,
+      "step": 1177
+    },
+    {
+      "epoch": 0.6165531691985333,
+      "loss_partial": 0.0517578125,
+      "step": 1177
+    },
+    {
+      "epoch": 0.6165531691985333,
+      "loss_partial": 0.0615234375,
+      "step": 1177
+    },
+    {
+      "epoch": 0.6170770036668413,
+      "grad_norm": 1.3273831605911255,
+      "learning_rate": 1.9822289524632002e-05,
+      "loss": 0.1114,
+      "step": 1178
+    },
+    {
+      "epoch": 0.6170770036668413,
+      "loss_partial": 0.04736328125,
+      "step": 1178
+    },
+    {
+      "epoch": 0.6170770036668413,
+      "loss_partial": 0.05419921875,
+      "step": 1178
+    },
+    {
+      "epoch": 0.6176008381351493,
+      "grad_norm": 1.6093419790267944,
+      "learning_rate": 1.9821430472864283e-05,
+      "loss": 0.1105,
+      "step": 1179
+    },
+    {
+      "epoch": 0.6176008381351493,
+      "loss_partial": 0.062255859375,
+      "step": 1179
+    },
+    {
+      "epoch": 0.6176008381351493,
+      "loss_partial": 0.068359375,
+      "step": 1179
+    },
+    {
+      "epoch": 0.6181246726034573,
+      "grad_norm": 1.7014034986495972,
+      "learning_rate": 1.982056936847833e-05,
+      "loss": 0.1129,
+      "step": 1180
+    },
+    {
+      "epoch": 0.6181246726034573,
+      "loss_partial": 0.0546875,
+      "step": 1180
+    },
+    {
+      "epoch": 0.6181246726034573,
+      "loss_partial": 0.052978515625,
+      "step": 1180
+    },
+    {
+      "epoch": 0.6186485070717653,
+      "grad_norm": 1.642912745475769,
+      "learning_rate": 1.98197062116541e-05,
+      "loss": 0.1107,
+      "step": 1181
+    },
+    {
+      "epoch": 0.6186485070717653,
+      "loss_partial": 0.06298828125,
+      "step": 1181
+    },
+    {
+      "epoch": 0.6186485070717653,
+      "loss_partial": 0.0556640625,
+      "step": 1181
+    },
+    {
+      "epoch": 0.6191723415400734,
+      "grad_norm": 2.282761573791504,
+      "learning_rate": 1.9818841002572e-05,
+      "loss": 0.1143,
+      "step": 1182
+    },
+    {
+      "epoch": 0.6191723415400734,
+      "loss_partial": 0.052734375,
+      "step": 1182
+    },
+    {
+      "epoch": 0.6191723415400734,
+      "loss_partial": 0.054931640625,
+      "step": 1182
+    },
+    {
+      "epoch": 0.6196961760083813,
+      "grad_norm": 1.1001726388931274,
+      "learning_rate": 1.981797374141285e-05,
+      "loss": 0.1086,
+      "step": 1183
+    },
+    {
+      "epoch": 0.6196961760083813,
+      "loss_partial": 0.0625,
+      "step": 1183
+    },
+    {
+      "epoch": 0.6196961760083813,
+      "loss_partial": 0.044677734375,
+      "step": 1183
+    },
+    {
+      "epoch": 0.6202200104766894,
+      "grad_norm": 1.4401575326919556,
+      "learning_rate": 1.9817104428357895e-05,
+      "loss": 0.1108,
+      "step": 1184
+    },
+    {
+      "epoch": 0.6202200104766894,
+      "loss_partial": 0.053466796875,
+      "step": 1184
+    },
+    {
+      "epoch": 0.6202200104766894,
+      "loss_partial": 0.04931640625,
+      "step": 1184
+    },
+    {
+      "epoch": 0.6207438449449973,
+      "grad_norm": 1.7119604349136353,
+      "learning_rate": 1.981623306358883e-05,
+      "loss": 0.1105,
+      "step": 1185
+    },
+    {
+      "epoch": 0.6207438449449973,
+      "loss_partial": 0.05224609375,
+      "step": 1185
+    },
+    {
+      "epoch": 0.6207438449449973,
+      "loss_partial": 0.056640625,
+      "step": 1185
+    },
+    {
+      "epoch": 0.6212676794133054,
+      "grad_norm": 1.6953892707824707,
+      "learning_rate": 1.981535964728775e-05,
+      "loss": 0.1123,
+      "step": 1186
+    },
+    {
+      "epoch": 0.6212676794133054,
+      "loss_partial": 0.051025390625,
+      "step": 1186
+    },
+    {
+      "epoch": 0.6212676794133054,
+      "loss_partial": 0.05126953125,
+      "step": 1186
+    },
+    {
+      "epoch": 0.6217915138816134,
+      "grad_norm": 1.866797685623169,
+      "learning_rate": 1.98144841796372e-05,
+      "loss": 0.1071,
+      "step": 1187
+    },
+    {
+      "epoch": 0.6217915138816134,
+      "loss_partial": 0.047607421875,
+      "step": 1187
+    },
+    {
+      "epoch": 0.6217915138816134,
+      "loss_partial": 0.05615234375,
+      "step": 1187
+    },
+    {
+      "epoch": 0.6223153483499214,
+      "grad_norm": 2.6280267238616943,
+      "learning_rate": 1.981360666082015e-05,
+      "loss": 0.1132,
+      "step": 1188
+    },
+    {
+      "epoch": 0.6223153483499214,
+      "loss_partial": 0.059326171875,
+      "step": 1188
+    },
+    {
+      "epoch": 0.6223153483499214,
+      "loss_partial": 0.044677734375,
+      "step": 1188
+    },
+    {
+      "epoch": 0.6228391828182295,
+      "grad_norm": 2.7790708541870117,
+      "learning_rate": 1.9812727091019994e-05,
+      "loss": 0.1117,
+      "step": 1189
+    },
+    {
+      "epoch": 0.6228391828182295,
+      "loss_partial": 0.056396484375,
+      "step": 1189
+    },
+    {
+      "epoch": 0.6228391828182295,
+      "loss_partial": 0.051513671875,
+      "step": 1189
+    },
+    {
+      "epoch": 0.6233630172865374,
+      "grad_norm": 2.689894199371338,
+      "learning_rate": 1.9811845470420552e-05,
+      "loss": 0.1124,
+      "step": 1190
+    },
+    {
+      "epoch": 0.6233630172865374,
+      "loss_partial": 0.059814453125,
+      "step": 1190
+    },
+    {
+      "epoch": 0.6233630172865374,
+      "loss_partial": 0.0478515625,
+      "step": 1190
+    },
+    {
+      "epoch": 0.6238868517548455,
+      "grad_norm": 2.227891206741333,
+      "learning_rate": 1.9810961799206086e-05,
+      "loss": 0.1089,
+      "step": 1191
+    },
+    {
+      "epoch": 0.6238868517548455,
+      "loss_partial": 0.060546875,
+      "step": 1191
+    },
+    {
+      "epoch": 0.6238868517548455,
+      "loss_partial": 0.048583984375,
+      "step": 1191
+    },
+    {
+      "epoch": 0.6244106862231534,
+      "grad_norm": 3.7835299968719482,
+      "learning_rate": 1.981007607756127e-05,
+      "loss": 0.1104,
+      "step": 1192
+    },
+    {
+      "epoch": 0.6244106862231534,
+      "loss_partial": 0.061767578125,
+      "step": 1192
+    },
+    {
+      "epoch": 0.6244106862231534,
+      "loss_partial": 0.0546875,
+      "step": 1192
+    },
+    {
+      "epoch": 0.6249345206914615,
+      "grad_norm": 2.8889667987823486,
+      "learning_rate": 1.980918830567122e-05,
+      "loss": 0.1115,
+      "step": 1193
+    },
+    {
+      "epoch": 0.6249345206914615,
+      "loss_partial": 0.056640625,
+      "step": 1193
+    },
+    {
+      "epoch": 0.6249345206914615,
+      "loss_partial": 0.056640625,
+      "step": 1193
+    },
+    {
+      "epoch": 0.6254583551597696,
+      "grad_norm": 1.5734423398971558,
+      "learning_rate": 1.9808298483721467e-05,
+      "loss": 0.1106,
+      "step": 1194
+    },
+    {
+      "epoch": 0.6254583551597696,
+      "loss_partial": 0.052734375,
+      "step": 1194
+    },
+    {
+      "epoch": 0.6254583551597696,
+      "loss_partial": 0.06298828125,
+      "step": 1194
+    },
+    {
+      "epoch": 0.6259821896280775,
+      "grad_norm": 1.9284616708755493,
+      "learning_rate": 1.980740661189799e-05,
+      "loss": 0.1122,
+      "step": 1195
+    },
+    {
+      "epoch": 0.6259821896280775,
+      "loss_partial": 0.052734375,
+      "step": 1195
+    },
+    {
+      "epoch": 0.6259821896280775,
+      "loss_partial": 0.05517578125,
+      "step": 1195
+    },
+    {
+      "epoch": 0.6265060240963856,
+      "grad_norm": 1.6593807935714722,
+      "learning_rate": 1.9806512690387177e-05,
+      "loss": 0.1113,
+      "step": 1196
+    },
+    {
+      "epoch": 0.6265060240963856,
+      "loss_partial": 0.057861328125,
+      "step": 1196
+    },
+    {
+      "epoch": 0.6265060240963856,
+      "loss_partial": 0.06201171875,
+      "step": 1196
+    },
+    {
+      "epoch": 0.6270298585646935,
+      "grad_norm": 1.6714040040969849,
+      "learning_rate": 1.9805616719375852e-05,
+      "loss": 0.1081,
+      "step": 1197
+    },
+    {
+      "epoch": 0.6270298585646935,
+      "loss_partial": 0.0615234375,
+      "step": 1197
+    },
+    {
+      "epoch": 0.6270298585646935,
+      "loss_partial": 0.059326171875,
+      "step": 1197
+    },
+    {
+      "epoch": 0.6275536930330016,
+      "grad_norm": 1.8769581317901611,
+      "learning_rate": 1.980471869905127e-05,
+      "loss": 0.11,
+      "step": 1198
+    },
+    {
+      "epoch": 0.6275536930330016,
+      "loss_partial": 0.06298828125,
+      "step": 1198
+    },
+    {
+      "epoch": 0.6275536930330016,
+      "loss_partial": 0.0615234375,
+      "step": 1198
+    },
+    {
+      "epoch": 0.6280775275013096,
+      "grad_norm": 1.3786115646362305,
+      "learning_rate": 1.9803818629601105e-05,
+      "loss": 0.1119,
+      "step": 1199
+    },
+    {
+      "epoch": 0.6280775275013096,
+      "loss_partial": 0.05908203125,
+      "step": 1199
+    },
+    {
+      "epoch": 0.6280775275013096,
+      "loss_partial": 0.055419921875,
+      "step": 1199
+    },
+    {
+      "epoch": 0.6286013619696176,
+      "grad_norm": 1.6352423429489136,
+      "learning_rate": 1.9802916511213478e-05,
+      "loss": 0.1112,
+      "step": 1200
+    },
+    {
+      "epoch": 0.6286013619696176,
+      "loss_partial": 0.054443359375,
+      "step": 1200
+    },
+    {
+      "epoch": 0.6286013619696176,
+      "loss_partial": 0.064453125,
+      "step": 1200
+    },
+    {
+      "epoch": 0.6291251964379256,
+      "grad_norm": 2.0656988620758057,
+      "learning_rate": 1.9802012344076917e-05,
+      "loss": 0.1162,
+      "step": 1201
+    },
+    {
+      "epoch": 0.6291251964379256,
+      "loss_partial": 0.05908203125,
+      "step": 1201
+    },
+    {
+      "epoch": 0.6291251964379256,
+      "loss_partial": 0.0537109375,
+      "step": 1201
+    },
+    {
+      "epoch": 0.6296490309062336,
+      "grad_norm": 2.3285839557647705,
+      "learning_rate": 1.980110612838039e-05,
+      "loss": 0.1093,
+      "step": 1202
+    },
+    {
+      "epoch": 0.6296490309062336,
+      "loss_partial": 0.05224609375,
+      "step": 1202
+    },
+    {
+      "epoch": 0.6296490309062336,
+      "loss_partial": 0.05517578125,
+      "step": 1202
+    },
+    {
+      "epoch": 0.6301728653745416,
+      "grad_norm": 2.0814898014068604,
+      "learning_rate": 1.980019786431329e-05,
+      "loss": 0.1066,
+      "step": 1203
+    },
+    {
+      "epoch": 0.6301728653745416,
+      "loss_partial": 0.056396484375,
+      "step": 1203
+    },
+    {
+      "epoch": 0.6301728653745416,
+      "loss_partial": 0.056640625,
+      "step": 1203
+    },
+    {
+      "epoch": 0.6306966998428497,
+      "grad_norm": 2.254425287246704,
+      "learning_rate": 1.979928755206544e-05,
+      "loss": 0.1078,
+      "step": 1204
+    },
+    {
+      "epoch": 0.6306966998428497,
+      "loss_partial": 0.05810546875,
+      "step": 1204
+    },
+    {
+      "epoch": 0.6306966998428497,
+      "loss_partial": 0.06103515625,
+      "step": 1204
+    },
+    {
+      "epoch": 0.6312205343111577,
+      "grad_norm": 3.6264147758483887,
+      "learning_rate": 1.979837519182709e-05,
+      "loss": 0.1118,
+      "step": 1205
+    },
+    {
+      "epoch": 0.6312205343111577,
+      "loss_partial": 0.040771484375,
+      "step": 1205
+    },
+    {
+      "epoch": 0.6312205343111577,
+      "loss_partial": 0.055908203125,
+      "step": 1205
+    },
+    {
+      "epoch": 0.6317443687794657,
+      "grad_norm": 3.1475841999053955,
+      "learning_rate": 1.9797460783788912e-05,
+      "loss": 0.1099,
+      "step": 1206
+    },
+    {
+      "epoch": 0.6317443687794657,
+      "loss_partial": 0.05029296875,
+      "step": 1206
+    },
+    {
+      "epoch": 0.6317443687794657,
+      "loss_partial": 0.059814453125,
+      "step": 1206
+    },
+    {
+      "epoch": 0.6322682032477737,
+      "grad_norm": 2.3656556606292725,
+      "learning_rate": 1.9796544328142017e-05,
+      "loss": 0.1118,
+      "step": 1207
+    },
+    {
+      "epoch": 0.6322682032477737,
+      "loss_partial": 0.05859375,
+      "step": 1207
+    },
+    {
+      "epoch": 0.6322682032477737,
+      "loss_partial": 0.06787109375,
+      "step": 1207
+    },
+    {
+      "epoch": 0.6327920377160817,
+      "grad_norm": 3.2136197090148926,
+      "learning_rate": 1.9795625825077937e-05,
+      "loss": 0.1146,
+      "step": 1208
+    },
+    {
+      "epoch": 0.6327920377160817,
+      "loss_partial": 0.054931640625,
+      "step": 1208
+    },
+    {
+      "epoch": 0.6327920377160817,
+      "loss_partial": 0.04833984375,
+      "step": 1208
+    },
+    {
+      "epoch": 0.6333158721843898,
+      "grad_norm": 3.147716999053955,
+      "learning_rate": 1.979470527478864e-05,
+      "loss": 0.1102,
+      "step": 1209
+    },
+    {
+      "epoch": 0.6333158721843898,
+      "loss_partial": 0.055908203125,
+      "step": 1209
+    },
+    {
+      "epoch": 0.6333158721843898,
+      "loss_partial": 0.052978515625,
+      "step": 1209
+    },
+    {
+      "epoch": 0.6338397066526977,
+      "grad_norm": 2.3302268981933594,
+      "learning_rate": 1.9793782677466503e-05,
+      "loss": 0.1087,
+      "step": 1210
+    },
+    {
+      "epoch": 0.6338397066526977,
+      "loss_partial": 0.05078125,
+      "step": 1210
+    },
+    {
+      "epoch": 0.6338397066526977,
+      "loss_partial": 0.047607421875,
+      "step": 1210
+    },
+    {
+      "epoch": 0.6343635411210058,
+      "grad_norm": 2.6347098350524902,
+      "learning_rate": 1.9792858033304353e-05,
+      "loss": 0.113,
+      "step": 1211
+    },
+    {
+      "epoch": 0.6343635411210058,
+      "loss_partial": 0.05712890625,
+      "step": 1211
+    },
+    {
+      "epoch": 0.6343635411210058,
+      "loss_partial": 0.05712890625,
+      "step": 1211
+    },
+    {
+      "epoch": 0.6348873755893137,
+      "grad_norm": 1.608556866645813,
+      "learning_rate": 1.979193134249543e-05,
+      "loss": 0.1164,
+      "step": 1212
+    },
+    {
+      "epoch": 0.6348873755893137,
+      "loss_partial": 0.0556640625,
+      "step": 1212
+    },
+    {
+      "epoch": 0.6348873755893137,
+      "loss_partial": 0.05126953125,
+      "step": 1212
+    },
+    {
+      "epoch": 0.6354112100576218,
+      "grad_norm": 3.0534708499908447,
+      "learning_rate": 1.9791002605233402e-05,
+      "loss": 0.1115,
+      "step": 1213
+    },
+    {
+      "epoch": 0.6354112100576218,
+      "loss_partial": 0.058349609375,
+      "step": 1213
+    },
+    {
+      "epoch": 0.6354112100576218,
+      "loss_partial": 0.059814453125,
+      "step": 1213
+    },
+    {
+      "epoch": 0.6359350445259299,
+      "grad_norm": 2.1022331714630127,
+      "learning_rate": 1.979007182171238e-05,
+      "loss": 0.1173,
+      "step": 1214
+    },
+    {
+      "epoch": 0.6359350445259299,
+      "loss_partial": 0.06591796875,
+      "step": 1214
+    },
+    {
+      "epoch": 0.6359350445259299,
+      "loss_partial": 0.0498046875,
+      "step": 1214
+    },
+    {
+      "epoch": 0.6364588789942378,
+      "grad_norm": 1.7828513383865356,
+      "learning_rate": 1.978913899212689e-05,
+      "loss": 0.1105,
+      "step": 1215
+    },
+    {
+      "epoch": 0.6364588789942378,
+      "loss_partial": 0.057861328125,
+      "step": 1215
+    },
+    {
+      "epoch": 0.6364588789942378,
+      "loss_partial": 0.05029296875,
+      "step": 1215
+    },
+    {
+      "epoch": 0.6369827134625459,
+      "grad_norm": 1.6297591924667358,
+      "learning_rate": 1.9788204116671882e-05,
+      "loss": 0.1117,
+      "step": 1216
+    },
+    {
+      "epoch": 0.6369827134625459,
+      "loss_partial": 0.054931640625,
+      "step": 1216
+    },
+    {
+      "epoch": 0.6369827134625459,
+      "loss_partial": 0.046875,
+      "step": 1216
+    },
+    {
+      "epoch": 0.6375065479308538,
+      "grad_norm": 2.1503522396087646,
+      "learning_rate": 1.9787267195542742e-05,
+      "loss": 0.1075,
+      "step": 1217
+    },
+    {
+      "epoch": 0.6375065479308538,
+      "loss_partial": 0.0595703125,
+      "step": 1217
+    },
+    {
+      "epoch": 0.6375065479308538,
+      "loss_partial": 0.054443359375,
+      "step": 1217
+    },
+    {
+      "epoch": 0.6380303823991619,
+      "grad_norm": 1.4996154308319092,
+      "learning_rate": 1.9786328228935283e-05,
+      "loss": 0.1125,
+      "step": 1218
+    },
+    {
+      "epoch": 0.6380303823991619,
+      "loss_partial": 0.056884765625,
+      "step": 1218
+    },
+    {
+      "epoch": 0.6380303823991619,
+      "loss_partial": 0.06591796875,
+      "step": 1218
+    },
+    {
+      "epoch": 0.6385542168674698,
+      "grad_norm": 2.5990612506866455,
+      "learning_rate": 1.9785387217045738e-05,
+      "loss": 0.1085,
+      "step": 1219
+    },
+    {
+      "epoch": 0.6385542168674698,
+      "loss_partial": 0.05712890625,
+      "step": 1219
+    },
+    {
+      "epoch": 0.6385542168674698,
+      "loss_partial": 0.050537109375,
+      "step": 1219
+    },
+    {
+      "epoch": 0.6390780513357779,
+      "grad_norm": 2.513157844543457,
+      "learning_rate": 1.9784444160070775e-05,
+      "loss": 0.1122,
+      "step": 1220
+    },
+    {
+      "epoch": 0.6390780513357779,
+      "loss_partial": 0.052490234375,
+      "step": 1220
+    },
+    {
+      "epoch": 0.6390780513357779,
+      "loss_partial": 0.05810546875,
+      "step": 1220
+    },
+    {
+      "epoch": 0.639601885804086,
+      "grad_norm": 2.2203428745269775,
+      "learning_rate": 1.9783499058207493e-05,
+      "loss": 0.111,
+      "step": 1221
+    },
+    {
+      "epoch": 0.639601885804086,
+      "loss_partial": 0.04931640625,
+      "step": 1221
+    },
+    {
+      "epoch": 0.639601885804086,
+      "loss_partial": 0.06298828125,
+      "step": 1221
+    },
+    {
+      "epoch": 0.6401257202723939,
+      "grad_norm": 2.337677478790283,
+      "learning_rate": 1.97825519116534e-05,
+      "loss": 0.1104,
+      "step": 1222
+    },
+    {
+      "epoch": 0.6401257202723939,
+      "loss_partial": 0.054443359375,
+      "step": 1222
+    },
+    {
+      "epoch": 0.6401257202723939,
+      "loss_partial": 0.06640625,
+      "step": 1222
+    },
+    {
+      "epoch": 0.640649554740702,
+      "grad_norm": 2.353423833847046,
+      "learning_rate": 1.9781602720606455e-05,
+      "loss": 0.1144,
+      "step": 1223
+    },
+    {
+      "epoch": 0.640649554740702,
+      "loss_partial": 0.05517578125,
+      "step": 1223
+    },
+    {
+      "epoch": 0.640649554740702,
+      "loss_partial": 0.044921875,
+      "step": 1223
+    },
+    {
+      "epoch": 0.6411733892090099,
+      "grad_norm": 2.022855281829834,
+      "learning_rate": 1.978065148526503e-05,
+      "loss": 0.107,
+      "step": 1224
+    },
+    {
+      "epoch": 0.6411733892090099,
+      "loss_partial": 0.047119140625,
+      "step": 1224
+    },
+    {
+      "epoch": 0.6411733892090099,
+      "loss_partial": 0.05517578125,
+      "step": 1224
+    },
+    {
+      "epoch": 0.641697223677318,
+      "grad_norm": 2.279515504837036,
+      "learning_rate": 1.9779698205827926e-05,
+      "loss": 0.1078,
+      "step": 1225
+    },
+    {
+      "epoch": 0.641697223677318,
+      "loss_partial": 0.049560546875,
+      "step": 1225
+    },
+    {
+      "epoch": 0.641697223677318,
+      "loss_partial": 0.05419921875,
+      "step": 1225
+    },
+    {
+      "epoch": 0.642221058145626,
+      "grad_norm": 1.943170189857483,
+      "learning_rate": 1.9778742882494374e-05,
+      "loss": 0.1091,
+      "step": 1226
+    },
+    {
+      "epoch": 0.642221058145626,
+      "loss_partial": 0.048828125,
+      "step": 1226
+    },
+    {
+      "epoch": 0.642221058145626,
+      "loss_partial": 0.057373046875,
+      "step": 1226
+    },
+    {
+      "epoch": 0.642744892613934,
+      "grad_norm": 2.4858853816986084,
+      "learning_rate": 1.9777785515464026e-05,
+      "loss": 0.1078,
+      "step": 1227
+    },
+    {
+      "epoch": 0.642744892613934,
+      "loss_partial": 0.05078125,
+      "step": 1227
+    },
+    {
+      "epoch": 0.642744892613934,
+      "loss_partial": 0.048095703125,
+      "step": 1227
+    },
+    {
+      "epoch": 0.643268727082242,
+      "grad_norm": 1.7941067218780518,
+      "learning_rate": 1.977682610493697e-05,
+      "loss": 0.1111,
+      "step": 1228
+    },
+    {
+      "epoch": 0.643268727082242,
+      "loss_partial": 0.060791015625,
+      "step": 1228
+    },
+    {
+      "epoch": 0.643268727082242,
+      "loss_partial": 0.0517578125,
+      "step": 1228
+    },
+    {
+      "epoch": 0.64379256155055,
+      "grad_norm": 2.6476078033447266,
+      "learning_rate": 1.9775864651113718e-05,
+      "loss": 0.1136,
+      "step": 1229
+    },
+    {
+      "epoch": 0.64379256155055,
+      "loss_partial": 0.053955078125,
+      "step": 1229
+    },
+    {
+      "epoch": 0.64379256155055,
+      "loss_partial": 0.0517578125,
+      "step": 1229
+    },
+    {
+      "epoch": 0.644316396018858,
+      "grad_norm": 2.678647041320801,
+      "learning_rate": 1.9774901154195205e-05,
+      "loss": 0.1109,
+      "step": 1230
+    },
+    {
+      "epoch": 0.644316396018858,
+      "loss_partial": 0.056640625,
+      "step": 1230
+    },
+    {
+      "epoch": 0.644316396018858,
+      "loss_partial": 0.0537109375,
+      "step": 1230
+    },
+    {
+      "epoch": 0.6448402304871661,
+      "grad_norm": 1.9492453336715698,
+      "learning_rate": 1.9773935614382795e-05,
+      "loss": 0.108,
+      "step": 1231
+    },
+    {
+      "epoch": 0.6448402304871661,
+      "loss_partial": 0.046875,
+      "step": 1231
+    },
+    {
+      "epoch": 0.6448402304871661,
+      "loss_partial": 0.049072265625,
+      "step": 1231
+    },
+    {
+      "epoch": 0.645364064955474,
+      "grad_norm": 2.47200345993042,
+      "learning_rate": 1.9772968031878283e-05,
+      "loss": 0.111,
+      "step": 1232
+    },
+    {
+      "epoch": 0.645364064955474,
+      "loss_partial": 0.0537109375,
+      "step": 1232
+    },
+    {
+      "epoch": 0.645364064955474,
+      "loss_partial": 0.06396484375,
+      "step": 1232
+    },
+    {
+      "epoch": 0.6458878994237821,
+      "grad_norm": 1.9653500318527222,
+      "learning_rate": 1.977199840688389e-05,
+      "loss": 0.1075,
+      "step": 1233
+    },
+    {
+      "epoch": 0.6458878994237821,
+      "loss_partial": 0.068359375,
+      "step": 1233
+    },
+    {
+      "epoch": 0.6458878994237821,
+      "loss_partial": 0.056396484375,
+      "step": 1233
+    },
+    {
+      "epoch": 0.6464117338920901,
+      "grad_norm": 2.4873218536376953,
+      "learning_rate": 1.9771026739602256e-05,
+      "loss": 0.1111,
+      "step": 1234
+    },
+    {
+      "epoch": 0.6464117338920901,
+      "loss_partial": 0.048583984375,
+      "step": 1234
+    },
+    {
+      "epoch": 0.6464117338920901,
+      "loss_partial": 0.04638671875,
+      "step": 1234
+    },
+    {
+      "epoch": 0.6469355683603981,
+      "grad_norm": 1.674932837486267,
+      "learning_rate": 1.977005303023645e-05,
+      "loss": 0.1116,
+      "step": 1235
+    },
+    {
+      "epoch": 0.6469355683603981,
+      "loss_partial": 0.05908203125,
+      "step": 1235
+    },
+    {
+      "epoch": 0.6469355683603981,
+      "loss_partial": 0.056396484375,
+      "step": 1235
+    },
+    {
+      "epoch": 0.6474594028287062,
+      "grad_norm": 1.3220069408416748,
+      "learning_rate": 1.9769077278989983e-05,
+      "loss": 0.107,
+      "step": 1236
+    },
+    {
+      "epoch": 0.6474594028287062,
+      "loss_partial": 0.06396484375,
+      "step": 1236
+    },
+    {
+      "epoch": 0.6474594028287062,
+      "loss_partial": 0.058349609375,
+      "step": 1236
+    },
+    {
+      "epoch": 0.6479832372970141,
+      "grad_norm": 1.652143120765686,
+      "learning_rate": 1.9768099486066776e-05,
+      "loss": 0.1133,
+      "step": 1237
+    },
+    {
+      "epoch": 0.6479832372970141,
+      "loss_partial": 0.053955078125,
+      "step": 1237
+    },
+    {
+      "epoch": 0.6479832372970141,
+      "loss_partial": 0.052490234375,
+      "step": 1237
+    },
+    {
+      "epoch": 0.6485070717653222,
+      "grad_norm": 1.8386820554733276,
+      "learning_rate": 1.976711965167118e-05,
+      "loss": 0.1093,
+      "step": 1238
+    },
+    {
+      "epoch": 0.6485070717653222,
+      "loss_partial": 0.05126953125,
+      "step": 1238
+    },
+    {
+      "epoch": 0.6485070717653222,
+      "loss_partial": 0.0478515625,
+      "step": 1238
+    },
+    {
+      "epoch": 0.6490309062336301,
+      "grad_norm": 1.7153189182281494,
+      "learning_rate": 1.9766137776007972e-05,
+      "loss": 0.1117,
+      "step": 1239
+    },
+    {
+      "epoch": 0.6490309062336301,
+      "loss_partial": 0.04931640625,
+      "step": 1239
+    },
+    {
+      "epoch": 0.6490309062336301,
+      "loss_partial": 0.0537109375,
+      "step": 1239
+    },
+    {
+      "epoch": 0.6495547407019382,
+      "grad_norm": 2.3683886528015137,
+      "learning_rate": 1.976515385928236e-05,
+      "loss": 0.1128,
+      "step": 1240
+    },
+    {
+      "epoch": 0.6495547407019382,
+      "loss_partial": 0.06640625,
+      "step": 1240
+    },
+    {
+      "epoch": 0.6495547407019382,
+      "loss_partial": 0.05615234375,
+      "step": 1240
+    },
+    {
+      "epoch": 0.6500785751702463,
+      "grad_norm": 2.976585626602173,
+      "learning_rate": 1.976416790169998e-05,
+      "loss": 0.1106,
+      "step": 1241
+    },
+    {
+      "epoch": 0.6500785751702463,
+      "loss_partial": 0.051513671875,
+      "step": 1241
+    },
+    {
+      "epoch": 0.6500785751702463,
+      "loss_partial": 0.05029296875,
+      "step": 1241
+    },
+    {
+      "epoch": 0.6506024096385542,
+      "grad_norm": 2.334139585494995,
+      "learning_rate": 1.9763179903466887e-05,
+      "loss": 0.1122,
+      "step": 1242
+    },
+    {
+      "epoch": 0.6506024096385542,
+      "loss_partial": 0.048583984375,
+      "step": 1242
+    },
+    {
+      "epoch": 0.6506024096385542,
+      "loss_partial": 0.05908203125,
+      "step": 1242
+    },
+    {
+      "epoch": 0.6511262441068623,
+      "grad_norm": 2.7279891967773438,
+      "learning_rate": 1.976218986478957e-05,
+      "loss": 0.1129,
+      "step": 1243
+    },
+    {
+      "epoch": 0.6511262441068623,
+      "loss_partial": 0.0517578125,
+      "step": 1243
+    },
+    {
+      "epoch": 0.6511262441068623,
+      "loss_partial": 0.051025390625,
+      "step": 1243
+    },
+    {
+      "epoch": 0.6516500785751702,
+      "grad_norm": 3.0185258388519287,
+      "learning_rate": 1.9761197785874934e-05,
+      "loss": 0.1104,
+      "step": 1244
+    },
+    {
+      "epoch": 0.6516500785751702,
+      "loss_partial": 0.05029296875,
+      "step": 1244
+    },
+    {
+      "epoch": 0.6516500785751702,
+      "loss_partial": 0.05615234375,
+      "step": 1244
+    },
+    {
+      "epoch": 0.6521739130434783,
+      "grad_norm": 2.0949275493621826,
+      "learning_rate": 1.9760203666930326e-05,
+      "loss": 0.1095,
+      "step": 1245
+    },
+    {
+      "epoch": 0.6521739130434783,
+      "loss_partial": 0.0595703125,
+      "step": 1245
+    },
+    {
+      "epoch": 0.6521739130434783,
+      "loss_partial": 0.05712890625,
+      "step": 1245
+    },
+    {
+      "epoch": 0.6526977475117863,
+      "grad_norm": 2.474268674850464,
+      "learning_rate": 1.9759207508163506e-05,
+      "loss": 0.1132,
+      "step": 1246
+    },
+    {
+      "epoch": 0.6526977475117863,
+      "loss_partial": 0.043212890625,
+      "step": 1246
+    },
+    {
+      "epoch": 0.6526977475117863,
+      "loss_partial": 0.05810546875,
+      "step": 1246
+    },
+    {
+      "epoch": 0.6532215819800943,
+      "grad_norm": 1.9430397748947144,
+      "learning_rate": 1.9758209309782668e-05,
+      "loss": 0.109,
+      "step": 1247
+    },
+    {
+      "epoch": 0.6532215819800943,
+      "loss_partial": 0.058837890625,
+      "step": 1247
+    },
+    {
+      "epoch": 0.6532215819800943,
+      "loss_partial": 0.056396484375,
+      "step": 1247
+    },
+    {
+      "epoch": 0.6537454164484023,
+      "grad_norm": 1.9523557424545288,
+      "learning_rate": 1.9757209071996427e-05,
+      "loss": 0.1094,
+      "step": 1248
+    },
+    {
+      "epoch": 0.6537454164484023,
+      "loss_partial": 0.04150390625,
+      "step": 1248
+    },
+    {
+      "epoch": 0.6537454164484023,
+      "loss_partial": 0.04638671875,
+      "step": 1248
+    },
+    {
+      "epoch": 0.6542692509167103,
+      "grad_norm": 2.61148738861084,
+      "learning_rate": 1.975620679501382e-05,
+      "loss": 0.1108,
+      "step": 1249
+    },
+    {
+      "epoch": 0.6542692509167103,
+      "loss_partial": 0.054443359375,
+      "step": 1249
+    },
+    {
+      "epoch": 0.6542692509167103,
+      "loss_partial": 0.0634765625,
+      "step": 1249
+    },
+    {
+      "epoch": 0.6547930853850183,
+      "grad_norm": 2.1955487728118896,
+      "learning_rate": 1.975520247904433e-05,
+      "loss": 0.1128,
+      "step": 1250
+    },
+    {
+      "epoch": 0.6547930853850183,
+      "loss_partial": 0.050048828125,
+      "step": 1250
+    },
+    {
+      "epoch": 0.6547930853850183,
+      "loss_partial": 0.0576171875,
+      "step": 1250
+    },
+    {
+      "epoch": 0.6553169198533263,
+      "grad_norm": 1.5941523313522339,
+      "learning_rate": 1.9754196124297847e-05,
+      "loss": 0.1063,
+      "step": 1251
+    },
+    {
+      "epoch": 0.6553169198533263,
+      "loss_partial": 0.05615234375,
+      "step": 1251
+    },
+    {
+      "epoch": 0.6553169198533263,
+      "loss_partial": 0.0625,
+      "step": 1251
+    },
+    {
+      "epoch": 0.6558407543216344,
+      "grad_norm": 2.088718891143799,
+      "learning_rate": 1.9753187730984687e-05,
+      "loss": 0.1096,
+      "step": 1252
+    },
+    {
+      "epoch": 0.6558407543216344,
+      "loss_partial": 0.04736328125,
+      "step": 1252
+    },
+    {
+      "epoch": 0.6558407543216344,
+      "loss_partial": 0.0556640625,
+      "step": 1252
+    },
+    {
+      "epoch": 0.6563645887899424,
+      "grad_norm": 1.79720938205719,
+      "learning_rate": 1.975217729931561e-05,
+      "loss": 0.113,
+      "step": 1253
+    },
+    {
+      "epoch": 0.6563645887899424,
+      "loss_partial": 0.0595703125,
+      "step": 1253
+    },
+    {
+      "epoch": 0.6563645887899424,
+      "loss_partial": 0.0478515625,
+      "step": 1253
+    },
+    {
+      "epoch": 0.6568884232582504,
+      "grad_norm": 1.8912652730941772,
+      "learning_rate": 1.9751164829501777e-05,
+      "loss": 0.1088,
+      "step": 1254
+    },
+    {
+      "epoch": 0.6568884232582504,
+      "loss_partial": 0.05517578125,
+      "step": 1254
+    },
+    {
+      "epoch": 0.6568884232582504,
+      "loss_partial": 0.058837890625,
+      "step": 1254
+    },
+    {
+      "epoch": 0.6574122577265584,
+      "grad_norm": 1.3539937734603882,
+      "learning_rate": 1.97501503217548e-05,
+      "loss": 0.1127,
+      "step": 1255
+    },
+    {
+      "epoch": 0.6574122577265584,
+      "loss_partial": 0.05078125,
+      "step": 1255
+    },
+    {
+      "epoch": 0.6574122577265584,
+      "loss_partial": 0.048828125,
+      "step": 1255
+    },
+    {
+      "epoch": 0.6579360921948664,
+      "grad_norm": 2.06777024269104,
+      "learning_rate": 1.974913377628669e-05,
+      "loss": 0.1105,
+      "step": 1256
+    },
+    {
+      "epoch": 0.6579360921948664,
+      "loss_partial": 0.046142578125,
+      "step": 1256
+    },
+    {
+      "epoch": 0.6579360921948664,
+      "loss_partial": 0.06103515625,
+      "step": 1256
+    },
+    {
+      "epoch": 0.6584599266631744,
+      "grad_norm": 2.5385804176330566,
+      "learning_rate": 1.9748115193309913e-05,
+      "loss": 0.1146,
+      "step": 1257
+    },
+    {
+      "epoch": 0.6584599266631744,
+      "loss_partial": 0.05517578125,
+      "step": 1257
+    },
+    {
+      "epoch": 0.6584599266631744,
+      "loss_partial": 0.059326171875,
+      "step": 1257
+    },
+    {
+      "epoch": 0.6589837611314825,
+      "grad_norm": 1.9256502389907837,
+      "learning_rate": 1.974709457303734e-05,
+      "loss": 0.1115,
+      "step": 1258
+    },
+    {
+      "epoch": 0.6589837611314825,
+      "loss_partial": 0.06103515625,
+      "step": 1258
+    },
+    {
+      "epoch": 0.6589837611314825,
+      "loss_partial": 0.05126953125,
+      "step": 1258
+    },
+    {
+      "epoch": 0.6595075955997904,
+      "grad_norm": 2.490783214569092,
+      "learning_rate": 1.974607191568228e-05,
+      "loss": 0.1141,
+      "step": 1259
+    },
+    {
+      "epoch": 0.6595075955997904,
+      "loss_partial": 0.06298828125,
+      "step": 1259
+    },
+    {
+      "epoch": 0.6595075955997904,
+      "loss_partial": 0.05224609375,
+      "step": 1259
+    },
+    {
+      "epoch": 0.6600314300680985,
+      "grad_norm": 2.3337318897247314,
+      "learning_rate": 1.974504722145845e-05,
+      "loss": 0.1122,
+      "step": 1260
+    },
+    {
+      "epoch": 0.6600314300680985,
+      "loss_partial": 0.0498046875,
+      "step": 1260
+    },
+    {
+      "epoch": 0.6600314300680985,
+      "loss_partial": 0.0546875,
+      "step": 1260
+    },
+    {
+      "epoch": 0.6605552645364065,
+      "grad_norm": 2.361201524734497,
+      "learning_rate": 1.9744020490580017e-05,
+      "loss": 0.1081,
+      "step": 1261
+    },
+    {
+      "epoch": 0.6605552645364065,
+      "loss_partial": 0.053955078125,
+      "step": 1261
+    },
+    {
+      "epoch": 0.6605552645364065,
+      "loss_partial": 0.057861328125,
+      "step": 1261
+    },
+    {
+      "epoch": 0.6610790990047145,
+      "grad_norm": 2.4962317943573,
+      "learning_rate": 1.9742991723261555e-05,
+      "loss": 0.1103,
+      "step": 1262
+    },
+    {
+      "epoch": 0.6610790990047145,
+      "loss_partial": 0.0703125,
+      "step": 1262
+    },
+    {
+      "epoch": 0.6610790990047145,
+      "loss_partial": 0.05322265625,
+      "step": 1262
+    },
+    {
+      "epoch": 0.6616029334730226,
+      "grad_norm": 2.063311815261841,
+      "learning_rate": 1.9741960919718073e-05,
+      "loss": 0.1108,
+      "step": 1263
+    },
+    {
+      "epoch": 0.6616029334730226,
+      "loss_partial": 0.056396484375,
+      "step": 1263
+    },
+    {
+      "epoch": 0.6616029334730226,
+      "loss_partial": 0.059814453125,
+      "step": 1263
+    },
+    {
+      "epoch": 0.6621267679413305,
+      "grad_norm": 3.1123411655426025,
+      "learning_rate": 1.9740928080165e-05,
+      "loss": 0.1127,
+      "step": 1264
+    },
+    {
+      "epoch": 0.6621267679413305,
+      "loss_partial": 0.0556640625,
+      "step": 1264
+    },
+    {
+      "epoch": 0.6621267679413305,
+      "loss_partial": 0.0478515625,
+      "step": 1264
+    },
+    {
+      "epoch": 0.6626506024096386,
+      "grad_norm": 1.2150295972824097,
+      "learning_rate": 1.9739893204818196e-05,
+      "loss": 0.106,
+      "step": 1265
+    },
+    {
+      "epoch": 0.6626506024096386,
+      "loss_partial": 0.05810546875,
+      "step": 1265
+    },
+    {
+      "epoch": 0.6626506024096386,
+      "loss_partial": 0.0546875,
+      "step": 1265
+    },
+    {
+      "epoch": 0.6631744368779465,
+      "grad_norm": 2.865570068359375,
+      "learning_rate": 1.973885629389394e-05,
+      "loss": 0.1149,
+      "step": 1266
+    },
+    {
+      "epoch": 0.6631744368779465,
+      "loss_partial": 0.05712890625,
+      "step": 1266
+    },
+    {
+      "epoch": 0.6631744368779465,
+      "loss_partial": 0.052978515625,
+      "step": 1266
+    },
+    {
+      "epoch": 0.6636982713462546,
+      "grad_norm": 1.6746282577514648,
+      "learning_rate": 1.9737817347608944e-05,
+      "loss": 0.1129,
+      "step": 1267
+    },
+    {
+      "epoch": 0.6636982713462546,
+      "loss_partial": 0.060546875,
+      "step": 1267
+    },
+    {
+      "epoch": 0.6636982713462546,
+      "loss_partial": 0.048095703125,
+      "step": 1267
+    },
+    {
+      "epoch": 0.6642221058145626,
+      "grad_norm": 1.6104837656021118,
+      "learning_rate": 1.9736776366180337e-05,
+      "loss": 0.1109,
+      "step": 1268
+    },
+    {
+      "epoch": 0.6642221058145626,
+      "loss_partial": 0.04638671875,
+      "step": 1268
+    },
+    {
+      "epoch": 0.6642221058145626,
+      "loss_partial": 0.054931640625,
+      "step": 1268
+    },
+    {
+      "epoch": 0.6647459402828706,
+      "grad_norm": 1.8405656814575195,
+      "learning_rate": 1.973573334982568e-05,
+      "loss": 0.1123,
+      "step": 1269
+    },
+    {
+      "epoch": 0.6647459402828706,
+      "loss_partial": 0.05908203125,
+      "step": 1269
+    },
+    {
+      "epoch": 0.6647459402828706,
+      "loss_partial": 0.05615234375,
+      "step": 1269
+    },
+    {
+      "epoch": 0.6652697747511787,
+      "grad_norm": 1.817315697669983,
+      "learning_rate": 1.9734688298762957e-05,
+      "loss": 0.1106,
+      "step": 1270
+    },
+    {
+      "epoch": 0.6652697747511787,
+      "loss_partial": 0.05029296875,
+      "step": 1270
+    },
+    {
+      "epoch": 0.6652697747511787,
+      "loss_partial": 0.060546875,
+      "step": 1270
+    },
+    {
+      "epoch": 0.6657936092194866,
+      "grad_norm": 1.4625240564346313,
+      "learning_rate": 1.973364121321058e-05,
+      "loss": 0.1069,
+      "step": 1271
+    },
+    {
+      "epoch": 0.6657936092194866,
+      "loss_partial": 0.0595703125,
+      "step": 1271
+    },
+    {
+      "epoch": 0.6657936092194866,
+      "loss_partial": 0.0546875,
+      "step": 1271
+    },
+    {
+      "epoch": 0.6663174436877947,
+      "grad_norm": 1.247323989868164,
+      "learning_rate": 1.9732592093387378e-05,
+      "loss": 0.1093,
+      "step": 1272
+    },
+    {
+      "epoch": 0.6663174436877947,
+      "loss_partial": 0.049072265625,
+      "step": 1272
+    },
+    {
+      "epoch": 0.6663174436877947,
+      "loss_partial": 0.04833984375,
+      "step": 1272
+    },
+    {
+      "epoch": 0.6668412781561027,
+      "grad_norm": 1.4699867963790894,
+      "learning_rate": 1.9731540939512613e-05,
+      "loss": 0.1099,
+      "step": 1273
+    },
+    {
+      "epoch": 0.6668412781561027,
+      "loss_partial": 0.053466796875,
+      "step": 1273
+    },
+    {
+      "epoch": 0.6668412781561027,
+      "loss_partial": 0.05322265625,
+      "step": 1273
+    },
+    {
+      "epoch": 0.6673651126244107,
+      "grad_norm": 1.6842588186264038,
+      "learning_rate": 1.9730487751805967e-05,
+      "loss": 0.1128,
+      "step": 1274
+    },
+    {
+      "epoch": 0.6673651126244107,
+      "loss_partial": 0.0595703125,
+      "step": 1274
+    },
+    {
+      "epoch": 0.6673651126244107,
+      "loss_partial": 0.04931640625,
+      "step": 1274
+    },
+    {
+      "epoch": 0.6678889470927187,
+      "grad_norm": 1.761732816696167,
+      "learning_rate": 1.972943253048756e-05,
+      "loss": 0.1104,
+      "step": 1275
+    },
+    {
+      "epoch": 0.6678889470927187,
+      "loss_partial": 0.07177734375,
+      "step": 1275
+    },
+    {
+      "epoch": 0.6678889470927187,
+      "loss_partial": 0.057373046875,
+      "step": 1275
+    },
+    {
+      "epoch": 0.6684127815610267,
+      "grad_norm": 1.5237033367156982,
+      "learning_rate": 1.9728375275777913e-05,
+      "loss": 0.108,
+      "step": 1276
+    },
+    {
+      "epoch": 0.6684127815610267,
+      "loss_partial": 0.05322265625,
+      "step": 1276
+    },
+    {
+      "epoch": 0.6684127815610267,
+      "loss_partial": 0.048095703125,
+      "step": 1276
+    },
+    {
+      "epoch": 0.6689366160293347,
+      "grad_norm": 1.3827919960021973,
+      "learning_rate": 1.9727315987897993e-05,
+      "loss": 0.1095,
+      "step": 1277
+    },
+    {
+      "epoch": 0.6689366160293347,
+      "loss_partial": 0.04345703125,
+      "step": 1277
+    },
+    {
+      "epoch": 0.6689366160293347,
+      "loss_partial": 0.058837890625,
+      "step": 1277
+    },
+    {
+      "epoch": 0.6694604504976427,
+      "grad_norm": 1.6955188512802124,
+      "learning_rate": 1.9726254667069182e-05,
+      "loss": 0.1133,
+      "step": 1278
+    },
+    {
+      "epoch": 0.6694604504976427,
+      "loss_partial": 0.0673828125,
+      "step": 1278
+    },
+    {
+      "epoch": 0.6694604504976427,
+      "loss_partial": 0.05810546875,
+      "step": 1278
+    },
+    {
+      "epoch": 0.6699842849659507,
+      "grad_norm": 2.4826011657714844,
+      "learning_rate": 1.972519131351329e-05,
+      "loss": 0.1095,
+      "step": 1279
+    },
+    {
+      "epoch": 0.6699842849659507,
+      "loss_partial": 0.06787109375,
+      "step": 1279
+    },
+    {
+      "epoch": 0.6699842849659507,
+      "loss_partial": 0.07080078125,
+      "step": 1279
+    },
+    {
+      "epoch": 0.6705081194342588,
+      "grad_norm": 1.6180325746536255,
+      "learning_rate": 1.9724125927452557e-05,
+      "loss": 0.1108,
+      "step": 1280
+    },
+    {
+      "epoch": 0.6705081194342588,
+      "loss_partial": 0.0634765625,
+      "step": 1280
+    },
+    {
+      "epoch": 0.6705081194342588,
+      "loss_partial": 0.056640625,
+      "step": 1280
+    },
+    {
+      "epoch": 0.6710319539025668,
+      "grad_norm": 2.3194479942321777,
+      "learning_rate": 1.9723058509109637e-05,
+      "loss": 0.1123,
+      "step": 1281
+    },
+    {
+      "epoch": 0.6710319539025668,
+      "loss_partial": 0.053466796875,
+      "step": 1281
+    },
+    {
+      "epoch": 0.6710319539025668,
+      "loss_partial": 0.0546875,
+      "step": 1281
+    },
+    {
+      "epoch": 0.6715557883708748,
+      "grad_norm": 2.4018094539642334,
+      "learning_rate": 1.9721989058707608e-05,
+      "loss": 0.1123,
+      "step": 1282
+    },
+    {
+      "epoch": 0.6715557883708748,
+      "loss_partial": 0.044921875,
+      "step": 1282
+    },
+    {
+      "epoch": 0.6715557883708748,
+      "loss_partial": 0.056640625,
+      "step": 1282
+    },
+    {
+      "epoch": 0.6720796228391828,
+      "grad_norm": 2.3503594398498535,
+      "learning_rate": 1.972091757646999e-05,
+      "loss": 0.1115,
+      "step": 1283
+    },
+    {
+      "epoch": 0.6720796228391828,
+      "loss_partial": 0.057373046875,
+      "step": 1283
+    },
+    {
+      "epoch": 0.6720796228391828,
+      "loss_partial": 0.04931640625,
+      "step": 1283
+    },
+    {
+      "epoch": 0.6726034573074908,
+      "grad_norm": 2.284161329269409,
+      "learning_rate": 1.9719844062620713e-05,
+      "loss": 0.1097,
+      "step": 1284
+    },
+    {
+      "epoch": 0.6726034573074908,
+      "loss_partial": 0.060791015625,
+      "step": 1284
+    },
+    {
+      "epoch": 0.6726034573074908,
+      "loss_partial": 0.05322265625,
+      "step": 1284
+    },
+    {
+      "epoch": 0.6731272917757989,
+      "grad_norm": 2.4994733333587646,
+      "learning_rate": 1.9718768517384127e-05,
+      "loss": 0.1114,
+      "step": 1285
+    },
+    {
+      "epoch": 0.6731272917757989,
+      "loss_partial": 0.04345703125,
+      "step": 1285
+    },
+    {
+      "epoch": 0.6731272917757989,
+      "loss_partial": 0.05029296875,
+      "step": 1285
+    },
+    {
+      "epoch": 0.6736511262441068,
+      "grad_norm": 1.862176537513733,
+      "learning_rate": 1.9717690940985024e-05,
+      "loss": 0.1091,
+      "step": 1286
+    },
+    {
+      "epoch": 0.6736511262441068,
+      "loss_partial": 0.057861328125,
+      "step": 1286
+    },
+    {
+      "epoch": 0.6736511262441068,
+      "loss_partial": 0.06005859375,
+      "step": 1286
+    },
+    {
+      "epoch": 0.6741749607124149,
+      "grad_norm": 2.063004970550537,
+      "learning_rate": 1.9716611333648603e-05,
+      "loss": 0.1119,
+      "step": 1287
+    },
+    {
+      "epoch": 0.6741749607124149,
+      "loss_partial": 0.058837890625,
+      "step": 1287
+    },
+    {
+      "epoch": 0.6741749607124149,
+      "loss_partial": 0.044921875,
+      "step": 1287
+    },
+    {
+      "epoch": 0.6746987951807228,
+      "grad_norm": 1.5611746311187744,
+      "learning_rate": 1.9715529695600503e-05,
+      "loss": 0.1097,
+      "step": 1288
+    },
+    {
+      "epoch": 0.6746987951807228,
+      "loss_partial": 0.06494140625,
+      "step": 1288
+    },
+    {
+      "epoch": 0.6746987951807228,
+      "loss_partial": 0.0498046875,
+      "step": 1288
+    },
+    {
+      "epoch": 0.6752226296490309,
+      "grad_norm": 2.1888675689697266,
+      "learning_rate": 1.9714446027066773e-05,
+      "loss": 0.1105,
+      "step": 1289
+    },
+    {
+      "epoch": 0.6752226296490309,
+      "loss_partial": 0.052734375,
+      "step": 1289
+    },
+    {
+      "epoch": 0.6752226296490309,
+      "loss_partial": 0.056640625,
+      "step": 1289
+    },
+    {
+      "epoch": 0.675746464117339,
+      "grad_norm": 2.4480974674224854,
+      "learning_rate": 1.9713360328273898e-05,
+      "loss": 0.1082,
+      "step": 1290
+    },
+    {
+      "epoch": 0.675746464117339,
+      "loss_partial": 0.051513671875,
+      "step": 1290
+    },
+    {
+      "epoch": 0.675746464117339,
+      "loss_partial": 0.052734375,
+      "step": 1290
+    },
+    {
+      "epoch": 0.6762702985856469,
+      "grad_norm": 2.4847490787506104,
+      "learning_rate": 1.9712272599448775e-05,
+      "loss": 0.1084,
+      "step": 1291
+    },
+    {
+      "epoch": 0.6762702985856469,
+      "loss_partial": 0.060546875,
+      "step": 1291
+    },
+    {
+      "epoch": 0.6762702985856469,
+      "loss_partial": 0.052734375,
+      "step": 1291
+    },
+    {
+      "epoch": 0.676794133053955,
+      "grad_norm": 1.4778414964675903,
+      "learning_rate": 1.9711182840818744e-05,
+      "loss": 0.1102,
+      "step": 1292
+    },
+    {
+      "epoch": 0.676794133053955,
+      "loss_partial": 0.060546875,
+      "step": 1292
+    },
+    {
+      "epoch": 0.676794133053955,
+      "loss_partial": 0.051025390625,
+      "step": 1292
+    },
+    {
+      "epoch": 0.6773179675222629,
+      "grad_norm": 1.7509950399398804,
+      "learning_rate": 1.9710091052611546e-05,
+      "loss": 0.1088,
+      "step": 1293
+    },
+    {
+      "epoch": 0.6773179675222629,
+      "loss_partial": 0.0673828125,
+      "step": 1293
+    },
+    {
+      "epoch": 0.6773179675222629,
+      "loss_partial": 0.046142578125,
+      "step": 1293
+    },
+    {
+      "epoch": 0.677841801990571,
+      "grad_norm": 1.7281299829483032,
+      "learning_rate": 1.970899723505537e-05,
+      "loss": 0.1104,
+      "step": 1294
+    },
+    {
+      "epoch": 0.677841801990571,
+      "loss_partial": 0.051025390625,
+      "step": 1294
+    },
+    {
+      "epoch": 0.677841801990571,
+      "loss_partial": 0.051025390625,
+      "step": 1294
+    },
+    {
+      "epoch": 0.678365636458879,
+      "grad_norm": 1.7940634489059448,
+      "learning_rate": 1.9707901388378803e-05,
+      "loss": 0.1122,
+      "step": 1295
+    },
+    {
+      "epoch": 0.678365636458879,
+      "loss_partial": 0.060791015625,
+      "step": 1295
+    },
+    {
+      "epoch": 0.678365636458879,
+      "loss_partial": 0.05615234375,
+      "step": 1295
+    },
+    {
+      "epoch": 0.678889470927187,
+      "grad_norm": 1.491577386856079,
+      "learning_rate": 1.970680351281088e-05,
+      "loss": 0.1153,
+      "step": 1296
+    },
+    {
+      "epoch": 0.678889470927187,
+      "loss_partial": 0.04443359375,
+      "step": 1296
+    },
+    {
+      "epoch": 0.678889470927187,
+      "loss_partial": 0.060791015625,
+      "step": 1296
+    },
+    {
+      "epoch": 0.679413305395495,
+      "grad_norm": 2.1085989475250244,
+      "learning_rate": 1.970570360858105e-05,
+      "loss": 0.1124,
+      "step": 1297
+    },
+    {
+      "epoch": 0.679413305395495,
+      "loss_partial": 0.048828125,
+      "step": 1297
+    },
+    {
+      "epoch": 0.679413305395495,
+      "loss_partial": 0.056884765625,
+      "step": 1297
+    },
+    {
+      "epoch": 0.679937139863803,
+      "grad_norm": 2.284278392791748,
+      "learning_rate": 1.9704601675919184e-05,
+      "loss": 0.1096,
+      "step": 1298
+    },
+    {
+      "epoch": 0.679937139863803,
+      "loss_partial": 0.053466796875,
+      "step": 1298
+    },
+    {
+      "epoch": 0.679937139863803,
+      "loss_partial": 0.056396484375,
+      "step": 1298
+    },
+    {
+      "epoch": 0.6804609743321111,
+      "grad_norm": 2.0597264766693115,
+      "learning_rate": 1.970349771505558e-05,
+      "loss": 0.1113,
+      "step": 1299
+    },
+    {
+      "epoch": 0.6804609743321111,
+      "loss_partial": 0.052734375,
+      "step": 1299
+    },
+    {
+      "epoch": 0.6804609743321111,
+      "loss_partial": 0.0498046875,
+      "step": 1299
+    },
+    {
+      "epoch": 0.6809848088004191,
+      "grad_norm": 2.457568407058716,
+      "learning_rate": 1.970239172622095e-05,
+      "loss": 0.1101,
+      "step": 1300
+    },
+    {
+      "epoch": 0.6809848088004191,
+      "loss_partial": 0.044677734375,
+      "step": 1300
+    },
+    {
+      "epoch": 0.6809848088004191,
+      "loss_partial": 0.05615234375,
+      "step": 1300
+    },
+    {
+      "epoch": 0.6815086432687271,
+      "grad_norm": 2.230663537979126,
+      "learning_rate": 1.9701283709646457e-05,
+      "loss": 0.1066,
+      "step": 1301
+    },
+    {
+      "epoch": 0.6815086432687271,
+      "loss_partial": 0.05517578125,
+      "step": 1301
+    },
+    {
+      "epoch": 0.6815086432687271,
+      "loss_partial": 0.05615234375,
+      "step": 1301
+    },
+    {
+      "epoch": 0.6820324777370351,
+      "grad_norm": 2.185753345489502,
+      "learning_rate": 1.9700173665563658e-05,
+      "loss": 0.1102,
+      "step": 1302
+    },
+    {
+      "epoch": 0.6820324777370351,
+      "loss_partial": 0.05322265625,
+      "step": 1302
+    },
+    {
+      "epoch": 0.6820324777370351,
+      "loss_partial": 0.057373046875,
+      "step": 1302
+    },
+    {
+      "epoch": 0.6825563122053431,
+      "grad_norm": 2.0424392223358154,
+      "learning_rate": 1.9699061594204542e-05,
+      "loss": 0.1106,
+      "step": 1303
+    },
+    {
+      "epoch": 0.6825563122053431,
+      "loss_partial": 0.0654296875,
+      "step": 1303
+    },
+    {
+      "epoch": 0.6825563122053431,
+      "loss_partial": 0.06005859375,
+      "step": 1303
+    },
+    {
+      "epoch": 0.6830801466736511,
+      "grad_norm": 2.335675001144409,
+      "learning_rate": 1.9697947495801535e-05,
+      "loss": 0.1138,
+      "step": 1304
+    },
+    {
+      "epoch": 0.6830801466736511,
+      "loss_partial": 0.0517578125,
+      "step": 1304
+    },
+    {
+      "epoch": 0.6830801466736511,
+      "loss_partial": 0.04833984375,
+      "step": 1304
+    },
+    {
+      "epoch": 0.6836039811419592,
+      "grad_norm": 3.748469591140747,
+      "learning_rate": 1.969683137058747e-05,
+      "loss": 0.1105,
+      "step": 1305
+    },
+    {
+      "epoch": 0.6836039811419592,
+      "loss_partial": 0.054931640625,
+      "step": 1305
+    },
+    {
+      "epoch": 0.6836039811419592,
+      "loss_partial": 0.049072265625,
+      "step": 1305
+    },
+    {
+      "epoch": 0.6841278156102671,
+      "grad_norm": 1.9523537158966064,
+      "learning_rate": 1.969571321879561e-05,
+      "loss": 0.1148,
+      "step": 1306
+    },
+    {
+      "epoch": 0.6841278156102671,
+      "loss_partial": 0.047607421875,
+      "step": 1306
+    },
+    {
+      "epoch": 0.6841278156102671,
+      "loss_partial": 0.056640625,
+      "step": 1306
+    },
+    {
+      "epoch": 0.6846516500785752,
+      "grad_norm": 2.6626439094543457,
+      "learning_rate": 1.9694593040659645e-05,
+      "loss": 0.1084,
+      "step": 1307
+    },
+    {
+      "epoch": 0.6846516500785752,
+      "loss_partial": 0.05078125,
+      "step": 1307
+    },
+    {
+      "epoch": 0.6846516500785752,
+      "loss_partial": 0.05712890625,
+      "step": 1307
+    },
+    {
+      "epoch": 0.6851754845468832,
+      "grad_norm": 2.2856762409210205,
+      "learning_rate": 1.9693470836413687e-05,
+      "loss": 0.1106,
+      "step": 1308
+    },
+    {
+      "epoch": 0.6851754845468832,
+      "loss_partial": 0.058349609375,
+      "step": 1308
+    },
+    {
+      "epoch": 0.6851754845468832,
+      "loss_partial": 0.06005859375,
+      "step": 1308
+    },
+    {
+      "epoch": 0.6856993190151912,
+      "grad_norm": 2.454688787460327,
+      "learning_rate": 1.9692346606292262e-05,
+      "loss": 0.11,
+      "step": 1309
+    },
+    {
+      "epoch": 0.6856993190151912,
+      "loss_partial": 0.052734375,
+      "step": 1309
+    },
+    {
+      "epoch": 0.6856993190151912,
+      "loss_partial": 0.0654296875,
+      "step": 1309
+    },
+    {
+      "epoch": 0.6862231534834992,
+      "grad_norm": 2.6033332347869873,
+      "learning_rate": 1.9691220350530335e-05,
+      "loss": 0.1113,
+      "step": 1310
+    },
+    {
+      "epoch": 0.6862231534834992,
+      "loss_partial": 0.07080078125,
+      "step": 1310
+    },
+    {
+      "epoch": 0.6862231534834992,
+      "loss_partial": 0.056640625,
+      "step": 1310
+    },
+    {
+      "epoch": 0.6867469879518072,
+      "grad_norm": 2.4072444438934326,
+      "learning_rate": 1.9690092069363282e-05,
+      "loss": 0.1168,
+      "step": 1311
+    },
+    {
+      "epoch": 0.6867469879518072,
+      "loss_partial": 0.052001953125,
+      "step": 1311
+    },
+    {
+      "epoch": 0.6867469879518072,
+      "loss_partial": 0.053466796875,
+      "step": 1311
+    },
+    {
+      "epoch": 0.6872708224201153,
+      "grad_norm": 1.6945581436157227,
+      "learning_rate": 1.9688961763026907e-05,
+      "loss": 0.1088,
+      "step": 1312
+    },
+    {
+      "epoch": 0.6872708224201153,
+      "loss_partial": 0.06005859375,
+      "step": 1312
+    },
+    {
+      "epoch": 0.6872708224201153,
+      "loss_partial": 0.05615234375,
+      "step": 1312
+    },
+    {
+      "epoch": 0.6877946568884232,
+      "grad_norm": 1.9479448795318604,
+      "learning_rate": 1.968782943175744e-05,
+      "loss": 0.111,
+      "step": 1313
+    },
+    {
+      "epoch": 0.6877946568884232,
+      "loss_partial": 0.052734375,
+      "step": 1313
+    },
+    {
+      "epoch": 0.6877946568884232,
+      "loss_partial": 0.052978515625,
+      "step": 1313
+    },
+    {
+      "epoch": 0.6883184913567313,
+      "grad_norm": 2.415311336517334,
+      "learning_rate": 1.968669507579153e-05,
+      "loss": 0.1104,
+      "step": 1314
+    },
+    {
+      "epoch": 0.6883184913567313,
+      "loss_partial": 0.0595703125,
+      "step": 1314
+    },
+    {
+      "epoch": 0.6883184913567313,
+      "loss_partial": 0.051513671875,
+      "step": 1314
+    },
+    {
+      "epoch": 0.6888423258250392,
+      "grad_norm": 1.4743764400482178,
+      "learning_rate": 1.9685558695366252e-05,
+      "loss": 0.1115,
+      "step": 1315
+    },
+    {
+      "epoch": 0.6888423258250392,
+      "loss_partial": 0.0556640625,
+      "step": 1315
+    },
+    {
+      "epoch": 0.6888423258250392,
+      "loss_partial": 0.046630859375,
+      "step": 1315
+    },
+    {
+      "epoch": 0.6893661602933473,
+      "grad_norm": 2.0882160663604736,
+      "learning_rate": 1.9684420290719094e-05,
+      "loss": 0.1125,
+      "step": 1316
+    },
+    {
+      "epoch": 0.6893661602933473,
+      "loss_partial": 0.046875,
+      "step": 1316
+    },
+    {
+      "epoch": 0.6893661602933473,
+      "loss_partial": 0.052490234375,
+      "step": 1316
+    },
+    {
+      "epoch": 0.6898899947616554,
+      "grad_norm": 2.094987154006958,
+      "learning_rate": 1.9683279862087986e-05,
+      "loss": 0.1099,
+      "step": 1317
+    },
+    {
+      "epoch": 0.6898899947616554,
+      "loss_partial": 0.056396484375,
+      "step": 1317
+    },
+    {
+      "epoch": 0.6898899947616554,
+      "loss_partial": 0.0478515625,
+      "step": 1317
+    },
+    {
+      "epoch": 0.6904138292299633,
+      "grad_norm": 1.786601185798645,
+      "learning_rate": 1.9682137409711267e-05,
+      "loss": 0.1078,
+      "step": 1318
+    },
+    {
+      "epoch": 0.6904138292299633,
+      "loss_partial": 0.05615234375,
+      "step": 1318
+    },
+    {
+      "epoch": 0.6904138292299633,
+      "loss_partial": 0.059814453125,
+      "step": 1318
+    },
+    {
+      "epoch": 0.6909376636982714,
+      "grad_norm": 2.180422306060791,
+      "learning_rate": 1.9680992933827703e-05,
+      "loss": 0.1128,
+      "step": 1319
+    },
+    {
+      "epoch": 0.6909376636982714,
+      "loss_partial": 0.051025390625,
+      "step": 1319
+    },
+    {
+      "epoch": 0.6909376636982714,
+      "loss_partial": 0.052001953125,
+      "step": 1319
+    },
+    {
+      "epoch": 0.6914614981665793,
+      "grad_norm": 1.8176689147949219,
+      "learning_rate": 1.967984643467648e-05,
+      "loss": 0.1098,
+      "step": 1320
+    },
+    {
+      "epoch": 0.6914614981665793,
+      "loss_partial": 0.059814453125,
+      "step": 1320
+    },
+    {
+      "epoch": 0.6914614981665793,
+      "loss_partial": 0.0576171875,
+      "step": 1320
+    },
+    {
+      "epoch": 0.6919853326348874,
+      "grad_norm": 1.2766164541244507,
+      "learning_rate": 1.967869791249721e-05,
+      "loss": 0.1137,
+      "step": 1321
+    },
+    {
+      "epoch": 0.6919853326348874,
+      "loss_partial": 0.051513671875,
+      "step": 1321
+    },
+    {
+      "epoch": 0.6919853326348874,
+      "loss_partial": 0.047119140625,
+      "step": 1321
+    },
+    {
+      "epoch": 0.6925091671031954,
+      "grad_norm": 1.5825501680374146,
+      "learning_rate": 1.9677547367529925e-05,
+      "loss": 0.1098,
+      "step": 1322
+    },
+    {
+      "epoch": 0.6925091671031954,
+      "loss_partial": 0.048095703125,
+      "step": 1322
+    },
+    {
+      "epoch": 0.6925091671031954,
+      "loss_partial": 0.06396484375,
+      "step": 1322
+    },
+    {
+      "epoch": 0.6930330015715034,
+      "grad_norm": 1.4702293872833252,
+      "learning_rate": 1.9676394800015087e-05,
+      "loss": 0.115,
+      "step": 1323
+    },
+    {
+      "epoch": 0.6930330015715034,
+      "loss_partial": 0.05908203125,
+      "step": 1323
+    },
+    {
+      "epoch": 0.6930330015715034,
+      "loss_partial": 0.0712890625,
+      "step": 1323
+    },
+    {
+      "epoch": 0.6935568360398114,
+      "grad_norm": 2.072833776473999,
+      "learning_rate": 1.9675240210193574e-05,
+      "loss": 0.1099,
+      "step": 1324
+    },
+    {
+      "epoch": 0.6935568360398114,
+      "loss_partial": 0.0537109375,
+      "step": 1324
+    },
+    {
+      "epoch": 0.6935568360398114,
+      "loss_partial": 0.0556640625,
+      "step": 1324
+    },
+    {
+      "epoch": 0.6940806705081194,
+      "grad_norm": 1.82939875125885,
+      "learning_rate": 1.9674083598306685e-05,
+      "loss": 0.1112,
+      "step": 1325
+    },
+    {
+      "epoch": 0.6940806705081194,
+      "loss_partial": 0.05517578125,
+      "step": 1325
+    },
+    {
+      "epoch": 0.6940806705081194,
+      "loss_partial": 0.05517578125,
+      "step": 1325
+    },
+    {
+      "epoch": 0.6946045049764275,
+      "grad_norm": 1.8554967641830444,
+      "learning_rate": 1.9672924964596148e-05,
+      "loss": 0.1124,
+      "step": 1326
+    },
+    {
+      "epoch": 0.6946045049764275,
+      "loss_partial": 0.0537109375,
+      "step": 1326
+    },
+    {
+      "epoch": 0.6946045049764275,
+      "loss_partial": 0.051025390625,
+      "step": 1326
+    },
+    {
+      "epoch": 0.6951283394447355,
+      "grad_norm": 2.184680938720703,
+      "learning_rate": 1.9671764309304108e-05,
+      "loss": 0.1146,
+      "step": 1327
+    },
+    {
+      "epoch": 0.6951283394447355,
+      "loss_partial": 0.058837890625,
+      "step": 1327
+    },
+    {
+      "epoch": 0.6951283394447355,
+      "loss_partial": 0.05322265625,
+      "step": 1327
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "grad_norm": 2.0312118530273438,
+      "learning_rate": 1.9670601632673137e-05,
+      "loss": 0.1094,
+      "step": 1328
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "loss_partial": 0.04931640625,
+      "step": 1328
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "loss_partial": 0.060791015625,
+      "step": 1328
+    },
+    {
+      "epoch": 0.6961760083813515,
+      "grad_norm": 2.3132386207580566,
+      "learning_rate": 1.9669436934946226e-05,
+      "loss": 0.1142,
+      "step": 1329
+    },
+    {
+      "epoch": 0.6961760083813515,
+      "loss_partial": 0.06396484375,
+      "step": 1329
+    },
+    {
+      "epoch": 0.6961760083813515,
+      "loss_partial": 0.050537109375,
+      "step": 1329
+    },
+    {
+      "epoch": 0.6966998428496595,
+      "grad_norm": 1.7250688076019287,
+      "learning_rate": 1.9668270216366788e-05,
+      "loss": 0.1106,
+      "step": 1330
+    },
+    {
+      "epoch": 0.6966998428496595,
+      "loss_partial": 0.05908203125,
+      "step": 1330
+    },
+    {
+      "epoch": 0.6966998428496595,
+      "loss_partial": 0.05810546875,
+      "step": 1330
+    },
+    {
+      "epoch": 0.6972236773179675,
+      "grad_norm": 1.8371845483779907,
+      "learning_rate": 1.966710147717866e-05,
+      "loss": 0.1107,
+      "step": 1331
+    },
+    {
+      "epoch": 0.6972236773179675,
+      "loss_partial": 0.05517578125,
+      "step": 1331
+    },
+    {
+      "epoch": 0.6972236773179675,
+      "loss_partial": 0.06640625,
+      "step": 1331
+    },
+    {
+      "epoch": 0.6977475117862756,
+      "grad_norm": 1.6731964349746704,
+      "learning_rate": 1.9665930717626106e-05,
+      "loss": 0.1135,
+      "step": 1332
+    },
+    {
+      "epoch": 0.6977475117862756,
+      "loss_partial": 0.05908203125,
+      "step": 1332
+    },
+    {
+      "epoch": 0.6977475117862756,
+      "loss_partial": 0.050048828125,
+      "step": 1332
+    },
+    {
+      "epoch": 0.6982713462545835,
+      "grad_norm": 2.1708552837371826,
+      "learning_rate": 1.9664757937953804e-05,
+      "loss": 0.1115,
+      "step": 1333
+    },
+    {
+      "epoch": 0.6982713462545835,
+      "loss_partial": 0.054443359375,
+      "step": 1333
+    },
+    {
+      "epoch": 0.6982713462545835,
+      "loss_partial": 0.048828125,
+      "step": 1333
+    },
+    {
+      "epoch": 0.6987951807228916,
+      "grad_norm": 1.4719958305358887,
+      "learning_rate": 1.9663583138406857e-05,
+      "loss": 0.1109,
+      "step": 1334
+    },
+    {
+      "epoch": 0.6987951807228916,
+      "loss_partial": 0.05517578125,
+      "step": 1334
+    },
+    {
+      "epoch": 0.6987951807228916,
+      "loss_partial": 0.05810546875,
+      "step": 1334
+    },
+    {
+      "epoch": 0.6993190151911995,
+      "grad_norm": 2.304205894470215,
+      "learning_rate": 1.966240631923079e-05,
+      "loss": 0.1107,
+      "step": 1335
+    },
+    {
+      "epoch": 0.6993190151911995,
+      "loss_partial": 0.052490234375,
+      "step": 1335
+    },
+    {
+      "epoch": 0.6993190151911995,
+      "loss_partial": 0.0537109375,
+      "step": 1335
+    },
+    {
+      "epoch": 0.6998428496595076,
+      "grad_norm": 1.9793440103530884,
+      "learning_rate": 1.9661227480671554e-05,
+      "loss": 0.1082,
+      "step": 1336
+    },
+    {
+      "epoch": 0.6998428496595076,
+      "loss_partial": 0.054443359375,
+      "step": 1336
+    },
+    {
+      "epoch": 0.6998428496595076,
+      "loss_partial": 0.057861328125,
+      "step": 1336
+    },
+    {
+      "epoch": 0.7003666841278157,
+      "grad_norm": 1.7633661031723022,
+      "learning_rate": 1.9660046622975518e-05,
+      "loss": 0.1124,
+      "step": 1337
+    },
+    {
+      "epoch": 0.7003666841278157,
+      "loss_partial": 0.057373046875,
+      "step": 1337
+    },
+    {
+      "epoch": 0.7003666841278157,
+      "loss_partial": 0.0703125,
+      "step": 1337
+    },
+    {
+      "epoch": 0.7008905185961236,
+      "grad_norm": 1.7213425636291504,
+      "learning_rate": 1.9658863746389476e-05,
+      "loss": 0.11,
+      "step": 1338
+    },
+    {
+      "epoch": 0.7008905185961236,
+      "loss_partial": 0.06396484375,
+      "step": 1338
+    },
+    {
+      "epoch": 0.7008905185961236,
+      "loss_partial": 0.059326171875,
+      "step": 1338
+    },
+    {
+      "epoch": 0.7014143530644317,
+      "grad_norm": 2.4434874057769775,
+      "learning_rate": 1.9657678851160637e-05,
+      "loss": 0.108,
+      "step": 1339
+    },
+    {
+      "epoch": 0.7014143530644317,
+      "loss_partial": 0.06494140625,
+      "step": 1339
+    },
+    {
+      "epoch": 0.7014143530644317,
+      "loss_partial": 0.058349609375,
+      "step": 1339
+    },
+    {
+      "epoch": 0.7019381875327396,
+      "grad_norm": 1.7264618873596191,
+      "learning_rate": 1.965649193753664e-05,
+      "loss": 0.1116,
+      "step": 1340
+    },
+    {
+      "epoch": 0.7019381875327396,
+      "loss_partial": 0.052734375,
+      "step": 1340
+    },
+    {
+      "epoch": 0.7019381875327396,
+      "loss_partial": 0.05224609375,
+      "step": 1340
+    },
+    {
+      "epoch": 0.7024620220010477,
+      "grad_norm": 2.5425665378570557,
+      "learning_rate": 1.9655303005765543e-05,
+      "loss": 0.1099,
+      "step": 1341
+    },
+    {
+      "epoch": 0.7024620220010477,
+      "loss_partial": 0.048583984375,
+      "step": 1341
+    },
+    {
+      "epoch": 0.7024620220010477,
+      "loss_partial": 0.0517578125,
+      "step": 1341
+    },
+    {
+      "epoch": 0.7029858564693556,
+      "grad_norm": 3.075477361679077,
+      "learning_rate": 1.9654112056095826e-05,
+      "loss": 0.1102,
+      "step": 1342
+    },
+    {
+      "epoch": 0.7029858564693556,
+      "loss_partial": 0.064453125,
+      "step": 1342
+    },
+    {
+      "epoch": 0.7029858564693556,
+      "loss_partial": 0.054931640625,
+      "step": 1342
+    },
+    {
+      "epoch": 0.7035096909376637,
+      "grad_norm": 2.608464479446411,
+      "learning_rate": 1.9652919088776386e-05,
+      "loss": 0.1109,
+      "step": 1343
+    },
+    {
+      "epoch": 0.7035096909376637,
+      "loss_partial": 0.058837890625,
+      "step": 1343
+    },
+    {
+      "epoch": 0.7035096909376637,
+      "loss_partial": 0.050537109375,
+      "step": 1343
+    },
+    {
+      "epoch": 0.7040335254059717,
+      "grad_norm": 2.4298620223999023,
+      "learning_rate": 1.965172410405655e-05,
+      "loss": 0.1098,
+      "step": 1344
+    },
+    {
+      "epoch": 0.7040335254059717,
+      "loss_partial": 0.055908203125,
+      "step": 1344
+    },
+    {
+      "epoch": 0.7040335254059717,
+      "loss_partial": 0.056640625,
+      "step": 1344
+    },
+    {
+      "epoch": 0.7045573598742797,
+      "grad_norm": 2.429960250854492,
+      "learning_rate": 1.9650527102186056e-05,
+      "loss": 0.1131,
+      "step": 1345
+    },
+    {
+      "epoch": 0.7045573598742797,
+      "loss_partial": 0.0712890625,
+      "step": 1345
+    },
+    {
+      "epoch": 0.7045573598742797,
+      "loss_partial": 0.053466796875,
+      "step": 1345
+    },
+    {
+      "epoch": 0.7050811943425878,
+      "grad_norm": 2.5479938983917236,
+      "learning_rate": 1.9649328083415078e-05,
+      "loss": 0.1145,
+      "step": 1346
+    },
+    {
+      "epoch": 0.7050811943425878,
+      "loss_partial": 0.054931640625,
+      "step": 1346
+    },
+    {
+      "epoch": 0.7050811943425878,
+      "loss_partial": 0.056640625,
+      "step": 1346
+    },
+    {
+      "epoch": 0.7056050288108957,
+      "grad_norm": 2.0816307067871094,
+      "learning_rate": 1.96481270479942e-05,
+      "loss": 0.109,
+      "step": 1347
+    },
+    {
+      "epoch": 0.7056050288108957,
+      "loss_partial": 0.0546875,
+      "step": 1347
+    },
+    {
+      "epoch": 0.7056050288108957,
+      "loss_partial": 0.053466796875,
+      "step": 1347
+    },
+    {
+      "epoch": 0.7061288632792038,
+      "grad_norm": 1.792782187461853,
+      "learning_rate": 1.9646923996174432e-05,
+      "loss": 0.1083,
+      "step": 1348
+    },
+    {
+      "epoch": 0.7061288632792038,
+      "loss_partial": 0.06201171875,
+      "step": 1348
+    },
+    {
+      "epoch": 0.7061288632792038,
+      "loss_partial": 0.05078125,
+      "step": 1348
+    },
+    {
+      "epoch": 0.7066526977475118,
+      "grad_norm": 2.455188274383545,
+      "learning_rate": 1.9645718928207204e-05,
+      "loss": 0.114,
+      "step": 1349
+    },
+    {
+      "epoch": 0.7066526977475118,
+      "loss_partial": 0.049072265625,
+      "step": 1349
+    },
+    {
+      "epoch": 0.7066526977475118,
+      "loss_partial": 0.041748046875,
+      "step": 1349
+    },
+    {
+      "epoch": 0.7071765322158198,
+      "grad_norm": 1.5130267143249512,
+      "learning_rate": 1.9644511844344368e-05,
+      "loss": 0.1087,
+      "step": 1350
+    },
+    {
+      "epoch": 0.7071765322158198,
+      "loss_partial": 0.055908203125,
+      "step": 1350
+    },
+    {
+      "epoch": 0.7071765322158198,
+      "loss_partial": 0.055908203125,
+      "step": 1350
+    },
+    {
+      "epoch": 0.7077003666841278,
+      "grad_norm": 2.0993497371673584,
+      "learning_rate": 1.9643302744838193e-05,
+      "loss": 0.1077,
+      "step": 1351
+    },
+    {
+      "epoch": 0.7077003666841278,
+      "loss_partial": 0.051025390625,
+      "step": 1351
+    },
+    {
+      "epoch": 0.7077003666841278,
+      "loss_partial": 0.050048828125,
+      "step": 1351
+    },
+    {
+      "epoch": 0.7082242011524358,
+      "grad_norm": 2.454514980316162,
+      "learning_rate": 1.964209162994138e-05,
+      "loss": 0.1086,
+      "step": 1352
+    },
+    {
+      "epoch": 0.7082242011524358,
+      "loss_partial": 0.05615234375,
+      "step": 1352
+    },
+    {
+      "epoch": 0.7082242011524358,
+      "loss_partial": 0.054443359375,
+      "step": 1352
+    },
+    {
+      "epoch": 0.7087480356207438,
+      "grad_norm": 2.5812788009643555,
+      "learning_rate": 1.9640878499907038e-05,
+      "loss": 0.1073,
+      "step": 1353
+    },
+    {
+      "epoch": 0.7087480356207438,
+      "loss_partial": 0.051513671875,
+      "step": 1353
+    },
+    {
+      "epoch": 0.7087480356207438,
+      "loss_partial": 0.06103515625,
+      "step": 1353
+    },
+    {
+      "epoch": 0.7092718700890519,
+      "grad_norm": 2.5292158126831055,
+      "learning_rate": 1.963966335498871e-05,
+      "loss": 0.1085,
+      "step": 1354
+    },
+    {
+      "epoch": 0.7092718700890519,
+      "loss_partial": 0.052734375,
+      "step": 1354
+    },
+    {
+      "epoch": 0.7092718700890519,
+      "loss_partial": 0.050048828125,
+      "step": 1354
+    },
+    {
+      "epoch": 0.7097957045573599,
+      "grad_norm": 2.444695234298706,
+      "learning_rate": 1.9638446195440348e-05,
+      "loss": 0.1085,
+      "step": 1355
+    },
+    {
+      "epoch": 0.7097957045573599,
+      "loss_partial": 0.04736328125,
+      "step": 1355
+    },
+    {
+      "epoch": 0.7097957045573599,
+      "loss_partial": 0.0498046875,
+      "step": 1355
+    },
+    {
+      "epoch": 0.7103195390256679,
+      "grad_norm": 1.6229244470596313,
+      "learning_rate": 1.9637227021516337e-05,
+      "loss": 0.1093,
+      "step": 1356
+    },
+    {
+      "epoch": 0.7103195390256679,
+      "loss_partial": 0.056884765625,
+      "step": 1356
+    },
+    {
+      "epoch": 0.7103195390256679,
+      "loss_partial": 0.05419921875,
+      "step": 1356
+    },
+    {
+      "epoch": 0.7108433734939759,
+      "grad_norm": 2.1555519104003906,
+      "learning_rate": 1.963600583347147e-05,
+      "loss": 0.1111,
+      "step": 1357
+    },
+    {
+      "epoch": 0.7108433734939759,
+      "loss_partial": 0.053466796875,
+      "step": 1357
+    },
+    {
+      "epoch": 0.7108433734939759,
+      "loss_partial": 0.056396484375,
+      "step": 1357
+    },
+    {
+      "epoch": 0.7113672079622839,
+      "grad_norm": 2.320502996444702,
+      "learning_rate": 1.963478263156097e-05,
+      "loss": 0.1104,
+      "step": 1358
+    },
+    {
+      "epoch": 0.7113672079622839,
+      "loss_partial": 0.06005859375,
+      "step": 1358
+    },
+    {
+      "epoch": 0.7113672079622839,
+      "loss_partial": 0.052001953125,
+      "step": 1358
+    },
+    {
+      "epoch": 0.711891042430592,
+      "grad_norm": 1.643439531326294,
+      "learning_rate": 1.9633557416040483e-05,
+      "loss": 0.1078,
+      "step": 1359
+    },
+    {
+      "epoch": 0.711891042430592,
+      "loss_partial": 0.0537109375,
+      "step": 1359
+    },
+    {
+      "epoch": 0.711891042430592,
+      "loss_partial": 0.061767578125,
+      "step": 1359
+    },
+    {
+      "epoch": 0.7124148768988999,
+      "grad_norm": 3.2242095470428467,
+      "learning_rate": 1.963233018716607e-05,
+      "loss": 0.1083,
+      "step": 1360
+    },
+    {
+      "epoch": 0.7124148768988999,
+      "loss_partial": 0.0537109375,
+      "step": 1360
+    },
+    {
+      "epoch": 0.7124148768988999,
+      "loss_partial": 0.05224609375,
+      "step": 1360
+    },
+    {
+      "epoch": 0.712938711367208,
+      "grad_norm": 2.039478302001953,
+      "learning_rate": 1.963110094519421e-05,
+      "loss": 0.1094,
+      "step": 1361
+    },
+    {
+      "epoch": 0.712938711367208,
+      "loss_partial": 0.0615234375,
+      "step": 1361
+    },
+    {
+      "epoch": 0.712938711367208,
+      "loss_partial": 0.046875,
+      "step": 1361
+    },
+    {
+      "epoch": 0.7134625458355159,
+      "grad_norm": 2.118431806564331,
+      "learning_rate": 1.9629869690381805e-05,
+      "loss": 0.109,
+      "step": 1362
+    },
+    {
+      "epoch": 0.7134625458355159,
+      "loss_partial": 0.055419921875,
+      "step": 1362
+    },
+    {
+      "epoch": 0.7134625458355159,
+      "loss_partial": 0.057861328125,
+      "step": 1362
+    },
+    {
+      "epoch": 0.713986380303824,
+      "grad_norm": 1.6009055376052856,
+      "learning_rate": 1.962863642298619e-05,
+      "loss": 0.1116,
+      "step": 1363
+    },
+    {
+      "epoch": 0.713986380303824,
+      "loss_partial": 0.06494140625,
+      "step": 1363
+    },
+    {
+      "epoch": 0.713986380303824,
+      "loss_partial": 0.061279296875,
+      "step": 1363
+    },
+    {
+      "epoch": 0.7145102147721321,
+      "grad_norm": 1.8644890785217285,
+      "learning_rate": 1.96274011432651e-05,
+      "loss": 0.1105,
+      "step": 1364
+    },
+    {
+      "epoch": 0.7145102147721321,
+      "loss_partial": 0.058349609375,
+      "step": 1364
+    },
+    {
+      "epoch": 0.7145102147721321,
+      "loss_partial": 0.0498046875,
+      "step": 1364
+    },
+    {
+      "epoch": 0.71503404924044,
+      "grad_norm": 1.9501521587371826,
+      "learning_rate": 1.962616385147671e-05,
+      "loss": 0.1089,
+      "step": 1365
+    },
+    {
+      "epoch": 0.71503404924044,
+      "loss_partial": 0.054443359375,
+      "step": 1365
+    },
+    {
+      "epoch": 0.71503404924044,
+      "loss_partial": 0.052490234375,
+      "step": 1365
+    },
+    {
+      "epoch": 0.7155578837087481,
+      "grad_norm": 2.1745121479034424,
+      "learning_rate": 1.9624924547879597e-05,
+      "loss": 0.1097,
+      "step": 1366
+    },
+    {
+      "epoch": 0.7155578837087481,
+      "loss_partial": 0.0732421875,
+      "step": 1366
+    },
+    {
+      "epoch": 0.7155578837087481,
+      "loss_partial": 0.0498046875,
+      "step": 1366
+    },
+    {
+      "epoch": 0.716081718177056,
+      "grad_norm": 2.2047717571258545,
+      "learning_rate": 1.962368323273277e-05,
+      "loss": 0.1142,
+      "step": 1367
+    },
+    {
+      "epoch": 0.716081718177056,
+      "loss_partial": 0.05517578125,
+      "step": 1367
+    },
+    {
+      "epoch": 0.716081718177056,
+      "loss_partial": 0.050537109375,
+      "step": 1367
+    },
+    {
+      "epoch": 0.7166055526453641,
+      "grad_norm": 2.1862120628356934,
+      "learning_rate": 1.962243990629567e-05,
+      "loss": 0.11,
+      "step": 1368
+    },
+    {
+      "epoch": 0.7166055526453641,
+      "loss_partial": 0.06396484375,
+      "step": 1368
+    },
+    {
+      "epoch": 0.7166055526453641,
+      "loss_partial": 0.05859375,
+      "step": 1368
+    },
+    {
+      "epoch": 0.7171293871136721,
+      "grad_norm": 1.920067548751831,
+      "learning_rate": 1.962119456882812e-05,
+      "loss": 0.1117,
+      "step": 1369
+    },
+    {
+      "epoch": 0.7171293871136721,
+      "loss_partial": 0.05419921875,
+      "step": 1369
+    },
+    {
+      "epoch": 0.7171293871136721,
+      "loss_partial": 0.0576171875,
+      "step": 1369
+    },
+    {
+      "epoch": 0.7176532215819801,
+      "grad_norm": 1.966367244720459,
+      "learning_rate": 1.961994722059041e-05,
+      "loss": 0.1104,
+      "step": 1370
+    },
+    {
+      "epoch": 0.7176532215819801,
+      "loss_partial": 0.050048828125,
+      "step": 1370
+    },
+    {
+      "epoch": 0.7176532215819801,
+      "loss_partial": 0.05615234375,
+      "step": 1370
+    },
+    {
+      "epoch": 0.7181770560502881,
+      "grad_norm": 2.3004627227783203,
+      "learning_rate": 1.9618697861843214e-05,
+      "loss": 0.1073,
+      "step": 1371
+    },
+    {
+      "epoch": 0.7181770560502881,
+      "loss_partial": 0.054931640625,
+      "step": 1371
+    },
+    {
+      "epoch": 0.7181770560502881,
+      "loss_partial": 0.06640625,
+      "step": 1371
+    },
+    {
+      "epoch": 0.7187008905185961,
+      "grad_norm": 2.2203333377838135,
+      "learning_rate": 1.9617446492847647e-05,
+      "loss": 0.1117,
+      "step": 1372
+    },
+    {
+      "epoch": 0.7187008905185961,
+      "loss_partial": 0.053955078125,
+      "step": 1372
+    },
+    {
+      "epoch": 0.7187008905185961,
+      "loss_partial": 0.055419921875,
+      "step": 1372
+    },
+    {
+      "epoch": 0.7192247249869042,
+      "grad_norm": 2.234727621078491,
+      "learning_rate": 1.9616193113865234e-05,
+      "loss": 0.109,
+      "step": 1373
+    },
+    {
+      "epoch": 0.7192247249869042,
+      "loss_partial": 0.056396484375,
+      "step": 1373
+    },
+    {
+      "epoch": 0.7192247249869042,
+      "loss_partial": 0.05712890625,
+      "step": 1373
+    },
+    {
+      "epoch": 0.7197485594552121,
+      "grad_norm": 2.0370235443115234,
+      "learning_rate": 1.9614937725157925e-05,
+      "loss": 0.1078,
+      "step": 1374
+    },
+    {
+      "epoch": 0.7197485594552121,
+      "loss_partial": 0.054931640625,
+      "step": 1374
+    },
+    {
+      "epoch": 0.7197485594552121,
+      "loss_partial": 0.060791015625,
+      "step": 1374
+    },
+    {
+      "epoch": 0.7202723939235202,
+      "grad_norm": 2.4532926082611084,
+      "learning_rate": 1.961368032698809e-05,
+      "loss": 0.1095,
+      "step": 1375
+    },
+    {
+      "epoch": 0.7202723939235202,
+      "loss_partial": 0.0546875,
+      "step": 1375
+    },
+    {
+      "epoch": 0.7202723939235202,
+      "loss_partial": 0.05908203125,
+      "step": 1375
+    },
+    {
+      "epoch": 0.7207962283918282,
+      "grad_norm": 2.427183151245117,
+      "learning_rate": 1.9612420919618514e-05,
+      "loss": 0.1089,
+      "step": 1376
+    },
+    {
+      "epoch": 0.7207962283918282,
+      "loss_partial": 0.060546875,
+      "step": 1376
+    },
+    {
+      "epoch": 0.7207962283918282,
+      "loss_partial": 0.05810546875,
+      "step": 1376
+    },
+    {
+      "epoch": 0.7213200628601362,
+      "grad_norm": 2.709961414337158,
+      "learning_rate": 1.961115950331241e-05,
+      "loss": 0.1163,
+      "step": 1377
+    },
+    {
+      "epoch": 0.7213200628601362,
+      "loss_partial": 0.049072265625,
+      "step": 1377
+    },
+    {
+      "epoch": 0.7213200628601362,
+      "loss_partial": 0.05078125,
+      "step": 1377
+    },
+    {
+      "epoch": 0.7218438973284442,
+      "grad_norm": 2.170250415802002,
+      "learning_rate": 1.96098960783334e-05,
+      "loss": 0.11,
+      "step": 1378
+    },
+    {
+      "epoch": 0.7218438973284442,
+      "loss_partial": 0.059326171875,
+      "step": 1378
+    },
+    {
+      "epoch": 0.7218438973284442,
+      "loss_partial": 0.05810546875,
+      "step": 1378
+    },
+    {
+      "epoch": 0.7223677317967522,
+      "grad_norm": 2.313993453979492,
+      "learning_rate": 1.9608630644945536e-05,
+      "loss": 0.1064,
+      "step": 1379
+    },
+    {
+      "epoch": 0.7223677317967522,
+      "loss_partial": 0.052490234375,
+      "step": 1379
+    },
+    {
+      "epoch": 0.7223677317967522,
+      "loss_partial": 0.0419921875,
+      "step": 1379
+    },
+    {
+      "epoch": 0.7228915662650602,
+      "grad_norm": 1.6695146560668945,
+      "learning_rate": 1.960736320341329e-05,
+      "loss": 0.1093,
+      "step": 1380
+    },
+    {
+      "epoch": 0.7228915662650602,
+      "loss_partial": 0.046142578125,
+      "step": 1380
+    },
+    {
+      "epoch": 0.7228915662650602,
+      "loss_partial": 0.048583984375,
+      "step": 1380
+    },
+    {
+      "epoch": 0.7234154007333683,
+      "grad_norm": 1.9107882976531982,
+      "learning_rate": 1.960609375400154e-05,
+      "loss": 0.1092,
+      "step": 1381
+    },
+    {
+      "epoch": 0.7234154007333683,
+      "loss_partial": 0.053466796875,
+      "step": 1381
+    },
+    {
+      "epoch": 0.7234154007333683,
+      "loss_partial": 0.06396484375,
+      "step": 1381
+    },
+    {
+      "epoch": 0.7239392352016762,
+      "grad_norm": 1.628051996231079,
+      "learning_rate": 1.9604822296975595e-05,
+      "loss": 0.1177,
+      "step": 1382
+    },
+    {
+      "epoch": 0.7239392352016762,
+      "loss_partial": 0.062255859375,
+      "step": 1382
+    },
+    {
+      "epoch": 0.7239392352016762,
+      "loss_partial": 0.054931640625,
+      "step": 1382
+    },
+    {
+      "epoch": 0.7244630696699843,
+      "grad_norm": 2.0720901489257812,
+      "learning_rate": 1.960354883260119e-05,
+      "loss": 0.1105,
+      "step": 1383
+    },
+    {
+      "epoch": 0.7244630696699843,
+      "loss_partial": 0.05322265625,
+      "step": 1383
+    },
+    {
+      "epoch": 0.7244630696699843,
+      "loss_partial": 0.06494140625,
+      "step": 1383
+    },
+    {
+      "epoch": 0.7249869041382923,
+      "grad_norm": 1.560210108757019,
+      "learning_rate": 1.9602273361144464e-05,
+      "loss": 0.1125,
+      "step": 1384
+    },
+    {
+      "epoch": 0.7249869041382923,
+      "loss_partial": 0.057861328125,
+      "step": 1384
+    },
+    {
+      "epoch": 0.7249869041382923,
+      "loss_partial": 0.056884765625,
+      "step": 1384
+    },
+    {
+      "epoch": 0.7255107386066003,
+      "grad_norm": 1.6694893836975098,
+      "learning_rate": 1.9600995882871982e-05,
+      "loss": 0.1095,
+      "step": 1385
+    },
+    {
+      "epoch": 0.7255107386066003,
+      "loss_partial": 0.0634765625,
+      "step": 1385
+    },
+    {
+      "epoch": 0.7255107386066003,
+      "loss_partial": 0.05859375,
+      "step": 1385
+    },
+    {
+      "epoch": 0.7260345730749084,
+      "grad_norm": 1.7989741563796997,
+      "learning_rate": 1.9599716398050735e-05,
+      "loss": 0.114,
+      "step": 1386
+    },
+    {
+      "epoch": 0.7260345730749084,
+      "loss_partial": 0.06298828125,
+      "step": 1386
+    },
+    {
+      "epoch": 0.7260345730749084,
+      "loss_partial": 0.04833984375,
+      "step": 1386
+    },
+    {
+      "epoch": 0.7265584075432163,
+      "grad_norm": 1.8033297061920166,
+      "learning_rate": 1.959843490694812e-05,
+      "loss": 0.1114,
+      "step": 1387
+    },
+    {
+      "epoch": 0.7265584075432163,
+      "loss_partial": 0.054443359375,
+      "step": 1387
+    },
+    {
+      "epoch": 0.7265584075432163,
+      "loss_partial": 0.060546875,
+      "step": 1387
+    },
+    {
+      "epoch": 0.7270822420115244,
+      "grad_norm": 1.8013675212860107,
+      "learning_rate": 1.9597151409831964e-05,
+      "loss": 0.1118,
+      "step": 1388
+    },
+    {
+      "epoch": 0.7270822420115244,
+      "loss_partial": 0.05322265625,
+      "step": 1388
+    },
+    {
+      "epoch": 0.7270822420115244,
+      "loss_partial": 0.052490234375,
+      "step": 1388
+    },
+    {
+      "epoch": 0.7276060764798323,
+      "grad_norm": 1.364149570465088,
+      "learning_rate": 1.9595865906970514e-05,
+      "loss": 0.1095,
+      "step": 1389
+    },
+    {
+      "epoch": 0.7276060764798323,
+      "loss_partial": 0.0537109375,
+      "step": 1389
+    },
+    {
+      "epoch": 0.7276060764798323,
+      "loss_partial": 0.0498046875,
+      "step": 1389
+    },
+    {
+      "epoch": 0.7281299109481404,
+      "grad_norm": 1.7287495136260986,
+      "learning_rate": 1.9594578398632428e-05,
+      "loss": 0.109,
+      "step": 1390
+    },
+    {
+      "epoch": 0.7281299109481404,
+      "loss_partial": 0.060302734375,
+      "step": 1390
+    },
+    {
+      "epoch": 0.7281299109481404,
+      "loss_partial": 0.0634765625,
+      "step": 1390
+    },
+    {
+      "epoch": 0.7286537454164485,
+      "grad_norm": 1.7031515836715698,
+      "learning_rate": 1.9593288885086785e-05,
+      "loss": 0.1095,
+      "step": 1391
+    },
+    {
+      "epoch": 0.7286537454164485,
+      "loss_partial": 0.05419921875,
+      "step": 1391
+    },
+    {
+      "epoch": 0.7286537454164485,
+      "loss_partial": 0.05078125,
+      "step": 1391
+    },
+    {
+      "epoch": 0.7291775798847564,
+      "grad_norm": 1.8953754901885986,
+      "learning_rate": 1.959199736660309e-05,
+      "loss": 0.1092,
+      "step": 1392
+    },
+    {
+      "epoch": 0.7291775798847564,
+      "loss_partial": 0.056396484375,
+      "step": 1392
+    },
+    {
+      "epoch": 0.7291775798847564,
+      "loss_partial": 0.046142578125,
+      "step": 1392
+    },
+    {
+      "epoch": 0.7297014143530645,
+      "grad_norm": 1.3551172018051147,
+      "learning_rate": 1.9590703843451262e-05,
+      "loss": 0.1063,
+      "step": 1393
+    },
+    {
+      "epoch": 0.7297014143530645,
+      "loss_partial": 0.056640625,
+      "step": 1393
+    },
+    {
+      "epoch": 0.7297014143530645,
+      "loss_partial": 0.055908203125,
+      "step": 1393
+    },
+    {
+      "epoch": 0.7302252488213724,
+      "grad_norm": 1.9933593273162842,
+      "learning_rate": 1.9589408315901633e-05,
+      "loss": 0.1089,
+      "step": 1394
+    },
+    {
+      "epoch": 0.7302252488213724,
+      "loss_partial": 0.05712890625,
+      "step": 1394
+    },
+    {
+      "epoch": 0.7302252488213724,
+      "loss_partial": 0.054931640625,
+      "step": 1394
+    },
+    {
+      "epoch": 0.7307490832896805,
+      "grad_norm": 1.801969289779663,
+      "learning_rate": 1.958811078422497e-05,
+      "loss": 0.1101,
+      "step": 1395
+    },
+    {
+      "epoch": 0.7307490832896805,
+      "loss_partial": 0.053466796875,
+      "step": 1395
+    },
+    {
+      "epoch": 0.7307490832896805,
+      "loss_partial": 0.048583984375,
+      "step": 1395
+    },
+    {
+      "epoch": 0.7312729177579885,
+      "grad_norm": 1.9214458465576172,
+      "learning_rate": 1.9586811248692445e-05,
+      "loss": 0.1123,
+      "step": 1396
+    },
+    {
+      "epoch": 0.7312729177579885,
+      "loss_partial": 0.044189453125,
+      "step": 1396
+    },
+    {
+      "epoch": 0.7312729177579885,
+      "loss_partial": 0.0625,
+      "step": 1396
+    },
+    {
+      "epoch": 0.7317967522262965,
+      "grad_norm": 1.3060280084609985,
+      "learning_rate": 1.9585509709575646e-05,
+      "loss": 0.1094,
+      "step": 1397
+    },
+    {
+      "epoch": 0.7317967522262965,
+      "loss_partial": 0.054931640625,
+      "step": 1397
+    },
+    {
+      "epoch": 0.7317967522262965,
+      "loss_partial": 0.05810546875,
+      "step": 1397
+    },
+    {
+      "epoch": 0.7323205866946045,
+      "grad_norm": 1.4379990100860596,
+      "learning_rate": 1.9584206167146598e-05,
+      "loss": 0.1089,
+      "step": 1398
+    },
+    {
+      "epoch": 0.7323205866946045,
+      "loss_partial": 0.05810546875,
+      "step": 1398
+    },
+    {
+      "epoch": 0.7323205866946045,
+      "loss_partial": 0.05615234375,
+      "step": 1398
+    },
+    {
+      "epoch": 0.7328444211629125,
+      "grad_norm": 1.8553816080093384,
+      "learning_rate": 1.9582900621677726e-05,
+      "loss": 0.1087,
+      "step": 1399
+    },
+    {
+      "epoch": 0.7328444211629125,
+      "loss_partial": 0.053955078125,
+      "step": 1399
+    },
+    {
+      "epoch": 0.7328444211629125,
+      "loss_partial": 0.050537109375,
+      "step": 1399
+    },
+    {
+      "epoch": 0.7333682556312205,
+      "grad_norm": 2.0846681594848633,
+      "learning_rate": 1.9581593073441883e-05,
+      "loss": 0.1123,
+      "step": 1400
+    },
+    {
+      "epoch": 0.7333682556312205,
+      "loss_partial": 0.052001953125,
+      "step": 1400
+    },
+    {
+      "epoch": 0.7333682556312205,
+      "loss_partial": 0.048583984375,
+      "step": 1400
+    },
+    {
+      "epoch": 0.7338920900995285,
+      "grad_norm": 1.1947726011276245,
+      "learning_rate": 1.958028352271234e-05,
+      "loss": 0.1113,
+      "step": 1401
+    },
+    {
+      "epoch": 0.7338920900995285,
+      "loss_partial": 0.056640625,
+      "step": 1401
+    },
+    {
+      "epoch": 0.7338920900995285,
+      "loss_partial": 0.055908203125,
+      "step": 1401
+    },
+    {
+      "epoch": 0.7344159245678366,
+      "grad_norm": 1.6170828342437744,
+      "learning_rate": 1.957897196976278e-05,
+      "loss": 0.1136,
+      "step": 1402
+    },
+    {
+      "epoch": 0.7344159245678366,
+      "loss_partial": 0.05615234375,
+      "step": 1402
+    },
+    {
+      "epoch": 0.7344159245678366,
+      "loss_partial": 0.052001953125,
+      "step": 1402
+    },
+    {
+      "epoch": 0.7349397590361446,
+      "grad_norm": 1.753695011138916,
+      "learning_rate": 1.9577658414867314e-05,
+      "loss": 0.1075,
+      "step": 1403
+    },
+    {
+      "epoch": 0.7349397590361446,
+      "loss_partial": 0.050048828125,
+      "step": 1403
+    },
+    {
+      "epoch": 0.7349397590361446,
+      "loss_partial": 0.0556640625,
+      "step": 1403
+    },
+    {
+      "epoch": 0.7354635935044526,
+      "grad_norm": 1.7336214780807495,
+      "learning_rate": 1.9576342858300467e-05,
+      "loss": 0.1112,
+      "step": 1404
+    },
+    {
+      "epoch": 0.7354635935044526,
+      "loss_partial": 0.058349609375,
+      "step": 1404
+    },
+    {
+      "epoch": 0.7354635935044526,
+      "loss_partial": 0.06103515625,
+      "step": 1404
+    },
+    {
+      "epoch": 0.7359874279727606,
+      "grad_norm": 2.359816074371338,
+      "learning_rate": 1.957502530033718e-05,
+      "loss": 0.1129,
+      "step": 1405
+    },
+    {
+      "epoch": 0.7359874279727606,
+      "loss_partial": 0.046630859375,
+      "step": 1405
+    },
+    {
+      "epoch": 0.7359874279727606,
+      "loss_partial": 0.055908203125,
+      "step": 1405
+    },
+    {
+      "epoch": 0.7365112624410686,
+      "grad_norm": 2.632108688354492,
+      "learning_rate": 1.957370574125282e-05,
+      "loss": 0.1116,
+      "step": 1406
+    },
+    {
+      "epoch": 0.7365112624410686,
+      "loss_partial": 0.058349609375,
+      "step": 1406
+    },
+    {
+      "epoch": 0.7365112624410686,
+      "loss_partial": 0.056640625,
+      "step": 1406
+    },
+    {
+      "epoch": 0.7370350969093766,
+      "grad_norm": 1.9512722492218018,
+      "learning_rate": 1.9572384181323152e-05,
+      "loss": 0.1117,
+      "step": 1407
+    },
+    {
+      "epoch": 0.7370350969093766,
+      "loss_partial": 0.052001953125,
+      "step": 1407
+    },
+    {
+      "epoch": 0.7370350969093766,
+      "loss_partial": 0.0576171875,
+      "step": 1407
+    },
+    {
+      "epoch": 0.7375589313776847,
+      "grad_norm": 2.2826874256134033,
+      "learning_rate": 1.9571060620824393e-05,
+      "loss": 0.1123,
+      "step": 1408
+    },
+    {
+      "epoch": 0.7375589313776847,
+      "loss_partial": 0.05859375,
+      "step": 1408
+    },
+    {
+      "epoch": 0.7375589313776847,
+      "loss_partial": 0.05126953125,
+      "step": 1408
+    },
+    {
+      "epoch": 0.7380827658459926,
+      "grad_norm": 2.327153444290161,
+      "learning_rate": 1.9569735060033148e-05,
+      "loss": 0.113,
+      "step": 1409
+    },
+    {
+      "epoch": 0.7380827658459926,
+      "loss_partial": 0.052978515625,
+      "step": 1409
+    },
+    {
+      "epoch": 0.7380827658459926,
+      "loss_partial": 0.0576171875,
+      "step": 1409
+    },
+    {
+      "epoch": 0.7386066003143007,
+      "grad_norm": 2.7572503089904785,
+      "learning_rate": 1.9568407499226453e-05,
+      "loss": 0.1101,
+      "step": 1410
+    },
+    {
+      "epoch": 0.7386066003143007,
+      "loss_partial": 0.048583984375,
+      "step": 1410
+    },
+    {
+      "epoch": 0.7386066003143007,
+      "loss_partial": 0.05712890625,
+      "step": 1410
+    },
+    {
+      "epoch": 0.7391304347826086,
+      "grad_norm": 2.113771677017212,
+      "learning_rate": 1.9567077938681754e-05,
+      "loss": 0.1144,
+      "step": 1411
+    },
+    {
+      "epoch": 0.7391304347826086,
+      "loss_partial": 0.046142578125,
+      "step": 1411
+    },
+    {
+      "epoch": 0.7391304347826086,
+      "loss_partial": 0.0595703125,
+      "step": 1411
+    },
+    {
+      "epoch": 0.7396542692509167,
+      "grad_norm": 2.1874797344207764,
+      "learning_rate": 1.9565746378676932e-05,
+      "loss": 0.1055,
+      "step": 1412
+    },
+    {
+      "epoch": 0.7396542692509167,
+      "loss_partial": 0.05908203125,
+      "step": 1412
+    },
+    {
+      "epoch": 0.7396542692509167,
+      "loss_partial": 0.048095703125,
+      "step": 1412
+    },
+    {
+      "epoch": 0.7401781037192248,
+      "grad_norm": 2.7077674865722656,
+      "learning_rate": 1.956441281949027e-05,
+      "loss": 0.1085,
+      "step": 1413
+    },
+    {
+      "epoch": 0.7401781037192248,
+      "loss_partial": 0.0556640625,
+      "step": 1413
+    },
+    {
+      "epoch": 0.7401781037192248,
+      "loss_partial": 0.05029296875,
+      "step": 1413
+    },
+    {
+      "epoch": 0.7407019381875327,
+      "grad_norm": 1.700701355934143,
+      "learning_rate": 1.9563077261400472e-05,
+      "loss": 0.1083,
+      "step": 1414
+    },
+    {
+      "epoch": 0.7407019381875327,
+      "loss_partial": 0.052734375,
+      "step": 1414
+    },
+    {
+      "epoch": 0.7407019381875327,
+      "loss_partial": 0.05322265625,
+      "step": 1414
+    },
+    {
+      "epoch": 0.7412257726558408,
+      "grad_norm": 2.145089864730835,
+      "learning_rate": 1.9561739704686668e-05,
+      "loss": 0.1096,
+      "step": 1415
+    },
+    {
+      "epoch": 0.7412257726558408,
+      "loss_partial": 0.053955078125,
+      "step": 1415
+    },
+    {
+      "epoch": 0.7412257726558408,
+      "loss_partial": 0.047119140625,
+      "step": 1415
+    },
+    {
+      "epoch": 0.7417496071241487,
+      "grad_norm": 1.6497000455856323,
+      "learning_rate": 1.9560400149628384e-05,
+      "loss": 0.1131,
+      "step": 1416
+    },
+    {
+      "epoch": 0.7417496071241487,
+      "loss_partial": 0.052978515625,
+      "step": 1416
+    },
+    {
+      "epoch": 0.7417496071241487,
+      "loss_partial": 0.061279296875,
+      "step": 1416
+    },
+    {
+      "epoch": 0.7422734415924568,
+      "grad_norm": 1.9650774002075195,
+      "learning_rate": 1.9559058596505595e-05,
+      "loss": 0.1147,
+      "step": 1417
+    },
+    {
+      "epoch": 0.7422734415924568,
+      "loss_partial": 0.049560546875,
+      "step": 1417
+    },
+    {
+      "epoch": 0.7422734415924568,
+      "loss_partial": 0.048583984375,
+      "step": 1417
+    },
+    {
+      "epoch": 0.7427972760607648,
+      "grad_norm": 2.1759629249572754,
+      "learning_rate": 1.955771504559867e-05,
+      "loss": 0.1107,
+      "step": 1418
+    },
+    {
+      "epoch": 0.7427972760607648,
+      "loss_partial": 0.045654296875,
+      "step": 1418
+    },
+    {
+      "epoch": 0.7427972760607648,
+      "loss_partial": 0.061279296875,
+      "step": 1418
+    },
+    {
+      "epoch": 0.7433211105290728,
+      "grad_norm": 2.452965497970581,
+      "learning_rate": 1.9556369497188404e-05,
+      "loss": 0.1107,
+      "step": 1419
+    },
+    {
+      "epoch": 0.7433211105290728,
+      "loss_partial": 0.058837890625,
+      "step": 1419
+    },
+    {
+      "epoch": 0.7433211105290728,
+      "loss_partial": 0.047607421875,
+      "step": 1419
+    },
+    {
+      "epoch": 0.7438449449973809,
+      "grad_norm": 1.7390244007110596,
+      "learning_rate": 1.9555021951556002e-05,
+      "loss": 0.1098,
+      "step": 1420
+    },
+    {
+      "epoch": 0.7438449449973809,
+      "loss_partial": 0.0703125,
+      "step": 1420
+    },
+    {
+      "epoch": 0.7438449449973809,
+      "loss_partial": 0.0634765625,
+      "step": 1420
+    },
+    {
+      "epoch": 0.7443687794656888,
+      "grad_norm": 1.8951908349990845,
+      "learning_rate": 1.9553672408983105e-05,
+      "loss": 0.113,
+      "step": 1421
+    },
+    {
+      "epoch": 0.7443687794656888,
+      "loss_partial": 0.05615234375,
+      "step": 1421
+    },
+    {
+      "epoch": 0.7443687794656888,
+      "loss_partial": 0.048828125,
+      "step": 1421
+    },
+    {
+      "epoch": 0.7448926139339969,
+      "grad_norm": 1.9780988693237305,
+      "learning_rate": 1.9552320869751754e-05,
+      "loss": 0.1096,
+      "step": 1422
+    },
+    {
+      "epoch": 0.7448926139339969,
+      "loss_partial": 0.04931640625,
+      "step": 1422
+    },
+    {
+      "epoch": 0.7448926139339969,
+      "loss_partial": 0.05517578125,
+      "step": 1422
+    },
+    {
+      "epoch": 0.7454164484023049,
+      "grad_norm": 1.918622612953186,
+      "learning_rate": 1.9550967334144407e-05,
+      "loss": 0.1082,
+      "step": 1423
+    },
+    {
+      "epoch": 0.7454164484023049,
+      "loss_partial": 0.050048828125,
+      "step": 1423
+    },
+    {
+      "epoch": 0.7454164484023049,
+      "loss_partial": 0.052734375,
+      "step": 1423
+    },
+    {
+      "epoch": 0.7459402828706129,
+      "grad_norm": 1.5721582174301147,
+      "learning_rate": 1.9549611802443947e-05,
+      "loss": 0.1138,
+      "step": 1424
+    },
+    {
+      "epoch": 0.7459402828706129,
+      "loss_partial": 0.055908203125,
+      "step": 1424
+    },
+    {
+      "epoch": 0.7459402828706129,
+      "loss_partial": 0.05517578125,
+      "step": 1424
+    },
+    {
+      "epoch": 0.7464641173389209,
+      "grad_norm": 2.62265682220459,
+      "learning_rate": 1.9548254274933676e-05,
+      "loss": 0.1109,
+      "step": 1425
+    },
+    {
+      "epoch": 0.7464641173389209,
+      "loss_partial": 0.06005859375,
+      "step": 1425
+    },
+    {
+      "epoch": 0.7464641173389209,
+      "loss_partial": 0.045654296875,
+      "step": 1425
+    },
+    {
+      "epoch": 0.7469879518072289,
+      "grad_norm": 3.056884765625,
+      "learning_rate": 1.9546894751897304e-05,
+      "loss": 0.1064,
+      "step": 1426
+    },
+    {
+      "epoch": 0.7469879518072289,
+      "loss_partial": 0.057861328125,
+      "step": 1426
+    },
+    {
+      "epoch": 0.7469879518072289,
+      "loss_partial": 0.054931640625,
+      "step": 1426
+    },
+    {
+      "epoch": 0.7475117862755369,
+      "grad_norm": 1.7513474225997925,
+      "learning_rate": 1.954553323361897e-05,
+      "loss": 0.1093,
+      "step": 1427
+    },
+    {
+      "epoch": 0.7475117862755369,
+      "loss_partial": 0.052978515625,
+      "step": 1427
+    },
+    {
+      "epoch": 0.7475117862755369,
+      "loss_partial": 0.04931640625,
+      "step": 1427
+    },
+    {
+      "epoch": 0.748035620743845,
+      "grad_norm": 2.212066888809204,
+      "learning_rate": 1.9544169720383216e-05,
+      "loss": 0.1095,
+      "step": 1428
+    },
+    {
+      "epoch": 0.748035620743845,
+      "loss_partial": 0.053955078125,
+      "step": 1428
+    },
+    {
+      "epoch": 0.748035620743845,
+      "loss_partial": 0.05859375,
+      "step": 1428
+    },
+    {
+      "epoch": 0.748559455212153,
+      "grad_norm": 1.4558281898498535,
+      "learning_rate": 1.954280421247501e-05,
+      "loss": 0.1068,
+      "step": 1429
+    },
+    {
+      "epoch": 0.748559455212153,
+      "loss_partial": 0.048828125,
+      "step": 1429
+    },
+    {
+      "epoch": 0.748559455212153,
+      "loss_partial": 0.054443359375,
+      "step": 1429
+    },
+    {
+      "epoch": 0.749083289680461,
+      "grad_norm": 2.8956024646759033,
+      "learning_rate": 1.9541436710179732e-05,
+      "loss": 0.1099,
+      "step": 1430
+    },
+    {
+      "epoch": 0.749083289680461,
+      "loss_partial": 0.06005859375,
+      "step": 1430
+    },
+    {
+      "epoch": 0.749083289680461,
+      "loss_partial": 0.053466796875,
+      "step": 1430
+    },
+    {
+      "epoch": 0.749607124148769,
+      "grad_norm": 2.5729734897613525,
+      "learning_rate": 1.9540067213783183e-05,
+      "loss": 0.1112,
+      "step": 1431
+    },
+    {
+      "epoch": 0.749607124148769,
+      "loss_partial": 0.054931640625,
+      "step": 1431
+    },
+    {
+      "epoch": 0.749607124148769,
+      "loss_partial": 0.047119140625,
+      "step": 1431
+    },
+    {
+      "epoch": 0.750130958617077,
+      "grad_norm": 2.2639172077178955,
+      "learning_rate": 1.9538695723571583e-05,
+      "loss": 0.112,
+      "step": 1432
+    },
+    {
+      "epoch": 0.750130958617077,
+      "loss_partial": 0.051025390625,
+      "step": 1432
+    },
+    {
+      "epoch": 0.750130958617077,
+      "loss_partial": 0.05029296875,
+      "step": 1432
+    },
+    {
+      "epoch": 0.750654793085385,
+      "grad_norm": 2.04811954498291,
+      "learning_rate": 1.9537322239831562e-05,
+      "loss": 0.1101,
+      "step": 1433
+    },
+    {
+      "epoch": 0.750654793085385,
+      "loss_partial": 0.052490234375,
+      "step": 1433
+    },
+    {
+      "epoch": 0.750654793085385,
+      "loss_partial": 0.0546875,
+      "step": 1433
+    },
+    {
+      "epoch": 0.751178627553693,
+      "grad_norm": 1.9048322439193726,
+      "learning_rate": 1.953594676285017e-05,
+      "loss": 0.1081,
+      "step": 1434
+    },
+    {
+      "epoch": 0.751178627553693,
+      "loss_partial": 0.050537109375,
+      "step": 1434
+    },
+    {
+      "epoch": 0.751178627553693,
+      "loss_partial": 0.0595703125,
+      "step": 1434
+    },
+    {
+      "epoch": 0.7517024620220011,
+      "grad_norm": 1.8795716762542725,
+      "learning_rate": 1.9534569292914872e-05,
+      "loss": 0.1072,
+      "step": 1435
+    },
+    {
+      "epoch": 0.7517024620220011,
+      "loss_partial": 0.057861328125,
+      "step": 1435
+    },
+    {
+      "epoch": 0.7517024620220011,
+      "loss_partial": 0.0673828125,
+      "step": 1435
+    },
+    {
+      "epoch": 0.752226296490309,
+      "grad_norm": 1.6702489852905273,
+      "learning_rate": 1.9533189830313555e-05,
+      "loss": 0.1107,
+      "step": 1436
+    },
+    {
+      "epoch": 0.752226296490309,
+      "loss_partial": 0.061767578125,
+      "step": 1436
+    },
+    {
+      "epoch": 0.752226296490309,
+      "loss_partial": 0.046875,
+      "step": 1436
+    },
+    {
+      "epoch": 0.7527501309586171,
+      "grad_norm": 1.8661490678787231,
+      "learning_rate": 1.9531808375334512e-05,
+      "loss": 0.1086,
+      "step": 1437
+    },
+    {
+      "epoch": 0.7527501309586171,
+      "loss_partial": 0.05859375,
+      "step": 1437
+    },
+    {
+      "epoch": 0.7527501309586171,
+      "loss_partial": 0.05224609375,
+      "step": 1437
+    },
+    {
+      "epoch": 0.753273965426925,
+      "grad_norm": 1.611624002456665,
+      "learning_rate": 1.9530424928266464e-05,
+      "loss": 0.1062,
+      "step": 1438
+    },
+    {
+      "epoch": 0.753273965426925,
+      "loss_partial": 0.06298828125,
+      "step": 1438
+    },
+    {
+      "epoch": 0.753273965426925,
+      "loss_partial": 0.055908203125,
+      "step": 1438
+    },
+    {
+      "epoch": 0.7537977998952331,
+      "grad_norm": 2.443760395050049,
+      "learning_rate": 1.9529039489398536e-05,
+      "loss": 0.1079,
+      "step": 1439
+    },
+    {
+      "epoch": 0.7537977998952331,
+      "loss_partial": 0.052734375,
+      "step": 1439
+    },
+    {
+      "epoch": 0.7537977998952331,
+      "loss_partial": 0.0546875,
+      "step": 1439
+    },
+    {
+      "epoch": 0.7543216343635412,
+      "grad_norm": 2.2579903602600098,
+      "learning_rate": 1.9527652059020285e-05,
+      "loss": 0.1068,
+      "step": 1440
+    },
+    {
+      "epoch": 0.7543216343635412,
+      "loss_partial": 0.048583984375,
+      "step": 1440
+    },
+    {
+      "epoch": 0.7543216343635412,
+      "loss_partial": 0.0498046875,
+      "step": 1440
+    },
+    {
+      "epoch": 0.7548454688318491,
+      "grad_norm": 2.7953057289123535,
+      "learning_rate": 1.952626263742167e-05,
+      "loss": 0.11,
+      "step": 1441
+    },
+    {
+      "epoch": 0.7548454688318491,
+      "loss_partial": 0.0576171875,
+      "step": 1441
+    },
+    {
+      "epoch": 0.7548454688318491,
+      "loss_partial": 0.052001953125,
+      "step": 1441
+    },
+    {
+      "epoch": 0.7553693033001572,
+      "grad_norm": 1.912420630455017,
+      "learning_rate": 1.952487122489307e-05,
+      "loss": 0.1079,
+      "step": 1442
+    },
+    {
+      "epoch": 0.7553693033001572,
+      "loss_partial": 0.05126953125,
+      "step": 1442
+    },
+    {
+      "epoch": 0.7553693033001572,
+      "loss_partial": 0.06298828125,
+      "step": 1442
+    },
+    {
+      "epoch": 0.7558931377684651,
+      "grad_norm": 2.2270047664642334,
+      "learning_rate": 1.9523477821725284e-05,
+      "loss": 0.1135,
+      "step": 1443
+    },
+    {
+      "epoch": 0.7558931377684651,
+      "loss_partial": 0.05712890625,
+      "step": 1443
+    },
+    {
+      "epoch": 0.7558931377684651,
+      "loss_partial": 0.054931640625,
+      "step": 1443
+    },
+    {
+      "epoch": 0.7564169722367732,
+      "grad_norm": 2.140596866607666,
+      "learning_rate": 1.9522082428209525e-05,
+      "loss": 0.1071,
+      "step": 1444
+    },
+    {
+      "epoch": 0.7564169722367732,
+      "loss_partial": 0.056884765625,
+      "step": 1444
+    },
+    {
+      "epoch": 0.7564169722367732,
+      "loss_partial": 0.056640625,
+      "step": 1444
+    },
+    {
+      "epoch": 0.7569408067050812,
+      "grad_norm": 2.31721830368042,
+      "learning_rate": 1.9520685044637423e-05,
+      "loss": 0.1083,
+      "step": 1445
+    },
+    {
+      "epoch": 0.7569408067050812,
+      "loss_partial": 0.05859375,
+      "step": 1445
+    },
+    {
+      "epoch": 0.7569408067050812,
+      "loss_partial": 0.06396484375,
+      "step": 1445
+    },
+    {
+      "epoch": 0.7574646411733892,
+      "grad_norm": 2.200930595397949,
+      "learning_rate": 1.951928567130102e-05,
+      "loss": 0.1099,
+      "step": 1446
+    },
+    {
+      "epoch": 0.7574646411733892,
+      "loss_partial": 0.06298828125,
+      "step": 1446
+    },
+    {
+      "epoch": 0.7574646411733892,
+      "loss_partial": 0.06494140625,
+      "step": 1446
+    },
+    {
+      "epoch": 0.7579884756416972,
+      "grad_norm": 2.099431276321411,
+      "learning_rate": 1.951788430849278e-05,
+      "loss": 0.1138,
+      "step": 1447
+    },
+    {
+      "epoch": 0.7579884756416972,
+      "loss_partial": 0.053955078125,
+      "step": 1447
+    },
+    {
+      "epoch": 0.7579884756416972,
+      "loss_partial": 0.056396484375,
+      "step": 1447
+    },
+    {
+      "epoch": 0.7585123101100052,
+      "grad_norm": 3.7274229526519775,
+      "learning_rate": 1.951648095650557e-05,
+      "loss": 0.1112,
+      "step": 1448
+    },
+    {
+      "epoch": 0.7585123101100052,
+      "loss_partial": 0.056640625,
+      "step": 1448
+    },
+    {
+      "epoch": 0.7585123101100052,
+      "loss_partial": 0.054931640625,
+      "step": 1448
+    },
+    {
+      "epoch": 0.7590361445783133,
+      "grad_norm": 3.009591579437256,
+      "learning_rate": 1.951507561563269e-05,
+      "loss": 0.1108,
+      "step": 1449
+    },
+    {
+      "epoch": 0.7590361445783133,
+      "loss_partial": 0.0654296875,
+      "step": 1449
+    },
+    {
+      "epoch": 0.7590361445783133,
+      "loss_partial": 0.05810546875,
+      "step": 1449
+    },
+    {
+      "epoch": 0.7595599790466213,
+      "grad_norm": 2.512664556503296,
+      "learning_rate": 1.9513668286167854e-05,
+      "loss": 0.1133,
+      "step": 1450
+    },
+    {
+      "epoch": 0.7595599790466213,
+      "loss_partial": 0.0615234375,
+      "step": 1450
+    },
+    {
+      "epoch": 0.7595599790466213,
+      "loss_partial": 0.046630859375,
+      "step": 1450
+    },
+    {
+      "epoch": 0.7600838135149293,
+      "grad_norm": 2.4440932273864746,
+      "learning_rate": 1.9512258968405167e-05,
+      "loss": 0.1075,
+      "step": 1451
+    },
+    {
+      "epoch": 0.7600838135149293,
+      "loss_partial": 0.052490234375,
+      "step": 1451
+    },
+    {
+      "epoch": 0.7600838135149293,
+      "loss_partial": 0.054443359375,
+      "step": 1451
+    },
+    {
+      "epoch": 0.7606076479832373,
+      "grad_norm": 1.927240252494812,
+      "learning_rate": 1.9510847662639182e-05,
+      "loss": 0.115,
+      "step": 1452
+    },
+    {
+      "epoch": 0.7606076479832373,
+      "loss_partial": 0.04931640625,
+      "step": 1452
+    },
+    {
+      "epoch": 0.7606076479832373,
+      "loss_partial": 0.055419921875,
+      "step": 1452
+    },
+    {
+      "epoch": 0.7611314824515453,
+      "grad_norm": 2.0151844024658203,
+      "learning_rate": 1.950943436916485e-05,
+      "loss": 0.1086,
+      "step": 1453
+    },
+    {
+      "epoch": 0.7611314824515453,
+      "loss_partial": 0.0615234375,
+      "step": 1453
+    },
+    {
+      "epoch": 0.7611314824515453,
+      "loss_partial": 0.056396484375,
+      "step": 1453
+    },
+    {
+      "epoch": 0.7616553169198533,
+      "grad_norm": 1.591888427734375,
+      "learning_rate": 1.9508019088277535e-05,
+      "loss": 0.1088,
+      "step": 1454
+    },
+    {
+      "epoch": 0.7616553169198533,
+      "loss_partial": 0.04736328125,
+      "step": 1454
+    },
+    {
+      "epoch": 0.7616553169198533,
+      "loss_partial": 0.052978515625,
+      "step": 1454
+    },
+    {
+      "epoch": 0.7621791513881614,
+      "grad_norm": 2.2603752613067627,
+      "learning_rate": 1.9506601820273032e-05,
+      "loss": 0.1093,
+      "step": 1455
+    },
+    {
+      "epoch": 0.7621791513881614,
+      "loss_partial": 0.06689453125,
+      "step": 1455
+    },
+    {
+      "epoch": 0.7621791513881614,
+      "loss_partial": 0.056640625,
+      "step": 1455
+    },
+    {
+      "epoch": 0.7627029858564693,
+      "grad_norm": 1.8843005895614624,
+      "learning_rate": 1.950518256544753e-05,
+      "loss": 0.1087,
+      "step": 1456
+    },
+    {
+      "epoch": 0.7627029858564693,
+      "loss_partial": 0.059326171875,
+      "step": 1456
+    },
+    {
+      "epoch": 0.7627029858564693,
+      "loss_partial": 0.056884765625,
+      "step": 1456
+    },
+    {
+      "epoch": 0.7632268203247774,
+      "grad_norm": 2.0132875442504883,
+      "learning_rate": 1.9503761324097653e-05,
+      "loss": 0.1137,
+      "step": 1457
+    },
+    {
+      "epoch": 0.7632268203247774,
+      "loss_partial": 0.051513671875,
+      "step": 1457
+    },
+    {
+      "epoch": 0.7632268203247774,
+      "loss_partial": 0.046630859375,
+      "step": 1457
+    },
+    {
+      "epoch": 0.7637506547930853,
+      "grad_norm": 1.704040288925171,
+      "learning_rate": 1.950233809652043e-05,
+      "loss": 0.109,
+      "step": 1458
+    },
+    {
+      "epoch": 0.7637506547930853,
+      "loss_partial": 0.06103515625,
+      "step": 1458
+    },
+    {
+      "epoch": 0.7637506547930853,
+      "loss_partial": 0.052001953125,
+      "step": 1458
+    },
+    {
+      "epoch": 0.7642744892613934,
+      "grad_norm": 1.649174690246582,
+      "learning_rate": 1.9500912883013302e-05,
+      "loss": 0.1074,
+      "step": 1459
+    },
+    {
+      "epoch": 0.7642744892613934,
+      "loss_partial": 0.06201171875,
+      "step": 1459
+    },
+    {
+      "epoch": 0.7642744892613934,
+      "loss_partial": 0.052490234375,
+      "step": 1459
+    },
+    {
+      "epoch": 0.7647983237297015,
+      "grad_norm": 1.5956083536148071,
+      "learning_rate": 1.9499485683874135e-05,
+      "loss": 0.1101,
+      "step": 1460
+    },
+    {
+      "epoch": 0.7647983237297015,
+      "loss_partial": 0.04931640625,
+      "step": 1460
+    },
+    {
+      "epoch": 0.7647983237297015,
+      "loss_partial": 0.07080078125,
+      "step": 1460
+    },
+    {
+      "epoch": 0.7653221581980094,
+      "grad_norm": 1.6502001285552979,
+      "learning_rate": 1.9498056499401202e-05,
+      "loss": 0.1086,
+      "step": 1461
+    },
+    {
+      "epoch": 0.7653221581980094,
+      "loss_partial": 0.044677734375,
+      "step": 1461
+    },
+    {
+      "epoch": 0.7653221581980094,
+      "loss_partial": 0.06640625,
+      "step": 1461
+    },
+    {
+      "epoch": 0.7658459926663175,
+      "grad_norm": 1.8507616519927979,
+      "learning_rate": 1.9496625329893198e-05,
+      "loss": 0.11,
+      "step": 1462
+    },
+    {
+      "epoch": 0.7658459926663175,
+      "loss_partial": 0.052490234375,
+      "step": 1462
+    },
+    {
+      "epoch": 0.7658459926663175,
+      "loss_partial": 0.05859375,
+      "step": 1462
+    },
+    {
+      "epoch": 0.7663698271346254,
+      "grad_norm": 1.3788691759109497,
+      "learning_rate": 1.9495192175649223e-05,
+      "loss": 0.1095,
+      "step": 1463
+    },
+    {
+      "epoch": 0.7663698271346254,
+      "loss_partial": 0.0546875,
+      "step": 1463
+    },
+    {
+      "epoch": 0.7663698271346254,
+      "loss_partial": 0.061767578125,
+      "step": 1463
+    },
+    {
+      "epoch": 0.7668936616029335,
+      "grad_norm": 2.0307817459106445,
+      "learning_rate": 1.9493757036968802e-05,
+      "loss": 0.1132,
+      "step": 1464
+    },
+    {
+      "epoch": 0.7668936616029335,
+      "loss_partial": 0.052001953125,
+      "step": 1464
+    },
+    {
+      "epoch": 0.7668936616029335,
+      "loss_partial": 0.054931640625,
+      "step": 1464
+    },
+    {
+      "epoch": 0.7674174960712414,
+      "grad_norm": 2.085106134414673,
+      "learning_rate": 1.9492319914151863e-05,
+      "loss": 0.1089,
+      "step": 1465
+    },
+    {
+      "epoch": 0.7674174960712414,
+      "loss_partial": 0.04638671875,
+      "step": 1465
+    },
+    {
+      "epoch": 0.7674174960712414,
+      "loss_partial": 0.052978515625,
+      "step": 1465
+    },
+    {
+      "epoch": 0.7679413305395495,
+      "grad_norm": 1.723461389541626,
+      "learning_rate": 1.9490880807498767e-05,
+      "loss": 0.1082,
+      "step": 1466
+    },
+    {
+      "epoch": 0.7679413305395495,
+      "loss_partial": 0.049072265625,
+      "step": 1466
+    },
+    {
+      "epoch": 0.7679413305395495,
+      "loss_partial": 0.048095703125,
+      "step": 1466
+    },
+    {
+      "epoch": 0.7684651650078576,
+      "grad_norm": 1.9330792427062988,
+      "learning_rate": 1.9489439717310268e-05,
+      "loss": 0.1044,
+      "step": 1467
+    },
+    {
+      "epoch": 0.7684651650078576,
+      "loss_partial": 0.059814453125,
+      "step": 1467
+    },
+    {
+      "epoch": 0.7684651650078576,
+      "loss_partial": 0.052001953125,
+      "step": 1467
+    },
+    {
+      "epoch": 0.7689889994761655,
+      "grad_norm": 1.46221923828125,
+      "learning_rate": 1.948799664388755e-05,
+      "loss": 0.109,
+      "step": 1468
+    },
+    {
+      "epoch": 0.7689889994761655,
+      "loss_partial": 0.050048828125,
+      "step": 1468
+    },
+    {
+      "epoch": 0.7689889994761655,
+      "loss_partial": 0.057861328125,
+      "step": 1468
+    },
+    {
+      "epoch": 0.7695128339444736,
+      "grad_norm": 2.038686513900757,
+      "learning_rate": 1.9486551587532207e-05,
+      "loss": 0.1076,
+      "step": 1469
+    },
+    {
+      "epoch": 0.7695128339444736,
+      "loss_partial": 0.060302734375,
+      "step": 1469
+    },
+    {
+      "epoch": 0.7695128339444736,
+      "loss_partial": 0.056640625,
+      "step": 1469
+    },
+    {
+      "epoch": 0.7700366684127815,
+      "grad_norm": 1.8221642971038818,
+      "learning_rate": 1.9485104548546243e-05,
+      "loss": 0.1118,
+      "step": 1470
+    },
+    {
+      "epoch": 0.7700366684127815,
+      "loss_partial": 0.0634765625,
+      "step": 1470
+    },
+    {
+      "epoch": 0.7700366684127815,
+      "loss_partial": 0.05126953125,
+      "step": 1470
+    },
+    {
+      "epoch": 0.7705605028810896,
+      "grad_norm": 2.1902923583984375,
+      "learning_rate": 1.948365552723208e-05,
+      "loss": 0.1118,
+      "step": 1471
+    },
+    {
+      "epoch": 0.7705605028810896,
+      "loss_partial": 0.061767578125,
+      "step": 1471
+    },
+    {
+      "epoch": 0.7705605028810896,
+      "loss_partial": 0.054443359375,
+      "step": 1471
+    },
+    {
+      "epoch": 0.7710843373493976,
+      "grad_norm": 2.281007766723633,
+      "learning_rate": 1.9482204523892564e-05,
+      "loss": 0.1124,
+      "step": 1472
+    },
+    {
+      "epoch": 0.7710843373493976,
+      "loss_partial": 0.052490234375,
+      "step": 1472
+    },
+    {
+      "epoch": 0.7710843373493976,
+      "loss_partial": 0.051025390625,
+      "step": 1472
+    },
+    {
+      "epoch": 0.7716081718177056,
+      "grad_norm": 2.1605746746063232,
+      "learning_rate": 1.9480751538830933e-05,
+      "loss": 0.1046,
+      "step": 1473
+    },
+    {
+      "epoch": 0.7716081718177056,
+      "loss_partial": 0.053955078125,
+      "step": 1473
+    },
+    {
+      "epoch": 0.7716081718177056,
+      "loss_partial": 0.04931640625,
+      "step": 1473
+    },
+    {
+      "epoch": 0.7721320062860136,
+      "grad_norm": 2.7116198539733887,
+      "learning_rate": 1.947929657235086e-05,
+      "loss": 0.1115,
+      "step": 1474
+    },
+    {
+      "epoch": 0.7721320062860136,
+      "loss_partial": 0.056396484375,
+      "step": 1474
+    },
+    {
+      "epoch": 0.7721320062860136,
+      "loss_partial": 0.046142578125,
+      "step": 1474
+    },
+    {
+      "epoch": 0.7726558407543216,
+      "grad_norm": 1.7660727500915527,
+      "learning_rate": 1.9477839624756423e-05,
+      "loss": 0.1098,
+      "step": 1475
+    },
+    {
+      "epoch": 0.7726558407543216,
+      "loss_partial": 0.05029296875,
+      "step": 1475
+    },
+    {
+      "epoch": 0.7726558407543216,
+      "loss_partial": 0.059814453125,
+      "step": 1475
+    },
+    {
+      "epoch": 0.7731796752226296,
+      "grad_norm": 2.2241899967193604,
+      "learning_rate": 1.9476380696352112e-05,
+      "loss": 0.1107,
+      "step": 1476
+    },
+    {
+      "epoch": 0.7731796752226296,
+      "loss_partial": 0.051513671875,
+      "step": 1476
+    },
+    {
+      "epoch": 0.7731796752226296,
+      "loss_partial": 0.04833984375,
+      "step": 1476
+    },
+    {
+      "epoch": 0.7737035096909377,
+      "grad_norm": 3.3476383686065674,
+      "learning_rate": 1.9474919787442835e-05,
+      "loss": 0.1044,
+      "step": 1477
+    },
+    {
+      "epoch": 0.7737035096909377,
+      "loss_partial": 0.048828125,
+      "step": 1477
+    },
+    {
+      "epoch": 0.7737035096909377,
+      "loss_partial": 0.05126953125,
+      "step": 1477
+    },
+    {
+      "epoch": 0.7742273441592457,
+      "grad_norm": 2.153017044067383,
+      "learning_rate": 1.9473456898333913e-05,
+      "loss": 0.1042,
+      "step": 1478
+    },
+    {
+      "epoch": 0.7742273441592457,
+      "loss_partial": 0.048828125,
+      "step": 1478
+    },
+    {
+      "epoch": 0.7742273441592457,
+      "loss_partial": 0.054443359375,
+      "step": 1478
+    },
+    {
+      "epoch": 0.7747511786275537,
+      "grad_norm": 1.8569903373718262,
+      "learning_rate": 1.947199202933108e-05,
+      "loss": 0.1093,
+      "step": 1479
+    },
+    {
+      "epoch": 0.7747511786275537,
+      "loss_partial": 0.04736328125,
+      "step": 1479
+    },
+    {
+      "epoch": 0.7747511786275537,
+      "loss_partial": 0.053466796875,
+      "step": 1479
+    },
+    {
+      "epoch": 0.7752750130958617,
+      "grad_norm": 2.20508074760437,
+      "learning_rate": 1.9470525180740488e-05,
+      "loss": 0.1091,
+      "step": 1480
+    },
+    {
+      "epoch": 0.7752750130958617,
+      "loss_partial": 0.04931640625,
+      "step": 1480
+    },
+    {
+      "epoch": 0.7752750130958617,
+      "loss_partial": 0.04931640625,
+      "step": 1480
+    },
+    {
+      "epoch": 0.7757988475641697,
+      "grad_norm": 1.9364609718322754,
+      "learning_rate": 1.94690563528687e-05,
+      "loss": 0.112,
+      "step": 1481
+    },
+    {
+      "epoch": 0.7757988475641697,
+      "loss_partial": 0.055419921875,
+      "step": 1481
+    },
+    {
+      "epoch": 0.7757988475641697,
+      "loss_partial": 0.056884765625,
+      "step": 1481
+    },
+    {
+      "epoch": 0.7763226820324778,
+      "grad_norm": 1.8053499460220337,
+      "learning_rate": 1.9467585546022688e-05,
+      "loss": 0.1116,
+      "step": 1482
+    },
+    {
+      "epoch": 0.7763226820324778,
+      "loss_partial": 0.0703125,
+      "step": 1482
+    },
+    {
+      "epoch": 0.7763226820324778,
+      "loss_partial": 0.040771484375,
+      "step": 1482
+    },
+    {
+      "epoch": 0.7768465165007857,
+      "grad_norm": 2.095268726348877,
+      "learning_rate": 1.946611276050984e-05,
+      "loss": 0.108,
+      "step": 1483
+    },
+    {
+      "epoch": 0.7768465165007857,
+      "loss_partial": 0.0537109375,
+      "step": 1483
+    },
+    {
+      "epoch": 0.7768465165007857,
+      "loss_partial": 0.0625,
+      "step": 1483
+    },
+    {
+      "epoch": 0.7773703509690938,
+      "grad_norm": 1.587651252746582,
+      "learning_rate": 1.9464637996637964e-05,
+      "loss": 0.1089,
+      "step": 1484
+    },
+    {
+      "epoch": 0.7773703509690938,
+      "loss_partial": 0.05517578125,
+      "step": 1484
+    },
+    {
+      "epoch": 0.7773703509690938,
+      "loss_partial": 0.05419921875,
+      "step": 1484
+    },
+    {
+      "epoch": 0.7778941854374017,
+      "grad_norm": 2.177755355834961,
+      "learning_rate": 1.946316125471527e-05,
+      "loss": 0.1102,
+      "step": 1485
+    },
+    {
+      "epoch": 0.7778941854374017,
+      "loss_partial": 0.051513671875,
+      "step": 1485
+    },
+    {
+      "epoch": 0.7778941854374017,
+      "loss_partial": 0.05078125,
+      "step": 1485
+    },
+    {
+      "epoch": 0.7784180199057098,
+      "grad_norm": 2.19185209274292,
+      "learning_rate": 1.9461682535050395e-05,
+      "loss": 0.1094,
+      "step": 1486
+    },
+    {
+      "epoch": 0.7784180199057098,
+      "loss_partial": 0.05810546875,
+      "step": 1486
+    },
+    {
+      "epoch": 0.7784180199057098,
+      "loss_partial": 0.053955078125,
+      "step": 1486
+    },
+    {
+      "epoch": 0.7789418543740179,
+      "grad_norm": 2.0522236824035645,
+      "learning_rate": 1.946020183795238e-05,
+      "loss": 0.1067,
+      "step": 1487
+    },
+    {
+      "epoch": 0.7789418543740179,
+      "loss_partial": 0.0556640625,
+      "step": 1487
+    },
+    {
+      "epoch": 0.7789418543740179,
+      "loss_partial": 0.045166015625,
+      "step": 1487
+    },
+    {
+      "epoch": 0.7794656888423258,
+      "grad_norm": 2.4846982955932617,
+      "learning_rate": 1.9458719163730682e-05,
+      "loss": 0.1044,
+      "step": 1488
+    },
+    {
+      "epoch": 0.7794656888423258,
+      "loss_partial": 0.056396484375,
+      "step": 1488
+    },
+    {
+      "epoch": 0.7794656888423258,
+      "loss_partial": 0.05419921875,
+      "step": 1488
+    },
+    {
+      "epoch": 0.7799895233106339,
+      "grad_norm": 2.234748125076294,
+      "learning_rate": 1.9457234512695166e-05,
+      "loss": 0.109,
+      "step": 1489
+    },
+    {
+      "epoch": 0.7799895233106339,
+      "loss_partial": 0.06591796875,
+      "step": 1489
+    },
+    {
+      "epoch": 0.7799895233106339,
+      "loss_partial": 0.05029296875,
+      "step": 1489
+    },
+    {
+      "epoch": 0.7805133577789418,
+      "grad_norm": 2.6226532459259033,
+      "learning_rate": 1.945574788515612e-05,
+      "loss": 0.1115,
+      "step": 1490
+    },
+    {
+      "epoch": 0.7805133577789418,
+      "loss_partial": 0.052978515625,
+      "step": 1490
+    },
+    {
+      "epoch": 0.7805133577789418,
+      "loss_partial": 0.052734375,
+      "step": 1490
+    },
+    {
+      "epoch": 0.7810371922472499,
+      "grad_norm": 2.307539939880371,
+      "learning_rate": 1.9454259281424243e-05,
+      "loss": 0.1076,
+      "step": 1491
+    },
+    {
+      "epoch": 0.7810371922472499,
+      "loss_partial": 0.059326171875,
+      "step": 1491
+    },
+    {
+      "epoch": 0.7810371922472499,
+      "loss_partial": 0.0478515625,
+      "step": 1491
+    },
+    {
+      "epoch": 0.7815610267155579,
+      "grad_norm": 2.795670986175537,
+      "learning_rate": 1.9452768701810636e-05,
+      "loss": 0.1113,
+      "step": 1492
+    },
+    {
+      "epoch": 0.7815610267155579,
+      "loss_partial": 0.049560546875,
+      "step": 1492
+    },
+    {
+      "epoch": 0.7815610267155579,
+      "loss_partial": 0.048095703125,
+      "step": 1492
+    },
+    {
+      "epoch": 0.7820848611838659,
+      "grad_norm": 1.7513505220413208,
+      "learning_rate": 1.9451276146626826e-05,
+      "loss": 0.1065,
+      "step": 1493
+    },
+    {
+      "epoch": 0.7820848611838659,
+      "loss_partial": 0.06494140625,
+      "step": 1493
+    },
+    {
+      "epoch": 0.7820848611838659,
+      "loss_partial": 0.053466796875,
+      "step": 1493
+    },
+    {
+      "epoch": 0.782608695652174,
+      "grad_norm": 2.0328524112701416,
+      "learning_rate": 1.9449781616184745e-05,
+      "loss": 0.1082,
+      "step": 1494
+    },
+    {
+      "epoch": 0.782608695652174,
+      "loss_partial": 0.0634765625,
+      "step": 1494
+    },
+    {
+      "epoch": 0.782608695652174,
+      "loss_partial": 0.05810546875,
+      "step": 1494
+    },
+    {
+      "epoch": 0.7831325301204819,
+      "grad_norm": 2.2137999534606934,
+      "learning_rate": 1.944828511079674e-05,
+      "loss": 0.1091,
+      "step": 1495
+    },
+    {
+      "epoch": 0.7831325301204819,
+      "loss_partial": 0.052978515625,
+      "step": 1495
+    },
+    {
+      "epoch": 0.7831325301204819,
+      "loss_partial": 0.05712890625,
+      "step": 1495
+    },
+    {
+      "epoch": 0.78365636458879,
+      "grad_norm": 1.7875397205352783,
+      "learning_rate": 1.9446786630775576e-05,
+      "loss": 0.1124,
+      "step": 1496
+    },
+    {
+      "epoch": 0.78365636458879,
+      "loss_partial": 0.0546875,
+      "step": 1496
+    },
+    {
+      "epoch": 0.78365636458879,
+      "loss_partial": 0.05810546875,
+      "step": 1496
+    },
+    {
+      "epoch": 0.7841801990570979,
+      "grad_norm": 1.4688472747802734,
+      "learning_rate": 1.9445286176434425e-05,
+      "loss": 0.1095,
+      "step": 1497
+    },
+    {
+      "epoch": 0.7841801990570979,
+      "loss_partial": 0.064453125,
+      "step": 1497
+    },
+    {
+      "epoch": 0.7841801990570979,
+      "loss_partial": 0.0576171875,
+      "step": 1497
+    },
+    {
+      "epoch": 0.784704033525406,
+      "grad_norm": 1.969114065170288,
+      "learning_rate": 1.9443783748086867e-05,
+      "loss": 0.1104,
+      "step": 1498
+    },
+    {
+      "epoch": 0.784704033525406,
+      "loss_partial": 0.052734375,
+      "step": 1498
+    },
+    {
+      "epoch": 0.784704033525406,
+      "loss_partial": 0.05517578125,
+      "step": 1498
+    },
+    {
+      "epoch": 0.785227867993714,
+      "grad_norm": 1.7321795225143433,
+      "learning_rate": 1.9442279346046902e-05,
+      "loss": 0.1108,
+      "step": 1499
+    },
+    {
+      "epoch": 0.785227867993714,
+      "loss_partial": 0.0595703125,
+      "step": 1499
+    },
+    {
+      "epoch": 0.785227867993714,
+      "loss_partial": 0.05517578125,
+      "step": 1499
+    },
+    {
+      "epoch": 0.785751702462022,
+      "grad_norm": 1.8938970565795898,
+      "learning_rate": 1.944077297062895e-05,
+      "loss": 0.1117,
+      "step": 1500
+    },
+    {
+      "epoch": 0.785751702462022,
+      "loss_partial": 0.056884765625,
+      "step": 1500
+    },
+    {
+      "epoch": 0.785751702462022,
+      "loss_partial": 0.054931640625,
+      "step": 1500
+    },
+    {
+      "epoch": 0.78627553693033,
+      "grad_norm": 1.29984450340271,
+      "learning_rate": 1.9439264622147824e-05,
+      "loss": 0.1133,
+      "step": 1501
+    },
+    {
+      "epoch": 0.78627553693033,
+      "loss_partial": 0.051025390625,
+      "step": 1501
+    },
+    {
+      "epoch": 0.78627553693033,
+      "loss_partial": 0.04638671875,
+      "step": 1501
+    },
+    {
+      "epoch": 0.786799371398638,
+      "grad_norm": 1.6237047910690308,
+      "learning_rate": 1.9437754300918764e-05,
+      "loss": 0.1143,
+      "step": 1502
+    },
+    {
+      "epoch": 0.786799371398638,
+      "loss_partial": 0.0537109375,
+      "step": 1502
+    },
+    {
+      "epoch": 0.786799371398638,
+      "loss_partial": 0.056640625,
+      "step": 1502
+    },
+    {
+      "epoch": 0.787323205866946,
+      "grad_norm": 1.3132293224334717,
+      "learning_rate": 1.9436242007257412e-05,
+      "loss": 0.1103,
+      "step": 1503
+    },
+    {
+      "epoch": 0.787323205866946,
+      "loss_partial": 0.049560546875,
+      "step": 1503
+    },
+    {
+      "epoch": 0.787323205866946,
+      "loss_partial": 0.048095703125,
+      "step": 1503
+    },
+    {
+      "epoch": 0.7878470403352541,
+      "grad_norm": 2.5199203491210938,
+      "learning_rate": 1.9434727741479838e-05,
+      "loss": 0.1088,
+      "step": 1504
+    },
+    {
+      "epoch": 0.7878470403352541,
+      "loss_partial": 0.05810546875,
+      "step": 1504
+    },
+    {
+      "epoch": 0.7878470403352541,
+      "loss_partial": 0.05224609375,
+      "step": 1504
+    },
+    {
+      "epoch": 0.788370874803562,
+      "grad_norm": 2.190986156463623,
+      "learning_rate": 1.9433211503902505e-05,
+      "loss": 0.1101,
+      "step": 1505
+    },
+    {
+      "epoch": 0.788370874803562,
+      "loss_partial": 0.051513671875,
+      "step": 1505
+    },
+    {
+      "epoch": 0.788370874803562,
+      "loss_partial": 0.052978515625,
+      "step": 1505
+    },
+    {
+      "epoch": 0.7888947092718701,
+      "grad_norm": 1.658591628074646,
+      "learning_rate": 1.9431693294842303e-05,
+      "loss": 0.1109,
+      "step": 1506
+    },
+    {
+      "epoch": 0.7888947092718701,
+      "loss_partial": 0.054443359375,
+      "step": 1506
+    },
+    {
+      "epoch": 0.7888947092718701,
+      "loss_partial": 0.0498046875,
+      "step": 1506
+    },
+    {
+      "epoch": 0.7894185437401781,
+      "grad_norm": 1.7574334144592285,
+      "learning_rate": 1.9430173114616523e-05,
+      "loss": 0.1015,
+      "step": 1507
+    },
+    {
+      "epoch": 0.7894185437401781,
+      "loss_partial": 0.0458984375,
+      "step": 1507
+    },
+    {
+      "epoch": 0.7894185437401781,
+      "loss_partial": 0.06494140625,
+      "step": 1507
+    },
+    {
+      "epoch": 0.7899423782084861,
+      "grad_norm": 2.1942055225372314,
+      "learning_rate": 1.942865096354288e-05,
+      "loss": 0.1077,
+      "step": 1508
+    },
+    {
+      "epoch": 0.7899423782084861,
+      "loss_partial": 0.06640625,
+      "step": 1508
+    },
+    {
+      "epoch": 0.7899423782084861,
+      "loss_partial": 0.059814453125,
+      "step": 1508
+    },
+    {
+      "epoch": 0.7904662126767942,
+      "grad_norm": 2.190659999847412,
+      "learning_rate": 1.942712684193949e-05,
+      "loss": 0.1123,
+      "step": 1509
+    },
+    {
+      "epoch": 0.7904662126767942,
+      "loss_partial": 0.051513671875,
+      "step": 1509
+    },
+    {
+      "epoch": 0.7904662126767942,
+      "loss_partial": 0.0576171875,
+      "step": 1509
+    },
+    {
+      "epoch": 0.7909900471451021,
+      "grad_norm": 2.4932050704956055,
+      "learning_rate": 1.9425600750124886e-05,
+      "loss": 0.1118,
+      "step": 1510
+    },
+    {
+      "epoch": 0.7909900471451021,
+      "loss_partial": 0.054443359375,
+      "step": 1510
+    },
+    {
+      "epoch": 0.7909900471451021,
+      "loss_partial": 0.052490234375,
+      "step": 1510
+    },
+    {
+      "epoch": 0.7915138816134102,
+      "grad_norm": 2.51346755027771,
+      "learning_rate": 1.9424072688418007e-05,
+      "loss": 0.1065,
+      "step": 1511
+    },
+    {
+      "epoch": 0.7915138816134102,
+      "loss_partial": 0.05810546875,
+      "step": 1511
+    },
+    {
+      "epoch": 0.7915138816134102,
+      "loss_partial": 0.04931640625,
+      "step": 1511
+    },
+    {
+      "epoch": 0.7920377160817181,
+      "grad_norm": 2.3414318561553955,
+      "learning_rate": 1.942254265713822e-05,
+      "loss": 0.1107,
+      "step": 1512
+    },
+    {
+      "epoch": 0.7920377160817181,
+      "loss_partial": 0.05712890625,
+      "step": 1512
+    },
+    {
+      "epoch": 0.7920377160817181,
+      "loss_partial": 0.047119140625,
+      "step": 1512
+    },
+    {
+      "epoch": 0.7925615505500262,
+      "grad_norm": 2.7344861030578613,
+      "learning_rate": 1.942101065660528e-05,
+      "loss": 0.108,
+      "step": 1513
+    },
+    {
+      "epoch": 0.7925615505500262,
+      "loss_partial": 0.044921875,
+      "step": 1513
+    },
+    {
+      "epoch": 0.7925615505500262,
+      "loss_partial": 0.055908203125,
+      "step": 1513
+    },
+    {
+      "epoch": 0.7930853850183343,
+      "grad_norm": 2.7437829971313477,
+      "learning_rate": 1.941947668713937e-05,
+      "loss": 0.1077,
+      "step": 1514
+    },
+    {
+      "epoch": 0.7930853850183343,
+      "loss_partial": 0.060302734375,
+      "step": 1514
+    },
+    {
+      "epoch": 0.7930853850183343,
+      "loss_partial": 0.041259765625,
+      "step": 1514
+    },
+    {
+      "epoch": 0.7936092194866422,
+      "grad_norm": 2.5693631172180176,
+      "learning_rate": 1.9417940749061083e-05,
+      "loss": 0.1057,
+      "step": 1515
+    },
+    {
+      "epoch": 0.7936092194866422,
+      "loss_partial": 0.0625,
+      "step": 1515
+    },
+    {
+      "epoch": 0.7936092194866422,
+      "loss_partial": 0.056640625,
+      "step": 1515
+    },
+    {
+      "epoch": 0.7941330539549503,
+      "grad_norm": 2.3571836948394775,
+      "learning_rate": 1.9416402842691417e-05,
+      "loss": 0.1094,
+      "step": 1516
+    },
+    {
+      "epoch": 0.7941330539549503,
+      "loss_partial": 0.050048828125,
+      "step": 1516
+    },
+    {
+      "epoch": 0.7941330539549503,
+      "loss_partial": 0.051025390625,
+      "step": 1516
+    },
+    {
+      "epoch": 0.7946568884232582,
+      "grad_norm": 1.914283275604248,
+      "learning_rate": 1.9414862968351788e-05,
+      "loss": 0.1088,
+      "step": 1517
+    },
+    {
+      "epoch": 0.7946568884232582,
+      "loss_partial": 0.05615234375,
+      "step": 1517
+    },
+    {
+      "epoch": 0.7946568884232582,
+      "loss_partial": 0.057861328125,
+      "step": 1517
+    },
+    {
+      "epoch": 0.7951807228915663,
+      "grad_norm": 2.9839892387390137,
+      "learning_rate": 1.9413321126364017e-05,
+      "loss": 0.119,
+      "step": 1518
+    },
+    {
+      "epoch": 0.7951807228915663,
+      "loss_partial": 0.0517578125,
+      "step": 1518
+    },
+    {
+      "epoch": 0.7951807228915663,
+      "loss_partial": 0.0517578125,
+      "step": 1518
+    },
+    {
+      "epoch": 0.7957045573598743,
+      "grad_norm": 2.782033681869507,
+      "learning_rate": 1.941177731705034e-05,
+      "loss": 0.1068,
+      "step": 1519
+    },
+    {
+      "epoch": 0.7957045573598743,
+      "loss_partial": 0.06982421875,
+      "step": 1519
+    },
+    {
+      "epoch": 0.7957045573598743,
+      "loss_partial": 0.0517578125,
+      "step": 1519
+    },
+    {
+      "epoch": 0.7962283918281823,
+      "grad_norm": 1.8742433786392212,
+      "learning_rate": 1.9410231540733403e-05,
+      "loss": 0.1132,
+      "step": 1520
+    },
+    {
+      "epoch": 0.7962283918281823,
+      "loss_partial": 0.05615234375,
+      "step": 1520
+    },
+    {
+      "epoch": 0.7962283918281823,
+      "loss_partial": 0.05908203125,
+      "step": 1520
+    },
+    {
+      "epoch": 0.7967522262964903,
+      "grad_norm": 2.372859239578247,
+      "learning_rate": 1.940868379773627e-05,
+      "loss": 0.113,
+      "step": 1521
+    },
+    {
+      "epoch": 0.7967522262964903,
+      "loss_partial": 0.060791015625,
+      "step": 1521
+    },
+    {
+      "epoch": 0.7967522262964903,
+      "loss_partial": 0.058837890625,
+      "step": 1521
+    },
+    {
+      "epoch": 0.7972760607647983,
+      "grad_norm": 1.8897730112075806,
+      "learning_rate": 1.94071340883824e-05,
+      "loss": 0.109,
+      "step": 1522
+    },
+    {
+      "epoch": 0.7972760607647983,
+      "loss_partial": 0.05029296875,
+      "step": 1522
+    },
+    {
+      "epoch": 0.7972760607647983,
+      "loss_partial": 0.057861328125,
+      "step": 1522
+    },
+    {
+      "epoch": 0.7977998952331063,
+      "grad_norm": 2.6519405841827393,
+      "learning_rate": 1.9405582412995684e-05,
+      "loss": 0.1097,
+      "step": 1523
+    },
+    {
+      "epoch": 0.7977998952331063,
+      "loss_partial": 0.05322265625,
+      "step": 1523
+    },
+    {
+      "epoch": 0.7977998952331063,
+      "loss_partial": 0.059814453125,
+      "step": 1523
+    },
+    {
+      "epoch": 0.7983237297014144,
+      "grad_norm": 1.8444929122924805,
+      "learning_rate": 1.94040287719004e-05,
+      "loss": 0.113,
+      "step": 1524
+    },
+    {
+      "epoch": 0.7983237297014144,
+      "loss_partial": 0.053955078125,
+      "step": 1524
+    },
+    {
+      "epoch": 0.7983237297014144,
+      "loss_partial": 0.05908203125,
+      "step": 1524
+    },
+    {
+      "epoch": 0.7988475641697224,
+      "grad_norm": 2.089500665664673,
+      "learning_rate": 1.9402473165421258e-05,
+      "loss": 0.1111,
+      "step": 1525
+    },
+    {
+      "epoch": 0.7988475641697224,
+      "loss_partial": 0.059326171875,
+      "step": 1525
+    },
+    {
+      "epoch": 0.7988475641697224,
+      "loss_partial": 0.0615234375,
+      "step": 1525
+    },
+    {
+      "epoch": 0.7993713986380304,
+      "grad_norm": 2.597020387649536,
+      "learning_rate": 1.9400915593883365e-05,
+      "loss": 0.1106,
+      "step": 1526
+    },
+    {
+      "epoch": 0.7993713986380304,
+      "loss_partial": 0.0634765625,
+      "step": 1526
+    },
+    {
+      "epoch": 0.7993713986380304,
+      "loss_partial": 0.05908203125,
+      "step": 1526
+    },
+    {
+      "epoch": 0.7998952331063384,
+      "grad_norm": 2.5150160789489746,
+      "learning_rate": 1.939935605761225e-05,
+      "loss": 0.1082,
+      "step": 1527
+    },
+    {
+      "epoch": 0.7998952331063384,
+      "loss_partial": 0.056884765625,
+      "step": 1527
+    },
+    {
+      "epoch": 0.7998952331063384,
+      "loss_partial": 0.054443359375,
+      "step": 1527
+    },
+    {
+      "epoch": 0.8004190675746464,
+      "grad_norm": 2.167982578277588,
+      "learning_rate": 1.939779455693384e-05,
+      "loss": 0.1138,
+      "step": 1528
+    },
+    {
+      "epoch": 0.8004190675746464,
+      "loss_partial": 0.0498046875,
+      "step": 1528
+    },
+    {
+      "epoch": 0.8004190675746464,
+      "loss_partial": 0.052978515625,
+      "step": 1528
+    },
+    {
+      "epoch": 0.8009429020429544,
+      "grad_norm": 2.6570675373077393,
+      "learning_rate": 1.939623109217448e-05,
+      "loss": 0.1091,
+      "step": 1529
+    },
+    {
+      "epoch": 0.8009429020429544,
+      "loss_partial": 0.053466796875,
+      "step": 1529
+    },
+    {
+      "epoch": 0.8009429020429544,
+      "loss_partial": 0.0546875,
+      "step": 1529
+    },
+    {
+      "epoch": 0.8014667365112624,
+      "grad_norm": 2.0474464893341064,
+      "learning_rate": 1.9394665663660933e-05,
+      "loss": 0.1054,
+      "step": 1530
+    },
+    {
+      "epoch": 0.8014667365112624,
+      "loss_partial": 0.050537109375,
+      "step": 1530
+    },
+    {
+      "epoch": 0.8014667365112624,
+      "loss_partial": 0.051025390625,
+      "step": 1530
+    },
+    {
+      "epoch": 0.8019905709795705,
+      "grad_norm": 2.0546209812164307,
+      "learning_rate": 1.9393098271720353e-05,
+      "loss": 0.1096,
+      "step": 1531
+    },
+    {
+      "epoch": 0.8019905709795705,
+      "loss_partial": 0.049560546875,
+      "step": 1531
+    },
+    {
+      "epoch": 0.8019905709795705,
+      "loss_partial": 0.054443359375,
+      "step": 1531
+    },
+    {
+      "epoch": 0.8025144054478784,
+      "grad_norm": 2.6939287185668945,
+      "learning_rate": 1.939152891668032e-05,
+      "loss": 0.107,
+      "step": 1532
+    },
+    {
+      "epoch": 0.8025144054478784,
+      "loss_partial": 0.06494140625,
+      "step": 1532
+    },
+    {
+      "epoch": 0.8025144054478784,
+      "loss_partial": 0.0576171875,
+      "step": 1532
+    },
+    {
+      "epoch": 0.8030382399161865,
+      "grad_norm": 2.150493860244751,
+      "learning_rate": 1.938995759886882e-05,
+      "loss": 0.1092,
+      "step": 1533
+    },
+    {
+      "epoch": 0.8030382399161865,
+      "loss_partial": 0.0556640625,
+      "step": 1533
+    },
+    {
+      "epoch": 0.8030382399161865,
+      "loss_partial": 0.05859375,
+      "step": 1533
+    },
+    {
+      "epoch": 0.8035620743844945,
+      "grad_norm": 1.888920783996582,
+      "learning_rate": 1.938838431861424e-05,
+      "loss": 0.1118,
+      "step": 1534
+    },
+    {
+      "epoch": 0.8035620743844945,
+      "loss_partial": 0.056640625,
+      "step": 1534
+    },
+    {
+      "epoch": 0.8035620743844945,
+      "loss_partial": 0.0546875,
+      "step": 1534
+    },
+    {
+      "epoch": 0.8040859088528025,
+      "grad_norm": 2.4314231872558594,
+      "learning_rate": 1.9386809076245403e-05,
+      "loss": 0.1098,
+      "step": 1535
+    },
+    {
+      "epoch": 0.8040859088528025,
+      "loss_partial": 0.07763671875,
+      "step": 1535
+    },
+    {
+      "epoch": 0.8040859088528025,
+      "loss_partial": 0.041748046875,
+      "step": 1535
+    },
+    {
+      "epoch": 0.8046097433211106,
+      "grad_norm": 1.9922120571136475,
+      "learning_rate": 1.9385231872091513e-05,
+      "loss": 0.112,
+      "step": 1536
+    },
+    {
+      "epoch": 0.8046097433211106,
+      "loss_partial": 0.051513671875,
+      "step": 1536
+    },
+    {
+      "epoch": 0.8046097433211106,
+      "loss_partial": 0.05908203125,
+      "step": 1536
+    },
+    {
+      "epoch": 0.8051335777894185,
+      "grad_norm": 1.929280400276184,
+      "learning_rate": 1.9383652706482196e-05,
+      "loss": 0.1086,
+      "step": 1537
+    },
+    {
+      "epoch": 0.8051335777894185,
+      "loss_partial": 0.0556640625,
+      "step": 1537
+    },
+    {
+      "epoch": 0.8051335777894185,
+      "loss_partial": 0.050537109375,
+      "step": 1537
+    },
+    {
+      "epoch": 0.8056574122577266,
+      "grad_norm": 2.271064281463623,
+      "learning_rate": 1.9382071579747492e-05,
+      "loss": 0.1098,
+      "step": 1538
+    },
+    {
+      "epoch": 0.8056574122577266,
+      "loss_partial": 0.05908203125,
+      "step": 1538
+    },
+    {
+      "epoch": 0.8056574122577266,
+      "loss_partial": 0.0546875,
+      "step": 1538
+    },
+    {
+      "epoch": 0.8061812467260345,
+      "grad_norm": 2.465940237045288,
+      "learning_rate": 1.9380488492217845e-05,
+      "loss": 0.1116,
+      "step": 1539
+    },
+    {
+      "epoch": 0.8061812467260345,
+      "loss_partial": 0.04833984375,
+      "step": 1539
+    },
+    {
+      "epoch": 0.8061812467260345,
+      "loss_partial": 0.053955078125,
+      "step": 1539
+    },
+    {
+      "epoch": 0.8067050811943426,
+      "grad_norm": 1.9096399545669556,
+      "learning_rate": 1.9378903444224105e-05,
+      "loss": 0.1077,
+      "step": 1540
+    },
+    {
+      "epoch": 0.8067050811943426,
+      "loss_partial": 0.056640625,
+      "step": 1540
+    },
+    {
+      "epoch": 0.8067050811943426,
+      "loss_partial": 0.0576171875,
+      "step": 1540
+    },
+    {
+      "epoch": 0.8072289156626506,
+      "grad_norm": 2.2710046768188477,
+      "learning_rate": 1.9377316436097548e-05,
+      "loss": 0.1125,
+      "step": 1541
+    },
+    {
+      "epoch": 0.8072289156626506,
+      "loss_partial": 0.060546875,
+      "step": 1541
+    },
+    {
+      "epoch": 0.8072289156626506,
+      "loss_partial": 0.046875,
+      "step": 1541
+    },
+    {
+      "epoch": 0.8077527501309586,
+      "grad_norm": 2.286670684814453,
+      "learning_rate": 1.9375727468169844e-05,
+      "loss": 0.1075,
+      "step": 1542
+    },
+    {
+      "epoch": 0.8077527501309586,
+      "loss_partial": 0.04931640625,
+      "step": 1542
+    },
+    {
+      "epoch": 0.8077527501309586,
+      "loss_partial": 0.04150390625,
+      "step": 1542
+    },
+    {
+      "epoch": 0.8082765845992667,
+      "grad_norm": 2.812387228012085,
+      "learning_rate": 1.9374136540773074e-05,
+      "loss": 0.1098,
+      "step": 1543
+    },
+    {
+      "epoch": 0.8082765845992667,
+      "loss_partial": 0.06298828125,
+      "step": 1543
+    },
+    {
+      "epoch": 0.8082765845992667,
+      "loss_partial": 0.0732421875,
+      "step": 1543
+    },
+    {
+      "epoch": 0.8088004190675746,
+      "grad_norm": 1.9383049011230469,
+      "learning_rate": 1.9372543654239736e-05,
+      "loss": 0.1111,
+      "step": 1544
+    },
+    {
+      "epoch": 0.8088004190675746,
+      "loss_partial": 0.055908203125,
+      "step": 1544
+    },
+    {
+      "epoch": 0.8088004190675746,
+      "loss_partial": 0.051025390625,
+      "step": 1544
+    },
+    {
+      "epoch": 0.8093242535358827,
+      "grad_norm": 2.641339063644409,
+      "learning_rate": 1.9370948808902735e-05,
+      "loss": 0.1099,
+      "step": 1545
+    },
+    {
+      "epoch": 0.8093242535358827,
+      "loss_partial": 0.0576171875,
+      "step": 1545
+    },
+    {
+      "epoch": 0.8093242535358827,
+      "loss_partial": 0.05322265625,
+      "step": 1545
+    },
+    {
+      "epoch": 0.8098480880041907,
+      "grad_norm": 2.0738608837127686,
+      "learning_rate": 1.9369352005095377e-05,
+      "loss": 0.1122,
+      "step": 1546
+    },
+    {
+      "epoch": 0.8098480880041907,
+      "loss_partial": 0.061279296875,
+      "step": 1546
+    },
+    {
+      "epoch": 0.8098480880041907,
+      "loss_partial": 0.05517578125,
+      "step": 1546
+    },
+    {
+      "epoch": 0.8103719224724987,
+      "grad_norm": 2.067356586456299,
+      "learning_rate": 1.936775324315139e-05,
+      "loss": 0.1111,
+      "step": 1547
+    },
+    {
+      "epoch": 0.8103719224724987,
+      "loss_partial": 0.04052734375,
+      "step": 1547
+    },
+    {
+      "epoch": 0.8103719224724987,
+      "loss_partial": 0.04736328125,
+      "step": 1547
+    },
+    {
+      "epoch": 0.8108957569408067,
+      "grad_norm": 1.9291993379592896,
+      "learning_rate": 1.9366152523404904e-05,
+      "loss": 0.1088,
+      "step": 1548
+    },
+    {
+      "epoch": 0.8108957569408067,
+      "loss_partial": 0.04541015625,
+      "step": 1548
+    },
+    {
+      "epoch": 0.8108957569408067,
+      "loss_partial": 0.048095703125,
+      "step": 1548
+    },
+    {
+      "epoch": 0.8114195914091147,
+      "grad_norm": 3.1326141357421875,
+      "learning_rate": 1.936454984619046e-05,
+      "loss": 0.109,
+      "step": 1549
+    },
+    {
+      "epoch": 0.8114195914091147,
+      "loss_partial": 0.048828125,
+      "step": 1549
+    },
+    {
+      "epoch": 0.8114195914091147,
+      "loss_partial": 0.049072265625,
+      "step": 1549
+    },
+    {
+      "epoch": 0.8119434258774227,
+      "grad_norm": 3.3319854736328125,
+      "learning_rate": 1.9362945211843005e-05,
+      "loss": 0.1089,
+      "step": 1550
+    },
+    {
+      "epoch": 0.8119434258774227,
+      "loss_partial": 0.054443359375,
+      "step": 1550
+    },
+    {
+      "epoch": 0.8119434258774227,
+      "loss_partial": 0.053466796875,
+      "step": 1550
+    },
+    {
+      "epoch": 0.8124672603457308,
+      "grad_norm": 2.662992238998413,
+      "learning_rate": 1.9361338620697898e-05,
+      "loss": 0.112,
+      "step": 1551
+    },
+    {
+      "epoch": 0.8124672603457308,
+      "loss_partial": 0.04931640625,
+      "step": 1551
+    },
+    {
+      "epoch": 0.8124672603457308,
+      "loss_partial": 0.050048828125,
+      "step": 1551
+    },
+    {
+      "epoch": 0.8129910948140388,
+      "grad_norm": 2.7667760848999023,
+      "learning_rate": 1.935973007309091e-05,
+      "loss": 0.1075,
+      "step": 1552
+    },
+    {
+      "epoch": 0.8129910948140388,
+      "loss_partial": 0.0546875,
+      "step": 1552
+    },
+    {
+      "epoch": 0.8129910948140388,
+      "loss_partial": 0.046630859375,
+      "step": 1552
+    },
+    {
+      "epoch": 0.8135149292823468,
+      "grad_norm": 1.9779690504074097,
+      "learning_rate": 1.9358119569358216e-05,
+      "loss": 0.1122,
+      "step": 1553
+    },
+    {
+      "epoch": 0.8135149292823468,
+      "loss_partial": 0.05810546875,
+      "step": 1553
+    },
+    {
+      "epoch": 0.8135149292823468,
+      "loss_partial": 0.06005859375,
+      "step": 1553
+    },
+    {
+      "epoch": 0.8140387637506548,
+      "grad_norm": 2.0052876472473145,
+      "learning_rate": 1.93565071098364e-05,
+      "loss": 0.1113,
+      "step": 1554
+    },
+    {
+      "epoch": 0.8140387637506548,
+      "loss_partial": 0.0654296875,
+      "step": 1554
+    },
+    {
+      "epoch": 0.8140387637506548,
+      "loss_partial": 0.06787109375,
+      "step": 1554
+    },
+    {
+      "epoch": 0.8145625982189628,
+      "grad_norm": 2.375040054321289,
+      "learning_rate": 1.935489269486246e-05,
+      "loss": 0.1122,
+      "step": 1555
+    },
+    {
+      "epoch": 0.8145625982189628,
+      "loss_partial": 0.052978515625,
+      "step": 1555
+    },
+    {
+      "epoch": 0.8145625982189628,
+      "loss_partial": 0.05224609375,
+      "step": 1555
+    },
+    {
+      "epoch": 0.8150864326872708,
+      "grad_norm": 1.4680832624435425,
+      "learning_rate": 1.935327632477379e-05,
+      "loss": 0.1131,
+      "step": 1556
+    },
+    {
+      "epoch": 0.8150864326872708,
+      "loss_partial": 0.0537109375,
+      "step": 1556
+    },
+    {
+      "epoch": 0.8150864326872708,
+      "loss_partial": 0.0546875,
+      "step": 1556
+    },
+    {
+      "epoch": 0.8156102671555788,
+      "grad_norm": 1.74882972240448,
+      "learning_rate": 1.935165799990821e-05,
+      "loss": 0.1104,
+      "step": 1557
+    },
+    {
+      "epoch": 0.8156102671555788,
+      "loss_partial": 0.0576171875,
+      "step": 1557
+    },
+    {
+      "epoch": 0.8156102671555788,
+      "loss_partial": 0.05224609375,
+      "step": 1557
+    },
+    {
+      "epoch": 0.8161341016238869,
+      "grad_norm": 1.5070552825927734,
+      "learning_rate": 1.935003772060394e-05,
+      "loss": 0.1122,
+      "step": 1558
+    },
+    {
+      "epoch": 0.8161341016238869,
+      "loss_partial": 0.061279296875,
+      "step": 1558
+    },
+    {
+      "epoch": 0.8161341016238869,
+      "loss_partial": 0.045166015625,
+      "step": 1558
+    },
+    {
+      "epoch": 0.8166579360921948,
+      "grad_norm": 1.9274085760116577,
+      "learning_rate": 1.9348415487199603e-05,
+      "loss": 0.1112,
+      "step": 1559
+    },
+    {
+      "epoch": 0.8166579360921948,
+      "loss_partial": 0.053466796875,
+      "step": 1559
+    },
+    {
+      "epoch": 0.8166579360921948,
+      "loss_partial": 0.0546875,
+      "step": 1559
+    },
+    {
+      "epoch": 0.8171817705605029,
+      "grad_norm": 1.5373152494430542,
+      "learning_rate": 1.9346791300034236e-05,
+      "loss": 0.1073,
+      "step": 1560
+    },
+    {
+      "epoch": 0.8171817705605029,
+      "loss_partial": 0.054931640625,
+      "step": 1560
+    },
+    {
+      "epoch": 0.8171817705605029,
+      "loss_partial": 0.057373046875,
+      "step": 1560
+    },
+    {
+      "epoch": 0.8177056050288108,
+      "grad_norm": 2.339134931564331,
+      "learning_rate": 1.9345165159447288e-05,
+      "loss": 0.1084,
+      "step": 1561
+    },
+    {
+      "epoch": 0.8177056050288108,
+      "loss_partial": 0.0693359375,
+      "step": 1561
+    },
+    {
+      "epoch": 0.8177056050288108,
+      "loss_partial": 0.0673828125,
+      "step": 1561
+    },
+    {
+      "epoch": 0.8182294394971189,
+      "grad_norm": 1.8955614566802979,
+      "learning_rate": 1.9343537065778607e-05,
+      "loss": 0.1118,
+      "step": 1562
+    },
+    {
+      "epoch": 0.8182294394971189,
+      "loss_partial": 0.0576171875,
+      "step": 1562
+    },
+    {
+      "epoch": 0.8182294394971189,
+      "loss_partial": 0.05078125,
+      "step": 1562
+    },
+    {
+      "epoch": 0.818753273965427,
+      "grad_norm": 2.19653058052063,
+      "learning_rate": 1.9341907019368463e-05,
+      "loss": 0.1122,
+      "step": 1563
+    },
+    {
+      "epoch": 0.818753273965427,
+      "loss_partial": 0.06005859375,
+      "step": 1563
+    },
+    {
+      "epoch": 0.818753273965427,
+      "loss_partial": 0.055908203125,
+      "step": 1563
+    },
+    {
+      "epoch": 0.8192771084337349,
+      "grad_norm": 2.2736406326293945,
+      "learning_rate": 1.9340275020557517e-05,
+      "loss": 0.1112,
+      "step": 1564
+    },
+    {
+      "epoch": 0.8192771084337349,
+      "loss_partial": 0.0615234375,
+      "step": 1564
+    },
+    {
+      "epoch": 0.8192771084337349,
+      "loss_partial": 0.0595703125,
+      "step": 1564
+    },
+    {
+      "epoch": 0.819800942902043,
+      "grad_norm": 1.8484874963760376,
+      "learning_rate": 1.933864106968685e-05,
+      "loss": 0.1099,
+      "step": 1565
+    },
+    {
+      "epoch": 0.819800942902043,
+      "loss_partial": 0.051025390625,
+      "step": 1565
+    },
+    {
+      "epoch": 0.819800942902043,
+      "loss_partial": 0.0537109375,
+      "step": 1565
+    },
+    {
+      "epoch": 0.8203247773703509,
+      "grad_norm": 1.7638715505599976,
+      "learning_rate": 1.9337005167097952e-05,
+      "loss": 0.1125,
+      "step": 1566
+    },
+    {
+      "epoch": 0.8203247773703509,
+      "loss_partial": 0.060302734375,
+      "step": 1566
+    },
+    {
+      "epoch": 0.8203247773703509,
+      "loss_partial": 0.07080078125,
+      "step": 1566
+    },
+    {
+      "epoch": 0.820848611838659,
+      "grad_norm": 1.6849236488342285,
+      "learning_rate": 1.933536731313271e-05,
+      "loss": 0.1093,
+      "step": 1567
+    },
+    {
+      "epoch": 0.820848611838659,
+      "loss_partial": 0.053466796875,
+      "step": 1567
+    },
+    {
+      "epoch": 0.820848611838659,
+      "loss_partial": 0.04931640625,
+      "step": 1567
+    },
+    {
+      "epoch": 0.821372446306967,
+      "grad_norm": 1.7824952602386475,
+      "learning_rate": 1.9333727508133427e-05,
+      "loss": 0.1049,
+      "step": 1568
+    },
+    {
+      "epoch": 0.821372446306967,
+      "loss_partial": 0.0576171875,
+      "step": 1568
+    },
+    {
+      "epoch": 0.821372446306967,
+      "loss_partial": 0.046875,
+      "step": 1568
+    },
+    {
+      "epoch": 0.821896280775275,
+      "grad_norm": 1.4895806312561035,
+      "learning_rate": 1.9332085752442813e-05,
+      "loss": 0.1052,
+      "step": 1569
+    },
+    {
+      "epoch": 0.821896280775275,
+      "loss_partial": 0.05322265625,
+      "step": 1569
+    },
+    {
+      "epoch": 0.821896280775275,
+      "loss_partial": 0.056884765625,
+      "step": 1569
+    },
+    {
+      "epoch": 0.822420115243583,
+      "grad_norm": 1.683141827583313,
+      "learning_rate": 1.9330442046403987e-05,
+      "loss": 0.11,
+      "step": 1570
+    },
+    {
+      "epoch": 0.822420115243583,
+      "loss_partial": 0.0625,
+      "step": 1570
+    },
+    {
+      "epoch": 0.822420115243583,
+      "loss_partial": 0.054931640625,
+      "step": 1570
+    },
+    {
+      "epoch": 0.822943949711891,
+      "grad_norm": 1.962478756904602,
+      "learning_rate": 1.9328796390360466e-05,
+      "loss": 0.1055,
+      "step": 1571
+    },
+    {
+      "epoch": 0.822943949711891,
+      "loss_partial": 0.057373046875,
+      "step": 1571
+    },
+    {
+      "epoch": 0.822943949711891,
+      "loss_partial": 0.055419921875,
+      "step": 1571
+    },
+    {
+      "epoch": 0.8234677841801991,
+      "grad_norm": 1.9215024709701538,
+      "learning_rate": 1.9327148784656192e-05,
+      "loss": 0.1087,
+      "step": 1572
+    },
+    {
+      "epoch": 0.8234677841801991,
+      "loss_partial": 0.0517578125,
+      "step": 1572
+    },
+    {
+      "epoch": 0.8234677841801991,
+      "loss_partial": 0.056396484375,
+      "step": 1572
+    },
+    {
+      "epoch": 0.8239916186485071,
+      "grad_norm": 2.789062023162842,
+      "learning_rate": 1.9325499229635494e-05,
+      "loss": 0.11,
+      "step": 1573
+    },
+    {
+      "epoch": 0.8239916186485071,
+      "loss_partial": 0.0556640625,
+      "step": 1573
+    },
+    {
+      "epoch": 0.8239916186485071,
+      "loss_partial": 0.047119140625,
+      "step": 1573
+    },
+    {
+      "epoch": 0.8245154531168151,
+      "grad_norm": 3.011937141418457,
+      "learning_rate": 1.932384772564313e-05,
+      "loss": 0.1082,
+      "step": 1574
+    },
+    {
+      "epoch": 0.8245154531168151,
+      "loss_partial": 0.062255859375,
+      "step": 1574
+    },
+    {
+      "epoch": 0.8245154531168151,
+      "loss_partial": 0.059326171875,
+      "step": 1574
+    },
+    {
+      "epoch": 0.8250392875851231,
+      "grad_norm": 1.6932064294815063,
+      "learning_rate": 1.9322194273024245e-05,
+      "loss": 0.1092,
+      "step": 1575
+    },
+    {
+      "epoch": 0.8250392875851231,
+      "loss_partial": 0.0517578125,
+      "step": 1575
+    },
+    {
+      "epoch": 0.8250392875851231,
+      "loss_partial": 0.060302734375,
+      "step": 1575
+    },
+    {
+      "epoch": 0.8255631220534311,
+      "grad_norm": 2.404902696609497,
+      "learning_rate": 1.9320538872124407e-05,
+      "loss": 0.1142,
+      "step": 1576
+    },
+    {
+      "epoch": 0.8255631220534311,
+      "loss_partial": 0.053955078125,
+      "step": 1576
+    },
+    {
+      "epoch": 0.8255631220534311,
+      "loss_partial": 0.050537109375,
+      "step": 1576
+    },
+    {
+      "epoch": 0.8260869565217391,
+      "grad_norm": 2.052438259124756,
+      "learning_rate": 1.931888152328958e-05,
+      "loss": 0.1072,
+      "step": 1577
+    },
+    {
+      "epoch": 0.8260869565217391,
+      "loss_partial": 0.0712890625,
+      "step": 1577
+    },
+    {
+      "epoch": 0.8260869565217391,
+      "loss_partial": 0.046875,
+      "step": 1577
+    },
+    {
+      "epoch": 0.8266107909900472,
+      "grad_norm": 1.858888030052185,
+      "learning_rate": 1.9317222226866135e-05,
+      "loss": 0.1082,
+      "step": 1578
+    },
+    {
+      "epoch": 0.8266107909900472,
+      "loss_partial": 0.0576171875,
+      "step": 1578
+    },
+    {
+      "epoch": 0.8266107909900472,
+      "loss_partial": 0.0595703125,
+      "step": 1578
+    },
+    {
+      "epoch": 0.8271346254583551,
+      "grad_norm": 2.4859888553619385,
+      "learning_rate": 1.931556098320087e-05,
+      "loss": 0.1124,
+      "step": 1579
+    },
+    {
+      "epoch": 0.8271346254583551,
+      "loss_partial": 0.05126953125,
+      "step": 1579
+    },
+    {
+      "epoch": 0.8271346254583551,
+      "loss_partial": 0.051025390625,
+      "step": 1579
+    },
+    {
+      "epoch": 0.8276584599266632,
+      "grad_norm": 1.9746712446212769,
+      "learning_rate": 1.9313897792640957e-05,
+      "loss": 0.107,
+      "step": 1580
+    },
+    {
+      "epoch": 0.8276584599266632,
+      "loss_partial": 0.057861328125,
+      "step": 1580
+    },
+    {
+      "epoch": 0.8276584599266632,
+      "loss_partial": 0.049560546875,
+      "step": 1580
+    },
+    {
+      "epoch": 0.8281822943949712,
+      "grad_norm": 1.90878427028656,
+      "learning_rate": 1.9312232655534006e-05,
+      "loss": 0.1119,
+      "step": 1581
+    },
+    {
+      "epoch": 0.8281822943949712,
+      "loss_partial": 0.0634765625,
+      "step": 1581
+    },
+    {
+      "epoch": 0.8281822943949712,
+      "loss_partial": 0.044921875,
+      "step": 1581
+    },
+    {
+      "epoch": 0.8287061288632792,
+      "grad_norm": 2.349569797515869,
+      "learning_rate": 1.9310565572228014e-05,
+      "loss": 0.1054,
+      "step": 1582
+    },
+    {
+      "epoch": 0.8287061288632792,
+      "loss_partial": 0.0498046875,
+      "step": 1582
+    },
+    {
+      "epoch": 0.8287061288632792,
+      "loss_partial": 0.0546875,
+      "step": 1582
+    },
+    {
+      "epoch": 0.8292299633315873,
+      "grad_norm": 1.8236908912658691,
+      "learning_rate": 1.930889654307139e-05,
+      "loss": 0.1104,
+      "step": 1583
+    },
+    {
+      "epoch": 0.8292299633315873,
+      "loss_partial": 0.0595703125,
+      "step": 1583
+    },
+    {
+      "epoch": 0.8292299633315873,
+      "loss_partial": 0.05615234375,
+      "step": 1583
+    },
+    {
+      "epoch": 0.8297537977998952,
+      "grad_norm": 1.8827102184295654,
+      "learning_rate": 1.9307225568412954e-05,
+      "loss": 0.1077,
+      "step": 1584
+    },
+    {
+      "epoch": 0.8297537977998952,
+      "loss_partial": 0.052978515625,
+      "step": 1584
+    },
+    {
+      "epoch": 0.8297537977998952,
+      "loss_partial": 0.0556640625,
+      "step": 1584
+    },
+    {
+      "epoch": 0.8302776322682033,
+      "grad_norm": 2.0577263832092285,
+      "learning_rate": 1.9305552648601926e-05,
+      "loss": 0.1074,
+      "step": 1585
+    },
+    {
+      "epoch": 0.8302776322682033,
+      "loss_partial": 0.054931640625,
+      "step": 1585
+    },
+    {
+      "epoch": 0.8302776322682033,
+      "loss_partial": 0.0595703125,
+      "step": 1585
+    },
+    {
+      "epoch": 0.8308014667365112,
+      "grad_norm": 1.6264798641204834,
+      "learning_rate": 1.930387778398794e-05,
+      "loss": 0.1085,
+      "step": 1586
+    },
+    {
+      "epoch": 0.8308014667365112,
+      "loss_partial": 0.06494140625,
+      "step": 1586
+    },
+    {
+      "epoch": 0.8308014667365112,
+      "loss_partial": 0.05126953125,
+      "step": 1586
+    },
+    {
+      "epoch": 0.8313253012048193,
+      "grad_norm": 1.6386927366256714,
+      "learning_rate": 1.9302200974921034e-05,
+      "loss": 0.1096,
+      "step": 1587
+    },
+    {
+      "epoch": 0.8313253012048193,
+      "loss_partial": 0.056640625,
+      "step": 1587
+    },
+    {
+      "epoch": 0.8313253012048193,
+      "loss_partial": 0.05810546875,
+      "step": 1587
+    },
+    {
+      "epoch": 0.8318491356731272,
+      "grad_norm": 2.058568239212036,
+      "learning_rate": 1.9300522221751646e-05,
+      "loss": 0.1101,
+      "step": 1588
+    },
+    {
+      "epoch": 0.8318491356731272,
+      "loss_partial": 0.046630859375,
+      "step": 1588
+    },
+    {
+      "epoch": 0.8318491356731272,
+      "loss_partial": 0.0537109375,
+      "step": 1588
+    },
+    {
+      "epoch": 0.8323729701414353,
+      "grad_norm": 2.0604565143585205,
+      "learning_rate": 1.9298841524830624e-05,
+      "loss": 0.1106,
+      "step": 1589
+    },
+    {
+      "epoch": 0.8323729701414353,
+      "loss_partial": 0.0625,
+      "step": 1589
+    },
+    {
+      "epoch": 0.8323729701414353,
+      "loss_partial": 0.05908203125,
+      "step": 1589
+    },
+    {
+      "epoch": 0.8328968046097434,
+      "grad_norm": 2.3084182739257812,
+      "learning_rate": 1.9297158884509226e-05,
+      "loss": 0.1076,
+      "step": 1590
+    },
+    {
+      "epoch": 0.8328968046097434,
+      "loss_partial": 0.060302734375,
+      "step": 1590
+    },
+    {
+      "epoch": 0.8328968046097434,
+      "loss_partial": 0.04931640625,
+      "step": 1590
+    },
+    {
+      "epoch": 0.8334206390780513,
+      "grad_norm": 2.464820146560669,
+      "learning_rate": 1.9295474301139116e-05,
+      "loss": 0.1064,
+      "step": 1591
+    },
+    {
+      "epoch": 0.8334206390780513,
+      "loss_partial": 0.056884765625,
+      "step": 1591
+    },
+    {
+      "epoch": 0.8334206390780513,
+      "loss_partial": 0.0517578125,
+      "step": 1591
+    },
+    {
+      "epoch": 0.8339444735463594,
+      "grad_norm": 1.2045111656188965,
+      "learning_rate": 1.9293787775072354e-05,
+      "loss": 0.1095,
+      "step": 1592
+    },
+    {
+      "epoch": 0.8339444735463594,
+      "loss_partial": 0.0625,
+      "step": 1592
+    },
+    {
+      "epoch": 0.8339444735463594,
+      "loss_partial": 0.046630859375,
+      "step": 1592
+    },
+    {
+      "epoch": 0.8344683080146673,
+      "grad_norm": 2.084503412246704,
+      "learning_rate": 1.9292099306661423e-05,
+      "loss": 0.1067,
+      "step": 1593
+    },
+    {
+      "epoch": 0.8344683080146673,
+      "loss_partial": 0.05078125,
+      "step": 1593
+    },
+    {
+      "epoch": 0.8344683080146673,
+      "loss_partial": 0.05126953125,
+      "step": 1593
+    },
+    {
+      "epoch": 0.8349921424829754,
+      "grad_norm": 1.727442741394043,
+      "learning_rate": 1.9290408896259195e-05,
+      "loss": 0.1067,
+      "step": 1594
+    },
+    {
+      "epoch": 0.8349921424829754,
+      "loss_partial": 0.058349609375,
+      "step": 1594
+    },
+    {
+      "epoch": 0.8349921424829754,
+      "loss_partial": 0.047607421875,
+      "step": 1594
+    },
+    {
+      "epoch": 0.8355159769512834,
+      "grad_norm": 1.8761154413223267,
+      "learning_rate": 1.9288716544218954e-05,
+      "loss": 0.1093,
+      "step": 1595
+    },
+    {
+      "epoch": 0.8355159769512834,
+      "loss_partial": 0.056884765625,
+      "step": 1595
+    },
+    {
+      "epoch": 0.8355159769512834,
+      "loss_partial": 0.05078125,
+      "step": 1595
+    },
+    {
+      "epoch": 0.8360398114195914,
+      "grad_norm": 1.8257992267608643,
+      "learning_rate": 1.9287022250894402e-05,
+      "loss": 0.1076,
+      "step": 1596
+    },
+    {
+      "epoch": 0.8360398114195914,
+      "loss_partial": 0.053955078125,
+      "step": 1596
+    },
+    {
+      "epoch": 0.8360398114195914,
+      "loss_partial": 0.06201171875,
+      "step": 1596
+    },
+    {
+      "epoch": 0.8365636458878994,
+      "grad_norm": 2.2513582706451416,
+      "learning_rate": 1.9285326016639624e-05,
+      "loss": 0.1099,
+      "step": 1597
+    },
+    {
+      "epoch": 0.8365636458878994,
+      "loss_partial": 0.047607421875,
+      "step": 1597
+    },
+    {
+      "epoch": 0.8365636458878994,
+      "loss_partial": 0.06298828125,
+      "step": 1597
+    },
+    {
+      "epoch": 0.8370874803562074,
+      "grad_norm": 2.0651493072509766,
+      "learning_rate": 1.928362784180913e-05,
+      "loss": 0.1111,
+      "step": 1598
+    },
+    {
+      "epoch": 0.8370874803562074,
+      "loss_partial": 0.054443359375,
+      "step": 1598
+    },
+    {
+      "epoch": 0.8370874803562074,
+      "loss_partial": 0.058349609375,
+      "step": 1598
+    },
+    {
+      "epoch": 0.8376113148245155,
+      "grad_norm": 1.7800319194793701,
+      "learning_rate": 1.9281927726757823e-05,
+      "loss": 0.1091,
+      "step": 1599
+    },
+    {
+      "epoch": 0.8376113148245155,
+      "loss_partial": 0.052001953125,
+      "step": 1599
+    },
+    {
+      "epoch": 0.8376113148245155,
+      "loss_partial": 0.0595703125,
+      "step": 1599
+    },
+    {
+      "epoch": 0.8381351492928235,
+      "grad_norm": 1.6605316400527954,
+      "learning_rate": 1.9280225671841018e-05,
+      "loss": 0.1052,
+      "step": 1600
+    },
+    {
+      "epoch": 0.8381351492928235,
+      "loss_partial": 0.06298828125,
+      "step": 1600
+    },
+    {
+      "epoch": 0.8381351492928235,
+      "loss_partial": 0.0791015625,
+      "step": 1600
+    },
+    {
+      "epoch": 0.8386589837611315,
+      "grad_norm": 1.4560900926589966,
+      "learning_rate": 1.927852167741444e-05,
+      "loss": 0.1119,
+      "step": 1601
+    },
+    {
+      "epoch": 0.8386589837611315,
+      "loss_partial": 0.05810546875,
+      "step": 1601
+    },
+    {
+      "epoch": 0.8386589837611315,
+      "loss_partial": 0.042236328125,
+      "step": 1601
+    },
+    {
+      "epoch": 0.8391828182294395,
+      "grad_norm": 2.100266456604004,
+      "learning_rate": 1.9276815743834203e-05,
+      "loss": 0.1107,
+      "step": 1602
+    },
+    {
+      "epoch": 0.8391828182294395,
+      "loss_partial": 0.050048828125,
+      "step": 1602
+    },
+    {
+      "epoch": 0.8391828182294395,
+      "loss_partial": 0.048095703125,
+      "step": 1602
+    },
+    {
+      "epoch": 0.8397066526977475,
+      "grad_norm": 1.4341669082641602,
+      "learning_rate": 1.9275107871456844e-05,
+      "loss": 0.1082,
+      "step": 1603
+    },
+    {
+      "epoch": 0.8397066526977475,
+      "loss_partial": 0.053955078125,
+      "step": 1603
+    },
+    {
+      "epoch": 0.8397066526977475,
+      "loss_partial": 0.051025390625,
+      "step": 1603
+    },
+    {
+      "epoch": 0.8402304871660555,
+      "grad_norm": 1.7683582305908203,
+      "learning_rate": 1.9273398060639294e-05,
+      "loss": 0.11,
+      "step": 1604
+    },
+    {
+      "epoch": 0.8402304871660555,
+      "loss_partial": 0.058349609375,
+      "step": 1604
+    },
+    {
+      "epoch": 0.8402304871660555,
+      "loss_partial": 0.06396484375,
+      "step": 1604
+    },
+    {
+      "epoch": 0.8407543216343636,
+      "grad_norm": 2.1519625186920166,
+      "learning_rate": 1.9271686311738898e-05,
+      "loss": 0.107,
+      "step": 1605
+    },
+    {
+      "epoch": 0.8407543216343636,
+      "loss_partial": 0.04931640625,
+      "step": 1605
+    },
+    {
+      "epoch": 0.8407543216343636,
+      "loss_partial": 0.052978515625,
+      "step": 1605
+    },
+    {
+      "epoch": 0.8412781561026715,
+      "grad_norm": 1.8938266038894653,
+      "learning_rate": 1.9269972625113392e-05,
+      "loss": 0.1105,
+      "step": 1606
+    },
+    {
+      "epoch": 0.8412781561026715,
+      "loss_partial": 0.052978515625,
+      "step": 1606
+    },
+    {
+      "epoch": 0.8412781561026715,
+      "loss_partial": 0.047119140625,
+      "step": 1606
+    },
+    {
+      "epoch": 0.8418019905709796,
+      "grad_norm": 1.815097689628601,
+      "learning_rate": 1.9268257001120935e-05,
+      "loss": 0.1109,
+      "step": 1607
+    },
+    {
+      "epoch": 0.8418019905709796,
+      "loss_partial": 0.05859375,
+      "step": 1607
+    },
+    {
+      "epoch": 0.8418019905709796,
+      "loss_partial": 0.059326171875,
+      "step": 1607
+    },
+    {
+      "epoch": 0.8423258250392875,
+      "grad_norm": 2.3827383518218994,
+      "learning_rate": 1.9266539440120077e-05,
+      "loss": 0.1089,
+      "step": 1608
+    },
+    {
+      "epoch": 0.8423258250392875,
+      "loss_partial": 0.05126953125,
+      "step": 1608
+    },
+    {
+      "epoch": 0.8423258250392875,
+      "loss_partial": 0.0546875,
+      "step": 1608
+    },
+    {
+      "epoch": 0.8428496595075956,
+      "grad_norm": 1.7825238704681396,
+      "learning_rate": 1.9264819942469777e-05,
+      "loss": 0.1118,
+      "step": 1609
+    },
+    {
+      "epoch": 0.8428496595075956,
+      "loss_partial": 0.0654296875,
+      "step": 1609
+    },
+    {
+      "epoch": 0.8428496595075956,
+      "loss_partial": 0.0615234375,
+      "step": 1609
+    },
+    {
+      "epoch": 0.8433734939759037,
+      "grad_norm": 2.321676254272461,
+      "learning_rate": 1.92630985085294e-05,
+      "loss": 0.1091,
+      "step": 1610
+    },
+    {
+      "epoch": 0.8433734939759037,
+      "loss_partial": 0.0458984375,
+      "step": 1610
+    },
+    {
+      "epoch": 0.8433734939759037,
+      "loss_partial": 0.056884765625,
+      "step": 1610
+    },
+    {
+      "epoch": 0.8438973284442116,
+      "grad_norm": 1.8623806238174438,
+      "learning_rate": 1.926137513865872e-05,
+      "loss": 0.1081,
+      "step": 1611
+    },
+    {
+      "epoch": 0.8438973284442116,
+      "loss_partial": 0.049560546875,
+      "step": 1611
+    },
+    {
+      "epoch": 0.8438973284442116,
+      "loss_partial": 0.05126953125,
+      "step": 1611
+    },
+    {
+      "epoch": 0.8444211629125197,
+      "grad_norm": 1.9532926082611084,
+      "learning_rate": 1.9259649833217903e-05,
+      "loss": 0.1116,
+      "step": 1612
+    },
+    {
+      "epoch": 0.8444211629125197,
+      "loss_partial": 0.0634765625,
+      "step": 1612
+    },
+    {
+      "epoch": 0.8444211629125197,
+      "loss_partial": 0.0546875,
+      "step": 1612
+    },
+    {
+      "epoch": 0.8449449973808276,
+      "grad_norm": 2.0811655521392822,
+      "learning_rate": 1.9257922592567534e-05,
+      "loss": 0.1082,
+      "step": 1613
+    },
+    {
+      "epoch": 0.8449449973808276,
+      "loss_partial": 0.0576171875,
+      "step": 1613
+    },
+    {
+      "epoch": 0.8449449973808276,
+      "loss_partial": 0.056640625,
+      "step": 1613
+    },
+    {
+      "epoch": 0.8454688318491357,
+      "grad_norm": 1.8560740947723389,
+      "learning_rate": 1.925619341706859e-05,
+      "loss": 0.1082,
+      "step": 1614
+    },
+    {
+      "epoch": 0.8454688318491357,
+      "loss_partial": 0.05908203125,
+      "step": 1614
+    },
+    {
+      "epoch": 0.8454688318491357,
+      "loss_partial": 0.05712890625,
+      "step": 1614
+    },
+    {
+      "epoch": 0.8459926663174437,
+      "grad_norm": 2.249019145965576,
+      "learning_rate": 1.925446230708246e-05,
+      "loss": 0.107,
+      "step": 1615
+    },
+    {
+      "epoch": 0.8459926663174437,
+      "loss_partial": 0.05419921875,
+      "step": 1615
+    },
+    {
+      "epoch": 0.8459926663174437,
+      "loss_partial": 0.05078125,
+      "step": 1615
+    },
+    {
+      "epoch": 0.8465165007857517,
+      "grad_norm": 1.9361449480056763,
+      "learning_rate": 1.925272926297094e-05,
+      "loss": 0.1098,
+      "step": 1616
+    },
+    {
+      "epoch": 0.8465165007857517,
+      "loss_partial": 0.060546875,
+      "step": 1616
+    },
+    {
+      "epoch": 0.8465165007857517,
+      "loss_partial": 0.053955078125,
+      "step": 1616
+    },
+    {
+      "epoch": 0.8470403352540598,
+      "grad_norm": 2.665325403213501,
+      "learning_rate": 1.9250994285096216e-05,
+      "loss": 0.1155,
+      "step": 1617
+    },
+    {
+      "epoch": 0.8470403352540598,
+      "loss_partial": 0.054931640625,
+      "step": 1617
+    },
+    {
+      "epoch": 0.8470403352540598,
+      "loss_partial": 0.06982421875,
+      "step": 1617
+    },
+    {
+      "epoch": 0.8475641697223677,
+      "grad_norm": 1.8120450973510742,
+      "learning_rate": 1.9249257373820893e-05,
+      "loss": 0.1111,
+      "step": 1618
+    },
+    {
+      "epoch": 0.8475641697223677,
+      "loss_partial": 0.05615234375,
+      "step": 1618
+    },
+    {
+      "epoch": 0.8475641697223677,
+      "loss_partial": 0.056640625,
+      "step": 1618
+    },
+    {
+      "epoch": 0.8480880041906758,
+      "grad_norm": 2.4133987426757812,
+      "learning_rate": 1.9247518529507977e-05,
+      "loss": 0.1109,
+      "step": 1619
+    },
+    {
+      "epoch": 0.8480880041906758,
+      "loss_partial": 0.05224609375,
+      "step": 1619
+    },
+    {
+      "epoch": 0.8480880041906758,
+      "loss_partial": 0.055419921875,
+      "step": 1619
+    },
+    {
+      "epoch": 0.8486118386589837,
+      "grad_norm": 3.459280014038086,
+      "learning_rate": 1.9245777752520874e-05,
+      "loss": 0.1096,
+      "step": 1620
+    },
+    {
+      "epoch": 0.8486118386589837,
+      "loss_partial": 0.054443359375,
+      "step": 1620
+    },
+    {
+      "epoch": 0.8486118386589837,
+      "loss_partial": 0.056396484375,
+      "step": 1620
+    },
+    {
+      "epoch": 0.8491356731272918,
+      "grad_norm": 2.784222364425659,
+      "learning_rate": 1.9244035043223393e-05,
+      "loss": 0.1113,
+      "step": 1621
+    },
+    {
+      "epoch": 0.8491356731272918,
+      "loss_partial": 0.0791015625,
+      "step": 1621
+    },
+    {
+      "epoch": 0.8491356731272918,
+      "loss_partial": 0.052001953125,
+      "step": 1621
+    },
+    {
+      "epoch": 0.8496595075955998,
+      "grad_norm": 1.9314414262771606,
+      "learning_rate": 1.9242290401979753e-05,
+      "loss": 0.1132,
+      "step": 1622
+    },
+    {
+      "epoch": 0.8496595075955998,
+      "loss_partial": 0.0458984375,
+      "step": 1622
+    },
+    {
+      "epoch": 0.8496595075955998,
+      "loss_partial": 0.060302734375,
+      "step": 1622
+    },
+    {
+      "epoch": 0.8501833420639078,
+      "grad_norm": 2.393173933029175,
+      "learning_rate": 1.924054382915457e-05,
+      "loss": 0.1105,
+      "step": 1623
+    },
+    {
+      "epoch": 0.8501833420639078,
+      "loss_partial": 0.053955078125,
+      "step": 1623
+    },
+    {
+      "epoch": 0.8501833420639078,
+      "loss_partial": 0.052734375,
+      "step": 1623
+    },
+    {
+      "epoch": 0.8507071765322158,
+      "grad_norm": 1.7557151317596436,
+      "learning_rate": 1.9238795325112867e-05,
+      "loss": 0.1113,
+      "step": 1624
+    },
+    {
+      "epoch": 0.8507071765322158,
+      "loss_partial": 0.05859375,
+      "step": 1624
+    },
+    {
+      "epoch": 0.8507071765322158,
+      "loss_partial": 0.058349609375,
+      "step": 1624
+    },
+    {
+      "epoch": 0.8512310110005238,
+      "grad_norm": 2.201958656311035,
+      "learning_rate": 1.9237044890220077e-05,
+      "loss": 0.1118,
+      "step": 1625
+    },
+    {
+      "epoch": 0.8512310110005238,
+      "loss_partial": 0.0517578125,
+      "step": 1625
+    },
+    {
+      "epoch": 0.8512310110005238,
+      "loss_partial": 0.064453125,
+      "step": 1625
+    },
+    {
+      "epoch": 0.8517548454688318,
+      "grad_norm": 2.249382257461548,
+      "learning_rate": 1.923529252484202e-05,
+      "loss": 0.1098,
+      "step": 1626
+    },
+    {
+      "epoch": 0.8517548454688318,
+      "loss_partial": 0.056640625,
+      "step": 1626
+    },
+    {
+      "epoch": 0.8517548454688318,
+      "loss_partial": 0.053466796875,
+      "step": 1626
+    },
+    {
+      "epoch": 0.8522786799371399,
+      "grad_norm": 2.320636510848999,
+      "learning_rate": 1.9233538229344937e-05,
+      "loss": 0.1082,
+      "step": 1627
+    },
+    {
+      "epoch": 0.8522786799371399,
+      "loss_partial": 0.0546875,
+      "step": 1627
+    },
+    {
+      "epoch": 0.8522786799371399,
+      "loss_partial": 0.0576171875,
+      "step": 1627
+    },
+    {
+      "epoch": 0.8528025144054479,
+      "grad_norm": 1.8088809251785278,
+      "learning_rate": 1.9231782004095464e-05,
+      "loss": 0.1128,
+      "step": 1628
+    },
+    {
+      "epoch": 0.8528025144054479,
+      "loss_partial": 0.06689453125,
+      "step": 1628
+    },
+    {
+      "epoch": 0.8528025144054479,
+      "loss_partial": 0.043701171875,
+      "step": 1628
+    },
+    {
+      "epoch": 0.8533263488737559,
+      "grad_norm": 2.011554479598999,
+      "learning_rate": 1.923002384946064e-05,
+      "loss": 0.1111,
+      "step": 1629
+    },
+    {
+      "epoch": 0.8533263488737559,
+      "loss_partial": 0.04541015625,
+      "step": 1629
+    },
+    {
+      "epoch": 0.8533263488737559,
+      "loss_partial": 0.05419921875,
+      "step": 1629
+    },
+    {
+      "epoch": 0.8538501833420639,
+      "grad_norm": 1.656667709350586,
+      "learning_rate": 1.9228263765807904e-05,
+      "loss": 0.1086,
+      "step": 1630
+    },
+    {
+      "epoch": 0.8538501833420639,
+      "loss_partial": 0.05419921875,
+      "step": 1630
+    },
+    {
+      "epoch": 0.8538501833420639,
+      "loss_partial": 0.05908203125,
+      "step": 1630
+    },
+    {
+      "epoch": 0.8543740178103719,
+      "grad_norm": 2.008493185043335,
+      "learning_rate": 1.922650175350511e-05,
+      "loss": 0.1074,
+      "step": 1631
+    },
+    {
+      "epoch": 0.8543740178103719,
+      "loss_partial": 0.057373046875,
+      "step": 1631
+    },
+    {
+      "epoch": 0.8543740178103719,
+      "loss_partial": 0.049072265625,
+      "step": 1631
+    },
+    {
+      "epoch": 0.85489785227868,
+      "grad_norm": 1.4536644220352173,
+      "learning_rate": 1.9224737812920505e-05,
+      "loss": 0.1092,
+      "step": 1632
+    },
+    {
+      "epoch": 0.85489785227868,
+      "loss_partial": 0.055908203125,
+      "step": 1632
+    },
+    {
+      "epoch": 0.85489785227868,
+      "loss_partial": 0.048828125,
+      "step": 1632
+    },
+    {
+      "epoch": 0.8554216867469879,
+      "grad_norm": 1.4529106616973877,
+      "learning_rate": 1.922297194442274e-05,
+      "loss": 0.1089,
+      "step": 1633
+    },
+    {
+      "epoch": 0.8554216867469879,
+      "loss_partial": 0.06201171875,
+      "step": 1633
+    },
+    {
+      "epoch": 0.8554216867469879,
+      "loss_partial": 0.05078125,
+      "step": 1633
+    },
+    {
+      "epoch": 0.855945521215296,
+      "grad_norm": 2.0191829204559326,
+      "learning_rate": 1.922120414838087e-05,
+      "loss": 0.1111,
+      "step": 1634
+    },
+    {
+      "epoch": 0.855945521215296,
+      "loss_partial": 0.047119140625,
+      "step": 1634
+    },
+    {
+      "epoch": 0.855945521215296,
+      "loss_partial": 0.068359375,
+      "step": 1634
+    },
+    {
+      "epoch": 0.8564693556836039,
+      "grad_norm": 2.0001566410064697,
+      "learning_rate": 1.9219434425164357e-05,
+      "loss": 0.1115,
+      "step": 1635
+    },
+    {
+      "epoch": 0.8564693556836039,
+      "loss_partial": 0.046875,
+      "step": 1635
+    },
+    {
+      "epoch": 0.8564693556836039,
+      "loss_partial": 0.056884765625,
+      "step": 1635
+    },
+    {
+      "epoch": 0.856993190151912,
+      "grad_norm": 1.8857927322387695,
+      "learning_rate": 1.921766277514306e-05,
+      "loss": 0.1073,
+      "step": 1636
+    },
+    {
+      "epoch": 0.856993190151912,
+      "loss_partial": 0.05810546875,
+      "step": 1636
+    },
+    {
+      "epoch": 0.856993190151912,
+      "loss_partial": 0.04638671875,
+      "step": 1636
+    },
+    {
+      "epoch": 0.8575170246202201,
+      "grad_norm": 1.5739461183547974,
+      "learning_rate": 1.9215889198687245e-05,
+      "loss": 0.1074,
+      "step": 1637
+    },
+    {
+      "epoch": 0.8575170246202201,
+      "loss_partial": 0.059326171875,
+      "step": 1637
+    },
+    {
+      "epoch": 0.8575170246202201,
+      "loss_partial": 0.05029296875,
+      "step": 1637
+    },
+    {
+      "epoch": 0.858040859088528,
+      "grad_norm": 2.1260793209075928,
+      "learning_rate": 1.9214113696167572e-05,
+      "loss": 0.1085,
+      "step": 1638
+    },
+    {
+      "epoch": 0.858040859088528,
+      "loss_partial": 0.04345703125,
+      "step": 1638
+    },
+    {
+      "epoch": 0.858040859088528,
+      "loss_partial": 0.051025390625,
+      "step": 1638
+    },
+    {
+      "epoch": 0.8585646935568361,
+      "grad_norm": 2.63079833984375,
+      "learning_rate": 1.921233626795512e-05,
+      "loss": 0.1091,
+      "step": 1639
+    },
+    {
+      "epoch": 0.8585646935568361,
+      "loss_partial": 0.0478515625,
+      "step": 1639
+    },
+    {
+      "epoch": 0.8585646935568361,
+      "loss_partial": 0.0576171875,
+      "step": 1639
+    },
+    {
+      "epoch": 0.859088528025144,
+      "grad_norm": 1.6363492012023926,
+      "learning_rate": 1.9210556914421352e-05,
+      "loss": 0.1088,
+      "step": 1640
+    },
+    {
+      "epoch": 0.859088528025144,
+      "loss_partial": 0.0595703125,
+      "step": 1640
+    },
+    {
+      "epoch": 0.859088528025144,
+      "loss_partial": 0.06591796875,
+      "step": 1640
+    },
+    {
+      "epoch": 0.8596123624934521,
+      "grad_norm": 1.9088526964187622,
+      "learning_rate": 1.9208775635938147e-05,
+      "loss": 0.1097,
+      "step": 1641
+    },
+    {
+      "epoch": 0.8596123624934521,
+      "loss_partial": 0.058837890625,
+      "step": 1641
+    },
+    {
+      "epoch": 0.8596123624934521,
+      "loss_partial": 0.06396484375,
+      "step": 1641
+    },
+    {
+      "epoch": 0.8601361969617601,
+      "grad_norm": 1.7899398803710938,
+      "learning_rate": 1.920699243287778e-05,
+      "loss": 0.1111,
+      "step": 1642
+    },
+    {
+      "epoch": 0.8601361969617601,
+      "loss_partial": 0.051025390625,
+      "step": 1642
+    },
+    {
+      "epoch": 0.8601361969617601,
+      "loss_partial": 0.05810546875,
+      "step": 1642
+    },
+    {
+      "epoch": 0.8606600314300681,
+      "grad_norm": 2.147757053375244,
+      "learning_rate": 1.9205207305612926e-05,
+      "loss": 0.1073,
+      "step": 1643
+    },
+    {
+      "epoch": 0.8606600314300681,
+      "loss_partial": 0.055908203125,
+      "step": 1643
+    },
+    {
+      "epoch": 0.8606600314300681,
+      "loss_partial": 0.05126953125,
+      "step": 1643
+    },
+    {
+      "epoch": 0.8611838658983761,
+      "grad_norm": 2.1444430351257324,
+      "learning_rate": 1.9203420254516675e-05,
+      "loss": 0.1078,
+      "step": 1644
+    },
+    {
+      "epoch": 0.8611838658983761,
+      "loss_partial": 0.04736328125,
+      "step": 1644
+    },
+    {
+      "epoch": 0.8611838658983761,
+      "loss_partial": 0.054931640625,
+      "step": 1644
+    },
+    {
+      "epoch": 0.8617077003666841,
+      "grad_norm": 1.553864598274231,
+      "learning_rate": 1.92016312799625e-05,
+      "loss": 0.1081,
+      "step": 1645
+    },
+    {
+      "epoch": 0.8617077003666841,
+      "loss_partial": 0.06494140625,
+      "step": 1645
+    },
+    {
+      "epoch": 0.8617077003666841,
+      "loss_partial": 0.050048828125,
+      "step": 1645
+    },
+    {
+      "epoch": 0.8622315348349922,
+      "grad_norm": 2.2698919773101807,
+      "learning_rate": 1.9199840382324284e-05,
+      "loss": 0.1112,
+      "step": 1646
+    },
+    {
+      "epoch": 0.8622315348349922,
+      "loss_partial": 0.052490234375,
+      "step": 1646
+    },
+    {
+      "epoch": 0.8622315348349922,
+      "loss_partial": 0.048095703125,
+      "step": 1646
+    },
+    {
+      "epoch": 0.8627553693033002,
+      "grad_norm": 1.8584504127502441,
+      "learning_rate": 1.919804756197632e-05,
+      "loss": 0.1088,
+      "step": 1647
+    },
+    {
+      "epoch": 0.8627553693033002,
+      "loss_partial": 0.052734375,
+      "step": 1647
+    },
+    {
+      "epoch": 0.8627553693033002,
+      "loss_partial": 0.0498046875,
+      "step": 1647
+    },
+    {
+      "epoch": 0.8632792037716082,
+      "grad_norm": 1.9600794315338135,
+      "learning_rate": 1.9196252819293296e-05,
+      "loss": 0.1082,
+      "step": 1648
+    },
+    {
+      "epoch": 0.8632792037716082,
+      "loss_partial": 0.06005859375,
+      "step": 1648
+    },
+    {
+      "epoch": 0.8632792037716082,
+      "loss_partial": 0.057373046875,
+      "step": 1648
+    },
+    {
+      "epoch": 0.8638030382399162,
+      "grad_norm": 1.7690579891204834,
+      "learning_rate": 1.9194456154650303e-05,
+      "loss": 0.1094,
+      "step": 1649
+    },
+    {
+      "epoch": 0.8638030382399162,
+      "loss_partial": 0.051025390625,
+      "step": 1649
+    },
+    {
+      "epoch": 0.8638030382399162,
+      "loss_partial": 0.068359375,
+      "step": 1649
+    },
+    {
+      "epoch": 0.8643268727082242,
+      "grad_norm": 1.9679245948791504,
+      "learning_rate": 1.9192657568422827e-05,
+      "loss": 0.1072,
+      "step": 1650
+    },
+    {
+      "epoch": 0.8643268727082242,
+      "loss_partial": 0.048095703125,
+      "step": 1650
+    },
+    {
+      "epoch": 0.8643268727082242,
+      "loss_partial": 0.05126953125,
+      "step": 1650
+    },
+    {
+      "epoch": 0.8648507071765322,
+      "grad_norm": 1.7143484354019165,
+      "learning_rate": 1.9190857060986764e-05,
+      "loss": 0.1083,
+      "step": 1651
+    },
+    {
+      "epoch": 0.8648507071765322,
+      "loss_partial": 0.052734375,
+      "step": 1651
+    },
+    {
+      "epoch": 0.8648507071765322,
+      "loss_partial": 0.052978515625,
+      "step": 1651
+    },
+    {
+      "epoch": 0.8653745416448402,
+      "grad_norm": 1.707980990409851,
+      "learning_rate": 1.9189054632718416e-05,
+      "loss": 0.1098,
+      "step": 1652
+    },
+    {
+      "epoch": 0.8653745416448402,
+      "loss_partial": 0.050048828125,
+      "step": 1652
+    },
+    {
+      "epoch": 0.8653745416448402,
+      "loss_partial": 0.055908203125,
+      "step": 1652
+    },
+    {
+      "epoch": 0.8658983761131482,
+      "grad_norm": 1.5869792699813843,
+      "learning_rate": 1.918725028399447e-05,
+      "loss": 0.1099,
+      "step": 1653
+    },
+    {
+      "epoch": 0.8658983761131482,
+      "loss_partial": 0.051513671875,
+      "step": 1653
+    },
+    {
+      "epoch": 0.8658983761131482,
+      "loss_partial": 0.044677734375,
+      "step": 1653
+    },
+    {
+      "epoch": 0.8664222105814563,
+      "grad_norm": 1.662856101989746,
+      "learning_rate": 1.9185444015192022e-05,
+      "loss": 0.1088,
+      "step": 1654
+    },
+    {
+      "epoch": 0.8664222105814563,
+      "loss_partial": 0.056640625,
+      "step": 1654
+    },
+    {
+      "epoch": 0.8664222105814563,
+      "loss_partial": 0.050537109375,
+      "step": 1654
+    },
+    {
+      "epoch": 0.8669460450497642,
+      "grad_norm": 1.7942402362823486,
+      "learning_rate": 1.918363582668858e-05,
+      "loss": 0.1099,
+      "step": 1655
+    },
+    {
+      "epoch": 0.8669460450497642,
+      "loss_partial": 0.051513671875,
+      "step": 1655
+    },
+    {
+      "epoch": 0.8669460450497642,
+      "loss_partial": 0.052978515625,
+      "step": 1655
+    },
+    {
+      "epoch": 0.8674698795180723,
+      "grad_norm": 2.390139579772949,
+      "learning_rate": 1.918182571886204e-05,
+      "loss": 0.1093,
+      "step": 1656
+    },
+    {
+      "epoch": 0.8674698795180723,
+      "loss_partial": 0.048583984375,
+      "step": 1656
+    },
+    {
+      "epoch": 0.8674698795180723,
+      "loss_partial": 0.053955078125,
+      "step": 1656
+    },
+    {
+      "epoch": 0.8679937139863803,
+      "grad_norm": 1.9993551969528198,
+      "learning_rate": 1.9180013692090704e-05,
+      "loss": 0.1111,
+      "step": 1657
+    },
+    {
+      "epoch": 0.8679937139863803,
+      "loss_partial": 0.0517578125,
+      "step": 1657
+    },
+    {
+      "epoch": 0.8679937139863803,
+      "loss_partial": 0.0478515625,
+      "step": 1657
+    },
+    {
+      "epoch": 0.8685175484546883,
+      "grad_norm": 1.6243770122528076,
+      "learning_rate": 1.9178199746753275e-05,
+      "loss": 0.1041,
+      "step": 1658
+    },
+    {
+      "epoch": 0.8685175484546883,
+      "loss_partial": 0.05908203125,
+      "step": 1658
+    },
+    {
+      "epoch": 0.8685175484546883,
+      "loss_partial": 0.05615234375,
+      "step": 1658
+    },
+    {
+      "epoch": 0.8690413829229964,
+      "grad_norm": 2.2548511028289795,
+      "learning_rate": 1.9176383883228852e-05,
+      "loss": 0.1132,
+      "step": 1659
+    },
+    {
+      "epoch": 0.8690413829229964,
+      "loss_partial": 0.06396484375,
+      "step": 1659
+    },
+    {
+      "epoch": 0.8690413829229964,
+      "loss_partial": 0.057861328125,
+      "step": 1659
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "grad_norm": 2.053457260131836,
+      "learning_rate": 1.9174566101896947e-05,
+      "loss": 0.1075,
+      "step": 1660
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "loss_partial": 0.049560546875,
+      "step": 1660
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "loss_partial": 0.0517578125,
+      "step": 1660
+    },
+    {
+      "epoch": 0.8700890518596124,
+      "grad_norm": 1.8575297594070435,
+      "learning_rate": 1.9172746403137457e-05,
+      "loss": 0.1065,
+      "step": 1661
+    },
+    {
+      "epoch": 0.8700890518596124,
+      "loss_partial": 0.059326171875,
+      "step": 1661
+    },
+    {
+      "epoch": 0.8700890518596124,
+      "loss_partial": 0.05029296875,
+      "step": 1661
+    },
+    {
+      "epoch": 0.8706128863279203,
+      "grad_norm": 2.0407707691192627,
+      "learning_rate": 1.9170924787330693e-05,
+      "loss": 0.1106,
+      "step": 1662
+    },
+    {
+      "epoch": 0.8706128863279203,
+      "loss_partial": 0.057861328125,
+      "step": 1662
+    },
+    {
+      "epoch": 0.8706128863279203,
+      "loss_partial": 0.0615234375,
+      "step": 1662
+    },
+    {
+      "epoch": 0.8711367207962284,
+      "grad_norm": 1.6480375528335571,
+      "learning_rate": 1.9169101254857355e-05,
+      "loss": 0.1121,
+      "step": 1663
+    },
+    {
+      "epoch": 0.8711367207962284,
+      "loss_partial": 0.055419921875,
+      "step": 1663
+    },
+    {
+      "epoch": 0.8711367207962284,
+      "loss_partial": 0.050048828125,
+      "step": 1663
+    },
+    {
+      "epoch": 0.8716605552645365,
+      "grad_norm": 2.178867816925049,
+      "learning_rate": 1.916727580609856e-05,
+      "loss": 0.1115,
+      "step": 1664
+    },
+    {
+      "epoch": 0.8716605552645365,
+      "loss_partial": 0.0498046875,
+      "step": 1664
+    },
+    {
+      "epoch": 0.8716605552645365,
+      "loss_partial": 0.053955078125,
+      "step": 1664
+    },
+    {
+      "epoch": 0.8721843897328444,
+      "grad_norm": 2.845778465270996,
+      "learning_rate": 1.9165448441435807e-05,
+      "loss": 0.1063,
+      "step": 1665
+    },
+    {
+      "epoch": 0.8721843897328444,
+      "loss_partial": 0.0556640625,
+      "step": 1665
+    },
+    {
+      "epoch": 0.8721843897328444,
+      "loss_partial": 0.049560546875,
+      "step": 1665
+    },
+    {
+      "epoch": 0.8727082242011525,
+      "grad_norm": 2.066629648208618,
+      "learning_rate": 1.9163619161251007e-05,
+      "loss": 0.1118,
+      "step": 1666
+    },
+    {
+      "epoch": 0.8727082242011525,
+      "loss_partial": 0.06005859375,
+      "step": 1666
+    },
+    {
+      "epoch": 0.8727082242011525,
+      "loss_partial": 0.06298828125,
+      "step": 1666
+    },
+    {
+      "epoch": 0.8732320586694604,
+      "grad_norm": 2.4555771350860596,
+      "learning_rate": 1.916178796592647e-05,
+      "loss": 0.1101,
+      "step": 1667
+    },
+    {
+      "epoch": 0.8732320586694604,
+      "loss_partial": 0.060546875,
+      "step": 1667
+    },
+    {
+      "epoch": 0.8732320586694604,
+      "loss_partial": 0.046630859375,
+      "step": 1667
+    },
+    {
+      "epoch": 0.8737558931377685,
+      "grad_norm": 1.983057975769043,
+      "learning_rate": 1.9159954855844906e-05,
+      "loss": 0.1084,
+      "step": 1668
+    },
+    {
+      "epoch": 0.8737558931377685,
+      "loss_partial": 0.06494140625,
+      "step": 1668
+    },
+    {
+      "epoch": 0.8737558931377685,
+      "loss_partial": 0.05224609375,
+      "step": 1668
+    },
+    {
+      "epoch": 0.8742797276060765,
+      "grad_norm": 1.984871506690979,
+      "learning_rate": 1.915811983138942e-05,
+      "loss": 0.1093,
+      "step": 1669
+    },
+    {
+      "epoch": 0.8742797276060765,
+      "loss_partial": 0.056396484375,
+      "step": 1669
+    },
+    {
+      "epoch": 0.8742797276060765,
+      "loss_partial": 0.048095703125,
+      "step": 1669
+    },
+    {
+      "epoch": 0.8748035620743845,
+      "grad_norm": 1.8746002912521362,
+      "learning_rate": 1.9156282892943518e-05,
+      "loss": 0.1051,
+      "step": 1670
+    },
+    {
+      "epoch": 0.8748035620743845,
+      "loss_partial": 0.056640625,
+      "step": 1670
+    },
+    {
+      "epoch": 0.8748035620743845,
+      "loss_partial": 0.0556640625,
+      "step": 1670
+    },
+    {
+      "epoch": 0.8753273965426925,
+      "grad_norm": 1.6507649421691895,
+      "learning_rate": 1.9154444040891113e-05,
+      "loss": 0.1111,
+      "step": 1671
+    },
+    {
+      "epoch": 0.8753273965426925,
+      "loss_partial": 0.0546875,
+      "step": 1671
+    },
+    {
+      "epoch": 0.8753273965426925,
+      "loss_partial": 0.052001953125,
+      "step": 1671
+    },
+    {
+      "epoch": 0.8758512310110005,
+      "grad_norm": 2.4615557193756104,
+      "learning_rate": 1.915260327561651e-05,
+      "loss": 0.1081,
+      "step": 1672
+    },
+    {
+      "epoch": 0.8758512310110005,
+      "loss_partial": 0.07275390625,
+      "step": 1672
+    },
+    {
+      "epoch": 0.8758512310110005,
+      "loss_partial": 0.052978515625,
+      "step": 1672
+    },
+    {
+      "epoch": 0.8763750654793085,
+      "grad_norm": 2.6518633365631104,
+      "learning_rate": 1.9150760597504424e-05,
+      "loss": 0.1097,
+      "step": 1673
+    },
+    {
+      "epoch": 0.8763750654793085,
+      "loss_partial": 0.05810546875,
+      "step": 1673
+    },
+    {
+      "epoch": 0.8763750654793085,
+      "loss_partial": 0.048095703125,
+      "step": 1673
+    },
+    {
+      "epoch": 0.8768988999476166,
+      "grad_norm": 1.6492210626602173,
+      "learning_rate": 1.914891600693996e-05,
+      "loss": 0.1083,
+      "step": 1674
+    },
+    {
+      "epoch": 0.8768988999476166,
+      "loss_partial": 0.056640625,
+      "step": 1674
+    },
+    {
+      "epoch": 0.8768988999476166,
+      "loss_partial": 0.058837890625,
+      "step": 1674
+    },
+    {
+      "epoch": 0.8774227344159246,
+      "grad_norm": 2.0648584365844727,
+      "learning_rate": 1.9147069504308624e-05,
+      "loss": 0.1135,
+      "step": 1675
+    },
+    {
+      "epoch": 0.8774227344159246,
+      "loss_partial": 0.06005859375,
+      "step": 1675
+    },
+    {
+      "epoch": 0.8774227344159246,
+      "loss_partial": 0.05224609375,
+      "step": 1675
+    },
+    {
+      "epoch": 0.8779465688842326,
+      "grad_norm": 2.0522682666778564,
+      "learning_rate": 1.9145221089996324e-05,
+      "loss": 0.1082,
+      "step": 1676
+    },
+    {
+      "epoch": 0.8779465688842326,
+      "loss_partial": 0.06640625,
+      "step": 1676
+    },
+    {
+      "epoch": 0.8779465688842326,
+      "loss_partial": 0.04638671875,
+      "step": 1676
+    },
+    {
+      "epoch": 0.8784704033525406,
+      "grad_norm": 1.73121976852417,
+      "learning_rate": 1.9143370764389374e-05,
+      "loss": 0.1085,
+      "step": 1677
+    },
+    {
+      "epoch": 0.8784704033525406,
+      "loss_partial": 0.056640625,
+      "step": 1677
+    },
+    {
+      "epoch": 0.8784704033525406,
+      "loss_partial": 0.06591796875,
+      "step": 1677
+    },
+    {
+      "epoch": 0.8789942378208486,
+      "grad_norm": 2.4886765480041504,
+      "learning_rate": 1.914151852787447e-05,
+      "loss": 0.1127,
+      "step": 1678
+    },
+    {
+      "epoch": 0.8789942378208486,
+      "loss_partial": 0.05712890625,
+      "step": 1678
+    },
+    {
+      "epoch": 0.8789942378208486,
+      "loss_partial": 0.057373046875,
+      "step": 1678
+    },
+    {
+      "epoch": 0.8795180722891566,
+      "grad_norm": 2.6789846420288086,
+      "learning_rate": 1.9139664380838723e-05,
+      "loss": 0.1123,
+      "step": 1679
+    },
+    {
+      "epoch": 0.8795180722891566,
+      "loss_partial": 0.05078125,
+      "step": 1679
+    },
+    {
+      "epoch": 0.8795180722891566,
+      "loss_partial": 0.054931640625,
+      "step": 1679
+    },
+    {
+      "epoch": 0.8800419067574646,
+      "grad_norm": 1.788970708847046,
+      "learning_rate": 1.913780832366964e-05,
+      "loss": 0.1099,
+      "step": 1680
+    },
+    {
+      "epoch": 0.8800419067574646,
+      "loss_partial": 0.048095703125,
+      "step": 1680
+    },
+    {
+      "epoch": 0.8800419067574646,
+      "loss_partial": 0.04931640625,
+      "step": 1680
+    },
+    {
+      "epoch": 0.8805657412257727,
+      "grad_norm": 1.9671379327774048,
+      "learning_rate": 1.9135950356755123e-05,
+      "loss": 0.1075,
+      "step": 1681
+    },
+    {
+      "epoch": 0.8805657412257727,
+      "loss_partial": 0.05712890625,
+      "step": 1681
+    },
+    {
+      "epoch": 0.8805657412257727,
+      "loss_partial": 0.0556640625,
+      "step": 1681
+    },
+    {
+      "epoch": 0.8810895756940806,
+      "grad_norm": 2.387554407119751,
+      "learning_rate": 1.913409048048348e-05,
+      "loss": 0.11,
+      "step": 1682
+    },
+    {
+      "epoch": 0.8810895756940806,
+      "loss_partial": 0.056640625,
+      "step": 1682
+    },
+    {
+      "epoch": 0.8810895756940806,
+      "loss_partial": 0.056396484375,
+      "step": 1682
+    },
+    {
+      "epoch": 0.8816134101623887,
+      "grad_norm": 2.603356122970581,
+      "learning_rate": 1.913222869524341e-05,
+      "loss": 0.1068,
+      "step": 1683
+    },
+    {
+      "epoch": 0.8816134101623887,
+      "loss_partial": 0.050537109375,
+      "step": 1683
+    },
+    {
+      "epoch": 0.8816134101623887,
+      "loss_partial": 0.052978515625,
+      "step": 1683
+    },
+    {
+      "epoch": 0.8821372446306966,
+      "grad_norm": 1.9369661808013916,
+      "learning_rate": 1.9130365001424008e-05,
+      "loss": 0.1088,
+      "step": 1684
+    },
+    {
+      "epoch": 0.8821372446306966,
+      "loss_partial": 0.05908203125,
+      "step": 1684
+    },
+    {
+      "epoch": 0.8821372446306966,
+      "loss_partial": 0.052490234375,
+      "step": 1684
+    },
+    {
+      "epoch": 0.8826610790990047,
+      "grad_norm": 1.8845618963241577,
+      "learning_rate": 1.9128499399414787e-05,
+      "loss": 0.1137,
+      "step": 1685
+    },
+    {
+      "epoch": 0.8826610790990047,
+      "loss_partial": 0.057373046875,
+      "step": 1685
+    },
+    {
+      "epoch": 0.8826610790990047,
+      "loss_partial": 0.053955078125,
+      "step": 1685
+    },
+    {
+      "epoch": 0.8831849135673128,
+      "grad_norm": 2.140108346939087,
+      "learning_rate": 1.912663188960564e-05,
+      "loss": 0.108,
+      "step": 1686
+    },
+    {
+      "epoch": 0.8831849135673128,
+      "loss_partial": 0.054931640625,
+      "step": 1686
+    },
+    {
+      "epoch": 0.8831849135673128,
+      "loss_partial": 0.060302734375,
+      "step": 1686
+    },
+    {
+      "epoch": 0.8837087480356207,
+      "grad_norm": 2.151358127593994,
+      "learning_rate": 1.9124762472386862e-05,
+      "loss": 0.106,
+      "step": 1687
+    },
+    {
+      "epoch": 0.8837087480356207,
+      "loss_partial": 0.047119140625,
+      "step": 1687
+    },
+    {
+      "epoch": 0.8837087480356207,
+      "loss_partial": 0.050048828125,
+      "step": 1687
+    },
+    {
+      "epoch": 0.8842325825039288,
+      "grad_norm": 2.212454319000244,
+      "learning_rate": 1.9122891148149157e-05,
+      "loss": 0.1078,
+      "step": 1688
+    },
+    {
+      "epoch": 0.8842325825039288,
+      "loss_partial": 0.044921875,
+      "step": 1688
+    },
+    {
+      "epoch": 0.8842325825039288,
+      "loss_partial": 0.05029296875,
+      "step": 1688
+    },
+    {
+      "epoch": 0.8847564169722367,
+      "grad_norm": 2.0445737838745117,
+      "learning_rate": 1.9121017917283615e-05,
+      "loss": 0.1125,
+      "step": 1689
+    },
+    {
+      "epoch": 0.8847564169722367,
+      "loss_partial": 0.056396484375,
+      "step": 1689
+    },
+    {
+      "epoch": 0.8847564169722367,
+      "loss_partial": 0.044921875,
+      "step": 1689
+    },
+    {
+      "epoch": 0.8852802514405448,
+      "grad_norm": 2.813995599746704,
+      "learning_rate": 1.9119142780181732e-05,
+      "loss": 0.108,
+      "step": 1690
+    },
+    {
+      "epoch": 0.8852802514405448,
+      "loss_partial": 0.0478515625,
+      "step": 1690
+    },
+    {
+      "epoch": 0.8852802514405448,
+      "loss_partial": 0.0478515625,
+      "step": 1690
+    },
+    {
+      "epoch": 0.8858040859088528,
+      "grad_norm": 2.217163562774658,
+      "learning_rate": 1.91172657372354e-05,
+      "loss": 0.1073,
+      "step": 1691
+    },
+    {
+      "epoch": 0.8858040859088528,
+      "loss_partial": 0.0517578125,
+      "step": 1691
+    },
+    {
+      "epoch": 0.8858040859088528,
+      "loss_partial": 0.056640625,
+      "step": 1691
+    },
+    {
+      "epoch": 0.8863279203771608,
+      "grad_norm": 1.7077726125717163,
+      "learning_rate": 1.9115386788836907e-05,
+      "loss": 0.1094,
+      "step": 1692
+    },
+    {
+      "epoch": 0.8863279203771608,
+      "loss_partial": 0.05078125,
+      "step": 1692
+    },
+    {
+      "epoch": 0.8863279203771608,
+      "loss_partial": 0.060302734375,
+      "step": 1692
+    },
+    {
+      "epoch": 0.8868517548454689,
+      "grad_norm": 2.1009104251861572,
+      "learning_rate": 1.9113505935378937e-05,
+      "loss": 0.1104,
+      "step": 1693
+    },
+    {
+      "epoch": 0.8868517548454689,
+      "loss_partial": 0.0546875,
+      "step": 1693
+    },
+    {
+      "epoch": 0.8868517548454689,
+      "loss_partial": 0.049560546875,
+      "step": 1693
+    },
+    {
+      "epoch": 0.8873755893137768,
+      "grad_norm": 1.6278587579727173,
+      "learning_rate": 1.911162317725459e-05,
+      "loss": 0.1098,
+      "step": 1694
+    },
+    {
+      "epoch": 0.8873755893137768,
+      "loss_partial": 0.06005859375,
+      "step": 1694
+    },
+    {
+      "epoch": 0.8873755893137768,
+      "loss_partial": 0.04833984375,
+      "step": 1694
+    },
+    {
+      "epoch": 0.8878994237820849,
+      "grad_norm": 2.138231039047241,
+      "learning_rate": 1.910973851485734e-05,
+      "loss": 0.106,
+      "step": 1695
+    },
+    {
+      "epoch": 0.8878994237820849,
+      "loss_partial": 0.048095703125,
+      "step": 1695
+    },
+    {
+      "epoch": 0.8878994237820849,
+      "loss_partial": 0.055908203125,
+      "step": 1695
+    },
+    {
+      "epoch": 0.8884232582503929,
+      "grad_norm": 1.4803087711334229,
+      "learning_rate": 1.9107851948581078e-05,
+      "loss": 0.1088,
+      "step": 1696
+    },
+    {
+      "epoch": 0.8884232582503929,
+      "loss_partial": 0.04931640625,
+      "step": 1696
+    },
+    {
+      "epoch": 0.8884232582503929,
+      "loss_partial": 0.050537109375,
+      "step": 1696
+    },
+    {
+      "epoch": 0.8889470927187009,
+      "grad_norm": 1.9424272775650024,
+      "learning_rate": 1.9105963478820077e-05,
+      "loss": 0.1094,
+      "step": 1697
+    },
+    {
+      "epoch": 0.8889470927187009,
+      "loss_partial": 0.05908203125,
+      "step": 1697
+    },
+    {
+      "epoch": 0.8889470927187009,
+      "loss_partial": 0.049560546875,
+      "step": 1697
+    },
+    {
+      "epoch": 0.8894709271870089,
+      "grad_norm": 1.970121145248413,
+      "learning_rate": 1.9104073105969017e-05,
+      "loss": 0.1119,
+      "step": 1698
+    },
+    {
+      "epoch": 0.8894709271870089,
+      "loss_partial": 0.052978515625,
+      "step": 1698
+    },
+    {
+      "epoch": 0.8894709271870089,
+      "loss_partial": 0.0595703125,
+      "step": 1698
+    },
+    {
+      "epoch": 0.8899947616553169,
+      "grad_norm": 2.4309237003326416,
+      "learning_rate": 1.9102180830422977e-05,
+      "loss": 0.1043,
+      "step": 1699
+    },
+    {
+      "epoch": 0.8899947616553169,
+      "loss_partial": 0.060546875,
+      "step": 1699
+    },
+    {
+      "epoch": 0.8899947616553169,
+      "loss_partial": 0.0595703125,
+      "step": 1699
+    },
+    {
+      "epoch": 0.8905185961236249,
+      "grad_norm": 2.853504180908203,
+      "learning_rate": 1.9100286652577428e-05,
+      "loss": 0.1075,
+      "step": 1700
+    },
+    {
+      "epoch": 0.8905185961236249,
+      "loss_partial": 0.05078125,
+      "step": 1700
+    },
+    {
+      "epoch": 0.8905185961236249,
+      "loss_partial": 0.05224609375,
+      "step": 1700
+    },
+    {
+      "epoch": 0.891042430591933,
+      "grad_norm": 2.120314121246338,
+      "learning_rate": 1.9098390572828246e-05,
+      "loss": 0.1068,
+      "step": 1701
+    },
+    {
+      "epoch": 0.891042430591933,
+      "loss_partial": 0.059326171875,
+      "step": 1701
+    },
+    {
+      "epoch": 0.891042430591933,
+      "loss_partial": 0.059814453125,
+      "step": 1701
+    },
+    {
+      "epoch": 0.891566265060241,
+      "grad_norm": 2.156907081604004,
+      "learning_rate": 1.9096492591571697e-05,
+      "loss": 0.1099,
+      "step": 1702
+    },
+    {
+      "epoch": 0.891566265060241,
+      "loss_partial": 0.045654296875,
+      "step": 1702
+    },
+    {
+      "epoch": 0.891566265060241,
+      "loss_partial": 0.05224609375,
+      "step": 1702
+    },
+    {
+      "epoch": 0.892090099528549,
+      "grad_norm": 2.1628854274749756,
+      "learning_rate": 1.9094592709204444e-05,
+      "loss": 0.1106,
+      "step": 1703
+    },
+    {
+      "epoch": 0.892090099528549,
+      "loss_partial": 0.05126953125,
+      "step": 1703
+    },
+    {
+      "epoch": 0.892090099528549,
+      "loss_partial": 0.05615234375,
+      "step": 1703
+    },
+    {
+      "epoch": 0.892613933996857,
+      "grad_norm": 1.7908169031143188,
+      "learning_rate": 1.9092690926123558e-05,
+      "loss": 0.109,
+      "step": 1704
+    },
+    {
+      "epoch": 0.892613933996857,
+      "loss_partial": 0.05712890625,
+      "step": 1704
+    },
+    {
+      "epoch": 0.892613933996857,
+      "loss_partial": 0.0556640625,
+      "step": 1704
+    },
+    {
+      "epoch": 0.893137768465165,
+      "grad_norm": 2.770514488220215,
+      "learning_rate": 1.9090787242726495e-05,
+      "loss": 0.1094,
+      "step": 1705
+    },
+    {
+      "epoch": 0.893137768465165,
+      "loss_partial": 0.05615234375,
+      "step": 1705
+    },
+    {
+      "epoch": 0.893137768465165,
+      "loss_partial": 0.06298828125,
+      "step": 1705
+    },
+    {
+      "epoch": 0.8936616029334731,
+      "grad_norm": 2.705792188644409,
+      "learning_rate": 1.9088881659411115e-05,
+      "loss": 0.1101,
+      "step": 1706
+    },
+    {
+      "epoch": 0.8936616029334731,
+      "loss_partial": 0.0654296875,
+      "step": 1706
+    },
+    {
+      "epoch": 0.8936616029334731,
+      "loss_partial": 0.05224609375,
+      "step": 1706
+    },
+    {
+      "epoch": 0.894185437401781,
+      "grad_norm": 2.437326669692993,
+      "learning_rate": 1.9086974176575667e-05,
+      "loss": 0.1113,
+      "step": 1707
+    },
+    {
+      "epoch": 0.894185437401781,
+      "loss_partial": 0.05615234375,
+      "step": 1707
+    },
+    {
+      "epoch": 0.894185437401781,
+      "loss_partial": 0.060546875,
+      "step": 1707
+    },
+    {
+      "epoch": 0.8947092718700891,
+      "grad_norm": 2.090681552886963,
+      "learning_rate": 1.9085064794618812e-05,
+      "loss": 0.1071,
+      "step": 1708
+    },
+    {
+      "epoch": 0.8947092718700891,
+      "loss_partial": 0.06591796875,
+      "step": 1708
+    },
+    {
+      "epoch": 0.8947092718700891,
+      "loss_partial": 0.0517578125,
+      "step": 1708
+    },
+    {
+      "epoch": 0.895233106338397,
+      "grad_norm": 1.737603783607483,
+      "learning_rate": 1.9083153513939594e-05,
+      "loss": 0.1097,
+      "step": 1709
+    },
+    {
+      "epoch": 0.895233106338397,
+      "loss_partial": 0.0634765625,
+      "step": 1709
+    },
+    {
+      "epoch": 0.895233106338397,
+      "loss_partial": 0.056884765625,
+      "step": 1709
+    },
+    {
+      "epoch": 0.8957569408067051,
+      "grad_norm": 1.9801112413406372,
+      "learning_rate": 1.9081240334937463e-05,
+      "loss": 0.109,
+      "step": 1710
+    },
+    {
+      "epoch": 0.8957569408067051,
+      "loss_partial": 0.049560546875,
+      "step": 1710
+    },
+    {
+      "epoch": 0.8957569408067051,
+      "loss_partial": 0.0576171875,
+      "step": 1710
+    },
+    {
+      "epoch": 0.896280775275013,
+      "grad_norm": 2.1818413734436035,
+      "learning_rate": 1.9079325258012257e-05,
+      "loss": 0.1072,
+      "step": 1711
+    },
+    {
+      "epoch": 0.896280775275013,
+      "loss_partial": 0.049072265625,
+      "step": 1711
+    },
+    {
+      "epoch": 0.896280775275013,
+      "loss_partial": 0.052490234375,
+      "step": 1711
+    },
+    {
+      "epoch": 0.8968046097433211,
+      "grad_norm": 2.387444496154785,
+      "learning_rate": 1.9077408283564215e-05,
+      "loss": 0.1097,
+      "step": 1712
+    },
+    {
+      "epoch": 0.8968046097433211,
+      "loss_partial": 0.059814453125,
+      "step": 1712
+    },
+    {
+      "epoch": 0.8968046097433211,
+      "loss_partial": 0.060302734375,
+      "step": 1712
+    },
+    {
+      "epoch": 0.8973284442116292,
+      "grad_norm": 1.5536004304885864,
+      "learning_rate": 1.9075489411993977e-05,
+      "loss": 0.111,
+      "step": 1713
+    },
+    {
+      "epoch": 0.8973284442116292,
+      "loss_partial": 0.051513671875,
+      "step": 1713
+    },
+    {
+      "epoch": 0.8973284442116292,
+      "loss_partial": 0.05078125,
+      "step": 1713
+    },
+    {
+      "epoch": 0.8978522786799371,
+      "grad_norm": 2.131378173828125,
+      "learning_rate": 1.9073568643702572e-05,
+      "loss": 0.1098,
+      "step": 1714
+    },
+    {
+      "epoch": 0.8978522786799371,
+      "loss_partial": 0.061279296875,
+      "step": 1714
+    },
+    {
+      "epoch": 0.8978522786799371,
+      "loss_partial": 0.06005859375,
+      "step": 1714
+    },
+    {
+      "epoch": 0.8983761131482452,
+      "grad_norm": 2.178826093673706,
+      "learning_rate": 1.907164597909143e-05,
+      "loss": 0.111,
+      "step": 1715
+    },
+    {
+      "epoch": 0.8983761131482452,
+      "loss_partial": 0.055908203125,
+      "step": 1715
+    },
+    {
+      "epoch": 0.8983761131482452,
+      "loss_partial": 0.05859375,
+      "step": 1715
+    },
+    {
+      "epoch": 0.8988999476165531,
+      "grad_norm": 2.26713490486145,
+      "learning_rate": 1.9069721418562377e-05,
+      "loss": 0.1099,
+      "step": 1716
+    },
+    {
+      "epoch": 0.8988999476165531,
+      "loss_partial": 0.044921875,
+      "step": 1716
+    },
+    {
+      "epoch": 0.8988999476165531,
+      "loss_partial": 0.049560546875,
+      "step": 1716
+    },
+    {
+      "epoch": 0.8994237820848612,
+      "grad_norm": 2.2869341373443604,
+      "learning_rate": 1.906779496251763e-05,
+      "loss": 0.11,
+      "step": 1717
+    },
+    {
+      "epoch": 0.8994237820848612,
+      "loss_partial": 0.055908203125,
+      "step": 1717
+    },
+    {
+      "epoch": 0.8994237820848612,
+      "loss_partial": 0.05126953125,
+      "step": 1717
+    },
+    {
+      "epoch": 0.8999476165531692,
+      "grad_norm": 1.8970179557800293,
+      "learning_rate": 1.9065866611359813e-05,
+      "loss": 0.1107,
+      "step": 1718
+    },
+    {
+      "epoch": 0.8999476165531692,
+      "loss_partial": 0.06640625,
+      "step": 1718
+    },
+    {
+      "epoch": 0.8999476165531692,
+      "loss_partial": 0.06689453125,
+      "step": 1718
+    },
+    {
+      "epoch": 0.9004714510214772,
+      "grad_norm": 2.1306285858154297,
+      "learning_rate": 1.9063936365491936e-05,
+      "loss": 0.1147,
+      "step": 1719
+    },
+    {
+      "epoch": 0.9004714510214772,
+      "loss_partial": 0.057373046875,
+      "step": 1719
+    },
+    {
+      "epoch": 0.9004714510214772,
+      "loss_partial": 0.051025390625,
+      "step": 1719
+    },
+    {
+      "epoch": 0.9009952854897852,
+      "grad_norm": 1.8629616498947144,
+      "learning_rate": 1.9062004225317404e-05,
+      "loss": 0.1091,
+      "step": 1720
+    },
+    {
+      "epoch": 0.9009952854897852,
+      "loss_partial": 0.05126953125,
+      "step": 1720
+    },
+    {
+      "epoch": 0.9009952854897852,
+      "loss_partial": 0.050048828125,
+      "step": 1720
+    },
+    {
+      "epoch": 0.9015191199580932,
+      "grad_norm": 2.5306336879730225,
+      "learning_rate": 1.906007019124003e-05,
+      "loss": 0.1102,
+      "step": 1721
+    },
+    {
+      "epoch": 0.9015191199580932,
+      "loss_partial": 0.044677734375,
+      "step": 1721
+    },
+    {
+      "epoch": 0.9015191199580932,
+      "loss_partial": 0.046142578125,
+      "step": 1721
+    },
+    {
+      "epoch": 0.9020429544264013,
+      "grad_norm": 2.1164424419403076,
+      "learning_rate": 1.9058134263664012e-05,
+      "loss": 0.1055,
+      "step": 1722
+    },
+    {
+      "epoch": 0.9020429544264013,
+      "loss_partial": 0.049072265625,
+      "step": 1722
+    },
+    {
+      "epoch": 0.9020429544264013,
+      "loss_partial": 0.05517578125,
+      "step": 1722
+    },
+    {
+      "epoch": 0.9025667888947093,
+      "grad_norm": 2.2969770431518555,
+      "learning_rate": 1.9056196442993943e-05,
+      "loss": 0.1106,
+      "step": 1723
+    },
+    {
+      "epoch": 0.9025667888947093,
+      "loss_partial": 0.052490234375,
+      "step": 1723
+    },
+    {
+      "epoch": 0.9025667888947093,
+      "loss_partial": 0.048583984375,
+      "step": 1723
+    },
+    {
+      "epoch": 0.9030906233630173,
+      "grad_norm": 3.0669808387756348,
+      "learning_rate": 1.9054256729634822e-05,
+      "loss": 0.1041,
+      "step": 1724
+    },
+    {
+      "epoch": 0.9030906233630173,
+      "loss_partial": 0.046875,
+      "step": 1724
+    },
+    {
+      "epoch": 0.9030906233630173,
+      "loss_partial": 0.056884765625,
+      "step": 1724
+    },
+    {
+      "epoch": 0.9036144578313253,
+      "grad_norm": 1.6878187656402588,
+      "learning_rate": 1.9052315123992035e-05,
+      "loss": 0.1091,
+      "step": 1725
+    },
+    {
+      "epoch": 0.9036144578313253,
+      "loss_partial": 0.054443359375,
+      "step": 1725
+    },
+    {
+      "epoch": 0.9036144578313253,
+      "loss_partial": 0.043701171875,
+      "step": 1725
+    },
+    {
+      "epoch": 0.9041382922996333,
+      "grad_norm": 2.3556339740753174,
+      "learning_rate": 1.9050371626471365e-05,
+      "loss": 0.1054,
+      "step": 1726
+    },
+    {
+      "epoch": 0.9041382922996333,
+      "loss_partial": 0.060302734375,
+      "step": 1726
+    },
+    {
+      "epoch": 0.9041382922996333,
+      "loss_partial": 0.0595703125,
+      "step": 1726
+    },
+    {
+      "epoch": 0.9046621267679413,
+      "grad_norm": 2.9915502071380615,
+      "learning_rate": 1.904842623747899e-05,
+      "loss": 0.1119,
+      "step": 1727
+    },
+    {
+      "epoch": 0.9046621267679413,
+      "loss_partial": 0.06103515625,
+      "step": 1727
+    },
+    {
+      "epoch": 0.9046621267679413,
+      "loss_partial": 0.046630859375,
+      "step": 1727
+    },
+    {
+      "epoch": 0.9051859612362494,
+      "grad_norm": 2.4803545475006104,
+      "learning_rate": 1.9046478957421484e-05,
+      "loss": 0.1036,
+      "step": 1728
+    },
+    {
+      "epoch": 0.9051859612362494,
+      "loss_partial": 0.05126953125,
+      "step": 1728
+    },
+    {
+      "epoch": 0.9051859612362494,
+      "loss_partial": 0.051513671875,
+      "step": 1728
+    },
+    {
+      "epoch": 0.9057097957045573,
+      "grad_norm": 2.1227309703826904,
+      "learning_rate": 1.904452978670582e-05,
+      "loss": 0.1084,
+      "step": 1729
+    },
+    {
+      "epoch": 0.9057097957045573,
+      "loss_partial": 0.053466796875,
+      "step": 1729
+    },
+    {
+      "epoch": 0.9057097957045573,
+      "loss_partial": 0.0625,
+      "step": 1729
+    },
+    {
+      "epoch": 0.9062336301728654,
+      "grad_norm": 3.5367751121520996,
+      "learning_rate": 1.9042578725739357e-05,
+      "loss": 0.1088,
+      "step": 1730
+    },
+    {
+      "epoch": 0.9062336301728654,
+      "loss_partial": 0.0712890625,
+      "step": 1730
+    },
+    {
+      "epoch": 0.9062336301728654,
+      "loss_partial": 0.054443359375,
+      "step": 1730
+    },
+    {
+      "epoch": 0.9067574646411733,
+      "grad_norm": 3.1747753620147705,
+      "learning_rate": 1.904062577492986e-05,
+      "loss": 0.1108,
+      "step": 1731
+    },
+    {
+      "epoch": 0.9067574646411733,
+      "loss_partial": 0.05029296875,
+      "step": 1731
+    },
+    {
+      "epoch": 0.9067574646411733,
+      "loss_partial": 0.0517578125,
+      "step": 1731
+    },
+    {
+      "epoch": 0.9072812991094814,
+      "grad_norm": 2.378612995147705,
+      "learning_rate": 1.9038670934685484e-05,
+      "loss": 0.11,
+      "step": 1732
+    },
+    {
+      "epoch": 0.9072812991094814,
+      "loss_partial": 0.052490234375,
+      "step": 1732
+    },
+    {
+      "epoch": 0.9072812991094814,
+      "loss_partial": 0.052978515625,
+      "step": 1732
+    },
+    {
+      "epoch": 0.9078051335777895,
+      "grad_norm": 2.4380006790161133,
+      "learning_rate": 1.9036714205414775e-05,
+      "loss": 0.1041,
+      "step": 1733
+    },
+    {
+      "epoch": 0.9078051335777895,
+      "loss_partial": 0.049560546875,
+      "step": 1733
+    },
+    {
+      "epoch": 0.9078051335777895,
+      "loss_partial": 0.0517578125,
+      "step": 1733
+    },
+    {
+      "epoch": 0.9083289680460974,
+      "grad_norm": 1.970903754234314,
+      "learning_rate": 1.9034755587526677e-05,
+      "loss": 0.1046,
+      "step": 1734
+    },
+    {
+      "epoch": 0.9083289680460974,
+      "loss_partial": 0.051513671875,
+      "step": 1734
+    },
+    {
+      "epoch": 0.9083289680460974,
+      "loss_partial": 0.059814453125,
+      "step": 1734
+    },
+    {
+      "epoch": 0.9088528025144055,
+      "grad_norm": 2.480837821960449,
+      "learning_rate": 1.9032795081430527e-05,
+      "loss": 0.1131,
+      "step": 1735
+    },
+    {
+      "epoch": 0.9088528025144055,
+      "loss_partial": 0.053466796875,
+      "step": 1735
+    },
+    {
+      "epoch": 0.9088528025144055,
+      "loss_partial": 0.052978515625,
+      "step": 1735
+    },
+    {
+      "epoch": 0.9093766369827134,
+      "grad_norm": 1.687261939048767,
+      "learning_rate": 1.9030832687536066e-05,
+      "loss": 0.1094,
+      "step": 1736
+    },
+    {
+      "epoch": 0.9093766369827134,
+      "loss_partial": 0.0458984375,
+      "step": 1736
+    },
+    {
+      "epoch": 0.9093766369827134,
+      "loss_partial": 0.050048828125,
+      "step": 1736
+    },
+    {
+      "epoch": 0.9099004714510215,
+      "grad_norm": 3.0373117923736572,
+      "learning_rate": 1.902886840625342e-05,
+      "loss": 0.1105,
+      "step": 1737
+    },
+    {
+      "epoch": 0.9099004714510215,
+      "loss_partial": 0.046630859375,
+      "step": 1737
+    },
+    {
+      "epoch": 0.9099004714510215,
+      "loss_partial": 0.0458984375,
+      "step": 1737
+    },
+    {
+      "epoch": 0.9104243059193295,
+      "grad_norm": 1.8329365253448486,
+      "learning_rate": 1.9026902237993105e-05,
+      "loss": 0.109,
+      "step": 1738
+    },
+    {
+      "epoch": 0.9104243059193295,
+      "loss_partial": 0.051025390625,
+      "step": 1738
+    },
+    {
+      "epoch": 0.9104243059193295,
+      "loss_partial": 0.04638671875,
+      "step": 1738
+    },
+    {
+      "epoch": 0.9109481403876375,
+      "grad_norm": 2.390298366546631,
+      "learning_rate": 1.9024934183166048e-05,
+      "loss": 0.1102,
+      "step": 1739
+    },
+    {
+      "epoch": 0.9109481403876375,
+      "loss_partial": 0.0595703125,
+      "step": 1739
+    },
+    {
+      "epoch": 0.9109481403876375,
+      "loss_partial": 0.057861328125,
+      "step": 1739
+    },
+    {
+      "epoch": 0.9114719748559456,
+      "grad_norm": 1.7328333854675293,
+      "learning_rate": 1.9022964242183554e-05,
+      "loss": 0.1076,
+      "step": 1740
+    },
+    {
+      "epoch": 0.9114719748559456,
+      "loss_partial": 0.050537109375,
+      "step": 1740
+    },
+    {
+      "epoch": 0.9114719748559456,
+      "loss_partial": 0.060546875,
+      "step": 1740
+    },
+    {
+      "epoch": 0.9119958093242535,
+      "grad_norm": 1.5134295225143433,
+      "learning_rate": 1.9020992415457326e-05,
+      "loss": 0.1073,
+      "step": 1741
+    },
+    {
+      "epoch": 0.9119958093242535,
+      "loss_partial": 0.06201171875,
+      "step": 1741
+    },
+    {
+      "epoch": 0.9119958093242535,
+      "loss_partial": 0.052001953125,
+      "step": 1741
+    },
+    {
+      "epoch": 0.9125196437925616,
+      "grad_norm": 1.8764076232910156,
+      "learning_rate": 1.901901870339947e-05,
+      "loss": 0.106,
+      "step": 1742
+    },
+    {
+      "epoch": 0.9125196437925616,
+      "loss_partial": 0.052978515625,
+      "step": 1742
+    },
+    {
+      "epoch": 0.9125196437925616,
+      "loss_partial": 0.059326171875,
+      "step": 1742
+    },
+    {
+      "epoch": 0.9130434782608695,
+      "grad_norm": 1.5072227716445923,
+      "learning_rate": 1.901704310642248e-05,
+      "loss": 0.106,
+      "step": 1743
+    },
+    {
+      "epoch": 0.9130434782608695,
+      "loss_partial": 0.05322265625,
+      "step": 1743
+    },
+    {
+      "epoch": 0.9130434782608695,
+      "loss_partial": 0.0517578125,
+      "step": 1743
+    },
+    {
+      "epoch": 0.9135673127291776,
+      "grad_norm": 1.8813716173171997,
+      "learning_rate": 1.9015065624939236e-05,
+      "loss": 0.1073,
+      "step": 1744
+    },
+    {
+      "epoch": 0.9135673127291776,
+      "loss_partial": 0.058837890625,
+      "step": 1744
+    },
+    {
+      "epoch": 0.9135673127291776,
+      "loss_partial": 0.053466796875,
+      "step": 1744
+    },
+    {
+      "epoch": 0.9140911471974856,
+      "grad_norm": 1.7198673486709595,
+      "learning_rate": 1.9013086259363028e-05,
+      "loss": 0.1121,
+      "step": 1745
+    },
+    {
+      "epoch": 0.9140911471974856,
+      "loss_partial": 0.06640625,
+      "step": 1745
+    },
+    {
+      "epoch": 0.9140911471974856,
+      "loss_partial": 0.046630859375,
+      "step": 1745
+    },
+    {
+      "epoch": 0.9146149816657936,
+      "grad_norm": 1.4824719429016113,
+      "learning_rate": 1.901110501010752e-05,
+      "loss": 0.1073,
+      "step": 1746
+    },
+    {
+      "epoch": 0.9146149816657936,
+      "loss_partial": 0.0556640625,
+      "step": 1746
+    },
+    {
+      "epoch": 0.9146149816657936,
+      "loss_partial": 0.051025390625,
+      "step": 1746
+    },
+    {
+      "epoch": 0.9151388161341016,
+      "grad_norm": 2.1680567264556885,
+      "learning_rate": 1.9009121877586793e-05,
+      "loss": 0.1086,
+      "step": 1747
+    },
+    {
+      "epoch": 0.9151388161341016,
+      "loss_partial": 0.048828125,
+      "step": 1747
+    },
+    {
+      "epoch": 0.9151388161341016,
+      "loss_partial": 0.05908203125,
+      "step": 1747
+    },
+    {
+      "epoch": 0.9156626506024096,
+      "grad_norm": 1.7723100185394287,
+      "learning_rate": 1.9007136862215306e-05,
+      "loss": 0.1088,
+      "step": 1748
+    },
+    {
+      "epoch": 0.9156626506024096,
+      "loss_partial": 0.051025390625,
+      "step": 1748
+    },
+    {
+      "epoch": 0.9156626506024096,
+      "loss_partial": 0.05322265625,
+      "step": 1748
+    },
+    {
+      "epoch": 0.9161864850707176,
+      "grad_norm": 1.850864052772522,
+      "learning_rate": 1.900514996440791e-05,
+      "loss": 0.1089,
+      "step": 1749
+    },
+    {
+      "epoch": 0.9161864850707176,
+      "loss_partial": 0.054443359375,
+      "step": 1749
+    },
+    {
+      "epoch": 0.9161864850707176,
+      "loss_partial": 0.048095703125,
+      "step": 1749
+    },
+    {
+      "epoch": 0.9167103195390257,
+      "grad_norm": 1.640876293182373,
+      "learning_rate": 1.9003161184579858e-05,
+      "loss": 0.1066,
+      "step": 1750
+    },
+    {
+      "epoch": 0.9167103195390257,
+      "loss_partial": 0.05419921875,
+      "step": 1750
+    },
+    {
+      "epoch": 0.9167103195390257,
+      "loss_partial": 0.06298828125,
+      "step": 1750
+    },
+    {
+      "epoch": 0.9172341540073337,
+      "grad_norm": 1.6866518259048462,
+      "learning_rate": 1.900117052314679e-05,
+      "loss": 0.1079,
+      "step": 1751
+    },
+    {
+      "epoch": 0.9172341540073337,
+      "loss_partial": 0.047119140625,
+      "step": 1751
+    },
+    {
+      "epoch": 0.9172341540073337,
+      "loss_partial": 0.051025390625,
+      "step": 1751
+    },
+    {
+      "epoch": 0.9177579884756417,
+      "grad_norm": 1.352431058883667,
+      "learning_rate": 1.8999177980524753e-05,
+      "loss": 0.1052,
+      "step": 1752
+    },
+    {
+      "epoch": 0.9177579884756417,
+      "loss_partial": 0.053955078125,
+      "step": 1752
+    },
+    {
+      "epoch": 0.9177579884756417,
+      "loss_partial": 0.04833984375,
+      "step": 1752
+    },
+    {
+      "epoch": 0.9182818229439497,
+      "grad_norm": 1.694630742073059,
+      "learning_rate": 1.899718355713016e-05,
+      "loss": 0.1049,
+      "step": 1753
+    },
+    {
+      "epoch": 0.9182818229439497,
+      "loss_partial": 0.04638671875,
+      "step": 1753
+    },
+    {
+      "epoch": 0.9182818229439497,
+      "loss_partial": 0.05615234375,
+      "step": 1753
+    },
+    {
+      "epoch": 0.9188056574122577,
+      "grad_norm": 2.042112112045288,
+      "learning_rate": 1.8995187253379842e-05,
+      "loss": 0.1072,
+      "step": 1754
+    },
+    {
+      "epoch": 0.9188056574122577,
+      "loss_partial": 0.059814453125,
+      "step": 1754
+    },
+    {
+      "epoch": 0.9188056574122577,
+      "loss_partial": 0.05126953125,
+      "step": 1754
+    },
+    {
+      "epoch": 0.9193294918805658,
+      "grad_norm": 2.522153377532959,
+      "learning_rate": 1.899318906969101e-05,
+      "loss": 0.103,
+      "step": 1755
+    },
+    {
+      "epoch": 0.9193294918805658,
+      "loss_partial": 0.058837890625,
+      "step": 1755
+    },
+    {
+      "epoch": 0.9193294918805658,
+      "loss_partial": 0.049560546875,
+      "step": 1755
+    },
+    {
+      "epoch": 0.9198533263488737,
+      "grad_norm": 1.687551736831665,
+      "learning_rate": 1.8991189006481278e-05,
+      "loss": 0.1132,
+      "step": 1756
+    },
+    {
+      "epoch": 0.9198533263488737,
+      "loss_partial": 0.0556640625,
+      "step": 1756
+    },
+    {
+      "epoch": 0.9198533263488737,
+      "loss_partial": 0.051513671875,
+      "step": 1756
+    },
+    {
+      "epoch": 0.9203771608171818,
+      "grad_norm": 2.2232096195220947,
+      "learning_rate": 1.8989187064168643e-05,
+      "loss": 0.108,
+      "step": 1757
+    },
+    {
+      "epoch": 0.9203771608171818,
+      "loss_partial": 0.0556640625,
+      "step": 1757
+    },
+    {
+      "epoch": 0.9203771608171818,
+      "loss_partial": 0.05810546875,
+      "step": 1757
+    },
+    {
+      "epoch": 0.9209009952854897,
+      "grad_norm": 2.4423346519470215,
+      "learning_rate": 1.8987183243171497e-05,
+      "loss": 0.1099,
+      "step": 1758
+    },
+    {
+      "epoch": 0.9209009952854897,
+      "loss_partial": 0.0576171875,
+      "step": 1758
+    },
+    {
+      "epoch": 0.9209009952854897,
+      "loss_partial": 0.0595703125,
+      "step": 1758
+    },
+    {
+      "epoch": 0.9214248297537978,
+      "grad_norm": 1.7283352613449097,
+      "learning_rate": 1.8985177543908634e-05,
+      "loss": 0.1093,
+      "step": 1759
+    },
+    {
+      "epoch": 0.9214248297537978,
+      "loss_partial": 0.05810546875,
+      "step": 1759
+    },
+    {
+      "epoch": 0.9214248297537978,
+      "loss_partial": 0.0478515625,
+      "step": 1759
+    },
+    {
+      "epoch": 0.9219486642221059,
+      "grad_norm": 1.7633919715881348,
+      "learning_rate": 1.898316996679922e-05,
+      "loss": 0.1083,
+      "step": 1760
+    },
+    {
+      "epoch": 0.9219486642221059,
+      "loss_partial": 0.051513671875,
+      "step": 1760
+    },
+    {
+      "epoch": 0.9219486642221059,
+      "loss_partial": 0.056640625,
+      "step": 1760
+    },
+    {
+      "epoch": 0.9224724986904138,
+      "grad_norm": 2.4764482975006104,
+      "learning_rate": 1.8981160512262837e-05,
+      "loss": 0.1074,
+      "step": 1761
+    },
+    {
+      "epoch": 0.9224724986904138,
+      "loss_partial": 0.05419921875,
+      "step": 1761
+    },
+    {
+      "epoch": 0.9224724986904138,
+      "loss_partial": 0.050537109375,
+      "step": 1761
+    },
+    {
+      "epoch": 0.9229963331587219,
+      "grad_norm": 1.7235963344573975,
+      "learning_rate": 1.8979149180719443e-05,
+      "loss": 0.1093,
+      "step": 1762
+    },
+    {
+      "epoch": 0.9229963331587219,
+      "loss_partial": 0.043701171875,
+      "step": 1762
+    },
+    {
+      "epoch": 0.9229963331587219,
+      "loss_partial": 0.06298828125,
+      "step": 1762
+    },
+    {
+      "epoch": 0.9235201676270298,
+      "grad_norm": 2.2665176391601562,
+      "learning_rate": 1.8977135972589397e-05,
+      "loss": 0.1058,
+      "step": 1763
+    },
+    {
+      "epoch": 0.9235201676270298,
+      "loss_partial": 0.05322265625,
+      "step": 1763
+    },
+    {
+      "epoch": 0.9235201676270298,
+      "loss_partial": 0.046630859375,
+      "step": 1763
+    },
+    {
+      "epoch": 0.9240440020953379,
+      "grad_norm": 1.2225086688995361,
+      "learning_rate": 1.8975120888293442e-05,
+      "loss": 0.1064,
+      "step": 1764
+    },
+    {
+      "epoch": 0.9240440020953379,
+      "loss_partial": 0.06103515625,
+      "step": 1764
+    },
+    {
+      "epoch": 0.9240440020953379,
+      "loss_partial": 0.05517578125,
+      "step": 1764
+    },
+    {
+      "epoch": 0.9245678365636459,
+      "grad_norm": 1.6794873476028442,
+      "learning_rate": 1.897310392825273e-05,
+      "loss": 0.1086,
+      "step": 1765
+    },
+    {
+      "epoch": 0.9245678365636459,
+      "loss_partial": 0.05078125,
+      "step": 1765
+    },
+    {
+      "epoch": 0.9245678365636459,
+      "loss_partial": 0.057373046875,
+      "step": 1765
+    },
+    {
+      "epoch": 0.9250916710319539,
+      "grad_norm": 2.127211093902588,
+      "learning_rate": 1.8971085092888778e-05,
+      "loss": 0.1084,
+      "step": 1766
+    },
+    {
+      "epoch": 0.9250916710319539,
+      "loss_partial": 0.055419921875,
+      "step": 1766
+    },
+    {
+      "epoch": 0.9250916710319539,
+      "loss_partial": 0.06201171875,
+      "step": 1766
+    },
+    {
+      "epoch": 0.925615505500262,
+      "grad_norm": 1.6367560625076294,
+      "learning_rate": 1.8969064382623518e-05,
+      "loss": 0.1077,
+      "step": 1767
+    },
+    {
+      "epoch": 0.925615505500262,
+      "loss_partial": 0.0517578125,
+      "step": 1767
+    },
+    {
+      "epoch": 0.925615505500262,
+      "loss_partial": 0.050537109375,
+      "step": 1767
+    },
+    {
+      "epoch": 0.9261393399685699,
+      "grad_norm": 2.5341994762420654,
+      "learning_rate": 1.8967041797879265e-05,
+      "loss": 0.1069,
+      "step": 1768
+    },
+    {
+      "epoch": 0.9261393399685699,
+      "loss_partial": 0.0458984375,
+      "step": 1768
+    },
+    {
+      "epoch": 0.9261393399685699,
+      "loss_partial": 0.05126953125,
+      "step": 1768
+    },
+    {
+      "epoch": 0.926663174436878,
+      "grad_norm": 2.015688896179199,
+      "learning_rate": 1.896501733907873e-05,
+      "loss": 0.1076,
+      "step": 1769
+    },
+    {
+      "epoch": 0.926663174436878,
+      "loss_partial": 0.052734375,
+      "step": 1769
+    },
+    {
+      "epoch": 0.926663174436878,
+      "loss_partial": 0.054931640625,
+      "step": 1769
+    },
+    {
+      "epoch": 0.927187008905186,
+      "grad_norm": 2.8347065448760986,
+      "learning_rate": 1.8962991006645007e-05,
+      "loss": 0.1103,
+      "step": 1770
+    },
+    {
+      "epoch": 0.927187008905186,
+      "loss_partial": 0.0439453125,
+      "step": 1770
+    },
+    {
+      "epoch": 0.927187008905186,
+      "loss_partial": 0.064453125,
+      "step": 1770
+    },
+    {
+      "epoch": 0.927710843373494,
+      "grad_norm": 1.8458117246627808,
+      "learning_rate": 1.896096280100159e-05,
+      "loss": 0.1086,
+      "step": 1771
+    },
+    {
+      "epoch": 0.927710843373494,
+      "loss_partial": 0.05615234375,
+      "step": 1771
+    },
+    {
+      "epoch": 0.927710843373494,
+      "loss_partial": 0.0458984375,
+      "step": 1771
+    },
+    {
+      "epoch": 0.928234677841802,
+      "grad_norm": 2.5986084938049316,
+      "learning_rate": 1.895893272257236e-05,
+      "loss": 0.1076,
+      "step": 1772
+    },
+    {
+      "epoch": 0.928234677841802,
+      "loss_partial": 0.0615234375,
+      "step": 1772
+    },
+    {
+      "epoch": 0.928234677841802,
+      "loss_partial": 0.0517578125,
+      "step": 1772
+    },
+    {
+      "epoch": 0.92875851231011,
+      "grad_norm": 1.9482117891311646,
+      "learning_rate": 1.8956900771781586e-05,
+      "loss": 0.1052,
+      "step": 1773
+    },
+    {
+      "epoch": 0.92875851231011,
+      "loss_partial": 0.0625,
+      "step": 1773
+    },
+    {
+      "epoch": 0.92875851231011,
+      "loss_partial": 0.04296875,
+      "step": 1773
+    },
+    {
+      "epoch": 0.929282346778418,
+      "grad_norm": 1.8071179389953613,
+      "learning_rate": 1.8954866949053947e-05,
+      "loss": 0.1083,
+      "step": 1774
+    },
+    {
+      "epoch": 0.929282346778418,
+      "loss_partial": 0.050048828125,
+      "step": 1774
+    },
+    {
+      "epoch": 0.929282346778418,
+      "loss_partial": 0.05615234375,
+      "step": 1774
+    },
+    {
+      "epoch": 0.929806181246726,
+      "grad_norm": 2.3403289318084717,
+      "learning_rate": 1.8952831254814484e-05,
+      "loss": 0.1069,
+      "step": 1775
+    },
+    {
+      "epoch": 0.929806181246726,
+      "loss_partial": 0.0654296875,
+      "step": 1775
+    },
+    {
+      "epoch": 0.929806181246726,
+      "loss_partial": 0.053466796875,
+      "step": 1775
+    },
+    {
+      "epoch": 0.930330015715034,
+      "grad_norm": 1.6689378023147583,
+      "learning_rate": 1.8950793689488657e-05,
+      "loss": 0.1057,
+      "step": 1776
+    },
+    {
+      "epoch": 0.930330015715034,
+      "loss_partial": 0.05078125,
+      "step": 1776
+    },
+    {
+      "epoch": 0.930330015715034,
+      "loss_partial": 0.06005859375,
+      "step": 1776
+    },
+    {
+      "epoch": 0.9308538501833421,
+      "grad_norm": 2.3858861923217773,
+      "learning_rate": 1.89487542535023e-05,
+      "loss": 0.1101,
+      "step": 1777
+    },
+    {
+      "epoch": 0.9308538501833421,
+      "loss_partial": 0.0576171875,
+      "step": 1777
+    },
+    {
+      "epoch": 0.9308538501833421,
+      "loss_partial": 0.06103515625,
+      "step": 1777
+    },
+    {
+      "epoch": 0.93137768465165,
+      "grad_norm": 2.770932674407959,
+      "learning_rate": 1.8946712947281637e-05,
+      "loss": 0.1085,
+      "step": 1778
+    },
+    {
+      "epoch": 0.93137768465165,
+      "loss_partial": 0.05224609375,
+      "step": 1778
+    },
+    {
+      "epoch": 0.93137768465165,
+      "loss_partial": 0.05078125,
+      "step": 1778
+    },
+    {
+      "epoch": 0.9319015191199581,
+      "grad_norm": 2.6954681873321533,
+      "learning_rate": 1.8944669771253294e-05,
+      "loss": 0.1071,
+      "step": 1779
+    },
+    {
+      "epoch": 0.9319015191199581,
+      "loss_partial": 0.058349609375,
+      "step": 1779
+    },
+    {
+      "epoch": 0.9319015191199581,
+      "loss_partial": 0.054931640625,
+      "step": 1779
+    },
+    {
+      "epoch": 0.9324253535882661,
+      "grad_norm": 1.801464557647705,
+      "learning_rate": 1.8942624725844277e-05,
+      "loss": 0.1108,
+      "step": 1780
+    },
+    {
+      "epoch": 0.9324253535882661,
+      "loss_partial": 0.04638671875,
+      "step": 1780
+    },
+    {
+      "epoch": 0.9324253535882661,
+      "loss_partial": 0.055419921875,
+      "step": 1780
+    },
+    {
+      "epoch": 0.9329491880565741,
+      "grad_norm": 2.1565868854522705,
+      "learning_rate": 1.8940577811482e-05,
+      "loss": 0.1095,
+      "step": 1781
+    },
+    {
+      "epoch": 0.9329491880565741,
+      "loss_partial": 0.055908203125,
+      "step": 1781
+    },
+    {
+      "epoch": 0.9329491880565741,
+      "loss_partial": 0.04931640625,
+      "step": 1781
+    },
+    {
+      "epoch": 0.9334730225248822,
+      "grad_norm": 1.8088918924331665,
+      "learning_rate": 1.893852902859424e-05,
+      "loss": 0.1075,
+      "step": 1782
+    },
+    {
+      "epoch": 0.9334730225248822,
+      "loss_partial": 0.054443359375,
+      "step": 1782
+    },
+    {
+      "epoch": 0.9334730225248822,
+      "loss_partial": 0.0498046875,
+      "step": 1782
+    },
+    {
+      "epoch": 0.9339968569931901,
+      "grad_norm": 1.9636609554290771,
+      "learning_rate": 1.8936478377609193e-05,
+      "loss": 0.1082,
+      "step": 1783
+    },
+    {
+      "epoch": 0.9339968569931901,
+      "loss_partial": 0.05126953125,
+      "step": 1783
+    },
+    {
+      "epoch": 0.9339968569931901,
+      "loss_partial": 0.06201171875,
+      "step": 1783
+    },
+    {
+      "epoch": 0.9345206914614982,
+      "grad_norm": 2.3588318824768066,
+      "learning_rate": 1.8934425858955422e-05,
+      "loss": 0.1075,
+      "step": 1784
+    },
+    {
+      "epoch": 0.9345206914614982,
+      "loss_partial": 0.054931640625,
+      "step": 1784
+    },
+    {
+      "epoch": 0.9345206914614982,
+      "loss_partial": 0.060302734375,
+      "step": 1784
+    },
+    {
+      "epoch": 0.9350445259298061,
+      "grad_norm": 1.5167567729949951,
+      "learning_rate": 1.8932371473061896e-05,
+      "loss": 0.1075,
+      "step": 1785
+    },
+    {
+      "epoch": 0.9350445259298061,
+      "loss_partial": 0.052734375,
+      "step": 1785
+    },
+    {
+      "epoch": 0.9350445259298061,
+      "loss_partial": 0.048583984375,
+      "step": 1785
+    },
+    {
+      "epoch": 0.9355683603981142,
+      "grad_norm": 2.3082330226898193,
+      "learning_rate": 1.893031522035797e-05,
+      "loss": 0.1078,
+      "step": 1786
+    },
+    {
+      "epoch": 0.9355683603981142,
+      "loss_partial": 0.0546875,
+      "step": 1786
+    },
+    {
+      "epoch": 0.9355683603981142,
+      "loss_partial": 0.06494140625,
+      "step": 1786
+    },
+    {
+      "epoch": 0.9360921948664223,
+      "grad_norm": 1.9305793046951294,
+      "learning_rate": 1.8928257101273386e-05,
+      "loss": 0.1087,
+      "step": 1787
+    },
+    {
+      "epoch": 0.9360921948664223,
+      "loss_partial": 0.0654296875,
+      "step": 1787
+    },
+    {
+      "epoch": 0.9360921948664223,
+      "loss_partial": 0.0498046875,
+      "step": 1787
+    },
+    {
+      "epoch": 0.9366160293347302,
+      "grad_norm": 2.1918540000915527,
+      "learning_rate": 1.8926197116238277e-05,
+      "loss": 0.109,
+      "step": 1788
+    },
+    {
+      "epoch": 0.9366160293347302,
+      "loss_partial": 0.050537109375,
+      "step": 1788
+    },
+    {
+      "epoch": 0.9366160293347302,
+      "loss_partial": 0.06396484375,
+      "step": 1788
+    },
+    {
+      "epoch": 0.9371398638030383,
+      "grad_norm": 2.1098742485046387,
+      "learning_rate": 1.892413526568317e-05,
+      "loss": 0.1127,
+      "step": 1789
+    },
+    {
+      "epoch": 0.9371398638030383,
+      "loss_partial": 0.046630859375,
+      "step": 1789
+    },
+    {
+      "epoch": 0.9371398638030383,
+      "loss_partial": 0.056884765625,
+      "step": 1789
+    },
+    {
+      "epoch": 0.9376636982713462,
+      "grad_norm": 1.8635860681533813,
+      "learning_rate": 1.8922071550038978e-05,
+      "loss": 0.1068,
+      "step": 1790
+    },
+    {
+      "epoch": 0.9376636982713462,
+      "loss_partial": 0.05859375,
+      "step": 1790
+    },
+    {
+      "epoch": 0.9376636982713462,
+      "loss_partial": 0.0595703125,
+      "step": 1790
+    },
+    {
+      "epoch": 0.9381875327396543,
+      "grad_norm": 2.103464126586914,
+      "learning_rate": 1.8920005969737002e-05,
+      "loss": 0.1086,
+      "step": 1791
+    },
+    {
+      "epoch": 0.9381875327396543,
+      "loss_partial": 0.068359375,
+      "step": 1791
+    },
+    {
+      "epoch": 0.9381875327396543,
+      "loss_partial": 0.06884765625,
+      "step": 1791
+    },
+    {
+      "epoch": 0.9387113672079623,
+      "grad_norm": 1.6077975034713745,
+      "learning_rate": 1.8917938525208943e-05,
+      "loss": 0.1083,
+      "step": 1792
+    },
+    {
+      "epoch": 0.9387113672079623,
+      "loss_partial": 0.058349609375,
+      "step": 1792
+    },
+    {
+      "epoch": 0.9387113672079623,
+      "loss_partial": 0.047607421875,
+      "step": 1792
+    },
+    {
+      "epoch": 0.9392352016762703,
+      "grad_norm": 1.8607968091964722,
+      "learning_rate": 1.8915869216886874e-05,
+      "loss": 0.1043,
+      "step": 1793
+    },
+    {
+      "epoch": 0.9392352016762703,
+      "loss_partial": 0.056884765625,
+      "step": 1793
+    },
+    {
+      "epoch": 0.9392352016762703,
+      "loss_partial": 0.060791015625,
+      "step": 1793
+    },
+    {
+      "epoch": 0.9397590361445783,
+      "grad_norm": 1.412713885307312,
+      "learning_rate": 1.8913798045203275e-05,
+      "loss": 0.1082,
+      "step": 1794
+    },
+    {
+      "epoch": 0.9397590361445783,
+      "loss_partial": 0.050537109375,
+      "step": 1794
+    },
+    {
+      "epoch": 0.9397590361445783,
+      "loss_partial": 0.04931640625,
+      "step": 1794
+    },
+    {
+      "epoch": 0.9402828706128863,
+      "grad_norm": 2.2237648963928223,
+      "learning_rate": 1.8911725010591008e-05,
+      "loss": 0.1062,
+      "step": 1795
+    },
+    {
+      "epoch": 0.9402828706128863,
+      "loss_partial": 0.052001953125,
+      "step": 1795
+    },
+    {
+      "epoch": 0.9402828706128863,
+      "loss_partial": 0.06201171875,
+      "step": 1795
+    },
+    {
+      "epoch": 0.9408067050811943,
+      "grad_norm": 1.5926851034164429,
+      "learning_rate": 1.890965011348332e-05,
+      "loss": 0.1074,
+      "step": 1796
+    },
+    {
+      "epoch": 0.9408067050811943,
+      "loss_partial": 0.051513671875,
+      "step": 1796
+    },
+    {
+      "epoch": 0.9408067050811943,
+      "loss_partial": 0.052978515625,
+      "step": 1796
+    },
+    {
+      "epoch": 0.9413305395495024,
+      "grad_norm": 2.148310661315918,
+      "learning_rate": 1.8907573354313853e-05,
+      "loss": 0.1058,
+      "step": 1797
+    },
+    {
+      "epoch": 0.9413305395495024,
+      "loss_partial": 0.060546875,
+      "step": 1797
+    },
+    {
+      "epoch": 0.9413305395495024,
+      "loss_partial": 0.050537109375,
+      "step": 1797
+    },
+    {
+      "epoch": 0.9418543740178104,
+      "grad_norm": 2.2512831687927246,
+      "learning_rate": 1.890549473351664e-05,
+      "loss": 0.1068,
+      "step": 1798
+    },
+    {
+      "epoch": 0.9418543740178104,
+      "loss_partial": 0.054443359375,
+      "step": 1798
+    },
+    {
+      "epoch": 0.9418543740178104,
+      "loss_partial": 0.050048828125,
+      "step": 1798
+    },
+    {
+      "epoch": 0.9423782084861184,
+      "grad_norm": 1.8886268138885498,
+      "learning_rate": 1.8903414251526092e-05,
+      "loss": 0.1083,
+      "step": 1799
+    },
+    {
+      "epoch": 0.9423782084861184,
+      "loss_partial": 0.05078125,
+      "step": 1799
+    },
+    {
+      "epoch": 0.9423782084861184,
+      "loss_partial": 0.052490234375,
+      "step": 1799
+    },
+    {
+      "epoch": 0.9429020429544264,
+      "grad_norm": 2.098553419113159,
+      "learning_rate": 1.8901331908777027e-05,
+      "loss": 0.1117,
+      "step": 1800
+    },
+    {
+      "epoch": 0.9429020429544264,
+      "loss_partial": 0.05078125,
+      "step": 1800
+    },
+    {
+      "epoch": 0.9429020429544264,
+      "loss_partial": 0.059326171875,
+      "step": 1800
+    },
+    {
+      "epoch": 0.9434258774227344,
+      "grad_norm": 2.2986795902252197,
+      "learning_rate": 1.8899247705704638e-05,
+      "loss": 0.1068,
+      "step": 1801
+    },
+    {
+      "epoch": 0.9434258774227344,
+      "loss_partial": 0.0439453125,
+      "step": 1801
+    },
+    {
+      "epoch": 0.9434258774227344,
+      "loss_partial": 0.0517578125,
+      "step": 1801
+    },
+    {
+      "epoch": 0.9439497118910424,
+      "grad_norm": 1.9781404733657837,
+      "learning_rate": 1.8897161642744507e-05,
+      "loss": 0.1044,
+      "step": 1802
+    },
+    {
+      "epoch": 0.9439497118910424,
+      "loss_partial": 0.052978515625,
+      "step": 1802
+    },
+    {
+      "epoch": 0.9439497118910424,
+      "loss_partial": 0.05419921875,
+      "step": 1802
+    },
+    {
+      "epoch": 0.9444735463593504,
+      "grad_norm": 2.5010969638824463,
+      "learning_rate": 1.889507372033261e-05,
+      "loss": 0.1084,
+      "step": 1803
+    },
+    {
+      "epoch": 0.9444735463593504,
+      "loss_partial": 0.045166015625,
+      "step": 1803
+    },
+    {
+      "epoch": 0.9444735463593504,
+      "loss_partial": 0.051025390625,
+      "step": 1803
+    },
+    {
+      "epoch": 0.9449973808276585,
+      "grad_norm": 1.7962809801101685,
+      "learning_rate": 1.889298393890532e-05,
+      "loss": 0.1059,
+      "step": 1804
+    },
+    {
+      "epoch": 0.9449973808276585,
+      "loss_partial": 0.053466796875,
+      "step": 1804
+    },
+    {
+      "epoch": 0.9449973808276585,
+      "loss_partial": 0.051025390625,
+      "step": 1804
+    },
+    {
+      "epoch": 0.9455212152959664,
+      "grad_norm": 1.6283377408981323,
+      "learning_rate": 1.8890892298899374e-05,
+      "loss": 0.108,
+      "step": 1805
+    },
+    {
+      "epoch": 0.9455212152959664,
+      "loss_partial": 0.05322265625,
+      "step": 1805
+    },
+    {
+      "epoch": 0.9455212152959664,
+      "loss_partial": 0.053466796875,
+      "step": 1805
+    },
+    {
+      "epoch": 0.9460450497642745,
+      "grad_norm": 1.7720096111297607,
+      "learning_rate": 1.8888798800751913e-05,
+      "loss": 0.1077,
+      "step": 1806
+    },
+    {
+      "epoch": 0.9460450497642745,
+      "loss_partial": 0.0625,
+      "step": 1806
+    },
+    {
+      "epoch": 0.9460450497642745,
+      "loss_partial": 0.044189453125,
+      "step": 1806
+    },
+    {
+      "epoch": 0.9465688842325825,
+      "grad_norm": 1.4806526899337769,
+      "learning_rate": 1.8886703444900472e-05,
+      "loss": 0.1053,
+      "step": 1807
+    },
+    {
+      "epoch": 0.9465688842325825,
+      "loss_partial": 0.06103515625,
+      "step": 1807
+    },
+    {
+      "epoch": 0.9465688842325825,
+      "loss_partial": 0.048583984375,
+      "step": 1807
+    },
+    {
+      "epoch": 0.9470927187008905,
+      "grad_norm": 2.3040378093719482,
+      "learning_rate": 1.8884606231782965e-05,
+      "loss": 0.1056,
+      "step": 1808
+    },
+    {
+      "epoch": 0.9470927187008905,
+      "loss_partial": 0.0537109375,
+      "step": 1808
+    },
+    {
+      "epoch": 0.9470927187008905,
+      "loss_partial": 0.052490234375,
+      "step": 1808
+    },
+    {
+      "epoch": 0.9476165531691986,
+      "grad_norm": 1.9298774003982544,
+      "learning_rate": 1.8882507161837696e-05,
+      "loss": 0.1115,
+      "step": 1809
+    },
+    {
+      "epoch": 0.9476165531691986,
+      "loss_partial": 0.045166015625,
+      "step": 1809
+    },
+    {
+      "epoch": 0.9476165531691986,
+      "loss_partial": 0.05029296875,
+      "step": 1809
+    },
+    {
+      "epoch": 0.9481403876375065,
+      "grad_norm": 2.0590200424194336,
+      "learning_rate": 1.8880406235503358e-05,
+      "loss": 0.1093,
+      "step": 1810
+    },
+    {
+      "epoch": 0.9481403876375065,
+      "loss_partial": 0.061279296875,
+      "step": 1810
+    },
+    {
+      "epoch": 0.9481403876375065,
+      "loss_partial": 0.05419921875,
+      "step": 1810
+    },
+    {
+      "epoch": 0.9486642221058146,
+      "grad_norm": 1.8292434215545654,
+      "learning_rate": 1.8878303453219032e-05,
+      "loss": 0.1102,
+      "step": 1811
+    },
+    {
+      "epoch": 0.9486642221058146,
+      "loss_partial": 0.0439453125,
+      "step": 1811
+    },
+    {
+      "epoch": 0.9486642221058146,
+      "loss_partial": 0.06591796875,
+      "step": 1811
+    },
+    {
+      "epoch": 0.9491880565741225,
+      "grad_norm": 2.117029905319214,
+      "learning_rate": 1.8876198815424187e-05,
+      "loss": 0.1058,
+      "step": 1812
+    },
+    {
+      "epoch": 0.9491880565741225,
+      "loss_partial": 0.046875,
+      "step": 1812
+    },
+    {
+      "epoch": 0.9491880565741225,
+      "loss_partial": 0.052734375,
+      "step": 1812
+    },
+    {
+      "epoch": 0.9497118910424306,
+      "grad_norm": 1.7883292436599731,
+      "learning_rate": 1.887409232255868e-05,
+      "loss": 0.1062,
+      "step": 1813
+    },
+    {
+      "epoch": 0.9497118910424306,
+      "loss_partial": 0.056640625,
+      "step": 1813
+    },
+    {
+      "epoch": 0.9497118910424306,
+      "loss_partial": 0.052001953125,
+      "step": 1813
+    },
+    {
+      "epoch": 0.9502357255107386,
+      "grad_norm": 2.4548025131225586,
+      "learning_rate": 1.8871983975062748e-05,
+      "loss": 0.1102,
+      "step": 1814
+    },
+    {
+      "epoch": 0.9502357255107386,
+      "loss_partial": 0.0546875,
+      "step": 1814
+    },
+    {
+      "epoch": 0.9502357255107386,
+      "loss_partial": 0.05224609375,
+      "step": 1814
+    },
+    {
+      "epoch": 0.9507595599790466,
+      "grad_norm": 2.687993288040161,
+      "learning_rate": 1.886987377337703e-05,
+      "loss": 0.1079,
+      "step": 1815
+    },
+    {
+      "epoch": 0.9507595599790466,
+      "loss_partial": 0.058349609375,
+      "step": 1815
+    },
+    {
+      "epoch": 0.9507595599790466,
+      "loss_partial": 0.047607421875,
+      "step": 1815
+    },
+    {
+      "epoch": 0.9512833944473547,
+      "grad_norm": 2.9336018562316895,
+      "learning_rate": 1.8867761717942543e-05,
+      "loss": 0.1073,
+      "step": 1816
+    },
+    {
+      "epoch": 0.9512833944473547,
+      "loss_partial": 0.057861328125,
+      "step": 1816
+    },
+    {
+      "epoch": 0.9512833944473547,
+      "loss_partial": 0.049560546875,
+      "step": 1816
+    },
+    {
+      "epoch": 0.9518072289156626,
+      "grad_norm": 2.6794586181640625,
+      "learning_rate": 1.8865647809200695e-05,
+      "loss": 0.108,
+      "step": 1817
+    },
+    {
+      "epoch": 0.9518072289156626,
+      "loss_partial": 0.05419921875,
+      "step": 1817
+    },
+    {
+      "epoch": 0.9518072289156626,
+      "loss_partial": 0.05322265625,
+      "step": 1817
+    },
+    {
+      "epoch": 0.9523310633839707,
+      "grad_norm": 1.8365200757980347,
+      "learning_rate": 1.8863532047593278e-05,
+      "loss": 0.1096,
+      "step": 1818
+    },
+    {
+      "epoch": 0.9523310633839707,
+      "loss_partial": 0.050537109375,
+      "step": 1818
+    },
+    {
+      "epoch": 0.9523310633839707,
+      "loss_partial": 0.05615234375,
+      "step": 1818
+    },
+    {
+      "epoch": 0.9528548978522787,
+      "grad_norm": 2.339616298675537,
+      "learning_rate": 1.886141443356247e-05,
+      "loss": 0.1072,
+      "step": 1819
+    },
+    {
+      "epoch": 0.9528548978522787,
+      "loss_partial": 0.04736328125,
+      "step": 1819
+    },
+    {
+      "epoch": 0.9528548978522787,
+      "loss_partial": 0.05908203125,
+      "step": 1819
+    },
+    {
+      "epoch": 0.9533787323205867,
+      "grad_norm": 1.8992325067520142,
+      "learning_rate": 1.885929496755085e-05,
+      "loss": 0.1086,
+      "step": 1820
+    },
+    {
+      "epoch": 0.9533787323205867,
+      "loss_partial": 0.053466796875,
+      "step": 1820
+    },
+    {
+      "epoch": 0.9533787323205867,
+      "loss_partial": 0.05322265625,
+      "step": 1820
+    },
+    {
+      "epoch": 0.9539025667888947,
+      "grad_norm": 2.141024112701416,
+      "learning_rate": 1.885717365000136e-05,
+      "loss": 0.1061,
+      "step": 1821
+    },
+    {
+      "epoch": 0.9539025667888947,
+      "loss_partial": 0.0556640625,
+      "step": 1821
+    },
+    {
+      "epoch": 0.9539025667888947,
+      "loss_partial": 0.0546875,
+      "step": 1821
+    },
+    {
+      "epoch": 0.9544264012572027,
+      "grad_norm": 2.4976258277893066,
+      "learning_rate": 1.885505048135735e-05,
+      "loss": 0.1125,
+      "step": 1822
+    },
+    {
+      "epoch": 0.9544264012572027,
+      "loss_partial": 0.052734375,
+      "step": 1822
+    },
+    {
+      "epoch": 0.9544264012572027,
+      "loss_partial": 0.053955078125,
+      "step": 1822
+    },
+    {
+      "epoch": 0.9549502357255107,
+      "grad_norm": 2.1863293647766113,
+      "learning_rate": 1.8852925462062546e-05,
+      "loss": 0.1075,
+      "step": 1823
+    },
+    {
+      "epoch": 0.9549502357255107,
+      "loss_partial": 0.056640625,
+      "step": 1823
+    },
+    {
+      "epoch": 0.9549502357255107,
+      "loss_partial": 0.053466796875,
+      "step": 1823
+    },
+    {
+      "epoch": 0.9554740701938188,
+      "grad_norm": 2.4875667095184326,
+      "learning_rate": 1.885079859256107e-05,
+      "loss": 0.1104,
+      "step": 1824
+    },
+    {
+      "epoch": 0.9554740701938188,
+      "loss_partial": 0.0537109375,
+      "step": 1824
+    },
+    {
+      "epoch": 0.9554740701938188,
+      "loss_partial": 0.058837890625,
+      "step": 1824
+    },
+    {
+      "epoch": 0.9559979046621268,
+      "grad_norm": 2.3518214225769043,
+      "learning_rate": 1.8848669873297416e-05,
+      "loss": 0.1111,
+      "step": 1825
+    },
+    {
+      "epoch": 0.9559979046621268,
+      "loss_partial": 0.056884765625,
+      "step": 1825
+    },
+    {
+      "epoch": 0.9559979046621268,
+      "loss_partial": 0.049560546875,
+      "step": 1825
+    },
+    {
+      "epoch": 0.9565217391304348,
+      "grad_norm": 1.7335600852966309,
+      "learning_rate": 1.8846539304716477e-05,
+      "loss": 0.1089,
+      "step": 1826
+    },
+    {
+      "epoch": 0.9565217391304348,
+      "loss_partial": 0.05419921875,
+      "step": 1826
+    },
+    {
+      "epoch": 0.9565217391304348,
+      "loss_partial": 0.057373046875,
+      "step": 1826
+    },
+    {
+      "epoch": 0.9570455735987428,
+      "grad_norm": 1.7384458780288696,
+      "learning_rate": 1.884440688726353e-05,
+      "loss": 0.1069,
+      "step": 1827
+    },
+    {
+      "epoch": 0.9570455735987428,
+      "loss_partial": 0.0546875,
+      "step": 1827
+    },
+    {
+      "epoch": 0.9570455735987428,
+      "loss_partial": 0.055419921875,
+      "step": 1827
+    },
+    {
+      "epoch": 0.9575694080670508,
+      "grad_norm": 2.065709352493286,
+      "learning_rate": 1.8842272621384235e-05,
+      "loss": 0.1052,
+      "step": 1828
+    },
+    {
+      "epoch": 0.9575694080670508,
+      "loss_partial": 0.06103515625,
+      "step": 1828
+    },
+    {
+      "epoch": 0.9575694080670508,
+      "loss_partial": 0.051513671875,
+      "step": 1828
+    },
+    {
+      "epoch": 0.9580932425353589,
+      "grad_norm": 2.380349636077881,
+      "learning_rate": 1.884013650752464e-05,
+      "loss": 0.113,
+      "step": 1829
+    },
+    {
+      "epoch": 0.9580932425353589,
+      "loss_partial": 0.05908203125,
+      "step": 1829
+    },
+    {
+      "epoch": 0.9580932425353589,
+      "loss_partial": 0.06005859375,
+      "step": 1829
+    },
+    {
+      "epoch": 0.9586170770036668,
+      "grad_norm": 2.3809309005737305,
+      "learning_rate": 1.8837998546131184e-05,
+      "loss": 0.1089,
+      "step": 1830
+    },
+    {
+      "epoch": 0.9586170770036668,
+      "loss_partial": 0.051025390625,
+      "step": 1830
+    },
+    {
+      "epoch": 0.9586170770036668,
+      "loss_partial": 0.044921875,
+      "step": 1830
+    },
+    {
+      "epoch": 0.9591409114719749,
+      "grad_norm": 2.236454963684082,
+      "learning_rate": 1.8835858737650682e-05,
+      "loss": 0.1103,
+      "step": 1831
+    },
+    {
+      "epoch": 0.9591409114719749,
+      "loss_partial": 0.047607421875,
+      "step": 1831
+    },
+    {
+      "epoch": 0.9591409114719749,
+      "loss_partial": 0.051513671875,
+      "step": 1831
+    },
+    {
+      "epoch": 0.9596647459402828,
+      "grad_norm": 1.7229716777801514,
+      "learning_rate": 1.8833717082530342e-05,
+      "loss": 0.1057,
+      "step": 1832
+    },
+    {
+      "epoch": 0.9596647459402828,
+      "loss_partial": 0.05810546875,
+      "step": 1832
+    },
+    {
+      "epoch": 0.9596647459402828,
+      "loss_partial": 0.054443359375,
+      "step": 1832
+    },
+    {
+      "epoch": 0.9601885804085909,
+      "grad_norm": 1.8469194173812866,
+      "learning_rate": 1.883157358121776e-05,
+      "loss": 0.1076,
+      "step": 1833
+    },
+    {
+      "epoch": 0.9601885804085909,
+      "loss_partial": 0.050537109375,
+      "step": 1833
+    },
+    {
+      "epoch": 0.9601885804085909,
+      "loss_partial": 0.0576171875,
+      "step": 1833
+    },
+    {
+      "epoch": 0.9607124148768988,
+      "grad_norm": 2.009244441986084,
+      "learning_rate": 1.882942823416091e-05,
+      "loss": 0.1084,
+      "step": 1834
+    },
+    {
+      "epoch": 0.9607124148768988,
+      "loss_partial": 0.0556640625,
+      "step": 1834
+    },
+    {
+      "epoch": 0.9607124148768988,
+      "loss_partial": 0.049560546875,
+      "step": 1834
+    },
+    {
+      "epoch": 0.9612362493452069,
+      "grad_norm": 2.4920082092285156,
+      "learning_rate": 1.882728104180816e-05,
+      "loss": 0.108,
+      "step": 1835
+    },
+    {
+      "epoch": 0.9612362493452069,
+      "loss_partial": 0.05419921875,
+      "step": 1835
+    },
+    {
+      "epoch": 0.9612362493452069,
+      "loss_partial": 0.059326171875,
+      "step": 1835
+    },
+    {
+      "epoch": 0.961760083813515,
+      "grad_norm": 2.3188939094543457,
+      "learning_rate": 1.8825132004608257e-05,
+      "loss": 0.111,
+      "step": 1836
+    },
+    {
+      "epoch": 0.961760083813515,
+      "loss_partial": 0.059814453125,
+      "step": 1836
+    },
+    {
+      "epoch": 0.961760083813515,
+      "loss_partial": 0.046142578125,
+      "step": 1836
+    },
+    {
+      "epoch": 0.9622839182818229,
+      "grad_norm": 1.784216046333313,
+      "learning_rate": 1.8822981123010343e-05,
+      "loss": 0.1089,
+      "step": 1837
+    },
+    {
+      "epoch": 0.9622839182818229,
+      "loss_partial": 0.051513671875,
+      "step": 1837
+    },
+    {
+      "epoch": 0.9622839182818229,
+      "loss_partial": 0.047119140625,
+      "step": 1837
+    },
+    {
+      "epoch": 0.962807752750131,
+      "grad_norm": 1.8911091089248657,
+      "learning_rate": 1.882082839746393e-05,
+      "loss": 0.1081,
+      "step": 1838
+    },
+    {
+      "epoch": 0.962807752750131,
+      "loss_partial": 0.05712890625,
+      "step": 1838
+    },
+    {
+      "epoch": 0.962807752750131,
+      "loss_partial": 0.05908203125,
+      "step": 1838
+    },
+    {
+      "epoch": 0.9633315872184389,
+      "grad_norm": 1.438779592514038,
+      "learning_rate": 1.8818673828418927e-05,
+      "loss": 0.1111,
+      "step": 1839
+    },
+    {
+      "epoch": 0.9633315872184389,
+      "loss_partial": 0.04833984375,
+      "step": 1839
+    },
+    {
+      "epoch": 0.9633315872184389,
+      "loss_partial": 0.052734375,
+      "step": 1839
+    },
+    {
+      "epoch": 0.963855421686747,
+      "grad_norm": 2.054701328277588,
+      "learning_rate": 1.881651741632563e-05,
+      "loss": 0.1074,
+      "step": 1840
+    },
+    {
+      "epoch": 0.963855421686747,
+      "loss_partial": 0.0615234375,
+      "step": 1840
+    },
+    {
+      "epoch": 0.963855421686747,
+      "loss_partial": 0.0615234375,
+      "step": 1840
+    },
+    {
+      "epoch": 0.964379256155055,
+      "grad_norm": 2.080084800720215,
+      "learning_rate": 1.881435916163471e-05,
+      "loss": 0.1098,
+      "step": 1841
+    },
+    {
+      "epoch": 0.964379256155055,
+      "loss_partial": 0.06689453125,
+      "step": 1841
+    },
+    {
+      "epoch": 0.964379256155055,
+      "loss_partial": 0.05615234375,
+      "step": 1841
+    },
+    {
+      "epoch": 0.964903090623363,
+      "grad_norm": 1.6214419603347778,
+      "learning_rate": 1.8812199064797234e-05,
+      "loss": 0.1095,
+      "step": 1842
+    },
+    {
+      "epoch": 0.964903090623363,
+      "loss_partial": 0.056396484375,
+      "step": 1842
+    },
+    {
+      "epoch": 0.964903090623363,
+      "loss_partial": 0.064453125,
+      "step": 1842
+    },
+    {
+      "epoch": 0.965426925091671,
+      "grad_norm": 1.675605058670044,
+      "learning_rate": 1.8810037126264646e-05,
+      "loss": 0.1129,
+      "step": 1843
+    },
+    {
+      "epoch": 0.965426925091671,
+      "loss_partial": 0.0478515625,
+      "step": 1843
+    },
+    {
+      "epoch": 0.965426925091671,
+      "loss_partial": 0.042724609375,
+      "step": 1843
+    },
+    {
+      "epoch": 0.965950759559979,
+      "grad_norm": 1.989422082901001,
+      "learning_rate": 1.8807873346488777e-05,
+      "loss": 0.1078,
+      "step": 1844
+    },
+    {
+      "epoch": 0.965950759559979,
+      "loss_partial": 0.052001953125,
+      "step": 1844
+    },
+    {
+      "epoch": 0.965950759559979,
+      "loss_partial": 0.051513671875,
+      "step": 1844
+    },
+    {
+      "epoch": 0.9664745940282871,
+      "grad_norm": 1.4270023107528687,
+      "learning_rate": 1.8805707725921846e-05,
+      "loss": 0.1075,
+      "step": 1845
+    },
+    {
+      "epoch": 0.9664745940282871,
+      "loss_partial": 0.053466796875,
+      "step": 1845
+    },
+    {
+      "epoch": 0.9664745940282871,
+      "loss_partial": 0.060302734375,
+      "step": 1845
+    },
+    {
+      "epoch": 0.9669984284965951,
+      "grad_norm": 2.7908031940460205,
+      "learning_rate": 1.880354026501645e-05,
+      "loss": 0.1138,
+      "step": 1846
+    },
+    {
+      "epoch": 0.9669984284965951,
+      "loss_partial": 0.0615234375,
+      "step": 1846
+    },
+    {
+      "epoch": 0.9669984284965951,
+      "loss_partial": 0.0478515625,
+      "step": 1846
+    },
+    {
+      "epoch": 0.9675222629649031,
+      "grad_norm": 2.250936269760132,
+      "learning_rate": 1.8801370964225584e-05,
+      "loss": 0.1063,
+      "step": 1847
+    },
+    {
+      "epoch": 0.9675222629649031,
+      "loss_partial": 0.04736328125,
+      "step": 1847
+    },
+    {
+      "epoch": 0.9675222629649031,
+      "loss_partial": 0.0478515625,
+      "step": 1847
+    },
+    {
+      "epoch": 0.9680460974332111,
+      "grad_norm": 1.6134675741195679,
+      "learning_rate": 1.8799199824002608e-05,
+      "loss": 0.1051,
+      "step": 1848
+    },
+    {
+      "epoch": 0.9680460974332111,
+      "loss_partial": 0.04833984375,
+      "step": 1848
+    },
+    {
+      "epoch": 0.9680460974332111,
+      "loss_partial": 0.0546875,
+      "step": 1848
+    },
+    {
+      "epoch": 0.9685699319015191,
+      "grad_norm": 2.4922120571136475,
+      "learning_rate": 1.8797026844801284e-05,
+      "loss": 0.108,
+      "step": 1849
+    },
+    {
+      "epoch": 0.9685699319015191,
+      "loss_partial": 0.05517578125,
+      "step": 1849
+    },
+    {
+      "epoch": 0.9685699319015191,
+      "loss_partial": 0.048828125,
+      "step": 1849
+    },
+    {
+      "epoch": 0.9690937663698271,
+      "grad_norm": 1.8368514776229858,
+      "learning_rate": 1.8794852027075744e-05,
+      "loss": 0.1075,
+      "step": 1850
+    },
+    {
+      "epoch": 0.9690937663698271,
+      "loss_partial": 0.04638671875,
+      "step": 1850
+    },
+    {
+      "epoch": 0.9690937663698271,
+      "loss_partial": 0.051025390625,
+      "step": 1850
+    },
+    {
+      "epoch": 0.9696176008381352,
+      "grad_norm": 2.7602155208587646,
+      "learning_rate": 1.879267537128052e-05,
+      "loss": 0.1116,
+      "step": 1851
+    },
+    {
+      "epoch": 0.9696176008381352,
+      "loss_partial": 0.05029296875,
+      "step": 1851
+    },
+    {
+      "epoch": 0.9696176008381352,
+      "loss_partial": 0.052490234375,
+      "step": 1851
+    },
+    {
+      "epoch": 0.9701414353064431,
+      "grad_norm": 2.368947744369507,
+      "learning_rate": 1.8790496877870518e-05,
+      "loss": 0.1074,
+      "step": 1852
+    },
+    {
+      "epoch": 0.9701414353064431,
+      "loss_partial": 0.06005859375,
+      "step": 1852
+    },
+    {
+      "epoch": 0.9701414353064431,
+      "loss_partial": 0.05517578125,
+      "step": 1852
+    },
+    {
+      "epoch": 0.9706652697747512,
+      "grad_norm": 1.6660678386688232,
+      "learning_rate": 1.8788316547301023e-05,
+      "loss": 0.1108,
+      "step": 1853
+    },
+    {
+      "epoch": 0.9706652697747512,
+      "loss_partial": 0.04833984375,
+      "step": 1853
+    },
+    {
+      "epoch": 0.9706652697747512,
+      "loss_partial": 0.05419921875,
+      "step": 1853
+    },
+    {
+      "epoch": 0.9711891042430592,
+      "grad_norm": 2.2710890769958496,
+      "learning_rate": 1.8786134380027718e-05,
+      "loss": 0.1039,
+      "step": 1854
+    },
+    {
+      "epoch": 0.9711891042430592,
+      "loss_partial": 0.05908203125,
+      "step": 1854
+    },
+    {
+      "epoch": 0.9711891042430592,
+      "loss_partial": 0.048583984375,
+      "step": 1854
+    },
+    {
+      "epoch": 0.9717129387113672,
+      "grad_norm": 2.170856237411499,
+      "learning_rate": 1.878395037650666e-05,
+      "loss": 0.1093,
+      "step": 1855
+    },
+    {
+      "epoch": 0.9717129387113672,
+      "loss_partial": 0.06494140625,
+      "step": 1855
+    },
+    {
+      "epoch": 0.9717129387113672,
+      "loss_partial": 0.06201171875,
+      "step": 1855
+    },
+    {
+      "epoch": 0.9722367731796753,
+      "grad_norm": 2.976073980331421,
+      "learning_rate": 1.8781764537194284e-05,
+      "loss": 0.1114,
+      "step": 1856
+    },
+    {
+      "epoch": 0.9722367731796753,
+      "loss_partial": 0.052490234375,
+      "step": 1856
+    },
+    {
+      "epoch": 0.9722367731796753,
+      "loss_partial": 0.059814453125,
+      "step": 1856
+    },
+    {
+      "epoch": 0.9727606076479832,
+      "grad_norm": 2.900702714920044,
+      "learning_rate": 1.877957686254743e-05,
+      "loss": 0.108,
+      "step": 1857
+    },
+    {
+      "epoch": 0.9727606076479832,
+      "loss_partial": 0.046875,
+      "step": 1857
+    },
+    {
+      "epoch": 0.9727606076479832,
+      "loss_partial": 0.055908203125,
+      "step": 1857
+    },
+    {
+      "epoch": 0.9732844421162913,
+      "grad_norm": 2.926008462905884,
+      "learning_rate": 1.87773873530233e-05,
+      "loss": 0.1125,
+      "step": 1858
+    },
+    {
+      "epoch": 0.9732844421162913,
+      "loss_partial": 0.052734375,
+      "step": 1858
+    },
+    {
+      "epoch": 0.9732844421162913,
+      "loss_partial": 0.048095703125,
+      "step": 1858
+    },
+    {
+      "epoch": 0.9738082765845992,
+      "grad_norm": 1.8229414224624634,
+      "learning_rate": 1.8775196009079497e-05,
+      "loss": 0.1096,
+      "step": 1859
+    },
+    {
+      "epoch": 0.9738082765845992,
+      "loss_partial": 0.058837890625,
+      "step": 1859
+    },
+    {
+      "epoch": 0.9738082765845992,
+      "loss_partial": 0.0498046875,
+      "step": 1859
+    },
+    {
+      "epoch": 0.9743321110529073,
+      "grad_norm": 2.347788095474243,
+      "learning_rate": 1.877300283117399e-05,
+      "loss": 0.1061,
+      "step": 1860
+    },
+    {
+      "epoch": 0.9743321110529073,
+      "loss_partial": 0.04931640625,
+      "step": 1860
+    },
+    {
+      "epoch": 0.9743321110529073,
+      "loss_partial": 0.049560546875,
+      "step": 1860
+    },
+    {
+      "epoch": 0.9748559455212154,
+      "grad_norm": 1.8553389310836792,
+      "learning_rate": 1.8770807819765136e-05,
+      "loss": 0.1086,
+      "step": 1861
+    },
+    {
+      "epoch": 0.9748559455212154,
+      "loss_partial": 0.05517578125,
+      "step": 1861
+    },
+    {
+      "epoch": 0.9748559455212154,
+      "loss_partial": 0.055419921875,
+      "step": 1861
+    },
+    {
+      "epoch": 0.9753797799895233,
+      "grad_norm": 2.0680437088012695,
+      "learning_rate": 1.8768610975311695e-05,
+      "loss": 0.11,
+      "step": 1862
+    },
+    {
+      "epoch": 0.9753797799895233,
+      "loss_partial": 0.06005859375,
+      "step": 1862
+    },
+    {
+      "epoch": 0.9753797799895233,
+      "loss_partial": 0.04736328125,
+      "step": 1862
+    },
+    {
+      "epoch": 0.9759036144578314,
+      "grad_norm": 1.6299288272857666,
+      "learning_rate": 1.8766412298272776e-05,
+      "loss": 0.1076,
+      "step": 1863
+    },
+    {
+      "epoch": 0.9759036144578314,
+      "loss_partial": 0.054931640625,
+      "step": 1863
+    },
+    {
+      "epoch": 0.9759036144578314,
+      "loss_partial": 0.052001953125,
+      "step": 1863
+    },
+    {
+      "epoch": 0.9764274489261393,
+      "grad_norm": 1.9708385467529297,
+      "learning_rate": 1.87642117891079e-05,
+      "loss": 0.1059,
+      "step": 1864
+    },
+    {
+      "epoch": 0.9764274489261393,
+      "loss_partial": 0.052001953125,
+      "step": 1864
+    },
+    {
+      "epoch": 0.9764274489261393,
+      "loss_partial": 0.0576171875,
+      "step": 1864
+    },
+    {
+      "epoch": 0.9769512833944474,
+      "grad_norm": 1.9434614181518555,
+      "learning_rate": 1.8762009448276953e-05,
+      "loss": 0.1095,
+      "step": 1865
+    },
+    {
+      "epoch": 0.9769512833944474,
+      "loss_partial": 0.056640625,
+      "step": 1865
+    },
+    {
+      "epoch": 0.9769512833944474,
+      "loss_partial": 0.059814453125,
+      "step": 1865
+    },
+    {
+      "epoch": 0.9774751178627553,
+      "grad_norm": 1.5843836069107056,
+      "learning_rate": 1.8759805276240216e-05,
+      "loss": 0.1095,
+      "step": 1866
+    },
+    {
+      "epoch": 0.9774751178627553,
+      "loss_partial": 0.059814453125,
+      "step": 1866
+    },
+    {
+      "epoch": 0.9774751178627553,
+      "loss_partial": 0.05322265625,
+      "step": 1866
+    },
+    {
+      "epoch": 0.9779989523310634,
+      "grad_norm": 1.7034488916397095,
+      "learning_rate": 1.8757599273458345e-05,
+      "loss": 0.1074,
+      "step": 1867
+    },
+    {
+      "epoch": 0.9779989523310634,
+      "loss_partial": 0.053955078125,
+      "step": 1867
+    },
+    {
+      "epoch": 0.9779989523310634,
+      "loss_partial": 0.057861328125,
+      "step": 1867
+    },
+    {
+      "epoch": 0.9785227867993714,
+      "grad_norm": 2.0541484355926514,
+      "learning_rate": 1.8755391440392384e-05,
+      "loss": 0.1056,
+      "step": 1868
+    },
+    {
+      "epoch": 0.9785227867993714,
+      "loss_partial": 0.06201171875,
+      "step": 1868
+    },
+    {
+      "epoch": 0.9785227867993714,
+      "loss_partial": 0.057861328125,
+      "step": 1868
+    },
+    {
+      "epoch": 0.9790466212676794,
+      "grad_norm": 1.5802745819091797,
+      "learning_rate": 1.875318177750375e-05,
+      "loss": 0.1115,
+      "step": 1869
+    },
+    {
+      "epoch": 0.9790466212676794,
+      "loss_partial": 0.0654296875,
+      "step": 1869
+    },
+    {
+      "epoch": 0.9790466212676794,
+      "loss_partial": 0.058837890625,
+      "step": 1869
+    },
+    {
+      "epoch": 0.9795704557359874,
+      "grad_norm": 1.7208153009414673,
+      "learning_rate": 1.8750970285254254e-05,
+      "loss": 0.1079,
+      "step": 1870
+    },
+    {
+      "epoch": 0.9795704557359874,
+      "loss_partial": 0.050537109375,
+      "step": 1870
+    },
+    {
+      "epoch": 0.9795704557359874,
+      "loss_partial": 0.05810546875,
+      "step": 1870
+    },
+    {
+      "epoch": 0.9800942902042954,
+      "grad_norm": 2.1742451190948486,
+      "learning_rate": 1.8748756964106083e-05,
+      "loss": 0.1066,
+      "step": 1871
+    },
+    {
+      "epoch": 0.9800942902042954,
+      "loss_partial": 0.05859375,
+      "step": 1871
+    },
+    {
+      "epoch": 0.9800942902042954,
+      "loss_partial": 0.059814453125,
+      "step": 1871
+    },
+    {
+      "epoch": 0.9806181246726035,
+      "grad_norm": 1.6119067668914795,
+      "learning_rate": 1.8746541814521806e-05,
+      "loss": 0.1101,
+      "step": 1872
+    },
+    {
+      "epoch": 0.9806181246726035,
+      "loss_partial": 0.057373046875,
+      "step": 1872
+    },
+    {
+      "epoch": 0.9806181246726035,
+      "loss_partial": 0.05126953125,
+      "step": 1872
+    },
+    {
+      "epoch": 0.9811419591409115,
+      "grad_norm": 1.2220423221588135,
+      "learning_rate": 1.8744324836964377e-05,
+      "loss": 0.1031,
+      "step": 1873
+    },
+    {
+      "epoch": 0.9811419591409115,
+      "loss_partial": 0.044921875,
+      "step": 1873
+    },
+    {
+      "epoch": 0.9811419591409115,
+      "loss_partial": 0.0556640625,
+      "step": 1873
+    },
+    {
+      "epoch": 0.9816657936092195,
+      "grad_norm": 1.3579752445220947,
+      "learning_rate": 1.874210603189713e-05,
+      "loss": 0.1093,
+      "step": 1874
+    },
+    {
+      "epoch": 0.9816657936092195,
+      "loss_partial": 0.0498046875,
+      "step": 1874
+    },
+    {
+      "epoch": 0.9816657936092195,
+      "loss_partial": 0.05517578125,
+      "step": 1874
+    },
+    {
+      "epoch": 0.9821896280775275,
+      "grad_norm": 1.8574680089950562,
+      "learning_rate": 1.8739885399783785e-05,
+      "loss": 0.1083,
+      "step": 1875
+    },
+    {
+      "epoch": 0.9821896280775275,
+      "loss_partial": 0.059326171875,
+      "step": 1875
+    },
+    {
+      "epoch": 0.9821896280775275,
+      "loss_partial": 0.05810546875,
+      "step": 1875
+    },
+    {
+      "epoch": 0.9827134625458355,
+      "grad_norm": 1.9050829410552979,
+      "learning_rate": 1.8737662941088432e-05,
+      "loss": 0.11,
+      "step": 1876
+    },
+    {
+      "epoch": 0.9827134625458355,
+      "loss_partial": 0.053955078125,
+      "step": 1876
+    },
+    {
+      "epoch": 0.9827134625458355,
+      "loss_partial": 0.05810546875,
+      "step": 1876
+    },
+    {
+      "epoch": 0.9832372970141435,
+      "grad_norm": 3.1190075874328613,
+      "learning_rate": 1.873543865627556e-05,
+      "loss": 0.1111,
+      "step": 1877
+    },
+    {
+      "epoch": 0.9832372970141435,
+      "loss_partial": 0.051513671875,
+      "step": 1877
+    },
+    {
+      "epoch": 0.9832372970141435,
+      "loss_partial": 0.060791015625,
+      "step": 1877
+    },
+    {
+      "epoch": 0.9837611314824516,
+      "grad_norm": 2.2750020027160645,
+      "learning_rate": 1.873321254581003e-05,
+      "loss": 0.109,
+      "step": 1878
+    },
+    {
+      "epoch": 0.9837611314824516,
+      "loss_partial": 0.0537109375,
+      "step": 1878
+    },
+    {
+      "epoch": 0.9837611314824516,
+      "loss_partial": 0.05029296875,
+      "step": 1878
+    },
+    {
+      "epoch": 0.9842849659507595,
+      "grad_norm": 2.3082289695739746,
+      "learning_rate": 1.8730984610157077e-05,
+      "loss": 0.1034,
+      "step": 1879
+    },
+    {
+      "epoch": 0.9842849659507595,
+      "loss_partial": 0.044677734375,
+      "step": 1879
+    },
+    {
+      "epoch": 0.9842849659507595,
+      "loss_partial": 0.050537109375,
+      "step": 1879
+    },
+    {
+      "epoch": 0.9848088004190676,
+      "grad_norm": 2.1417596340179443,
+      "learning_rate": 1.8728754849782335e-05,
+      "loss": 0.1082,
+      "step": 1880
+    },
+    {
+      "epoch": 0.9848088004190676,
+      "loss_partial": 0.049560546875,
+      "step": 1880
+    },
+    {
+      "epoch": 0.9848088004190676,
+      "loss_partial": 0.05126953125,
+      "step": 1880
+    },
+    {
+      "epoch": 0.9853326348873755,
+      "grad_norm": 2.354125499725342,
+      "learning_rate": 1.8726523265151803e-05,
+      "loss": 0.1107,
+      "step": 1881
+    },
+    {
+      "epoch": 0.9853326348873755,
+      "loss_partial": 0.05908203125,
+      "step": 1881
+    },
+    {
+      "epoch": 0.9853326348873755,
+      "loss_partial": 0.050537109375,
+      "step": 1881
+    },
+    {
+      "epoch": 0.9858564693556836,
+      "grad_norm": 2.5063998699188232,
+      "learning_rate": 1.872428985673187e-05,
+      "loss": 0.1126,
+      "step": 1882
+    },
+    {
+      "epoch": 0.9858564693556836,
+      "loss_partial": 0.05859375,
+      "step": 1882
+    },
+    {
+      "epoch": 0.9858564693556836,
+      "loss_partial": 0.053955078125,
+      "step": 1882
+    },
+    {
+      "epoch": 0.9863803038239917,
+      "grad_norm": 2.1827504634857178,
+      "learning_rate": 1.872205462498931e-05,
+      "loss": 0.1095,
+      "step": 1883
+    },
+    {
+      "epoch": 0.9863803038239917,
+      "loss_partial": 0.056884765625,
+      "step": 1883
+    },
+    {
+      "epoch": 0.9863803038239917,
+      "loss_partial": 0.0595703125,
+      "step": 1883
+    },
+    {
+      "epoch": 0.9869041382922996,
+      "grad_norm": 1.7854148149490356,
+      "learning_rate": 1.871981757039127e-05,
+      "loss": 0.106,
+      "step": 1884
+    },
+    {
+      "epoch": 0.9869041382922996,
+      "loss_partial": 0.0615234375,
+      "step": 1884
+    },
+    {
+      "epoch": 0.9869041382922996,
+      "loss_partial": 0.048828125,
+      "step": 1884
+    },
+    {
+      "epoch": 0.9874279727606077,
+      "grad_norm": 2.867206573486328,
+      "learning_rate": 1.871757869340528e-05,
+      "loss": 0.1107,
+      "step": 1885
+    },
+    {
+      "epoch": 0.9874279727606077,
+      "loss_partial": 0.050537109375,
+      "step": 1885
+    },
+    {
+      "epoch": 0.9874279727606077,
+      "loss_partial": 0.0654296875,
+      "step": 1885
+    },
+    {
+      "epoch": 0.9879518072289156,
+      "grad_norm": 2.4689009189605713,
+      "learning_rate": 1.8715337994499247e-05,
+      "loss": 0.1095,
+      "step": 1886
+    },
+    {
+      "epoch": 0.9879518072289156,
+      "loss_partial": 0.05419921875,
+      "step": 1886
+    },
+    {
+      "epoch": 0.9879518072289156,
+      "loss_partial": 0.051513671875,
+      "step": 1886
+    },
+    {
+      "epoch": 0.9884756416972237,
+      "grad_norm": 2.7699084281921387,
+      "learning_rate": 1.8713095474141473e-05,
+      "loss": 0.1088,
+      "step": 1887
+    },
+    {
+      "epoch": 0.9884756416972237,
+      "loss_partial": 0.05810546875,
+      "step": 1887
+    },
+    {
+      "epoch": 0.9884756416972237,
+      "loss_partial": 0.06298828125,
+      "step": 1887
+    },
+    {
+      "epoch": 0.9889994761655317,
+      "grad_norm": 1.5514811277389526,
+      "learning_rate": 1.8710851132800617e-05,
+      "loss": 0.1107,
+      "step": 1888
+    },
+    {
+      "epoch": 0.9889994761655317,
+      "loss_partial": 0.060302734375,
+      "step": 1888
+    },
+    {
+      "epoch": 0.9889994761655317,
+      "loss_partial": 0.052734375,
+      "step": 1888
+    },
+    {
+      "epoch": 0.9895233106338397,
+      "grad_norm": 2.6406726837158203,
+      "learning_rate": 1.8708604970945746e-05,
+      "loss": 0.1076,
+      "step": 1889
+    },
+    {
+      "epoch": 0.9895233106338397,
+      "loss_partial": 0.05615234375,
+      "step": 1889
+    },
+    {
+      "epoch": 0.9895233106338397,
+      "loss_partial": 0.05224609375,
+      "step": 1889
+    },
+    {
+      "epoch": 0.9900471451021478,
+      "grad_norm": 1.9152324199676514,
+      "learning_rate": 1.870635698904629e-05,
+      "loss": 0.1098,
+      "step": 1890
+    },
+    {
+      "epoch": 0.9900471451021478,
+      "loss_partial": 0.052734375,
+      "step": 1890
+    },
+    {
+      "epoch": 0.9900471451021478,
+      "loss_partial": 0.050048828125,
+      "step": 1890
+    },
+    {
+      "epoch": 0.9905709795704557,
+      "grad_norm": 2.355679750442505,
+      "learning_rate": 1.8704107187572057e-05,
+      "loss": 0.1116,
+      "step": 1891
+    },
+    {
+      "epoch": 0.9905709795704557,
+      "loss_partial": 0.044921875,
+      "step": 1891
+    },
+    {
+      "epoch": 0.9905709795704557,
+      "loss_partial": 0.052978515625,
+      "step": 1891
+    },
+    {
+      "epoch": 0.9910948140387638,
+      "grad_norm": 1.925642967224121,
+      "learning_rate": 1.8701855566993248e-05,
+      "loss": 0.1092,
+      "step": 1892
+    },
+    {
+      "epoch": 0.9910948140387638,
+      "loss_partial": 0.056884765625,
+      "step": 1892
+    },
+    {
+      "epoch": 0.9910948140387638,
+      "loss_partial": 0.052978515625,
+      "step": 1892
+    },
+    {
+      "epoch": 0.9916186485070718,
+      "grad_norm": 2.4415385723114014,
+      "learning_rate": 1.8699602127780438e-05,
+      "loss": 0.1104,
+      "step": 1893
+    },
+    {
+      "epoch": 0.9916186485070718,
+      "loss_partial": 0.050537109375,
+      "step": 1893
+    },
+    {
+      "epoch": 0.9916186485070718,
+      "loss_partial": 0.055419921875,
+      "step": 1893
+    },
+    {
+      "epoch": 0.9921424829753798,
+      "grad_norm": 1.809395670890808,
+      "learning_rate": 1.869734687040458e-05,
+      "loss": 0.1093,
+      "step": 1894
+    },
+    {
+      "epoch": 0.9921424829753798,
+      "loss_partial": 0.054443359375,
+      "step": 1894
+    },
+    {
+      "epoch": 0.9921424829753798,
+      "loss_partial": 0.05078125,
+      "step": 1894
+    },
+    {
+      "epoch": 0.9926663174436878,
+      "grad_norm": 1.9443732500076294,
+      "learning_rate": 1.869508979533701e-05,
+      "loss": 0.1118,
+      "step": 1895
+    },
+    {
+      "epoch": 0.9926663174436878,
+      "loss_partial": 0.05224609375,
+      "step": 1895
+    },
+    {
+      "epoch": 0.9926663174436878,
+      "loss_partial": 0.048095703125,
+      "step": 1895
+    },
+    {
+      "epoch": 0.9931901519119958,
+      "grad_norm": 1.3859320878982544,
+      "learning_rate": 1.8692830903049435e-05,
+      "loss": 0.1079,
+      "step": 1896
+    },
+    {
+      "epoch": 0.9931901519119958,
+      "loss_partial": 0.05322265625,
+      "step": 1896
+    },
+    {
+      "epoch": 0.9931901519119958,
+      "loss_partial": 0.048828125,
+      "step": 1896
+    },
+    {
+      "epoch": 0.9937139863803038,
+      "grad_norm": 1.658262848854065,
+      "learning_rate": 1.869057019401396e-05,
+      "loss": 0.1081,
+      "step": 1897
+    },
+    {
+      "epoch": 0.9937139863803038,
+      "loss_partial": 0.056640625,
+      "step": 1897
+    },
+    {
+      "epoch": 0.9937139863803038,
+      "loss_partial": 0.0546875,
+      "step": 1897
+    },
+    {
+      "epoch": 0.9942378208486118,
+      "grad_norm": 2.091254472732544,
+      "learning_rate": 1.8688307668703057e-05,
+      "loss": 0.1106,
+      "step": 1898
+    },
+    {
+      "epoch": 0.9942378208486118,
+      "loss_partial": 0.053955078125,
+      "step": 1898
+    },
+    {
+      "epoch": 0.9942378208486118,
+      "loss_partial": 0.054443359375,
+      "step": 1898
+    },
+    {
+      "epoch": 0.9947616553169198,
+      "grad_norm": 1.5976815223693848,
+      "learning_rate": 1.8686043327589577e-05,
+      "loss": 0.1067,
+      "step": 1899
+    },
+    {
+      "epoch": 0.9947616553169198,
+      "loss_partial": 0.0478515625,
+      "step": 1899
+    },
+    {
+      "epoch": 0.9947616553169198,
+      "loss_partial": 0.056640625,
+      "step": 1899
+    },
+    {
+      "epoch": 0.9952854897852279,
+      "grad_norm": 1.6617566347122192,
+      "learning_rate": 1.8683777171146754e-05,
+      "loss": 0.1064,
+      "step": 1900
+    },
+    {
+      "epoch": 0.9952854897852279,
+      "loss_partial": 0.0478515625,
+      "step": 1900
+    },
+    {
+      "epoch": 0.9952854897852279,
+      "loss_partial": 0.05224609375,
+      "step": 1900
+    },
+    {
+      "epoch": 0.9958093242535359,
+      "grad_norm": 1.9904876947402954,
+      "learning_rate": 1.86815091998482e-05,
+      "loss": 0.105,
+      "step": 1901
+    },
+    {
+      "epoch": 0.9958093242535359,
+      "loss_partial": 0.048583984375,
+      "step": 1901
+    },
+    {
+      "epoch": 0.9958093242535359,
+      "loss_partial": 0.05224609375,
+      "step": 1901
+    },
+    {
+      "epoch": 0.9963331587218439,
+      "grad_norm": 1.6518065929412842,
+      "learning_rate": 1.867923941416791e-05,
+      "loss": 0.1056,
+      "step": 1902
+    },
+    {
+      "epoch": 0.9963331587218439,
+      "loss_partial": 0.055908203125,
+      "step": 1902
+    },
+    {
+      "epoch": 0.9963331587218439,
+      "loss_partial": 0.049560546875,
+      "step": 1902
+    },
+    {
+      "epoch": 0.9968569931901519,
+      "grad_norm": 1.9470348358154297,
+      "learning_rate": 1.8676967814580253e-05,
+      "loss": 0.1063,
+      "step": 1903
+    },
+    {
+      "epoch": 0.9968569931901519,
+      "loss_partial": 0.0634765625,
+      "step": 1903
+    },
+    {
+      "epoch": 0.9968569931901519,
+      "loss_partial": 0.0546875,
+      "step": 1903
+    },
+    {
+      "epoch": 0.9973808276584599,
+      "grad_norm": 1.8201767206192017,
+      "learning_rate": 1.867469440155998e-05,
+      "loss": 0.1086,
+      "step": 1904
+    },
+    {
+      "epoch": 0.9973808276584599,
+      "loss_partial": 0.053466796875,
+      "step": 1904
+    },
+    {
+      "epoch": 0.9973808276584599,
+      "loss_partial": 0.04638671875,
+      "step": 1904
+    },
+    {
+      "epoch": 0.997904662126768,
+      "grad_norm": 1.709432601928711,
+      "learning_rate": 1.8672419175582215e-05,
+      "loss": 0.1067,
+      "step": 1905
+    },
+    {
+      "epoch": 0.997904662126768,
+      "loss_partial": 0.04296875,
+      "step": 1905
+    },
+    {
+      "epoch": 0.997904662126768,
+      "loss_partial": 0.0634765625,
+      "step": 1905
+    },
+    {
+      "epoch": 0.9984284965950759,
+      "grad_norm": 1.7894378900527954,
+      "learning_rate": 1.8670142137122475e-05,
+      "loss": 0.107,
+      "step": 1906
+    },
+    {
+      "epoch": 0.9984284965950759,
+      "loss_partial": 0.05419921875,
+      "step": 1906
+    },
+    {
+      "epoch": 0.9984284965950759,
+      "loss_partial": 0.054931640625,
+      "step": 1906
+    },
+    {
+      "epoch": 0.998952331063384,
+      "grad_norm": 2.0385565757751465,
+      "learning_rate": 1.8667863286656644e-05,
+      "loss": 0.11,
+      "step": 1907
+    },
+    {
+      "epoch": 0.998952331063384,
+      "loss_partial": 0.051513671875,
+      "step": 1907
+    },
+    {
+      "epoch": 0.998952331063384,
+      "loss_partial": 0.048583984375,
+      "step": 1907
+    },
+    {
+      "epoch": 0.9994761655316919,
+      "grad_norm": 2.7804648876190186,
+      "learning_rate": 1.8665582624660985e-05,
+      "loss": 0.1035,
+      "step": 1908
+    },
+    {
+      "epoch": 0.9994761655316919,
+      "loss_partial": 0.055908203125,
+      "step": 1908
+    },
+    {
+      "epoch": 0.9994761655316919,
+      "loss_partial": 0.055908203125,
+      "step": 1908
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.8493319749832153,
+      "learning_rate": 1.8663300151612143e-05,
+      "loss": 0.1125,
+      "step": 1909
+    },
+    {
+      "epoch": 1.0,
+      "loss_partial": 0.056640625,
+      "step": 1909
+    },
+    {
+      "epoch": 1.0,
+      "loss_partial": 0.050537109375,
+      "step": 1909
+    },
+    {
+      "epoch": 1.000523834468308,
+      "grad_norm": 1.9675685167312622,
+      "learning_rate": 1.8661015867987144e-05,
+      "loss": 0.1087,
+      "step": 1910
+    },
+    {
+      "epoch": 1.000523834468308,
+      "loss_partial": 0.050048828125,
+      "step": 1910
+    },
+    {
+      "epoch": 1.000523834468308,
+      "loss_partial": 0.0517578125,
+      "step": 1910
+    },
+    {
+      "epoch": 1.0010476689366161,
+      "grad_norm": 2.5741331577301025,
+      "learning_rate": 1.8658729774263385e-05,
+      "loss": 0.1084,
+      "step": 1911
+    },
+    {
+      "epoch": 1.0010476689366161,
+      "loss_partial": 0.0537109375,
+      "step": 1911
+    },
+    {
+      "epoch": 1.0010476689366161,
+      "loss_partial": 0.051025390625,
+      "step": 1911
+    },
+    {
+      "epoch": 1.001571503404924,
+      "grad_norm": 1.8408257961273193,
+      "learning_rate": 1.8656441870918648e-05,
+      "loss": 0.1053,
+      "step": 1912
+    },
+    {
+      "epoch": 1.001571503404924,
+      "loss_partial": 0.05712890625,
+      "step": 1912
+    },
+    {
+      "epoch": 1.001571503404924,
+      "loss_partial": 0.0615234375,
+      "step": 1912
+    },
+    {
+      "epoch": 1.002095337873232,
+      "grad_norm": 1.5339144468307495,
+      "learning_rate": 1.865415215843109e-05,
+      "loss": 0.1081,
+      "step": 1913
+    },
+    {
+      "epoch": 1.002095337873232,
+      "loss_partial": 0.05126953125,
+      "step": 1913
+    },
+    {
+      "epoch": 1.002095337873232,
+      "loss_partial": 0.05029296875,
+      "step": 1913
+    },
+    {
+      "epoch": 1.00261917234154,
+      "grad_norm": 1.8214960098266602,
+      "learning_rate": 1.8651860637279245e-05,
+      "loss": 0.1071,
+      "step": 1914
+    },
+    {
+      "epoch": 1.00261917234154,
+      "loss_partial": 0.056884765625,
+      "step": 1914
+    },
+    {
+      "epoch": 1.00261917234154,
+      "loss_partial": 0.052490234375,
+      "step": 1914
+    },
+    {
+      "epoch": 1.0031430068098481,
+      "grad_norm": 1.810804843902588,
+      "learning_rate": 1.8649567307942028e-05,
+      "loss": 0.1073,
+      "step": 1915
+    },
+    {
+      "epoch": 1.0031430068098481,
+      "loss_partial": 0.054443359375,
+      "step": 1915
+    },
+    {
+      "epoch": 1.0031430068098481,
+      "loss_partial": 0.06396484375,
+      "step": 1915
+    },
+    {
+      "epoch": 1.0036668412781562,
+      "grad_norm": 2.2838568687438965,
+      "learning_rate": 1.8647272170898733e-05,
+      "loss": 0.1059,
+      "step": 1916
+    },
+    {
+      "epoch": 1.0036668412781562,
+      "loss_partial": 0.05810546875,
+      "step": 1916
+    },
+    {
+      "epoch": 1.0036668412781562,
+      "loss_partial": 0.05126953125,
+      "step": 1916
+    },
+    {
+      "epoch": 1.004190675746464,
+      "grad_norm": 2.372927665710449,
+      "learning_rate": 1.8644975226629025e-05,
+      "loss": 0.1073,
+      "step": 1917
+    },
+    {
+      "epoch": 1.004190675746464,
+      "loss_partial": 0.060546875,
+      "step": 1917
+    },
+    {
+      "epoch": 1.004190675746464,
+      "loss_partial": 0.053466796875,
+      "step": 1917
+    },
+    {
+      "epoch": 1.004714510214772,
+      "grad_norm": 1.5935304164886475,
+      "learning_rate": 1.8642676475612952e-05,
+      "loss": 0.1074,
+      "step": 1918
+    },
+    {
+      "epoch": 1.004714510214772,
+      "loss_partial": 0.05712890625,
+      "step": 1918
+    },
+    {
+      "epoch": 1.004714510214772,
+      "loss_partial": 0.04833984375,
+      "step": 1918
+    },
+    {
+      "epoch": 1.0052383446830802,
+      "grad_norm": 1.5919743776321411,
+      "learning_rate": 1.8640375918330943e-05,
+      "loss": 0.1056,
+      "step": 1919
+    },
+    {
+      "epoch": 1.0052383446830802,
+      "loss_partial": 0.06103515625,
+      "step": 1919
+    },
+    {
+      "epoch": 1.0052383446830802,
+      "loss_partial": 0.040283203125,
+      "step": 1919
+    },
+    {
+      "epoch": 1.0057621791513882,
+      "grad_norm": 1.5906802415847778,
+      "learning_rate": 1.8638073555263802e-05,
+      "loss": 0.1095,
+      "step": 1920
+    },
+    {
+      "epoch": 1.0057621791513882,
+      "loss_partial": 0.04931640625,
+      "step": 1920
+    },
+    {
+      "epoch": 1.0057621791513882,
+      "loss_partial": 0.049560546875,
+      "step": 1920
+    },
+    {
+      "epoch": 1.0062860136196963,
+      "grad_norm": 2.295048713684082,
+      "learning_rate": 1.8635769386892697e-05,
+      "loss": 0.106,
+      "step": 1921
+    },
+    {
+      "epoch": 1.0062860136196963,
+      "loss_partial": 0.0556640625,
+      "step": 1921
+    },
+    {
+      "epoch": 1.0062860136196963,
+      "loss_partial": 0.05908203125,
+      "step": 1921
+    },
+    {
+      "epoch": 1.0068098480880041,
+      "grad_norm": 2.3985795974731445,
+      "learning_rate": 1.8633463413699195e-05,
+      "loss": 0.1116,
+      "step": 1922
+    },
+    {
+      "epoch": 1.0068098480880041,
+      "loss_partial": 0.051513671875,
+      "step": 1922
+    },
+    {
+      "epoch": 1.0068098480880041,
+      "loss_partial": 0.04931640625,
+      "step": 1922
+    },
+    {
+      "epoch": 1.0073336825563122,
+      "grad_norm": 1.2337228059768677,
+      "learning_rate": 1.863115563616523e-05,
+      "loss": 0.1046,
+      "step": 1923
+    },
+    {
+      "epoch": 1.0073336825563122,
+      "loss_partial": 0.054931640625,
+      "step": 1923
+    },
+    {
+      "epoch": 1.0073336825563122,
+      "loss_partial": 0.0546875,
+      "step": 1923
+    },
+    {
+      "epoch": 1.0078575170246202,
+      "grad_norm": 2.5527420043945312,
+      "learning_rate": 1.8628846054773107e-05,
+      "loss": 0.1148,
+      "step": 1924
+    },
+    {
+      "epoch": 1.0078575170246202,
+      "loss_partial": 0.052978515625,
+      "step": 1924
+    },
+    {
+      "epoch": 1.0078575170246202,
+      "loss_partial": 0.058349609375,
+      "step": 1924
+    },
+    {
+      "epoch": 1.0083813514929283,
+      "grad_norm": 1.924911618232727,
+      "learning_rate": 1.862653467000552e-05,
+      "loss": 0.1101,
+      "step": 1925
+    },
+    {
+      "epoch": 1.0083813514929283,
+      "loss_partial": 0.0576171875,
+      "step": 1925
+    },
+    {
+      "epoch": 1.0083813514929283,
+      "loss_partial": 0.05029296875,
+      "step": 1925
+    },
+    {
+      "epoch": 1.0089051859612364,
+      "grad_norm": 2.0954065322875977,
+      "learning_rate": 1.862422148234553e-05,
+      "loss": 0.1079,
+      "step": 1926
+    },
+    {
+      "epoch": 1.0089051859612364,
+      "loss_partial": 0.053466796875,
+      "step": 1926
+    },
+    {
+      "epoch": 1.0089051859612364,
+      "loss_partial": 0.0517578125,
+      "step": 1926
+    },
+    {
+      "epoch": 1.0094290204295442,
+      "grad_norm": 1.8477749824523926,
+      "learning_rate": 1.862190649227658e-05,
+      "loss": 0.1078,
+      "step": 1927
+    },
+    {
+      "epoch": 1.0094290204295442,
+      "loss_partial": 0.07470703125,
+      "step": 1927
+    },
+    {
+      "epoch": 1.0094290204295442,
+      "loss_partial": 0.064453125,
+      "step": 1927
+    },
+    {
+      "epoch": 1.0099528548978522,
+      "grad_norm": 1.7281605005264282,
+      "learning_rate": 1.861958970028249e-05,
+      "loss": 0.11,
+      "step": 1928
+    },
+    {
+      "epoch": 1.0099528548978522,
+      "loss_partial": 0.06005859375,
+      "step": 1928
+    },
+    {
+      "epoch": 1.0099528548978522,
+      "loss_partial": 0.04150390625,
+      "step": 1928
+    },
+    {
+      "epoch": 1.0104766893661603,
+      "grad_norm": 2.1269819736480713,
+      "learning_rate": 1.8617271106847456e-05,
+      "loss": 0.1082,
+      "step": 1929
+    },
+    {
+      "epoch": 1.0104766893661603,
+      "loss_partial": 0.06689453125,
+      "step": 1929
+    },
+    {
+      "epoch": 1.0104766893661603,
+      "loss_partial": 0.0517578125,
+      "step": 1929
+    },
+    {
+      "epoch": 1.0110005238344684,
+      "grad_norm": 2.784015655517578,
+      "learning_rate": 1.8614950712456048e-05,
+      "loss": 0.113,
+      "step": 1930
+    },
+    {
+      "epoch": 1.0110005238344684,
+      "loss_partial": 0.052978515625,
+      "step": 1930
+    },
+    {
+      "epoch": 1.0110005238344684,
+      "loss_partial": 0.0595703125,
+      "step": 1930
+    },
+    {
+      "epoch": 1.0115243583027764,
+      "grad_norm": 2.366584539413452,
+      "learning_rate": 1.8612628517593213e-05,
+      "loss": 0.1089,
+      "step": 1931
+    },
+    {
+      "epoch": 1.0115243583027764,
+      "loss_partial": 0.05419921875,
+      "step": 1931
+    },
+    {
+      "epoch": 1.0115243583027764,
+      "loss_partial": 0.060302734375,
+      "step": 1931
+    },
+    {
+      "epoch": 1.0120481927710843,
+      "grad_norm": 3.2974634170532227,
+      "learning_rate": 1.8610304522744277e-05,
+      "loss": 0.1103,
+      "step": 1932
+    },
+    {
+      "epoch": 1.0120481927710843,
+      "loss_partial": 0.04931640625,
+      "step": 1932
+    },
+    {
+      "epoch": 1.0120481927710843,
+      "loss_partial": 0.0546875,
+      "step": 1932
+    },
+    {
+      "epoch": 1.0125720272393923,
+      "grad_norm": 2.4301300048828125,
+      "learning_rate": 1.860797872839494e-05,
+      "loss": 0.1074,
+      "step": 1933
+    },
+    {
+      "epoch": 1.0125720272393923,
+      "loss_partial": 0.0625,
+      "step": 1933
+    },
+    {
+      "epoch": 1.0125720272393923,
+      "loss_partial": 0.052978515625,
+      "step": 1933
+    },
+    {
+      "epoch": 1.0130958617077004,
+      "grad_norm": 2.002899646759033,
+      "learning_rate": 1.860565113503128e-05,
+      "loss": 0.1082,
+      "step": 1934
+    },
+    {
+      "epoch": 1.0130958617077004,
+      "loss_partial": 0.053955078125,
+      "step": 1934
+    },
+    {
+      "epoch": 1.0130958617077004,
+      "loss_partial": 0.055908203125,
+      "step": 1934
+    },
+    {
+      "epoch": 1.0136196961760084,
+      "grad_norm": 2.1796605587005615,
+      "learning_rate": 1.8603321743139748e-05,
+      "loss": 0.1118,
+      "step": 1935
+    },
+    {
+      "epoch": 1.0136196961760084,
+      "loss_partial": 0.0654296875,
+      "step": 1935
+    },
+    {
+      "epoch": 1.0136196961760084,
+      "loss_partial": 0.050048828125,
+      "step": 1935
+    },
+    {
+      "epoch": 1.0141435306443165,
+      "grad_norm": 2.019866466522217,
+      "learning_rate": 1.8600990553207175e-05,
+      "loss": 0.1073,
+      "step": 1936
+    },
+    {
+      "epoch": 1.0141435306443165,
+      "loss_partial": 0.0537109375,
+      "step": 1936
+    },
+    {
+      "epoch": 1.0141435306443165,
+      "loss_partial": 0.053466796875,
+      "step": 1936
+    },
+    {
+      "epoch": 1.0146673651126243,
+      "grad_norm": 1.8585461378097534,
+      "learning_rate": 1.859865756572076e-05,
+      "loss": 0.109,
+      "step": 1937
+    },
+    {
+      "epoch": 1.0146673651126243,
+      "loss_partial": 0.054443359375,
+      "step": 1937
+    },
+    {
+      "epoch": 1.0146673651126243,
+      "loss_partial": 0.0625,
+      "step": 1937
+    },
+    {
+      "epoch": 1.0151911995809324,
+      "grad_norm": 2.289822578430176,
+      "learning_rate": 1.859632278116809e-05,
+      "loss": 0.1076,
+      "step": 1938
+    },
+    {
+      "epoch": 1.0151911995809324,
+      "loss_partial": 0.05712890625,
+      "step": 1938
+    },
+    {
+      "epoch": 1.0151911995809324,
+      "loss_partial": 0.056640625,
+      "step": 1938
+    },
+    {
+      "epoch": 1.0157150340492405,
+      "grad_norm": 1.857613205909729,
+      "learning_rate": 1.8593986200037114e-05,
+      "loss": 0.1078,
+      "step": 1939
+    },
+    {
+      "epoch": 1.0157150340492405,
+      "loss_partial": 0.059814453125,
+      "step": 1939
+    },
+    {
+      "epoch": 1.0157150340492405,
+      "loss_partial": 0.052490234375,
+      "step": 1939
+    },
+    {
+      "epoch": 1.0162388685175485,
+      "grad_norm": 1.9429391622543335,
+      "learning_rate": 1.8591647822816165e-05,
+      "loss": 0.1124,
+      "step": 1940
+    },
+    {
+      "epoch": 1.0162388685175485,
+      "loss_partial": 0.0576171875,
+      "step": 1940
+    },
+    {
+      "epoch": 1.0162388685175485,
+      "loss_partial": 0.052490234375,
+      "step": 1940
+    },
+    {
+      "epoch": 1.0167627029858566,
+      "grad_norm": 2.2086093425750732,
+      "learning_rate": 1.8589307649993952e-05,
+      "loss": 0.1077,
+      "step": 1941
+    },
+    {
+      "epoch": 1.0167627029858566,
+      "loss_partial": 0.051513671875,
+      "step": 1941
+    },
+    {
+      "epoch": 1.0167627029858566,
+      "loss_partial": 0.050537109375,
+      "step": 1941
+    },
+    {
+      "epoch": 1.0172865374541644,
+      "grad_norm": 1.76565420627594,
+      "learning_rate": 1.8586965682059553e-05,
+      "loss": 0.1061,
+      "step": 1942
+    },
+    {
+      "epoch": 1.0172865374541644,
+      "loss_partial": 0.048828125,
+      "step": 1942
+    },
+    {
+      "epoch": 1.0172865374541644,
+      "loss_partial": 0.0546875,
+      "step": 1942
+    },
+    {
+      "epoch": 1.0178103719224725,
+      "grad_norm": 1.9355261325836182,
+      "learning_rate": 1.858462191950243e-05,
+      "loss": 0.1029,
+      "step": 1943
+    },
+    {
+      "epoch": 1.0178103719224725,
+      "loss_partial": 0.0634765625,
+      "step": 1943
+    },
+    {
+      "epoch": 1.0178103719224725,
+      "loss_partial": 0.052001953125,
+      "step": 1943
+    },
+    {
+      "epoch": 1.0183342063907805,
+      "grad_norm": 1.6928691864013672,
+      "learning_rate": 1.8582276362812412e-05,
+      "loss": 0.1123,
+      "step": 1944
+    },
+    {
+      "epoch": 1.0183342063907805,
+      "loss_partial": 0.05517578125,
+      "step": 1944
+    },
+    {
+      "epoch": 1.0183342063907805,
+      "loss_partial": 0.045654296875,
+      "step": 1944
+    },
+    {
+      "epoch": 1.0188580408590886,
+      "grad_norm": 1.9871079921722412,
+      "learning_rate": 1.85799290124797e-05,
+      "loss": 0.1059,
+      "step": 1945
+    },
+    {
+      "epoch": 1.0188580408590886,
+      "loss_partial": 0.050537109375,
+      "step": 1945
+    },
+    {
+      "epoch": 1.0188580408590886,
+      "loss_partial": 0.05419921875,
+      "step": 1945
+    },
+    {
+      "epoch": 1.0193818753273964,
+      "grad_norm": 1.8050694465637207,
+      "learning_rate": 1.8577579868994884e-05,
+      "loss": 0.1082,
+      "step": 1946
+    },
+    {
+      "epoch": 1.0193818753273964,
+      "loss_partial": 0.04541015625,
+      "step": 1946
+    },
+    {
+      "epoch": 1.0193818753273964,
+      "loss_partial": 0.05615234375,
+      "step": 1946
+    },
+    {
+      "epoch": 1.0199057097957045,
+      "grad_norm": 2.1731057167053223,
+      "learning_rate": 1.857522893284892e-05,
+      "loss": 0.108,
+      "step": 1947
+    },
+    {
+      "epoch": 1.0199057097957045,
+      "loss_partial": 0.0634765625,
+      "step": 1947
+    },
+    {
+      "epoch": 1.0199057097957045,
+      "loss_partial": 0.058349609375,
+      "step": 1947
+    },
+    {
+      "epoch": 1.0204295442640126,
+      "grad_norm": 1.952982783317566,
+      "learning_rate": 1.857287620453313e-05,
+      "loss": 0.1131,
+      "step": 1948
+    },
+    {
+      "epoch": 1.0204295442640126,
+      "loss_partial": 0.0654296875,
+      "step": 1948
+    },
+    {
+      "epoch": 1.0204295442640126,
+      "loss_partial": 0.051513671875,
+      "step": 1948
+    },
+    {
+      "epoch": 1.0209533787323206,
+      "grad_norm": 2.165214776992798,
+      "learning_rate": 1.8570521684539233e-05,
+      "loss": 0.1107,
+      "step": 1949
+    },
+    {
+      "epoch": 1.0209533787323206,
+      "loss_partial": 0.047119140625,
+      "step": 1949
+    },
+    {
+      "epoch": 1.0209533787323206,
+      "loss_partial": 0.055908203125,
+      "step": 1949
+    },
+    {
+      "epoch": 1.0214772132006287,
+      "grad_norm": 1.7470977306365967,
+      "learning_rate": 1.85681653733593e-05,
+      "loss": 0.1044,
+      "step": 1950
+    },
+    {
+      "epoch": 1.0214772132006287,
+      "loss_partial": 0.05908203125,
+      "step": 1950
+    },
+    {
+      "epoch": 1.0214772132006287,
+      "loss_partial": 0.053466796875,
+      "step": 1950
+    },
+    {
+      "epoch": 1.0220010476689365,
+      "grad_norm": 1.9910192489624023,
+      "learning_rate": 1.8565807271485786e-05,
+      "loss": 0.1109,
+      "step": 1951
+    },
+    {
+      "epoch": 1.0220010476689365,
+      "loss_partial": 0.04833984375,
+      "step": 1951
+    },
+    {
+      "epoch": 1.0220010476689365,
+      "loss_partial": 0.057373046875,
+      "step": 1951
+    },
+    {
+      "epoch": 1.0225248821372446,
+      "grad_norm": 1.9853991270065308,
+      "learning_rate": 1.8563447379411523e-05,
+      "loss": 0.1089,
+      "step": 1952
+    },
+    {
+      "epoch": 1.0225248821372446,
+      "loss_partial": 0.060791015625,
+      "step": 1952
+    },
+    {
+      "epoch": 1.0225248821372446,
+      "loss_partial": 0.04638671875,
+      "step": 1952
+    },
+    {
+      "epoch": 1.0230487166055526,
+      "grad_norm": 2.1558268070220947,
+      "learning_rate": 1.8561085697629715e-05,
+      "loss": 0.1104,
+      "step": 1953
+    },
+    {
+      "epoch": 1.0230487166055526,
+      "loss_partial": 0.05517578125,
+      "step": 1953
+    },
+    {
+      "epoch": 1.0230487166055526,
+      "loss_partial": 0.053466796875,
+      "step": 1953
+    },
+    {
+      "epoch": 1.0235725510738607,
+      "grad_norm": 2.300410032272339,
+      "learning_rate": 1.8558722226633934e-05,
+      "loss": 0.106,
+      "step": 1954
+    },
+    {
+      "epoch": 1.0235725510738607,
+      "loss_partial": 0.056640625,
+      "step": 1954
+    },
+    {
+      "epoch": 1.0235725510738607,
+      "loss_partial": 0.0576171875,
+      "step": 1954
+    },
+    {
+      "epoch": 1.0240963855421688,
+      "grad_norm": 1.9295134544372559,
+      "learning_rate": 1.8556356966918132e-05,
+      "loss": 0.107,
+      "step": 1955
+    },
+    {
+      "epoch": 1.0240963855421688,
+      "loss_partial": 0.061279296875,
+      "step": 1955
+    },
+    {
+      "epoch": 1.0240963855421688,
+      "loss_partial": 0.0478515625,
+      "step": 1955
+    },
+    {
+      "epoch": 1.0246202200104766,
+      "grad_norm": 2.3252112865448,
+      "learning_rate": 1.8553989918976638e-05,
+      "loss": 0.1082,
+      "step": 1956
+    },
+    {
+      "epoch": 1.0246202200104766,
+      "loss_partial": 0.05224609375,
+      "step": 1956
+    },
+    {
+      "epoch": 1.0246202200104766,
+      "loss_partial": 0.053955078125,
+      "step": 1956
+    },
+    {
+      "epoch": 1.0251440544787846,
+      "grad_norm": 2.508312702178955,
+      "learning_rate": 1.8551621083304147e-05,
+      "loss": 0.1064,
+      "step": 1957
+    },
+    {
+      "epoch": 1.0251440544787846,
+      "loss_partial": 0.048095703125,
+      "step": 1957
+    },
+    {
+      "epoch": 1.0251440544787846,
+      "loss_partial": 0.05224609375,
+      "step": 1957
+    },
+    {
+      "epoch": 1.0256678889470927,
+      "grad_norm": 2.0748839378356934,
+      "learning_rate": 1.8549250460395728e-05,
+      "loss": 0.1115,
+      "step": 1958
+    },
+    {
+      "epoch": 1.0256678889470927,
+      "loss_partial": 0.059326171875,
+      "step": 1958
+    },
+    {
+      "epoch": 1.0256678889470927,
+      "loss_partial": 0.049560546875,
+      "step": 1958
+    },
+    {
+      "epoch": 1.0261917234154008,
+      "grad_norm": 2.492793321609497,
+      "learning_rate": 1.854687805074683e-05,
+      "loss": 0.1123,
+      "step": 1959
+    },
+    {
+      "epoch": 1.0261917234154008,
+      "loss_partial": 0.05419921875,
+      "step": 1959
+    },
+    {
+      "epoch": 1.0261917234154008,
+      "loss_partial": 0.05859375,
+      "step": 1959
+    },
+    {
+      "epoch": 1.0267155578837088,
+      "grad_norm": 2.0071115493774414,
+      "learning_rate": 1.8544503854853274e-05,
+      "loss": 0.11,
+      "step": 1960
+    },
+    {
+      "epoch": 1.0267155578837088,
+      "loss_partial": 0.0517578125,
+      "step": 1960
+    },
+    {
+      "epoch": 1.0267155578837088,
+      "loss_partial": 0.058837890625,
+      "step": 1960
+    },
+    {
+      "epoch": 1.0272393923520167,
+      "grad_norm": 2.334285259246826,
+      "learning_rate": 1.854212787321125e-05,
+      "loss": 0.1091,
+      "step": 1961
+    },
+    {
+      "epoch": 1.0272393923520167,
+      "loss_partial": 0.059326171875,
+      "step": 1961
+    },
+    {
+      "epoch": 1.0272393923520167,
+      "loss_partial": 0.05224609375,
+      "step": 1961
+    },
+    {
+      "epoch": 1.0277632268203247,
+      "grad_norm": 2.5955018997192383,
+      "learning_rate": 1.8539750106317324e-05,
+      "loss": 0.1084,
+      "step": 1962
+    },
+    {
+      "epoch": 1.0277632268203247,
+      "loss_partial": 0.055419921875,
+      "step": 1962
+    },
+    {
+      "epoch": 1.0277632268203247,
+      "loss_partial": 0.050537109375,
+      "step": 1962
+    },
+    {
+      "epoch": 1.0282870612886328,
+      "grad_norm": 2.2115724086761475,
+      "learning_rate": 1.853737055466843e-05,
+      "loss": 0.1085,
+      "step": 1963
+    },
+    {
+      "epoch": 1.0282870612886328,
+      "loss_partial": 0.052734375,
+      "step": 1963
+    },
+    {
+      "epoch": 1.0282870612886328,
+      "loss_partial": 0.05078125,
+      "step": 1963
+    },
+    {
+      "epoch": 1.0288108957569408,
+      "grad_norm": 2.163494348526001,
+      "learning_rate": 1.8534989218761886e-05,
+      "loss": 0.1078,
+      "step": 1964
+    },
+    {
+      "epoch": 1.0288108957569408,
+      "loss_partial": 0.047119140625,
+      "step": 1964
+    },
+    {
+      "epoch": 1.0288108957569408,
+      "loss_partial": 0.05859375,
+      "step": 1964
+    },
+    {
+      "epoch": 1.029334730225249,
+      "grad_norm": 2.0461831092834473,
+      "learning_rate": 1.8532606099095376e-05,
+      "loss": 0.1068,
+      "step": 1965
+    },
+    {
+      "epoch": 1.029334730225249,
+      "loss_partial": 0.05322265625,
+      "step": 1965
+    },
+    {
+      "epoch": 1.029334730225249,
+      "loss_partial": 0.047607421875,
+      "step": 1965
+    },
+    {
+      "epoch": 1.0298585646935567,
+      "grad_norm": 2.022643804550171,
+      "learning_rate": 1.8530221196166954e-05,
+      "loss": 0.1089,
+      "step": 1966
+    },
+    {
+      "epoch": 1.0298585646935567,
+      "loss_partial": 0.048828125,
+      "step": 1966
+    },
+    {
+      "epoch": 1.0298585646935567,
+      "loss_partial": 0.052490234375,
+      "step": 1966
+    },
+    {
+      "epoch": 1.0303823991618648,
+      "grad_norm": 1.7395230531692505,
+      "learning_rate": 1.852783451047505e-05,
+      "loss": 0.1015,
+      "step": 1967
+    },
+    {
+      "epoch": 1.0303823991618648,
+      "loss_partial": 0.051513671875,
+      "step": 1967
+    },
+    {
+      "epoch": 1.0303823991618648,
+      "loss_partial": 0.052734375,
+      "step": 1967
+    },
+    {
+      "epoch": 1.0309062336301729,
+      "grad_norm": 1.8619463443756104,
+      "learning_rate": 1.852544604251847e-05,
+      "loss": 0.1087,
+      "step": 1968
+    },
+    {
+      "epoch": 1.0309062336301729,
+      "loss_partial": 0.0556640625,
+      "step": 1968
+    },
+    {
+      "epoch": 1.0309062336301729,
+      "loss_partial": 0.05712890625,
+      "step": 1968
+    },
+    {
+      "epoch": 1.031430068098481,
+      "grad_norm": 1.65328049659729,
+      "learning_rate": 1.8523055792796387e-05,
+      "loss": 0.1045,
+      "step": 1969
+    },
+    {
+      "epoch": 1.031430068098481,
+      "loss_partial": 0.0595703125,
+      "step": 1969
+    },
+    {
+      "epoch": 1.031430068098481,
+      "loss_partial": 0.057373046875,
+      "step": 1969
+    },
+    {
+      "epoch": 1.031953902566789,
+      "grad_norm": 1.48519766330719,
+      "learning_rate": 1.8520663761808344e-05,
+      "loss": 0.1061,
+      "step": 1970
+    },
+    {
+      "epoch": 1.031953902566789,
+      "loss_partial": 0.056884765625,
+      "step": 1970
+    },
+    {
+      "epoch": 1.031953902566789,
+      "loss_partial": 0.06640625,
+      "step": 1970
+    },
+    {
+      "epoch": 1.0324777370350968,
+      "grad_norm": 1.6295140981674194,
+      "learning_rate": 1.851826995005427e-05,
+      "loss": 0.114,
+      "step": 1971
+    },
+    {
+      "epoch": 1.0324777370350968,
+      "loss_partial": 0.07080078125,
+      "step": 1971
+    },
+    {
+      "epoch": 1.0324777370350968,
+      "loss_partial": 0.064453125,
+      "step": 1971
+    },
+    {
+      "epoch": 1.0330015715034049,
+      "grad_norm": 1.4543534517288208,
+      "learning_rate": 1.851587435803445e-05,
+      "loss": 0.1035,
+      "step": 1972
+    },
+    {
+      "epoch": 1.0330015715034049,
+      "loss_partial": 0.048828125,
+      "step": 1972
+    },
+    {
+      "epoch": 1.0330015715034049,
+      "loss_partial": 0.060791015625,
+      "step": 1972
+    },
+    {
+      "epoch": 1.033525405971713,
+      "grad_norm": 1.6347049474716187,
+      "learning_rate": 1.8513476986249547e-05,
+      "loss": 0.1116,
+      "step": 1973
+    },
+    {
+      "epoch": 1.033525405971713,
+      "loss_partial": 0.051025390625,
+      "step": 1973
+    },
+    {
+      "epoch": 1.033525405971713,
+      "loss_partial": 0.053466796875,
+      "step": 1973
+    },
+    {
+      "epoch": 1.034049240440021,
+      "grad_norm": 1.5969042778015137,
+      "learning_rate": 1.8511077835200605e-05,
+      "loss": 0.1111,
+      "step": 1974
+    },
+    {
+      "epoch": 1.034049240440021,
+      "loss_partial": 0.060791015625,
+      "step": 1974
+    },
+    {
+      "epoch": 1.034049240440021,
+      "loss_partial": 0.051513671875,
+      "step": 1974
+    },
+    {
+      "epoch": 1.034573074908329,
+      "grad_norm": 2.5898776054382324,
+      "learning_rate": 1.850867690538902e-05,
+      "loss": 0.1072,
+      "step": 1975
+    },
+    {
+      "epoch": 1.034573074908329,
+      "loss_partial": 0.056396484375,
+      "step": 1975
+    },
+    {
+      "epoch": 1.034573074908329,
+      "loss_partial": 0.052978515625,
+      "step": 1975
+    },
+    {
+      "epoch": 1.035096909376637,
+      "grad_norm": 1.4479007720947266,
+      "learning_rate": 1.8506274197316584e-05,
+      "loss": 0.1068,
+      "step": 1976
+    },
+    {
+      "epoch": 1.035096909376637,
+      "loss_partial": 0.052490234375,
+      "step": 1976
+    },
+    {
+      "epoch": 1.035096909376637,
+      "loss_partial": 0.060302734375,
+      "step": 1976
+    },
+    {
+      "epoch": 1.035620743844945,
+      "grad_norm": 2.2100722789764404,
+      "learning_rate": 1.850386971148544e-05,
+      "loss": 0.1079,
+      "step": 1977
+    },
+    {
+      "epoch": 1.035620743844945,
+      "loss_partial": 0.051025390625,
+      "step": 1977
+    },
+    {
+      "epoch": 1.035620743844945,
+      "loss_partial": 0.046630859375,
+      "step": 1977
+    },
+    {
+      "epoch": 1.036144578313253,
+      "grad_norm": 1.796998143196106,
+      "learning_rate": 1.8501463448398108e-05,
+      "loss": 0.1064,
+      "step": 1978
+    },
+    {
+      "epoch": 1.036144578313253,
+      "loss_partial": 0.050537109375,
+      "step": 1978
+    },
+    {
+      "epoch": 1.036144578313253,
+      "loss_partial": 0.05029296875,
+      "step": 1978
+    },
+    {
+      "epoch": 1.036668412781561,
+      "grad_norm": 2.0845189094543457,
+      "learning_rate": 1.8499055408557493e-05,
+      "loss": 0.1063,
+      "step": 1979
+    },
+    {
+      "epoch": 1.036668412781561,
+      "loss_partial": 0.05322265625,
+      "step": 1979
+    },
+    {
+      "epoch": 1.036668412781561,
+      "loss_partial": 0.048095703125,
+      "step": 1979
+    },
+    {
+      "epoch": 1.0371922472498691,
+      "grad_norm": 2.4544739723205566,
+      "learning_rate": 1.849664559246685e-05,
+      "loss": 0.1092,
+      "step": 1980
+    },
+    {
+      "epoch": 1.0371922472498691,
+      "loss_partial": 0.0517578125,
+      "step": 1980
+    },
+    {
+      "epoch": 1.0371922472498691,
+      "loss_partial": 0.04931640625,
+      "step": 1980
+    },
+    {
+      "epoch": 1.037716081718177,
+      "grad_norm": 1.4931844472885132,
+      "learning_rate": 1.8494234000629823e-05,
+      "loss": 0.1071,
+      "step": 1981
+    },
+    {
+      "epoch": 1.037716081718177,
+      "loss_partial": 0.05126953125,
+      "step": 1981
+    },
+    {
+      "epoch": 1.037716081718177,
+      "loss_partial": 0.061767578125,
+      "step": 1981
+    },
+    {
+      "epoch": 1.038239916186485,
+      "grad_norm": 2.280266523361206,
+      "learning_rate": 1.849182063355042e-05,
+      "loss": 0.1094,
+      "step": 1982
+    },
+    {
+      "epoch": 1.038239916186485,
+      "loss_partial": 0.04931640625,
+      "step": 1982
+    },
+    {
+      "epoch": 1.038239916186485,
+      "loss_partial": 0.053466796875,
+      "step": 1982
+    },
+    {
+      "epoch": 1.038763750654793,
+      "grad_norm": 2.2842464447021484,
+      "learning_rate": 1.8489405491733007e-05,
+      "loss": 0.1096,
+      "step": 1983
+    },
+    {
+      "epoch": 1.038763750654793,
+      "loss_partial": 0.05322265625,
+      "step": 1983
+    },
+    {
+      "epoch": 1.038763750654793,
+      "loss_partial": 0.06005859375,
+      "step": 1983
+    },
+    {
+      "epoch": 1.0392875851231012,
+      "grad_norm": 2.2694032192230225,
+      "learning_rate": 1.8486988575682346e-05,
+      "loss": 0.1107,
+      "step": 1984
+    },
+    {
+      "epoch": 1.0392875851231012,
+      "loss_partial": 0.05029296875,
+      "step": 1984
+    },
+    {
+      "epoch": 1.0392875851231012,
+      "loss_partial": 0.05712890625,
+      "step": 1984
+    },
+    {
+      "epoch": 1.0398114195914092,
+      "grad_norm": 2.193092107772827,
+      "learning_rate": 1.848456988590356e-05,
+      "loss": 0.108,
+      "step": 1985
+    },
+    {
+      "epoch": 1.0398114195914092,
+      "loss_partial": 0.0498046875,
+      "step": 1985
+    },
+    {
+      "epoch": 1.0398114195914092,
+      "loss_partial": 0.05859375,
+      "step": 1985
+    },
+    {
+      "epoch": 1.040335254059717,
+      "grad_norm": 2.273711681365967,
+      "learning_rate": 1.8482149422902132e-05,
+      "loss": 0.1031,
+      "step": 1986
+    },
+    {
+      "epoch": 1.040335254059717,
+      "loss_partial": 0.058349609375,
+      "step": 1986
+    },
+    {
+      "epoch": 1.040335254059717,
+      "loss_partial": 0.05126953125,
+      "step": 1986
+    },
+    {
+      "epoch": 1.0408590885280251,
+      "grad_norm": 2.3498806953430176,
+      "learning_rate": 1.8479727187183923e-05,
+      "loss": 0.1104,
+      "step": 1987
+    },
+    {
+      "epoch": 1.0408590885280251,
+      "loss_partial": 0.05419921875,
+      "step": 1987
+    },
+    {
+      "epoch": 1.0408590885280251,
+      "loss_partial": 0.052001953125,
+      "step": 1987
+    },
+    {
+      "epoch": 1.0413829229963332,
+      "grad_norm": 3.001042604446411,
+      "learning_rate": 1.8477303179255175e-05,
+      "loss": 0.1033,
+      "step": 1988
+    },
+    {
+      "epoch": 1.0413829229963332,
+      "loss_partial": 0.05712890625,
+      "step": 1988
+    },
+    {
+      "epoch": 1.0413829229963332,
+      "loss_partial": 0.0556640625,
+      "step": 1988
+    },
+    {
+      "epoch": 1.0419067574646412,
+      "grad_norm": 2.063737154006958,
+      "learning_rate": 1.847487739962248e-05,
+      "loss": 0.1084,
+      "step": 1989
+    },
+    {
+      "epoch": 1.0419067574646412,
+      "loss_partial": 0.054931640625,
+      "step": 1989
+    },
+    {
+      "epoch": 1.0419067574646412,
+      "loss_partial": 0.06494140625,
+      "step": 1989
+    },
+    {
+      "epoch": 1.0424305919329493,
+      "grad_norm": 2.149235725402832,
+      "learning_rate": 1.8472449848792818e-05,
+      "loss": 0.1111,
+      "step": 1990
+    },
+    {
+      "epoch": 1.0424305919329493,
+      "loss_partial": 0.05419921875,
+      "step": 1990
+    },
+    {
+      "epoch": 1.0424305919329493,
+      "loss_partial": 0.049560546875,
+      "step": 1990
+    },
+    {
+      "epoch": 1.0429544264012571,
+      "grad_norm": 2.0831031799316406,
+      "learning_rate": 1.8470020527273526e-05,
+      "loss": 0.105,
+      "step": 1991
+    },
+    {
+      "epoch": 1.0429544264012571,
+      "loss_partial": 0.049560546875,
+      "step": 1991
+    },
+    {
+      "epoch": 1.0429544264012571,
+      "loss_partial": 0.056640625,
+      "step": 1991
+    },
+    {
+      "epoch": 1.0434782608695652,
+      "grad_norm": 2.1881814002990723,
+      "learning_rate": 1.8467589435572323e-05,
+      "loss": 0.1069,
+      "step": 1992
+    },
+    {
+      "epoch": 1.0434782608695652,
+      "loss_partial": 0.052001953125,
+      "step": 1992
+    },
+    {
+      "epoch": 1.0434782608695652,
+      "loss_partial": 0.046630859375,
+      "step": 1992
+    },
+    {
+      "epoch": 1.0440020953378732,
+      "grad_norm": 1.6097415685653687,
+      "learning_rate": 1.8465156574197293e-05,
+      "loss": 0.1045,
+      "step": 1993
+    },
+    {
+      "epoch": 1.0440020953378732,
+      "loss_partial": 0.05810546875,
+      "step": 1993
+    },
+    {
+      "epoch": 1.0440020953378732,
+      "loss_partial": 0.061279296875,
+      "step": 1993
+    },
+    {
+      "epoch": 1.0445259298061813,
+      "grad_norm": 1.4588899612426758,
+      "learning_rate": 1.8462721943656885e-05,
+      "loss": 0.1076,
+      "step": 1994
+    },
+    {
+      "epoch": 1.0445259298061813,
+      "loss_partial": 0.04931640625,
+      "step": 1994
+    },
+    {
+      "epoch": 1.0445259298061813,
+      "loss_partial": 0.060546875,
+      "step": 1994
+    },
+    {
+      "epoch": 1.0450497642744894,
+      "grad_norm": 1.5345954895019531,
+      "learning_rate": 1.846028554445992e-05,
+      "loss": 0.108,
+      "step": 1995
+    },
+    {
+      "epoch": 1.0450497642744894,
+      "loss_partial": 0.056396484375,
+      "step": 1995
+    },
+    {
+      "epoch": 1.0450497642744894,
+      "loss_partial": 0.0498046875,
+      "step": 1995
+    },
+    {
+      "epoch": 1.0455735987427972,
+      "grad_norm": 2.1383249759674072,
+      "learning_rate": 1.8457847377115595e-05,
+      "loss": 0.111,
+      "step": 1996
+    },
+    {
+      "epoch": 1.0455735987427972,
+      "loss_partial": 0.052734375,
+      "step": 1996
+    },
+    {
+      "epoch": 1.0455735987427972,
+      "loss_partial": 0.059326171875,
+      "step": 1996
+    },
+    {
+      "epoch": 1.0460974332111053,
+      "grad_norm": 2.170590400695801,
+      "learning_rate": 1.8455407442133467e-05,
+      "loss": 0.1094,
+      "step": 1997
+    },
+    {
+      "epoch": 1.0460974332111053,
+      "loss_partial": 0.07275390625,
+      "step": 1997
+    },
+    {
+      "epoch": 1.0460974332111053,
+      "loss_partial": 0.061279296875,
+      "step": 1997
+    },
+    {
+      "epoch": 1.0466212676794133,
+      "grad_norm": 2.007880210876465,
+      "learning_rate": 1.8452965740023476e-05,
+      "loss": 0.1096,
+      "step": 1998
+    },
+    {
+      "epoch": 1.0466212676794133,
+      "loss_partial": 0.052734375,
+      "step": 1998
+    },
+    {
+      "epoch": 1.0466212676794133,
+      "loss_partial": 0.05078125,
+      "step": 1998
+    },
+    {
+      "epoch": 1.0471451021477214,
+      "grad_norm": 1.4512600898742676,
+      "learning_rate": 1.845052227129591e-05,
+      "loss": 0.1094,
+      "step": 1999
+    },
+    {
+      "epoch": 1.0471451021477214,
+      "loss_partial": 0.042236328125,
+      "step": 1999
+    },
+    {
+      "epoch": 1.0471451021477214,
+      "loss_partial": 0.05908203125,
+      "step": 1999
+    },
+    {
+      "epoch": 1.0476689366160294,
+      "grad_norm": 2.1997969150543213,
+      "learning_rate": 1.844807703646145e-05,
+      "loss": 0.1096,
+      "step": 2000
+    },
+    {
+      "epoch": 1.0476689366160294,
+      "loss_partial": 0.051025390625,
+      "step": 2000
+    },
+    {
+      "epoch": 1.0476689366160294,
+      "loss_partial": 0.052490234375,
+      "step": 2000
+    },
+    {
+      "epoch": 1.0481927710843373,
+      "grad_norm": 1.5814090967178345,
+      "learning_rate": 1.844563003603113e-05,
+      "loss": 0.1071,
+      "step": 2001
+    },
+    {
+      "epoch": 1.0481927710843373,
+      "loss_partial": 0.04638671875,
+      "step": 2001
+    },
+    {
+      "epoch": 1.0481927710843373,
+      "loss_partial": 0.0546875,
+      "step": 2001
+    },
+    {
+      "epoch": 1.0487166055526453,
+      "grad_norm": 1.7493927478790283,
+      "learning_rate": 1.8443181270516358e-05,
+      "loss": 0.109,
+      "step": 2002
+    },
+    {
+      "epoch": 1.0487166055526453,
+      "loss_partial": 0.045166015625,
+      "step": 2002
+    },
+    {
+      "epoch": 1.0487166055526453,
+      "loss_partial": 0.060791015625,
+      "step": 2002
+    },
+    {
+      "epoch": 1.0492404400209534,
+      "grad_norm": 2.249873161315918,
+      "learning_rate": 1.8440730740428914e-05,
+      "loss": 0.1071,
+      "step": 2003
+    },
+    {
+      "epoch": 1.0492404400209534,
+      "loss_partial": 0.059326171875,
+      "step": 2003
+    },
+    {
+      "epoch": 1.0492404400209534,
+      "loss_partial": 0.05419921875,
+      "step": 2003
+    },
+    {
+      "epoch": 1.0497642744892615,
+      "grad_norm": 2.032864570617676,
+      "learning_rate": 1.843827844628094e-05,
+      "loss": 0.1099,
+      "step": 2004
+    },
+    {
+      "epoch": 1.0497642744892615,
+      "loss_partial": 0.055419921875,
+      "step": 2004
+    },
+    {
+      "epoch": 1.0497642744892615,
+      "loss_partial": 0.048583984375,
+      "step": 2004
+    },
+    {
+      "epoch": 1.0502881089575693,
+      "grad_norm": 1.4331167936325073,
+      "learning_rate": 1.843582438858495e-05,
+      "loss": 0.1106,
+      "step": 2005
+    },
+    {
+      "epoch": 1.0502881089575693,
+      "loss_partial": 0.054443359375,
+      "step": 2005
+    },
+    {
+      "epoch": 1.0502881089575693,
+      "loss_partial": 0.0595703125,
+      "step": 2005
+    },
+    {
+      "epoch": 1.0508119434258774,
+      "grad_norm": 2.41422963142395,
+      "learning_rate": 1.8433368567853827e-05,
+      "loss": 0.1084,
+      "step": 2006
+    },
+    {
+      "epoch": 1.0508119434258774,
+      "loss_partial": 0.0478515625,
+      "step": 2006
+    },
+    {
+      "epoch": 1.0508119434258774,
+      "loss_partial": 0.051513671875,
+      "step": 2006
+    },
+    {
+      "epoch": 1.0513357778941854,
+      "grad_norm": 2.780153274536133,
+      "learning_rate": 1.8430910984600826e-05,
+      "loss": 0.1076,
+      "step": 2007
+    },
+    {
+      "epoch": 1.0513357778941854,
+      "loss_partial": 0.046875,
+      "step": 2007
+    },
+    {
+      "epoch": 1.0513357778941854,
+      "loss_partial": 0.057861328125,
+      "step": 2007
+    },
+    {
+      "epoch": 1.0518596123624935,
+      "grad_norm": 2.1829450130462646,
+      "learning_rate": 1.8428451639339565e-05,
+      "loss": 0.1071,
+      "step": 2008
+    },
+    {
+      "epoch": 1.0518596123624935,
+      "loss_partial": 0.049560546875,
+      "step": 2008
+    },
+    {
+      "epoch": 1.0518596123624935,
+      "loss_partial": 0.056884765625,
+      "step": 2008
+    },
+    {
+      "epoch": 1.0523834468308015,
+      "grad_norm": 2.107097864151001,
+      "learning_rate": 1.842599053258403e-05,
+      "loss": 0.1111,
+      "step": 2009
+    },
+    {
+      "epoch": 1.0523834468308015,
+      "loss_partial": 0.042724609375,
+      "step": 2009
+    },
+    {
+      "epoch": 1.0523834468308015,
+      "loss_partial": 0.050537109375,
+      "step": 2009
+    },
+    {
+      "epoch": 1.0529072812991094,
+      "grad_norm": 2.4606943130493164,
+      "learning_rate": 1.842352766484858e-05,
+      "loss": 0.1083,
+      "step": 2010
+    },
+    {
+      "epoch": 1.0529072812991094,
+      "loss_partial": 0.050048828125,
+      "step": 2010
+    },
+    {
+      "epoch": 1.0529072812991094,
+      "loss_partial": 0.061767578125,
+      "step": 2010
+    },
+    {
+      "epoch": 1.0534311157674174,
+      "grad_norm": 2.1042346954345703,
+      "learning_rate": 1.8421063036647935e-05,
+      "loss": 0.1091,
+      "step": 2011
+    },
+    {
+      "epoch": 1.0534311157674174,
+      "loss_partial": 0.054931640625,
+      "step": 2011
+    },
+    {
+      "epoch": 1.0534311157674174,
+      "loss_partial": 0.04833984375,
+      "step": 2011
+    },
+    {
+      "epoch": 1.0539549502357255,
+      "grad_norm": 3.2033979892730713,
+      "learning_rate": 1.8418596648497192e-05,
+      "loss": 0.1065,
+      "step": 2012
+    },
+    {
+      "epoch": 1.0539549502357255,
+      "loss_partial": 0.0595703125,
+      "step": 2012
+    },
+    {
+      "epoch": 1.0539549502357255,
+      "loss_partial": 0.053466796875,
+      "step": 2012
+    },
+    {
+      "epoch": 1.0544787847040336,
+      "grad_norm": 1.7153382301330566,
+      "learning_rate": 1.8416128500911807e-05,
+      "loss": 0.1106,
+      "step": 2013
+    },
+    {
+      "epoch": 1.0544787847040336,
+      "loss_partial": 0.05078125,
+      "step": 2013
+    },
+    {
+      "epoch": 1.0544787847040336,
+      "loss_partial": 0.056396484375,
+      "step": 2013
+    },
+    {
+      "epoch": 1.0550026191723416,
+      "grad_norm": 2.415418863296509,
+      "learning_rate": 1.841365859440761e-05,
+      "loss": 0.1091,
+      "step": 2014
+    },
+    {
+      "epoch": 1.0550026191723416,
+      "loss_partial": 0.049072265625,
+      "step": 2014
+    },
+    {
+      "epoch": 1.0550026191723416,
+      "loss_partial": 0.051513671875,
+      "step": 2014
+    },
+    {
+      "epoch": 1.0555264536406495,
+      "grad_norm": 2.1495258808135986,
+      "learning_rate": 1.8411186929500796e-05,
+      "loss": 0.1067,
+      "step": 2015
+    },
+    {
+      "epoch": 1.0555264536406495,
+      "loss_partial": 0.06591796875,
+      "step": 2015
+    },
+    {
+      "epoch": 1.0555264536406495,
+      "loss_partial": 0.05419921875,
+      "step": 2015
+    },
+    {
+      "epoch": 1.0560502881089575,
+      "grad_norm": 2.006580352783203,
+      "learning_rate": 1.8408713506707924e-05,
+      "loss": 0.1069,
+      "step": 2016
+    },
+    {
+      "epoch": 1.0560502881089575,
+      "loss_partial": 0.04736328125,
+      "step": 2016
+    },
+    {
+      "epoch": 1.0560502881089575,
+      "loss_partial": 0.048828125,
+      "step": 2016
+    },
+    {
+      "epoch": 1.0565741225772656,
+      "grad_norm": 2.4095280170440674,
+      "learning_rate": 1.840623832654593e-05,
+      "loss": 0.1096,
+      "step": 2017
+    },
+    {
+      "epoch": 1.0565741225772656,
+      "loss_partial": 0.048828125,
+      "step": 2017
+    },
+    {
+      "epoch": 1.0565741225772656,
+      "loss_partial": 0.04833984375,
+      "step": 2017
+    },
+    {
+      "epoch": 1.0570979570455736,
+      "grad_norm": 2.7967119216918945,
+      "learning_rate": 1.8403761389532107e-05,
+      "loss": 0.107,
+      "step": 2018
+    },
+    {
+      "epoch": 1.0570979570455736,
+      "loss_partial": 0.044921875,
+      "step": 2018
+    },
+    {
+      "epoch": 1.0570979570455736,
+      "loss_partial": 0.04833984375,
+      "step": 2018
+    },
+    {
+      "epoch": 1.0576217915138817,
+      "grad_norm": 2.7510950565338135,
+      "learning_rate": 1.840128269618412e-05,
+      "loss": 0.1102,
+      "step": 2019
+    },
+    {
+      "epoch": 1.0576217915138817,
+      "loss_partial": 0.05615234375,
+      "step": 2019
+    },
+    {
+      "epoch": 1.0576217915138817,
+      "loss_partial": 0.053955078125,
+      "step": 2019
+    },
+    {
+      "epoch": 1.0581456259821895,
+      "grad_norm": 1.9318242073059082,
+      "learning_rate": 1.839880224702e-05,
+      "loss": 0.1087,
+      "step": 2020
+    },
+    {
+      "epoch": 1.0581456259821895,
+      "loss_partial": 0.052490234375,
+      "step": 2020
+    },
+    {
+      "epoch": 1.0581456259821895,
+      "loss_partial": 0.051513671875,
+      "step": 2020
+    },
+    {
+      "epoch": 1.0586694604504976,
+      "grad_norm": 2.34673810005188,
+      "learning_rate": 1.8396320042558154e-05,
+      "loss": 0.1071,
+      "step": 2021
+    },
+    {
+      "epoch": 1.0586694604504976,
+      "loss_partial": 0.0478515625,
+      "step": 2021
+    },
+    {
+      "epoch": 1.0586694604504976,
+      "loss_partial": 0.0517578125,
+      "step": 2021
+    },
+    {
+      "epoch": 1.0591932949188057,
+      "grad_norm": 2.5286240577697754,
+      "learning_rate": 1.8393836083317337e-05,
+      "loss": 0.1064,
+      "step": 2022
+    },
+    {
+      "epoch": 1.0591932949188057,
+      "loss_partial": 0.04931640625,
+      "step": 2022
+    },
+    {
+      "epoch": 1.0591932949188057,
+      "loss_partial": 0.047607421875,
+      "step": 2022
+    },
+    {
+      "epoch": 1.0597171293871137,
+      "grad_norm": 2.2751708030700684,
+      "learning_rate": 1.8391350369816683e-05,
+      "loss": 0.1057,
+      "step": 2023
+    },
+    {
+      "epoch": 1.0597171293871137,
+      "loss_partial": 0.04833984375,
+      "step": 2023
+    },
+    {
+      "epoch": 1.0597171293871137,
+      "loss_partial": 0.049560546875,
+      "step": 2023
+    },
+    {
+      "epoch": 1.0602409638554218,
+      "grad_norm": 2.3614277839660645,
+      "learning_rate": 1.838886290257569e-05,
+      "loss": 0.1043,
+      "step": 2024
+    },
+    {
+      "epoch": 1.0602409638554218,
+      "loss_partial": 0.0498046875,
+      "step": 2024
+    },
+    {
+      "epoch": 1.0602409638554218,
+      "loss_partial": 0.05126953125,
+      "step": 2024
+    },
+    {
+      "epoch": 1.0607647983237296,
+      "grad_norm": 2.2964603900909424,
+      "learning_rate": 1.838637368211423e-05,
+      "loss": 0.1064,
+      "step": 2025
+    },
+    {
+      "epoch": 1.0607647983237296,
+      "loss_partial": 0.057373046875,
+      "step": 2025
+    },
+    {
+      "epoch": 1.0607647983237296,
+      "loss_partial": 0.057861328125,
+      "step": 2025
+    },
+    {
+      "epoch": 1.0612886327920377,
+      "grad_norm": 1.902483582496643,
+      "learning_rate": 1.838388270895253e-05,
+      "loss": 0.1074,
+      "step": 2026
+    },
+    {
+      "epoch": 1.0612886327920377,
+      "loss_partial": 0.04931640625,
+      "step": 2026
+    },
+    {
+      "epoch": 1.0612886327920377,
+      "loss_partial": 0.053466796875,
+      "step": 2026
+    },
+    {
+      "epoch": 1.0618124672603457,
+      "grad_norm": 1.8079845905303955,
+      "learning_rate": 1.8381389983611186e-05,
+      "loss": 0.11,
+      "step": 2027
+    },
+    {
+      "epoch": 1.0618124672603457,
+      "loss_partial": 0.05029296875,
+      "step": 2027
+    },
+    {
+      "epoch": 1.0618124672603457,
+      "loss_partial": 0.05419921875,
+      "step": 2027
+    },
+    {
+      "epoch": 1.0623363017286538,
+      "grad_norm": 2.0115585327148438,
+      "learning_rate": 1.8378895506611166e-05,
+      "loss": 0.111,
+      "step": 2028
+    },
+    {
+      "epoch": 1.0623363017286538,
+      "loss_partial": 0.045654296875,
+      "step": 2028
+    },
+    {
+      "epoch": 1.0623363017286538,
+      "loss_partial": 0.050048828125,
+      "step": 2028
+    },
+    {
+      "epoch": 1.0628601361969618,
+      "grad_norm": 2.291337728500366,
+      "learning_rate": 1.8376399278473794e-05,
+      "loss": 0.1055,
+      "step": 2029
+    },
+    {
+      "epoch": 1.0628601361969618,
+      "loss_partial": 0.053466796875,
+      "step": 2029
+    },
+    {
+      "epoch": 1.0628601361969618,
+      "loss_partial": 0.0458984375,
+      "step": 2029
+    },
+    {
+      "epoch": 1.0633839706652697,
+      "grad_norm": 1.8899606466293335,
+      "learning_rate": 1.837390129972078e-05,
+      "loss": 0.1061,
+      "step": 2030
+    },
+    {
+      "epoch": 1.0633839706652697,
+      "loss_partial": 0.0439453125,
+      "step": 2030
+    },
+    {
+      "epoch": 1.0633839706652697,
+      "loss_partial": 0.051513671875,
+      "step": 2030
+    },
+    {
+      "epoch": 1.0639078051335777,
+      "grad_norm": 1.8553169965744019,
+      "learning_rate": 1.8371401570874166e-05,
+      "loss": 0.1098,
+      "step": 2031
+    },
+    {
+      "epoch": 1.0639078051335777,
+      "loss_partial": 0.056640625,
+      "step": 2031
+    },
+    {
+      "epoch": 1.0639078051335777,
+      "loss_partial": 0.053955078125,
+      "step": 2031
+    },
+    {
+      "epoch": 1.0644316396018858,
+      "grad_norm": 1.903170108795166,
+      "learning_rate": 1.83689000924564e-05,
+      "loss": 0.1083,
+      "step": 2032
+    },
+    {
+      "epoch": 1.0644316396018858,
+      "loss_partial": 0.04833984375,
+      "step": 2032
+    },
+    {
+      "epoch": 1.0644316396018858,
+      "loss_partial": 0.05712890625,
+      "step": 2032
+    },
+    {
+      "epoch": 1.0649554740701939,
+      "grad_norm": 1.901611328125,
+      "learning_rate": 1.8366396864990262e-05,
+      "loss": 0.1083,
+      "step": 2033
+    },
+    {
+      "epoch": 1.0649554740701939,
+      "loss_partial": 0.06201171875,
+      "step": 2033
+    },
+    {
+      "epoch": 1.0649554740701939,
+      "loss_partial": 0.05029296875,
+      "step": 2033
+    },
+    {
+      "epoch": 1.065479308538502,
+      "grad_norm": 1.4390127658843994,
+      "learning_rate": 1.8363891888998914e-05,
+      "loss": 0.1089,
+      "step": 2034
+    },
+    {
+      "epoch": 1.065479308538502,
+      "loss_partial": 0.050537109375,
+      "step": 2034
+    },
+    {
+      "epoch": 1.065479308538502,
+      "loss_partial": 0.046142578125,
+      "step": 2034
+    },
+    {
+      "epoch": 1.0660031430068098,
+      "grad_norm": 1.7281957864761353,
+      "learning_rate": 1.8361385165005886e-05,
+      "loss": 0.1093,
+      "step": 2035
+    },
+    {
+      "epoch": 1.0660031430068098,
+      "loss_partial": 0.05322265625,
+      "step": 2035
+    },
+    {
+      "epoch": 1.0660031430068098,
+      "loss_partial": 0.058349609375,
+      "step": 2035
+    },
+    {
+      "epoch": 1.0665269774751178,
+      "grad_norm": 1.807129979133606,
+      "learning_rate": 1.8358876693535062e-05,
+      "loss": 0.1105,
+      "step": 2036
+    },
+    {
+      "epoch": 1.0665269774751178,
+      "loss_partial": 0.0498046875,
+      "step": 2036
+    },
+    {
+      "epoch": 1.0665269774751178,
+      "loss_partial": 0.0595703125,
+      "step": 2036
+    },
+    {
+      "epoch": 1.0670508119434259,
+      "grad_norm": 1.9864380359649658,
+      "learning_rate": 1.8356366475110697e-05,
+      "loss": 0.1083,
+      "step": 2037
+    },
+    {
+      "epoch": 1.0670508119434259,
+      "loss_partial": 0.053466796875,
+      "step": 2037
+    },
+    {
+      "epoch": 1.0670508119434259,
+      "loss_partial": 0.044189453125,
+      "step": 2037
+    },
+    {
+      "epoch": 1.067574646411734,
+      "grad_norm": 1.7908475399017334,
+      "learning_rate": 1.835385451025742e-05,
+      "loss": 0.1055,
+      "step": 2038
+    },
+    {
+      "epoch": 1.067574646411734,
+      "loss_partial": 0.05419921875,
+      "step": 2038
+    },
+    {
+      "epoch": 1.067574646411734,
+      "loss_partial": 0.053466796875,
+      "step": 2038
+    },
+    {
+      "epoch": 1.068098480880042,
+      "grad_norm": 1.7487932443618774,
+      "learning_rate": 1.83513407995002e-05,
+      "loss": 0.1088,
+      "step": 2039
+    },
+    {
+      "epoch": 1.068098480880042,
+      "loss_partial": 0.05517578125,
+      "step": 2039
+    },
+    {
+      "epoch": 1.068098480880042,
+      "loss_partial": 0.052734375,
+      "step": 2039
+    },
+    {
+      "epoch": 1.0686223153483498,
+      "grad_norm": 1.9031282663345337,
+      "learning_rate": 1.83488253433644e-05,
+      "loss": 0.1091,
+      "step": 2040
+    },
+    {
+      "epoch": 1.0686223153483498,
+      "loss_partial": 0.060302734375,
+      "step": 2040
+    },
+    {
+      "epoch": 1.0686223153483498,
+      "loss_partial": 0.05419921875,
+      "step": 2040
+    },
+    {
+      "epoch": 1.069146149816658,
+      "grad_norm": 1.801949143409729,
+      "learning_rate": 1.834630814237573e-05,
+      "loss": 0.1058,
+      "step": 2041
+    },
+    {
+      "epoch": 1.069146149816658,
+      "loss_partial": 0.045654296875,
+      "step": 2041
+    },
+    {
+      "epoch": 1.069146149816658,
+      "loss_partial": 0.052490234375,
+      "step": 2041
+    },
+    {
+      "epoch": 1.069669984284966,
+      "grad_norm": 2.0468380451202393,
+      "learning_rate": 1.8343789197060272e-05,
+      "loss": 0.1056,
+      "step": 2042
+    },
+    {
+      "epoch": 1.069669984284966,
+      "loss_partial": 0.05517578125,
+      "step": 2042
+    },
+    {
+      "epoch": 1.069669984284966,
+      "loss_partial": 0.060546875,
+      "step": 2042
+    },
+    {
+      "epoch": 1.070193818753274,
+      "grad_norm": 2.394196033477783,
+      "learning_rate": 1.8341268507944468e-05,
+      "loss": 0.1105,
+      "step": 2043
+    },
+    {
+      "epoch": 1.070193818753274,
+      "loss_partial": 0.050537109375,
+      "step": 2043
+    },
+    {
+      "epoch": 1.070193818753274,
+      "loss_partial": 0.0458984375,
+      "step": 2043
+    },
+    {
+      "epoch": 1.070717653221582,
+      "grad_norm": 1.7609448432922363,
+      "learning_rate": 1.8338746075555126e-05,
+      "loss": 0.1079,
+      "step": 2044
+    },
+    {
+      "epoch": 1.070717653221582,
+      "loss_partial": 0.056884765625,
+      "step": 2044
+    },
+    {
+      "epoch": 1.070717653221582,
+      "loss_partial": 0.046875,
+      "step": 2044
+    },
+    {
+      "epoch": 1.07124148768989,
+      "grad_norm": 2.6458165645599365,
+      "learning_rate": 1.833622190041942e-05,
+      "loss": 0.1092,
+      "step": 2045
+    },
+    {
+      "epoch": 1.07124148768989,
+      "loss_partial": 0.059326171875,
+      "step": 2045
+    },
+    {
+      "epoch": 1.07124148768989,
+      "loss_partial": 0.07080078125,
+      "step": 2045
+    },
+    {
+      "epoch": 1.071765322158198,
+      "grad_norm": 2.84541916847229,
+      "learning_rate": 1.8333695983064887e-05,
+      "loss": 0.1113,
+      "step": 2046
+    },
+    {
+      "epoch": 1.071765322158198,
+      "loss_partial": 0.041748046875,
+      "step": 2046
+    },
+    {
+      "epoch": 1.071765322158198,
+      "loss_partial": 0.06298828125,
+      "step": 2046
+    },
+    {
+      "epoch": 1.072289156626506,
+      "grad_norm": 1.451730728149414,
+      "learning_rate": 1.8331168324019424e-05,
+      "loss": 0.1081,
+      "step": 2047
+    },
+    {
+      "epoch": 1.072289156626506,
+      "loss_partial": 0.056396484375,
+      "step": 2047
+    },
+    {
+      "epoch": 1.072289156626506,
+      "loss_partial": 0.05615234375,
+      "step": 2047
+    },
+    {
+      "epoch": 1.072812991094814,
+      "grad_norm": 2.333106756210327,
+      "learning_rate": 1.83286389238113e-05,
+      "loss": 0.1072,
+      "step": 2048
+    },
+    {
+      "epoch": 1.072812991094814,
+      "loss_partial": 0.05419921875,
+      "step": 2048
+    },
+    {
+      "epoch": 1.072812991094814,
+      "loss_partial": 0.04541015625,
+      "step": 2048
+    },
+    {
+      "epoch": 1.0733368255631222,
+      "grad_norm": 3.0156409740448,
+      "learning_rate": 1.8326107782969148e-05,
+      "loss": 0.1049,
+      "step": 2049
+    },
+    {
+      "epoch": 1.0733368255631222,
+      "loss_partial": 0.04345703125,
+      "step": 2049
+    },
+    {
+      "epoch": 1.0733368255631222,
+      "loss_partial": 0.068359375,
+      "step": 2049
+    },
+    {
+      "epoch": 1.07386066003143,
+      "grad_norm": 1.503769040107727,
+      "learning_rate": 1.8323574902021954e-05,
+      "loss": 0.1108,
+      "step": 2050
+    },
+    {
+      "epoch": 1.07386066003143,
+      "loss_partial": 0.060791015625,
+      "step": 2050
+    },
+    {
+      "epoch": 1.07386066003143,
+      "loss_partial": 0.047119140625,
+      "step": 2050
+    },
+    {
+      "epoch": 1.074384494499738,
+      "grad_norm": 2.3072938919067383,
+      "learning_rate": 1.8321040281499076e-05,
+      "loss": 0.1068,
+      "step": 2051
+    },
+    {
+      "epoch": 1.074384494499738,
+      "loss_partial": 0.047119140625,
+      "step": 2051
+    },
+    {
+      "epoch": 1.074384494499738,
+      "loss_partial": 0.04931640625,
+      "step": 2051
+    },
+    {
+      "epoch": 1.0749083289680461,
+      "grad_norm": 1.6696970462799072,
+      "learning_rate": 1.831850392193024e-05,
+      "loss": 0.1035,
+      "step": 2052
+    },
+    {
+      "epoch": 1.0749083289680461,
+      "loss_partial": 0.04443359375,
+      "step": 2052
+    },
+    {
+      "epoch": 1.0749083289680461,
+      "loss_partial": 0.05517578125,
+      "step": 2052
+    },
+    {
+      "epoch": 1.0754321634363542,
+      "grad_norm": 2.316779851913452,
+      "learning_rate": 1.8315965823845518e-05,
+      "loss": 0.1054,
+      "step": 2053
+    },
+    {
+      "epoch": 1.0754321634363542,
+      "loss_partial": 0.045166015625,
+      "step": 2053
+    },
+    {
+      "epoch": 1.0754321634363542,
+      "loss_partial": 0.04638671875,
+      "step": 2053
+    },
+    {
+      "epoch": 1.075955997904662,
+      "grad_norm": 1.4219074249267578,
+      "learning_rate": 1.8313425987775367e-05,
+      "loss": 0.1104,
+      "step": 2054
+    },
+    {
+      "epoch": 1.075955997904662,
+      "loss_partial": 0.044921875,
+      "step": 2054
+    },
+    {
+      "epoch": 1.075955997904662,
+      "loss_partial": 0.051025390625,
+      "step": 2054
+    },
+    {
+      "epoch": 1.07647983237297,
+      "grad_norm": 3.422001600265503,
+      "learning_rate": 1.831088441425059e-05,
+      "loss": 0.1074,
+      "step": 2055
+    },
+    {
+      "epoch": 1.07647983237297,
+      "loss_partial": 0.060791015625,
+      "step": 2055
+    },
+    {
+      "epoch": 1.07647983237297,
+      "loss_partial": 0.05322265625,
+      "step": 2055
+    },
+    {
+      "epoch": 1.0770036668412781,
+      "grad_norm": 1.7003962993621826,
+      "learning_rate": 1.8308341103802364e-05,
+      "loss": 0.1082,
+      "step": 2056
+    },
+    {
+      "epoch": 1.0770036668412781,
+      "loss_partial": 0.05810546875,
+      "step": 2056
+    },
+    {
+      "epoch": 1.0770036668412781,
+      "loss_partial": 0.0517578125,
+      "step": 2056
+    },
+    {
+      "epoch": 1.0775275013095862,
+      "grad_norm": 1.934804916381836,
+      "learning_rate": 1.830579605696223e-05,
+      "loss": 0.106,
+      "step": 2057
+    },
+    {
+      "epoch": 1.0775275013095862,
+      "loss_partial": 0.058349609375,
+      "step": 2057
+    },
+    {
+      "epoch": 1.0775275013095862,
+      "loss_partial": 0.04931640625,
+      "step": 2057
+    },
+    {
+      "epoch": 1.0780513357778942,
+      "grad_norm": 2.002735137939453,
+      "learning_rate": 1.8303249274262077e-05,
+      "loss": 0.1069,
+      "step": 2058
+    },
+    {
+      "epoch": 1.0780513357778942,
+      "loss_partial": 0.055908203125,
+      "step": 2058
+    },
+    {
+      "epoch": 1.0780513357778942,
+      "loss_partial": 0.052001953125,
+      "step": 2058
+    },
+    {
+      "epoch": 1.0785751702462023,
+      "grad_norm": 1.6808267831802368,
+      "learning_rate": 1.8300700756234172e-05,
+      "loss": 0.1068,
+      "step": 2059
+    },
+    {
+      "epoch": 1.0785751702462023,
+      "loss_partial": 0.0546875,
+      "step": 2059
+    },
+    {
+      "epoch": 1.0785751702462023,
+      "loss_partial": 0.0537109375,
+      "step": 2059
+    },
+    {
+      "epoch": 1.0790990047145101,
+      "grad_norm": 1.8601694107055664,
+      "learning_rate": 1.8298150503411136e-05,
+      "loss": 0.1107,
+      "step": 2060
+    },
+    {
+      "epoch": 1.0790990047145101,
+      "loss_partial": 0.0712890625,
+      "step": 2060
+    },
+    {
+      "epoch": 1.0790990047145101,
+      "loss_partial": 0.062255859375,
+      "step": 2060
+    },
+    {
+      "epoch": 1.0796228391828182,
+      "grad_norm": 1.5255259275436401,
+      "learning_rate": 1.8295598516325964e-05,
+      "loss": 0.1108,
+      "step": 2061
+    },
+    {
+      "epoch": 1.0796228391828182,
+      "loss_partial": 0.0546875,
+      "step": 2061
+    },
+    {
+      "epoch": 1.0796228391828182,
+      "loss_partial": 0.054443359375,
+      "step": 2061
+    },
+    {
+      "epoch": 1.0801466736511263,
+      "grad_norm": 1.9409679174423218,
+      "learning_rate": 1.8293044795512e-05,
+      "loss": 0.1099,
+      "step": 2062
+    },
+    {
+      "epoch": 1.0801466736511263,
+      "loss_partial": 0.055419921875,
+      "step": 2062
+    },
+    {
+      "epoch": 1.0801466736511263,
+      "loss_partial": 0.048828125,
+      "step": 2062
+    },
+    {
+      "epoch": 1.0806705081194343,
+      "grad_norm": 1.6678564548492432,
+      "learning_rate": 1.8290489341502953e-05,
+      "loss": 0.1028,
+      "step": 2063
+    },
+    {
+      "epoch": 1.0806705081194343,
+      "loss_partial": 0.0546875,
+      "step": 2063
+    },
+    {
+      "epoch": 1.0806705081194343,
+      "loss_partial": 0.0625,
+      "step": 2063
+    },
+    {
+      "epoch": 1.0811943425877422,
+      "grad_norm": 2.032353639602661,
+      "learning_rate": 1.8287932154832902e-05,
+      "loss": 0.1071,
+      "step": 2064
+    },
+    {
+      "epoch": 1.0811943425877422,
+      "loss_partial": 0.060302734375,
+      "step": 2064
+    },
+    {
+      "epoch": 1.0811943425877422,
+      "loss_partial": 0.052734375,
+      "step": 2064
+    },
+    {
+      "epoch": 1.0817181770560502,
+      "grad_norm": 2.203803777694702,
+      "learning_rate": 1.828537323603628e-05,
+      "loss": 0.1105,
+      "step": 2065
+    },
+    {
+      "epoch": 1.0817181770560502,
+      "loss_partial": 0.056884765625,
+      "step": 2065
+    },
+    {
+      "epoch": 1.0817181770560502,
+      "loss_partial": 0.05029296875,
+      "step": 2065
+    },
+    {
+      "epoch": 1.0822420115243583,
+      "grad_norm": 1.8826078176498413,
+      "learning_rate": 1.828281258564789e-05,
+      "loss": 0.1042,
+      "step": 2066
+    },
+    {
+      "epoch": 1.0822420115243583,
+      "loss_partial": 0.049560546875,
+      "step": 2066
+    },
+    {
+      "epoch": 1.0822420115243583,
+      "loss_partial": 0.060791015625,
+      "step": 2066
+    },
+    {
+      "epoch": 1.0827658459926663,
+      "grad_norm": 2.0256686210632324,
+      "learning_rate": 1.8280250204202888e-05,
+      "loss": 0.1091,
+      "step": 2067
+    },
+    {
+      "epoch": 1.0827658459926663,
+      "loss_partial": 0.045654296875,
+      "step": 2067
+    },
+    {
+      "epoch": 1.0827658459926663,
+      "loss_partial": 0.0595703125,
+      "step": 2067
+    },
+    {
+      "epoch": 1.0832896804609744,
+      "grad_norm": 2.132505416870117,
+      "learning_rate": 1.8277686092236793e-05,
+      "loss": 0.1084,
+      "step": 2068
+    },
+    {
+      "epoch": 1.0832896804609744,
+      "loss_partial": 0.04443359375,
+      "step": 2068
+    },
+    {
+      "epoch": 1.0832896804609744,
+      "loss_partial": 0.04638671875,
+      "step": 2068
+    },
+    {
+      "epoch": 1.0838135149292825,
+      "grad_norm": 1.7343984842300415,
+      "learning_rate": 1.82751202502855e-05,
+      "loss": 0.1062,
+      "step": 2069
+    },
+    {
+      "epoch": 1.0838135149292825,
+      "loss_partial": 0.060791015625,
+      "step": 2069
+    },
+    {
+      "epoch": 1.0838135149292825,
+      "loss_partial": 0.054443359375,
+      "step": 2069
+    },
+    {
+      "epoch": 1.0843373493975903,
+      "grad_norm": 1.9925252199172974,
+      "learning_rate": 1.827255267888524e-05,
+      "loss": 0.112,
+      "step": 2070
+    },
+    {
+      "epoch": 1.0843373493975903,
+      "loss_partial": 0.054931640625,
+      "step": 2070
+    },
+    {
+      "epoch": 1.0843373493975903,
+      "loss_partial": 0.05126953125,
+      "step": 2070
+    },
+    {
+      "epoch": 1.0848611838658984,
+      "grad_norm": 1.8451999425888062,
+      "learning_rate": 1.8269983378572627e-05,
+      "loss": 0.108,
+      "step": 2071
+    },
+    {
+      "epoch": 1.0848611838658984,
+      "loss_partial": 0.05810546875,
+      "step": 2071
+    },
+    {
+      "epoch": 1.0848611838658984,
+      "loss_partial": 0.0556640625,
+      "step": 2071
+    },
+    {
+      "epoch": 1.0853850183342064,
+      "grad_norm": 2.1209263801574707,
+      "learning_rate": 1.826741234988463e-05,
+      "loss": 0.1073,
+      "step": 2072
+    },
+    {
+      "epoch": 1.0853850183342064,
+      "loss_partial": 0.049072265625,
+      "step": 2072
+    },
+    {
+      "epoch": 1.0853850183342064,
+      "loss_partial": 0.046875,
+      "step": 2072
+    },
+    {
+      "epoch": 1.0859088528025145,
+      "grad_norm": 2.0308310985565186,
+      "learning_rate": 1.8264839593358575e-05,
+      "loss": 0.1076,
+      "step": 2073
+    },
+    {
+      "epoch": 1.0859088528025145,
+      "loss_partial": 0.055908203125,
+      "step": 2073
+    },
+    {
+      "epoch": 1.0859088528025145,
+      "loss_partial": 0.05078125,
+      "step": 2073
+    },
+    {
+      "epoch": 1.0864326872708223,
+      "grad_norm": 2.328953742980957,
+      "learning_rate": 1.8262265109532155e-05,
+      "loss": 0.1053,
+      "step": 2074
+    },
+    {
+      "epoch": 1.0864326872708223,
+      "loss_partial": 0.06005859375,
+      "step": 2074
+    },
+    {
+      "epoch": 1.0864326872708223,
+      "loss_partial": 0.047119140625,
+      "step": 2074
+    },
+    {
+      "epoch": 1.0869565217391304,
+      "grad_norm": 2.469057083129883,
+      "learning_rate": 1.825968889894342e-05,
+      "loss": 0.1076,
+      "step": 2075
+    },
+    {
+      "epoch": 1.0869565217391304,
+      "loss_partial": 0.0537109375,
+      "step": 2075
+    },
+    {
+      "epoch": 1.0869565217391304,
+      "loss_partial": 0.0498046875,
+      "step": 2075
+    },
+    {
+      "epoch": 1.0874803562074384,
+      "grad_norm": 2.6005544662475586,
+      "learning_rate": 1.825711096213078e-05,
+      "loss": 0.1032,
+      "step": 2076
+    },
+    {
+      "epoch": 1.0874803562074384,
+      "loss_partial": 0.062255859375,
+      "step": 2076
+    },
+    {
+      "epoch": 1.0874803562074384,
+      "loss_partial": 0.055419921875,
+      "step": 2076
+    },
+    {
+      "epoch": 1.0880041906757465,
+      "grad_norm": 2.1837363243103027,
+      "learning_rate": 1.8254531299633007e-05,
+      "loss": 0.1063,
+      "step": 2077
+    },
+    {
+      "epoch": 1.0880041906757465,
+      "loss_partial": 0.06298828125,
+      "step": 2077
+    },
+    {
+      "epoch": 1.0880041906757465,
+      "loss_partial": 0.061767578125,
+      "step": 2077
+    },
+    {
+      "epoch": 1.0885280251440546,
+      "grad_norm": 1.9536652565002441,
+      "learning_rate": 1.825194991198924e-05,
+      "loss": 0.1144,
+      "step": 2078
+    },
+    {
+      "epoch": 1.0885280251440546,
+      "loss_partial": 0.047607421875,
+      "step": 2078
+    },
+    {
+      "epoch": 1.0885280251440546,
+      "loss_partial": 0.048583984375,
+      "step": 2078
+    },
+    {
+      "epoch": 1.0890518596123624,
+      "grad_norm": 2.1024832725524902,
+      "learning_rate": 1.8249366799738966e-05,
+      "loss": 0.1059,
+      "step": 2079
+    },
+    {
+      "epoch": 1.0890518596123624,
+      "loss_partial": 0.0517578125,
+      "step": 2079
+    },
+    {
+      "epoch": 1.0890518596123624,
+      "loss_partial": 0.04833984375,
+      "step": 2079
+    },
+    {
+      "epoch": 1.0895756940806705,
+      "grad_norm": 1.7422316074371338,
+      "learning_rate": 1.8246781963422048e-05,
+      "loss": 0.1084,
+      "step": 2080
+    },
+    {
+      "epoch": 1.0895756940806705,
+      "loss_partial": 0.050537109375,
+      "step": 2080
+    },
+    {
+      "epoch": 1.0895756940806705,
+      "loss_partial": 0.052001953125,
+      "step": 2080
+    },
+    {
+      "epoch": 1.0900995285489785,
+      "grad_norm": 2.2757949829101562,
+      "learning_rate": 1.8244195403578694e-05,
+      "loss": 0.109,
+      "step": 2081
+    },
+    {
+      "epoch": 1.0900995285489785,
+      "loss_partial": 0.048828125,
+      "step": 2081
+    },
+    {
+      "epoch": 1.0900995285489785,
+      "loss_partial": 0.052978515625,
+      "step": 2081
+    },
+    {
+      "epoch": 1.0906233630172866,
+      "grad_norm": 1.660848617553711,
+      "learning_rate": 1.824160712074948e-05,
+      "loss": 0.1041,
+      "step": 2082
+    },
+    {
+      "epoch": 1.0906233630172866,
+      "loss_partial": 0.052490234375,
+      "step": 2082
+    },
+    {
+      "epoch": 1.0906233630172866,
+      "loss_partial": 0.05712890625,
+      "step": 2082
+    },
+    {
+      "epoch": 1.0911471974855946,
+      "grad_norm": 1.4183409214019775,
+      "learning_rate": 1.8239017115475347e-05,
+      "loss": 0.1078,
+      "step": 2083
+    },
+    {
+      "epoch": 1.0911471974855946,
+      "loss_partial": 0.057861328125,
+      "step": 2083
+    },
+    {
+      "epoch": 1.0911471974855946,
+      "loss_partial": 0.050537109375,
+      "step": 2083
+    },
+    {
+      "epoch": 1.0916710319539025,
+      "grad_norm": 2.1728992462158203,
+      "learning_rate": 1.8236425388297578e-05,
+      "loss": 0.1107,
+      "step": 2084
+    },
+    {
+      "epoch": 1.0916710319539025,
+      "loss_partial": 0.042724609375,
+      "step": 2084
+    },
+    {
+      "epoch": 1.0916710319539025,
+      "loss_partial": 0.06787109375,
+      "step": 2084
+    },
+    {
+      "epoch": 1.0921948664222105,
+      "grad_norm": 2.292278289794922,
+      "learning_rate": 1.8233831939757843e-05,
+      "loss": 0.1077,
+      "step": 2085
+    },
+    {
+      "epoch": 1.0921948664222105,
+      "loss_partial": 0.06005859375,
+      "step": 2085
+    },
+    {
+      "epoch": 1.0921948664222105,
+      "loss_partial": 0.06005859375,
+      "step": 2085
+    },
+    {
+      "epoch": 1.0927187008905186,
+      "grad_norm": 1.3043426275253296,
+      "learning_rate": 1.8231236770398144e-05,
+      "loss": 0.1096,
+      "step": 2086
+    },
+    {
+      "epoch": 1.0927187008905186,
+      "loss_partial": 0.054931640625,
+      "step": 2086
+    },
+    {
+      "epoch": 1.0927187008905186,
+      "loss_partial": 0.056884765625,
+      "step": 2086
+    },
+    {
+      "epoch": 1.0932425353588267,
+      "grad_norm": 2.0700788497924805,
+      "learning_rate": 1.8228639880760864e-05,
+      "loss": 0.1092,
+      "step": 2087
+    },
+    {
+      "epoch": 1.0932425353588267,
+      "loss_partial": 0.06591796875,
+      "step": 2087
+    },
+    {
+      "epoch": 1.0932425353588267,
+      "loss_partial": 0.0693359375,
+      "step": 2087
+    },
+    {
+      "epoch": 1.0937663698271347,
+      "grad_norm": 2.351933240890503,
+      "learning_rate": 1.822604127138873e-05,
+      "loss": 0.1163,
+      "step": 2088
+    },
+    {
+      "epoch": 1.0937663698271347,
+      "loss_partial": 0.0478515625,
+      "step": 2088
+    },
+    {
+      "epoch": 1.0937663698271347,
+      "loss_partial": 0.052490234375,
+      "step": 2088
+    },
+    {
+      "epoch": 1.0942902042954425,
+      "grad_norm": 1.527948021888733,
+      "learning_rate": 1.822344094282484e-05,
+      "loss": 0.1049,
+      "step": 2089
+    },
+    {
+      "epoch": 1.0942902042954425,
+      "loss_partial": 0.05859375,
+      "step": 2089
+    },
+    {
+      "epoch": 1.0942902042954425,
+      "loss_partial": 0.05224609375,
+      "step": 2089
+    },
+    {
+      "epoch": 1.0948140387637506,
+      "grad_norm": 2.209233283996582,
+      "learning_rate": 1.8220838895612646e-05,
+      "loss": 0.1071,
+      "step": 2090
+    },
+    {
+      "epoch": 1.0948140387637506,
+      "loss_partial": 0.05859375,
+      "step": 2090
+    },
+    {
+      "epoch": 1.0948140387637506,
+      "loss_partial": 0.04736328125,
+      "step": 2090
+    },
+    {
+      "epoch": 1.0953378732320587,
+      "grad_norm": 1.5201343297958374,
+      "learning_rate": 1.8218235130295955e-05,
+      "loss": 0.1081,
+      "step": 2091
+    },
+    {
+      "epoch": 1.0953378732320587,
+      "loss_partial": 0.048095703125,
+      "step": 2091
+    },
+    {
+      "epoch": 1.0953378732320587,
+      "loss_partial": 0.0517578125,
+      "step": 2091
+    },
+    {
+      "epoch": 1.0958617077003667,
+      "grad_norm": 1.9935383796691895,
+      "learning_rate": 1.8215629647418945e-05,
+      "loss": 0.1086,
+      "step": 2092
+    },
+    {
+      "epoch": 1.0958617077003667,
+      "loss_partial": 0.05615234375,
+      "step": 2092
+    },
+    {
+      "epoch": 1.0958617077003667,
+      "loss_partial": 0.051025390625,
+      "step": 2092
+    },
+    {
+      "epoch": 1.0963855421686748,
+      "grad_norm": 1.7858370542526245,
+      "learning_rate": 1.8213022447526144e-05,
+      "loss": 0.1089,
+      "step": 2093
+    },
+    {
+      "epoch": 1.0963855421686748,
+      "loss_partial": 0.052001953125,
+      "step": 2093
+    },
+    {
+      "epoch": 1.0963855421686748,
+      "loss_partial": 0.050537109375,
+      "step": 2093
+    },
+    {
+      "epoch": 1.0969093766369826,
+      "grad_norm": 1.704903483390808,
+      "learning_rate": 1.8210413531162435e-05,
+      "loss": 0.107,
+      "step": 2094
+    },
+    {
+      "epoch": 1.0969093766369826,
+      "loss_partial": 0.05859375,
+      "step": 2094
+    },
+    {
+      "epoch": 1.0969093766369826,
+      "loss_partial": 0.056396484375,
+      "step": 2094
+    },
+    {
+      "epoch": 1.0974332111052907,
+      "grad_norm": 2.043867826461792,
+      "learning_rate": 1.8207802898873076e-05,
+      "loss": 0.1093,
+      "step": 2095
+    },
+    {
+      "epoch": 1.0974332111052907,
+      "loss_partial": 0.064453125,
+      "step": 2095
+    },
+    {
+      "epoch": 1.0974332111052907,
+      "loss_partial": 0.0595703125,
+      "step": 2095
+    },
+    {
+      "epoch": 1.0979570455735987,
+      "grad_norm": 2.2597575187683105,
+      "learning_rate": 1.820519055120366e-05,
+      "loss": 0.1093,
+      "step": 2096
+    },
+    {
+      "epoch": 1.0979570455735987,
+      "loss_partial": 0.04736328125,
+      "step": 2096
+    },
+    {
+      "epoch": 1.0979570455735987,
+      "loss_partial": 0.061767578125,
+      "step": 2096
+    },
+    {
+      "epoch": 1.0984808800419068,
+      "grad_norm": 1.7826290130615234,
+      "learning_rate": 1.8202576488700164e-05,
+      "loss": 0.1086,
+      "step": 2097
+    },
+    {
+      "epoch": 1.0984808800419068,
+      "loss_partial": 0.055419921875,
+      "step": 2097
+    },
+    {
+      "epoch": 1.0984808800419068,
+      "loss_partial": 0.0546875,
+      "step": 2097
+    },
+    {
+      "epoch": 1.0990047145102149,
+      "grad_norm": 2.396639347076416,
+      "learning_rate": 1.8199960711908904e-05,
+      "loss": 0.1064,
+      "step": 2098
+    },
+    {
+      "epoch": 1.0990047145102149,
+      "loss_partial": 0.060302734375,
+      "step": 2098
+    },
+    {
+      "epoch": 1.0990047145102149,
+      "loss_partial": 0.0673828125,
+      "step": 2098
+    },
+    {
+      "epoch": 1.0995285489785227,
+      "grad_norm": 1.849340796470642,
+      "learning_rate": 1.8197343221376557e-05,
+      "loss": 0.1073,
+      "step": 2099
+    },
+    {
+      "epoch": 1.0995285489785227,
+      "loss_partial": 0.05517578125,
+      "step": 2099
+    },
+    {
+      "epoch": 1.0995285489785227,
+      "loss_partial": 0.060791015625,
+      "step": 2099
+    },
+    {
+      "epoch": 1.1000523834468308,
+      "grad_norm": 2.605283260345459,
+      "learning_rate": 1.819472401765017e-05,
+      "loss": 0.1081,
+      "step": 2100
+    },
+    {
+      "epoch": 1.1000523834468308,
+      "loss_partial": 0.046630859375,
+      "step": 2100
+    },
+    {
+      "epoch": 1.1000523834468308,
+      "loss_partial": 0.044921875,
+      "step": 2100
+    },
+    {
+      "epoch": 1.1005762179151388,
+      "grad_norm": 2.0239226818084717,
+      "learning_rate": 1.8192103101277144e-05,
+      "loss": 0.1075,
+      "step": 2101
+    },
+    {
+      "epoch": 1.1005762179151388,
+      "loss_partial": 0.06591796875,
+      "step": 2101
+    },
+    {
+      "epoch": 1.1005762179151388,
+      "loss_partial": 0.06591796875,
+      "step": 2101
+    },
+    {
+      "epoch": 1.1011000523834469,
+      "grad_norm": 2.1477112770080566,
+      "learning_rate": 1.818948047280522e-05,
+      "loss": 0.1067,
+      "step": 2102
+    },
+    {
+      "epoch": 1.1011000523834469,
+      "loss_partial": 0.059814453125,
+      "step": 2102
+    },
+    {
+      "epoch": 1.1011000523834469,
+      "loss_partial": 0.052978515625,
+      "step": 2102
+    },
+    {
+      "epoch": 1.101623886851755,
+      "grad_norm": 1.9162724018096924,
+      "learning_rate": 1.8186856132782525e-05,
+      "loss": 0.1078,
+      "step": 2103
+    },
+    {
+      "epoch": 1.101623886851755,
+      "loss_partial": 0.050537109375,
+      "step": 2103
+    },
+    {
+      "epoch": 1.101623886851755,
+      "loss_partial": 0.045166015625,
+      "step": 2103
+    },
+    {
+      "epoch": 1.1021477213200628,
+      "grad_norm": 1.979694128036499,
+      "learning_rate": 1.818423008175752e-05,
+      "loss": 0.1074,
+      "step": 2104
+    },
+    {
+      "epoch": 1.1021477213200628,
+      "loss_partial": 0.042724609375,
+      "step": 2104
+    },
+    {
+      "epoch": 1.1021477213200628,
+      "loss_partial": 0.051025390625,
+      "step": 2104
+    },
+    {
+      "epoch": 1.1026715557883708,
+      "grad_norm": 1.908992052078247,
+      "learning_rate": 1.818160232027904e-05,
+      "loss": 0.1084,
+      "step": 2105
+    },
+    {
+      "epoch": 1.1026715557883708,
+      "loss_partial": 0.07470703125,
+      "step": 2105
+    },
+    {
+      "epoch": 1.1026715557883708,
+      "loss_partial": 0.061279296875,
+      "step": 2105
+    },
+    {
+      "epoch": 1.103195390256679,
+      "grad_norm": 1.7697923183441162,
+      "learning_rate": 1.817897284889627e-05,
+      "loss": 0.1126,
+      "step": 2106
+    },
+    {
+      "epoch": 1.103195390256679,
+      "loss_partial": 0.0576171875,
+      "step": 2106
+    },
+    {
+      "epoch": 1.103195390256679,
+      "loss_partial": 0.051025390625,
+      "step": 2106
+    },
+    {
+      "epoch": 1.103719224724987,
+      "grad_norm": 2.3765056133270264,
+      "learning_rate": 1.817634166815875e-05,
+      "loss": 0.1094,
+      "step": 2107
+    },
+    {
+      "epoch": 1.103719224724987,
+      "loss_partial": 0.05908203125,
+      "step": 2107
+    },
+    {
+      "epoch": 1.103719224724987,
+      "loss_partial": 0.046875,
+      "step": 2107
+    },
+    {
+      "epoch": 1.104243059193295,
+      "grad_norm": 2.0784401893615723,
+      "learning_rate": 1.8173708778616382e-05,
+      "loss": 0.1108,
+      "step": 2108
+    },
+    {
+      "epoch": 1.104243059193295,
+      "loss_partial": 0.051513671875,
+      "step": 2108
+    },
+    {
+      "epoch": 1.104243059193295,
+      "loss_partial": 0.05615234375,
+      "step": 2108
+    },
+    {
+      "epoch": 1.1047668936616029,
+      "grad_norm": 1.5009461641311646,
+      "learning_rate": 1.8171074180819425e-05,
+      "loss": 0.1091,
+      "step": 2109
+    },
+    {
+      "epoch": 1.1047668936616029,
+      "loss_partial": 0.05322265625,
+      "step": 2109
+    },
+    {
+      "epoch": 1.1047668936616029,
+      "loss_partial": 0.048828125,
+      "step": 2109
+    },
+    {
+      "epoch": 1.105290728129911,
+      "grad_norm": 1.7534716129302979,
+      "learning_rate": 1.8168437875318493e-05,
+      "loss": 0.1041,
+      "step": 2110
+    },
+    {
+      "epoch": 1.105290728129911,
+      "loss_partial": 0.0634765625,
+      "step": 2110
+    },
+    {
+      "epoch": 1.105290728129911,
+      "loss_partial": 0.052734375,
+      "step": 2110
+    },
+    {
+      "epoch": 1.105814562598219,
+      "grad_norm": 1.397887945175171,
+      "learning_rate": 1.8165799862664556e-05,
+      "loss": 0.1104,
+      "step": 2111
+    },
+    {
+      "epoch": 1.105814562598219,
+      "loss_partial": 0.0517578125,
+      "step": 2111
+    },
+    {
+      "epoch": 1.105814562598219,
+      "loss_partial": 0.054931640625,
+      "step": 2111
+    },
+    {
+      "epoch": 1.106338397066527,
+      "grad_norm": 1.5021854639053345,
+      "learning_rate": 1.8163160143408942e-05,
+      "loss": 0.1077,
+      "step": 2112
+    },
+    {
+      "epoch": 1.106338397066527,
+      "loss_partial": 0.052490234375,
+      "step": 2112
+    },
+    {
+      "epoch": 1.106338397066527,
+      "loss_partial": 0.051025390625,
+      "step": 2112
+    },
+    {
+      "epoch": 1.106862231534835,
+      "grad_norm": 1.2918471097946167,
+      "learning_rate": 1.816051871810334e-05,
+      "loss": 0.1106,
+      "step": 2113
+    },
+    {
+      "epoch": 1.106862231534835,
+      "loss_partial": 0.0693359375,
+      "step": 2113
+    },
+    {
+      "epoch": 1.106862231534835,
+      "loss_partial": 0.05615234375,
+      "step": 2113
+    },
+    {
+      "epoch": 1.107386066003143,
+      "grad_norm": 1.6528501510620117,
+      "learning_rate": 1.8157875587299787e-05,
+      "loss": 0.1059,
+      "step": 2114
+    },
+    {
+      "epoch": 1.107386066003143,
+      "loss_partial": 0.04443359375,
+      "step": 2114
+    },
+    {
+      "epoch": 1.107386066003143,
+      "loss_partial": 0.06884765625,
+      "step": 2114
+    },
+    {
+      "epoch": 1.107909900471451,
+      "grad_norm": 1.8016159534454346,
+      "learning_rate": 1.8155230751550685e-05,
+      "loss": 0.1088,
+      "step": 2115
+    },
+    {
+      "epoch": 1.107909900471451,
+      "loss_partial": 0.051025390625,
+      "step": 2115
+    },
+    {
+      "epoch": 1.107909900471451,
+      "loss_partial": 0.052490234375,
+      "step": 2115
+    },
+    {
+      "epoch": 1.108433734939759,
+      "grad_norm": 2.4045281410217285,
+      "learning_rate": 1.8152584211408784e-05,
+      "loss": 0.1047,
+      "step": 2116
+    },
+    {
+      "epoch": 1.108433734939759,
+      "loss_partial": 0.053466796875,
+      "step": 2116
+    },
+    {
+      "epoch": 1.108433734939759,
+      "loss_partial": 0.048095703125,
+      "step": 2116
+    },
+    {
+      "epoch": 1.1089575694080671,
+      "grad_norm": 1.7509621381759644,
+      "learning_rate": 1.81499359674272e-05,
+      "loss": 0.1096,
+      "step": 2117
+    },
+    {
+      "epoch": 1.1089575694080671,
+      "loss_partial": 0.04345703125,
+      "step": 2117
+    },
+    {
+      "epoch": 1.1089575694080671,
+      "loss_partial": 0.050537109375,
+      "step": 2117
+    },
+    {
+      "epoch": 1.1094814038763752,
+      "grad_norm": 1.458978295326233,
+      "learning_rate": 1.8147286020159394e-05,
+      "loss": 0.1061,
+      "step": 2118
+    },
+    {
+      "epoch": 1.1094814038763752,
+      "loss_partial": 0.04296875,
+      "step": 2118
+    },
+    {
+      "epoch": 1.1094814038763752,
+      "loss_partial": 0.0595703125,
+      "step": 2118
+    },
+    {
+      "epoch": 1.110005238344683,
+      "grad_norm": 1.6624882221221924,
+      "learning_rate": 1.8144634370159194e-05,
+      "loss": 0.1128,
+      "step": 2119
+    },
+    {
+      "epoch": 1.110005238344683,
+      "loss_partial": 0.0498046875,
+      "step": 2119
+    },
+    {
+      "epoch": 1.110005238344683,
+      "loss_partial": 0.0537109375,
+      "step": 2119
+    },
+    {
+      "epoch": 1.110529072812991,
+      "grad_norm": 1.3391633033752441,
+      "learning_rate": 1.8141981017980773e-05,
+      "loss": 0.1059,
+      "step": 2120
+    },
+    {
+      "epoch": 1.110529072812991,
+      "loss_partial": 0.0673828125,
+      "step": 2120
+    },
+    {
+      "epoch": 1.110529072812991,
+      "loss_partial": 0.05908203125,
+      "step": 2120
+    },
+    {
+      "epoch": 1.1110529072812991,
+      "grad_norm": 1.7584360837936401,
+      "learning_rate": 1.8139325964178667e-05,
+      "loss": 0.1093,
+      "step": 2121
+    },
+    {
+      "epoch": 1.1110529072812991,
+      "loss_partial": 0.0615234375,
+      "step": 2121
+    },
+    {
+      "epoch": 1.1110529072812991,
+      "loss_partial": 0.06591796875,
+      "step": 2121
+    },
+    {
+      "epoch": 1.1115767417496072,
+      "grad_norm": 1.271456003189087,
+      "learning_rate": 1.8136669209307768e-05,
+      "loss": 0.1099,
+      "step": 2122
+    },
+    {
+      "epoch": 1.1115767417496072,
+      "loss_partial": 0.0537109375,
+      "step": 2122
+    },
+    {
+      "epoch": 1.1115767417496072,
+      "loss_partial": 0.060791015625,
+      "step": 2122
+    },
+    {
+      "epoch": 1.112100576217915,
+      "grad_norm": 1.827226161956787,
+      "learning_rate": 1.8134010753923318e-05,
+      "loss": 0.1069,
+      "step": 2123
+    },
+    {
+      "epoch": 1.112100576217915,
+      "loss_partial": 0.057373046875,
+      "step": 2123
+    },
+    {
+      "epoch": 1.112100576217915,
+      "loss_partial": 0.0615234375,
+      "step": 2123
+    },
+    {
+      "epoch": 1.112624410686223,
+      "grad_norm": 1.8040052652359009,
+      "learning_rate": 1.8131350598580925e-05,
+      "loss": 0.1104,
+      "step": 2124
+    },
+    {
+      "epoch": 1.112624410686223,
+      "loss_partial": 0.054443359375,
+      "step": 2124
+    },
+    {
+      "epoch": 1.112624410686223,
+      "loss_partial": 0.045654296875,
+      "step": 2124
+    },
+    {
+      "epoch": 1.1131482451545311,
+      "grad_norm": 1.812445878982544,
+      "learning_rate": 1.8128688743836537e-05,
+      "loss": 0.1043,
+      "step": 2125
+    },
+    {
+      "epoch": 1.1131482451545311,
+      "loss_partial": 0.050048828125,
+      "step": 2125
+    },
+    {
+      "epoch": 1.1131482451545311,
+      "loss_partial": 0.058837890625,
+      "step": 2125
+    },
+    {
+      "epoch": 1.1136720796228392,
+      "grad_norm": 2.1591360569000244,
+      "learning_rate": 1.8126025190246467e-05,
+      "loss": 0.1063,
+      "step": 2126
+    },
+    {
+      "epoch": 1.1136720796228392,
+      "loss_partial": 0.047607421875,
+      "step": 2126
+    },
+    {
+      "epoch": 1.1136720796228392,
+      "loss_partial": 0.055419921875,
+      "step": 2126
+    },
+    {
+      "epoch": 1.1141959140911473,
+      "grad_norm": 2.1146719455718994,
+      "learning_rate": 1.812335993836738e-05,
+      "loss": 0.1086,
+      "step": 2127
+    },
+    {
+      "epoch": 1.1141959140911473,
+      "loss_partial": 0.052001953125,
+      "step": 2127
+    },
+    {
+      "epoch": 1.1141959140911473,
+      "loss_partial": 0.061279296875,
+      "step": 2127
+    },
+    {
+      "epoch": 1.1147197485594553,
+      "grad_norm": 1.9405502080917358,
+      "learning_rate": 1.81206929887563e-05,
+      "loss": 0.1064,
+      "step": 2128
+    },
+    {
+      "epoch": 1.1147197485594553,
+      "loss_partial": 0.05078125,
+      "step": 2128
+    },
+    {
+      "epoch": 1.1147197485594553,
+      "loss_partial": 0.0556640625,
+      "step": 2128
+    },
+    {
+      "epoch": 1.1152435830277632,
+      "grad_norm": 1.5561691522598267,
+      "learning_rate": 1.8118024341970603e-05,
+      "loss": 0.1049,
+      "step": 2129
+    },
+    {
+      "epoch": 1.1152435830277632,
+      "loss_partial": 0.055908203125,
+      "step": 2129
+    },
+    {
+      "epoch": 1.1152435830277632,
+      "loss_partial": 0.050537109375,
+      "step": 2129
+    },
+    {
+      "epoch": 1.1157674174960712,
+      "grad_norm": 2.0904293060302734,
+      "learning_rate": 1.811535399856802e-05,
+      "loss": 0.1048,
+      "step": 2130
+    },
+    {
+      "epoch": 1.1157674174960712,
+      "loss_partial": 0.045654296875,
+      "step": 2130
+    },
+    {
+      "epoch": 1.1157674174960712,
+      "loss_partial": 0.048828125,
+      "step": 2130
+    },
+    {
+      "epoch": 1.1162912519643793,
+      "grad_norm": 2.416969060897827,
+      "learning_rate": 1.8112681959106636e-05,
+      "loss": 0.1046,
+      "step": 2131
+    },
+    {
+      "epoch": 1.1162912519643793,
+      "loss_partial": 0.055908203125,
+      "step": 2131
+    },
+    {
+      "epoch": 1.1162912519643793,
+      "loss_partial": 0.055908203125,
+      "step": 2131
+    },
+    {
+      "epoch": 1.1168150864326873,
+      "grad_norm": 2.2100203037261963,
+      "learning_rate": 1.8110008224144887e-05,
+      "loss": 0.1121,
+      "step": 2132
+    },
+    {
+      "epoch": 1.1168150864326873,
+      "loss_partial": 0.0576171875,
+      "step": 2132
+    },
+    {
+      "epoch": 1.1168150864326873,
+      "loss_partial": 0.0498046875,
+      "step": 2132
+    },
+    {
+      "epoch": 1.1173389209009952,
+      "grad_norm": 2.1731250286102295,
+      "learning_rate": 1.810733279424157e-05,
+      "loss": 0.1053,
+      "step": 2133
+    },
+    {
+      "epoch": 1.1173389209009952,
+      "loss_partial": 0.06591796875,
+      "step": 2133
+    },
+    {
+      "epoch": 1.1173389209009952,
+      "loss_partial": 0.053466796875,
+      "step": 2133
+    },
+    {
+      "epoch": 1.1178627553693032,
+      "grad_norm": 1.7137327194213867,
+      "learning_rate": 1.8104655669955837e-05,
+      "loss": 0.1073,
+      "step": 2134
+    },
+    {
+      "epoch": 1.1178627553693032,
+      "loss_partial": 0.0517578125,
+      "step": 2134
+    },
+    {
+      "epoch": 1.1178627553693032,
+      "loss_partial": 0.0546875,
+      "step": 2134
+    },
+    {
+      "epoch": 1.1183865898376113,
+      "grad_norm": 2.0392050743103027,
+      "learning_rate": 1.8101976851847186e-05,
+      "loss": 0.1067,
+      "step": 2135
+    },
+    {
+      "epoch": 1.1183865898376113,
+      "loss_partial": 0.053955078125,
+      "step": 2135
+    },
+    {
+      "epoch": 1.1183865898376113,
+      "loss_partial": 0.053955078125,
+      "step": 2135
+    },
+    {
+      "epoch": 1.1189104243059194,
+      "grad_norm": 1.392767071723938,
+      "learning_rate": 1.8099296340475468e-05,
+      "loss": 0.1096,
+      "step": 2136
+    },
+    {
+      "epoch": 1.1189104243059194,
+      "loss_partial": 0.051025390625,
+      "step": 2136
+    },
+    {
+      "epoch": 1.1189104243059194,
+      "loss_partial": 0.047607421875,
+      "step": 2136
+    },
+    {
+      "epoch": 1.1194342587742274,
+      "grad_norm": 2.067511796951294,
+      "learning_rate": 1.8096614136400904e-05,
+      "loss": 0.109,
+      "step": 2137
+    },
+    {
+      "epoch": 1.1194342587742274,
+      "loss_partial": 0.051025390625,
+      "step": 2137
+    },
+    {
+      "epoch": 1.1194342587742274,
+      "loss_partial": 0.044677734375,
+      "step": 2137
+    },
+    {
+      "epoch": 1.1199580932425353,
+      "grad_norm": 1.767757773399353,
+      "learning_rate": 1.8093930240184054e-05,
+      "loss": 0.1095,
+      "step": 2138
+    },
+    {
+      "epoch": 1.1199580932425353,
+      "loss_partial": 0.05322265625,
+      "step": 2138
+    },
+    {
+      "epoch": 1.1199580932425353,
+      "loss_partial": 0.05712890625,
+      "step": 2138
+    },
+    {
+      "epoch": 1.1204819277108433,
+      "grad_norm": 1.5657871961593628,
+      "learning_rate": 1.8091244652385834e-05,
+      "loss": 0.1075,
+      "step": 2139
+    },
+    {
+      "epoch": 1.1204819277108433,
+      "loss_partial": 0.05029296875,
+      "step": 2139
+    },
+    {
+      "epoch": 1.1204819277108433,
+      "loss_partial": 0.06201171875,
+      "step": 2139
+    },
+    {
+      "epoch": 1.1210057621791514,
+      "grad_norm": 1.5593136548995972,
+      "learning_rate": 1.8088557373567517e-05,
+      "loss": 0.106,
+      "step": 2140
+    },
+    {
+      "epoch": 1.1210057621791514,
+      "loss_partial": 0.06103515625,
+      "step": 2140
+    },
+    {
+      "epoch": 1.1210057621791514,
+      "loss_partial": 0.045654296875,
+      "step": 2140
+    },
+    {
+      "epoch": 1.1215295966474594,
+      "grad_norm": 1.2747602462768555,
+      "learning_rate": 1.8085868404290728e-05,
+      "loss": 0.11,
+      "step": 2141
+    },
+    {
+      "epoch": 1.1215295966474594,
+      "loss_partial": 0.0517578125,
+      "step": 2141
+    },
+    {
+      "epoch": 1.1215295966474594,
+      "loss_partial": 0.0556640625,
+      "step": 2141
+    },
+    {
+      "epoch": 1.1220534311157675,
+      "grad_norm": 1.3937636613845825,
+      "learning_rate": 1.8083177745117442e-05,
+      "loss": 0.1085,
+      "step": 2142
+    },
+    {
+      "epoch": 1.1220534311157675,
+      "loss_partial": 0.052978515625,
+      "step": 2142
+    },
+    {
+      "epoch": 1.1220534311157675,
+      "loss_partial": 0.0595703125,
+      "step": 2142
+    },
+    {
+      "epoch": 1.1225772655840753,
+      "grad_norm": 1.4946409463882446,
+      "learning_rate": 1.808048539660999e-05,
+      "loss": 0.1083,
+      "step": 2143
+    },
+    {
+      "epoch": 1.1225772655840753,
+      "loss_partial": 0.05859375,
+      "step": 2143
+    },
+    {
+      "epoch": 1.1225772655840753,
+      "loss_partial": 0.046875,
+      "step": 2143
+    },
+    {
+      "epoch": 1.1231011000523834,
+      "grad_norm": 1.8207412958145142,
+      "learning_rate": 1.807779135933106e-05,
+      "loss": 0.1096,
+      "step": 2144
+    },
+    {
+      "epoch": 1.1231011000523834,
+      "loss_partial": 0.04638671875,
+      "step": 2144
+    },
+    {
+      "epoch": 1.1231011000523834,
+      "loss_partial": 0.05859375,
+      "step": 2144
+    },
+    {
+      "epoch": 1.1236249345206915,
+      "grad_norm": 1.441483974456787,
+      "learning_rate": 1.807509563384369e-05,
+      "loss": 0.1064,
+      "step": 2145
+    },
+    {
+      "epoch": 1.1236249345206915,
+      "loss_partial": 0.05029296875,
+      "step": 2145
+    },
+    {
+      "epoch": 1.1236249345206915,
+      "loss_partial": 0.05908203125,
+      "step": 2145
+    },
+    {
+      "epoch": 1.1241487689889995,
+      "grad_norm": 1.750182032585144,
+      "learning_rate": 1.8072398220711268e-05,
+      "loss": 0.1074,
+      "step": 2146
+    },
+    {
+      "epoch": 1.1241487689889995,
+      "loss_partial": 0.04833984375,
+      "step": 2146
+    },
+    {
+      "epoch": 1.1241487689889995,
+      "loss_partial": 0.05517578125,
+      "step": 2146
+    },
+    {
+      "epoch": 1.1246726034573076,
+      "grad_norm": 1.5668977499008179,
+      "learning_rate": 1.8069699120497533e-05,
+      "loss": 0.1114,
+      "step": 2147
+    },
+    {
+      "epoch": 1.1246726034573076,
+      "loss_partial": 0.06298828125,
+      "step": 2147
+    },
+    {
+      "epoch": 1.1246726034573076,
+      "loss_partial": 0.04833984375,
+      "step": 2147
+    },
+    {
+      "epoch": 1.1251964379256154,
+      "grad_norm": 1.6792019605636597,
+      "learning_rate": 1.8066998333766587e-05,
+      "loss": 0.1075,
+      "step": 2148
+    },
+    {
+      "epoch": 1.1251964379256154,
+      "loss_partial": 0.049072265625,
+      "step": 2148
+    },
+    {
+      "epoch": 1.1251964379256154,
+      "loss_partial": 0.05615234375,
+      "step": 2148
+    },
+    {
+      "epoch": 1.1257202723939235,
+      "grad_norm": 1.337951898574829,
+      "learning_rate": 1.806429586108287e-05,
+      "loss": 0.1078,
+      "step": 2149
+    },
+    {
+      "epoch": 1.1257202723939235,
+      "loss_partial": 0.0419921875,
+      "step": 2149
+    },
+    {
+      "epoch": 1.1257202723939235,
+      "loss_partial": 0.0458984375,
+      "step": 2149
+    },
+    {
+      "epoch": 1.1262441068622315,
+      "grad_norm": 2.191188335418701,
+      "learning_rate": 1.806159170301119e-05,
+      "loss": 0.1107,
+      "step": 2150
+    },
+    {
+      "epoch": 1.1262441068622315,
+      "loss_partial": 0.06884765625,
+      "step": 2150
+    },
+    {
+      "epoch": 1.1262441068622315,
+      "loss_partial": 0.04736328125,
+      "step": 2150
+    },
+    {
+      "epoch": 1.1267679413305396,
+      "grad_norm": 1.4685513973236084,
+      "learning_rate": 1.805888586011669e-05,
+      "loss": 0.1072,
+      "step": 2151
+    },
+    {
+      "epoch": 1.1267679413305396,
+      "loss_partial": 0.06298828125,
+      "step": 2151
+    },
+    {
+      "epoch": 1.1267679413305396,
+      "loss_partial": 0.0546875,
+      "step": 2151
+    },
+    {
+      "epoch": 1.1272917757988477,
+      "grad_norm": 1.9280421733856201,
+      "learning_rate": 1.8056178332964882e-05,
+      "loss": 0.1088,
+      "step": 2152
+    },
+    {
+      "epoch": 1.1272917757988477,
+      "loss_partial": 0.055419921875,
+      "step": 2152
+    },
+    {
+      "epoch": 1.1272917757988477,
+      "loss_partial": 0.052001953125,
+      "step": 2152
+    },
+    {
+      "epoch": 1.1278156102671555,
+      "grad_norm": 1.2932316064834595,
+      "learning_rate": 1.805346912212162e-05,
+      "loss": 0.1082,
+      "step": 2153
+    },
+    {
+      "epoch": 1.1278156102671555,
+      "loss_partial": 0.05712890625,
+      "step": 2153
+    },
+    {
+      "epoch": 1.1278156102671555,
+      "loss_partial": 0.051025390625,
+      "step": 2153
+    },
+    {
+      "epoch": 1.1283394447354635,
+      "grad_norm": 1.922957181930542,
+      "learning_rate": 1.8050758228153117e-05,
+      "loss": 0.1122,
+      "step": 2154
+    },
+    {
+      "epoch": 1.1283394447354635,
+      "loss_partial": 0.05322265625,
+      "step": 2154
+    },
+    {
+      "epoch": 1.1283394447354635,
+      "loss_partial": 0.049560546875,
+      "step": 2154
+    },
+    {
+      "epoch": 1.1288632792037716,
+      "grad_norm": 1.6878767013549805,
+      "learning_rate": 1.8048045651625926e-05,
+      "loss": 0.1093,
+      "step": 2155
+    },
+    {
+      "epoch": 1.1288632792037716,
+      "loss_partial": 0.04736328125,
+      "step": 2155
+    },
+    {
+      "epoch": 1.1288632792037716,
+      "loss_partial": 0.05810546875,
+      "step": 2155
+    },
+    {
+      "epoch": 1.1293871136720797,
+      "grad_norm": 1.579218864440918,
+      "learning_rate": 1.804533139310696e-05,
+      "loss": 0.1059,
+      "step": 2156
+    },
+    {
+      "epoch": 1.1293871136720797,
+      "loss_partial": 0.059326171875,
+      "step": 2156
+    },
+    {
+      "epoch": 1.1293871136720797,
+      "loss_partial": 0.0595703125,
+      "step": 2156
+    },
+    {
+      "epoch": 1.1299109481403877,
+      "grad_norm": 1.8525301218032837,
+      "learning_rate": 1.8042615453163484e-05,
+      "loss": 0.1088,
+      "step": 2157
+    },
+    {
+      "epoch": 1.1299109481403877,
+      "loss_partial": 0.057861328125,
+      "step": 2157
+    },
+    {
+      "epoch": 1.1299109481403877,
+      "loss_partial": 0.056640625,
+      "step": 2157
+    },
+    {
+      "epoch": 1.1304347826086956,
+      "grad_norm": 0.991853654384613,
+      "learning_rate": 1.8039897832363114e-05,
+      "loss": 0.1054,
+      "step": 2158
+    },
+    {
+      "epoch": 1.1304347826086956,
+      "loss_partial": 0.0546875,
+      "step": 2158
+    },
+    {
+      "epoch": 1.1304347826086956,
+      "loss_partial": 0.0556640625,
+      "step": 2158
+    },
+    {
+      "epoch": 1.1309586170770036,
+      "grad_norm": 1.519075870513916,
+      "learning_rate": 1.803717853127381e-05,
+      "loss": 0.103,
+      "step": 2159
+    },
+    {
+      "epoch": 1.1309586170770036,
+      "loss_partial": 0.05419921875,
+      "step": 2159
+    },
+    {
+      "epoch": 1.1309586170770036,
+      "loss_partial": 0.056884765625,
+      "step": 2159
+    },
+    {
+      "epoch": 1.1314824515453117,
+      "grad_norm": 1.9036487340927124,
+      "learning_rate": 1.80344575504639e-05,
+      "loss": 0.1069,
+      "step": 2160
+    },
+    {
+      "epoch": 1.1314824515453117,
+      "loss_partial": 0.05810546875,
+      "step": 2160
+    },
+    {
+      "epoch": 1.1314824515453117,
+      "loss_partial": 0.0537109375,
+      "step": 2160
+    },
+    {
+      "epoch": 1.1320062860136197,
+      "grad_norm": 1.4712125062942505,
+      "learning_rate": 1.8031734890502043e-05,
+      "loss": 0.1096,
+      "step": 2161
+    },
+    {
+      "epoch": 1.1320062860136197,
+      "loss_partial": 0.046142578125,
+      "step": 2161
+    },
+    {
+      "epoch": 1.1320062860136197,
+      "loss_partial": 0.047607421875,
+      "step": 2161
+    },
+    {
+      "epoch": 1.1325301204819278,
+      "grad_norm": 2.0649912357330322,
+      "learning_rate": 1.8029010551957257e-05,
+      "loss": 0.1025,
+      "step": 2162
+    },
+    {
+      "epoch": 1.1325301204819278,
+      "loss_partial": 0.0546875,
+      "step": 2162
+    },
+    {
+      "epoch": 1.1325301204819278,
+      "loss_partial": 0.055908203125,
+      "step": 2162
+    },
+    {
+      "epoch": 1.1330539549502356,
+      "grad_norm": 2.121832847595215,
+      "learning_rate": 1.8026284535398918e-05,
+      "loss": 0.11,
+      "step": 2163
+    },
+    {
+      "epoch": 1.1330539549502356,
+      "loss_partial": 0.05517578125,
+      "step": 2163
+    },
+    {
+      "epoch": 1.1330539549502356,
+      "loss_partial": 0.05029296875,
+      "step": 2163
+    },
+    {
+      "epoch": 1.1335777894185437,
+      "grad_norm": 2.5854811668395996,
+      "learning_rate": 1.8023556841396747e-05,
+      "loss": 0.1054,
+      "step": 2164
+    },
+    {
+      "epoch": 1.1335777894185437,
+      "loss_partial": 0.054443359375,
+      "step": 2164
+    },
+    {
+      "epoch": 1.1335777894185437,
+      "loss_partial": 0.053466796875,
+      "step": 2164
+    },
+    {
+      "epoch": 1.1341016238868518,
+      "grad_norm": 2.7397985458374023,
+      "learning_rate": 1.802082747052081e-05,
+      "loss": 0.1079,
+      "step": 2165
+    },
+    {
+      "epoch": 1.1341016238868518,
+      "loss_partial": 0.050537109375,
+      "step": 2165
+    },
+    {
+      "epoch": 1.1341016238868518,
+      "loss_partial": 0.060302734375,
+      "step": 2165
+    },
+    {
+      "epoch": 1.1346254583551598,
+      "grad_norm": 2.206690549850464,
+      "learning_rate": 1.8018096423341533e-05,
+      "loss": 0.1089,
+      "step": 2166
+    },
+    {
+      "epoch": 1.1346254583551598,
+      "loss_partial": 0.04638671875,
+      "step": 2166
+    },
+    {
+      "epoch": 1.1346254583551598,
+      "loss_partial": 0.057373046875,
+      "step": 2166
+    },
+    {
+      "epoch": 1.1351492928234679,
+      "grad_norm": 3.6900229454040527,
+      "learning_rate": 1.801536370042968e-05,
+      "loss": 0.1074,
+      "step": 2167
+    },
+    {
+      "epoch": 1.1351492928234679,
+      "loss_partial": 0.060546875,
+      "step": 2167
+    },
+    {
+      "epoch": 1.1351492928234679,
+      "loss_partial": 0.04736328125,
+      "step": 2167
+    },
+    {
+      "epoch": 1.1356731272917757,
+      "grad_norm": 1.9463980197906494,
+      "learning_rate": 1.8012629302356383e-05,
+      "loss": 0.1118,
+      "step": 2168
+    },
+    {
+      "epoch": 1.1356731272917757,
+      "loss_partial": 0.050537109375,
+      "step": 2168
+    },
+    {
+      "epoch": 1.1356731272917757,
+      "loss_partial": 0.049560546875,
+      "step": 2168
+    },
+    {
+      "epoch": 1.1361969617600838,
+      "grad_norm": 2.1752874851226807,
+      "learning_rate": 1.8009893229693115e-05,
+      "loss": 0.1068,
+      "step": 2169
+    },
+    {
+      "epoch": 1.1361969617600838,
+      "loss_partial": 0.0625,
+      "step": 2169
+    },
+    {
+      "epoch": 1.1361969617600838,
+      "loss_partial": 0.06689453125,
+      "step": 2169
+    },
+    {
+      "epoch": 1.1367207962283918,
+      "grad_norm": 2.305771589279175,
+      "learning_rate": 1.8007155483011685e-05,
+      "loss": 0.1093,
+      "step": 2170
+    },
+    {
+      "epoch": 1.1367207962283918,
+      "loss_partial": 0.05859375,
+      "step": 2170
+    },
+    {
+      "epoch": 1.1367207962283918,
+      "loss_partial": 0.05419921875,
+      "step": 2170
+    },
+    {
+      "epoch": 1.1372446306967,
+      "grad_norm": 2.240382194519043,
+      "learning_rate": 1.800441606288428e-05,
+      "loss": 0.1073,
+      "step": 2171
+    },
+    {
+      "epoch": 1.1372446306967,
+      "loss_partial": 0.052734375,
+      "step": 2171
+    },
+    {
+      "epoch": 1.1372446306967,
+      "loss_partial": 0.050048828125,
+      "step": 2171
+    },
+    {
+      "epoch": 1.1377684651650077,
+      "grad_norm": 2.4260973930358887,
+      "learning_rate": 1.8001674969883414e-05,
+      "loss": 0.1075,
+      "step": 2172
+    },
+    {
+      "epoch": 1.1377684651650077,
+      "loss_partial": 0.05615234375,
+      "step": 2172
+    },
+    {
+      "epoch": 1.1377684651650077,
+      "loss_partial": 0.052734375,
+      "step": 2172
+    },
+    {
+      "epoch": 1.1382922996333158,
+      "grad_norm": 1.9068950414657593,
+      "learning_rate": 1.799893220458196e-05,
+      "loss": 0.1074,
+      "step": 2173
+    },
+    {
+      "epoch": 1.1382922996333158,
+      "loss_partial": 0.05615234375,
+      "step": 2173
+    },
+    {
+      "epoch": 1.1382922996333158,
+      "loss_partial": 0.0537109375,
+      "step": 2173
+    },
+    {
+      "epoch": 1.1388161341016239,
+      "grad_norm": 2.001939058303833,
+      "learning_rate": 1.799618776755314e-05,
+      "loss": 0.1074,
+      "step": 2174
+    },
+    {
+      "epoch": 1.1388161341016239,
+      "loss_partial": 0.0517578125,
+      "step": 2174
+    },
+    {
+      "epoch": 1.1388161341016239,
+      "loss_partial": 0.05029296875,
+      "step": 2174
+    },
+    {
+      "epoch": 1.139339968569932,
+      "grad_norm": 2.3125369548797607,
+      "learning_rate": 1.799344165937052e-05,
+      "loss": 0.1068,
+      "step": 2175
+    },
+    {
+      "epoch": 1.139339968569932,
+      "loss_partial": 0.054443359375,
+      "step": 2175
+    },
+    {
+      "epoch": 1.139339968569932,
+      "loss_partial": 0.048828125,
+      "step": 2175
+    },
+    {
+      "epoch": 1.13986380303824,
+      "grad_norm": 1.6157275438308716,
+      "learning_rate": 1.7990693880608027e-05,
+      "loss": 0.1101,
+      "step": 2176
+    },
+    {
+      "epoch": 1.13986380303824,
+      "loss_partial": 0.0517578125,
+      "step": 2176
+    },
+    {
+      "epoch": 1.13986380303824,
+      "loss_partial": 0.05078125,
+      "step": 2176
+    },
+    {
+      "epoch": 1.140387637506548,
+      "grad_norm": 2.0123472213745117,
+      "learning_rate": 1.7987944431839922e-05,
+      "loss": 0.1083,
+      "step": 2177
+    },
+    {
+      "epoch": 1.140387637506548,
+      "loss_partial": 0.055419921875,
+      "step": 2177
+    },
+    {
+      "epoch": 1.140387637506548,
+      "loss_partial": 0.0751953125,
+      "step": 2177
+    },
+    {
+      "epoch": 1.1409114719748559,
+      "grad_norm": 2.3551106452941895,
+      "learning_rate": 1.798519331364083e-05,
+      "loss": 0.1126,
+      "step": 2178
+    },
+    {
+      "epoch": 1.1409114719748559,
+      "loss_partial": 0.048583984375,
+      "step": 2178
+    },
+    {
+      "epoch": 1.1409114719748559,
+      "loss_partial": 0.055419921875,
+      "step": 2178
+    },
+    {
+      "epoch": 1.141435306443164,
+      "grad_norm": 2.202491283416748,
+      "learning_rate": 1.7982440526585715e-05,
+      "loss": 0.1046,
+      "step": 2179
+    },
+    {
+      "epoch": 1.141435306443164,
+      "loss_partial": 0.055419921875,
+      "step": 2179
+    },
+    {
+      "epoch": 1.141435306443164,
+      "loss_partial": 0.053466796875,
+      "step": 2179
+    },
+    {
+      "epoch": 1.141959140911472,
+      "grad_norm": 1.6969348192214966,
+      "learning_rate": 1.7979686071249886e-05,
+      "loss": 0.1063,
+      "step": 2180
+    },
+    {
+      "epoch": 1.141959140911472,
+      "loss_partial": 0.058349609375,
+      "step": 2180
+    },
+    {
+      "epoch": 1.141959140911472,
+      "loss_partial": 0.05126953125,
+      "step": 2180
+    },
+    {
+      "epoch": 1.14248297537978,
+      "grad_norm": 2.2540297508239746,
+      "learning_rate": 1.797692994820902e-05,
+      "loss": 0.1096,
+      "step": 2181
+    },
+    {
+      "epoch": 1.14248297537978,
+      "loss_partial": 0.061279296875,
+      "step": 2181
+    },
+    {
+      "epoch": 1.14248297537978,
+      "loss_partial": 0.05322265625,
+      "step": 2181
+    },
+    {
+      "epoch": 1.143006809848088,
+      "grad_norm": 1.448001503944397,
+      "learning_rate": 1.797417215803912e-05,
+      "loss": 0.1079,
+      "step": 2182
+    },
+    {
+      "epoch": 1.143006809848088,
+      "loss_partial": 0.049072265625,
+      "step": 2182
+    },
+    {
+      "epoch": 1.143006809848088,
+      "loss_partial": 0.048583984375,
+      "step": 2182
+    },
+    {
+      "epoch": 1.143530644316396,
+      "grad_norm": 1.639121651649475,
+      "learning_rate": 1.797141270131655e-05,
+      "loss": 0.1042,
+      "step": 2183
+    },
+    {
+      "epoch": 1.143530644316396,
+      "loss_partial": 0.0537109375,
+      "step": 2183
+    },
+    {
+      "epoch": 1.143530644316396,
+      "loss_partial": 0.055419921875,
+      "step": 2183
+    },
+    {
+      "epoch": 1.144054478784704,
+      "grad_norm": 2.160104274749756,
+      "learning_rate": 1.796865157861802e-05,
+      "loss": 0.1043,
+      "step": 2184
+    },
+    {
+      "epoch": 1.144054478784704,
+      "loss_partial": 0.056884765625,
+      "step": 2184
+    },
+    {
+      "epoch": 1.144054478784704,
+      "loss_partial": 0.053955078125,
+      "step": 2184
+    },
+    {
+      "epoch": 1.144578313253012,
+      "grad_norm": 1.9554346799850464,
+      "learning_rate": 1.7965888790520583e-05,
+      "loss": 0.1048,
+      "step": 2185
+    },
+    {
+      "epoch": 1.144578313253012,
+      "loss_partial": 0.0595703125,
+      "step": 2185
+    },
+    {
+      "epoch": 1.144578313253012,
+      "loss_partial": 0.050537109375,
+      "step": 2185
+    },
+    {
+      "epoch": 1.1451021477213201,
+      "grad_norm": 2.4492552280426025,
+      "learning_rate": 1.796312433760165e-05,
+      "loss": 0.1082,
+      "step": 2186
+    },
+    {
+      "epoch": 1.1451021477213201,
+      "loss_partial": 0.050048828125,
+      "step": 2186
+    },
+    {
+      "epoch": 1.1451021477213201,
+      "loss_partial": 0.045654296875,
+      "step": 2186
+    },
+    {
+      "epoch": 1.1456259821896282,
+      "grad_norm": 1.7733778953552246,
+      "learning_rate": 1.7960358220438976e-05,
+      "loss": 0.1078,
+      "step": 2187
+    },
+    {
+      "epoch": 1.1456259821896282,
+      "loss_partial": 0.06201171875,
+      "step": 2187
+    },
+    {
+      "epoch": 1.1456259821896282,
+      "loss_partial": 0.054931640625,
+      "step": 2187
+    },
+    {
+      "epoch": 1.146149816657936,
+      "grad_norm": 2.729982614517212,
+      "learning_rate": 1.7957590439610657e-05,
+      "loss": 0.1079,
+      "step": 2188
+    },
+    {
+      "epoch": 1.146149816657936,
+      "loss_partial": 0.046142578125,
+      "step": 2188
+    },
+    {
+      "epoch": 1.146149816657936,
+      "loss_partial": 0.04931640625,
+      "step": 2188
+    },
+    {
+      "epoch": 1.146673651126244,
+      "grad_norm": 1.3292596340179443,
+      "learning_rate": 1.7954820995695145e-05,
+      "loss": 0.1036,
+      "step": 2189
+    },
+    {
+      "epoch": 1.146673651126244,
+      "loss_partial": 0.0595703125,
+      "step": 2189
+    },
+    {
+      "epoch": 1.146673651126244,
+      "loss_partial": 0.04296875,
+      "step": 2189
+    },
+    {
+      "epoch": 1.1471974855945521,
+      "grad_norm": 1.9943974018096924,
+      "learning_rate": 1.7952049889271235e-05,
+      "loss": 0.1086,
+      "step": 2190
+    },
+    {
+      "epoch": 1.1471974855945521,
+      "loss_partial": 0.05078125,
+      "step": 2190
+    },
+    {
+      "epoch": 1.1471974855945521,
+      "loss_partial": 0.04833984375,
+      "step": 2190
+    },
+    {
+      "epoch": 1.1477213200628602,
+      "grad_norm": 1.757269263267517,
+      "learning_rate": 1.7949277120918075e-05,
+      "loss": 0.1079,
+      "step": 2191
+    },
+    {
+      "epoch": 1.1477213200628602,
+      "loss_partial": 0.05322265625,
+      "step": 2191
+    },
+    {
+      "epoch": 1.1477213200628602,
+      "loss_partial": 0.046630859375,
+      "step": 2191
+    },
+    {
+      "epoch": 1.148245154531168,
+      "grad_norm": 2.634608507156372,
+      "learning_rate": 1.794650269121515e-05,
+      "loss": 0.1039,
+      "step": 2192
+    },
+    {
+      "epoch": 1.148245154531168,
+      "loss_partial": 0.0546875,
+      "step": 2192
+    },
+    {
+      "epoch": 1.148245154531168,
+      "loss_partial": 0.06103515625,
+      "step": 2192
+    },
+    {
+      "epoch": 1.148768988999476,
+      "grad_norm": 1.344098687171936,
+      "learning_rate": 1.7943726600742302e-05,
+      "loss": 0.1094,
+      "step": 2193
+    },
+    {
+      "epoch": 1.148768988999476,
+      "loss_partial": 0.0625,
+      "step": 2193
+    },
+    {
+      "epoch": 1.148768988999476,
+      "loss_partial": 0.050537109375,
+      "step": 2193
+    },
+    {
+      "epoch": 1.1492928234677842,
+      "grad_norm": 2.095181465148926,
+      "learning_rate": 1.794094885007972e-05,
+      "loss": 0.1088,
+      "step": 2194
+    },
+    {
+      "epoch": 1.1492928234677842,
+      "loss_partial": 0.04150390625,
+      "step": 2194
+    },
+    {
+      "epoch": 1.1492928234677842,
+      "loss_partial": 0.0498046875,
+      "step": 2194
+    },
+    {
+      "epoch": 1.1498166579360922,
+      "grad_norm": 1.8836239576339722,
+      "learning_rate": 1.793816943980793e-05,
+      "loss": 0.1028,
+      "step": 2195
+    },
+    {
+      "epoch": 1.1498166579360922,
+      "loss_partial": 0.043212890625,
+      "step": 2195
+    },
+    {
+      "epoch": 1.1498166579360922,
+      "loss_partial": 0.046142578125,
+      "step": 2195
+    },
+    {
+      "epoch": 1.1503404924044003,
+      "grad_norm": 1.548179030418396,
+      "learning_rate": 1.7935388370507822e-05,
+      "loss": 0.1066,
+      "step": 2196
+    },
+    {
+      "epoch": 1.1503404924044003,
+      "loss_partial": 0.057373046875,
+      "step": 2196
+    },
+    {
+      "epoch": 1.1503404924044003,
+      "loss_partial": 0.048095703125,
+      "step": 2196
+    },
+    {
+      "epoch": 1.1508643268727083,
+      "grad_norm": 2.2852623462677,
+      "learning_rate": 1.7932605642760607e-05,
+      "loss": 0.1035,
+      "step": 2197
+    },
+    {
+      "epoch": 1.1508643268727083,
+      "loss_partial": 0.05517578125,
+      "step": 2197
+    },
+    {
+      "epoch": 1.1508643268727083,
+      "loss_partial": 0.05029296875,
+      "step": 2197
+    },
+    {
+      "epoch": 1.1513881613410162,
+      "grad_norm": 1.997554898262024,
+      "learning_rate": 1.7929821257147873e-05,
+      "loss": 0.1042,
+      "step": 2198
+    },
+    {
+      "epoch": 1.1513881613410162,
+      "loss_partial": 0.04296875,
+      "step": 2198
+    },
+    {
+      "epoch": 1.1513881613410162,
+      "loss_partial": 0.052734375,
+      "step": 2198
+    },
+    {
+      "epoch": 1.1519119958093242,
+      "grad_norm": 2.2841150760650635,
+      "learning_rate": 1.7927035214251528e-05,
+      "loss": 0.1097,
+      "step": 2199
+    },
+    {
+      "epoch": 1.1519119958093242,
+      "loss_partial": 0.055419921875,
+      "step": 2199
+    },
+    {
+      "epoch": 1.1519119958093242,
+      "loss_partial": 0.047119140625,
+      "step": 2199
+    },
+    {
+      "epoch": 1.1524358302776323,
+      "grad_norm": 1.9996417760849,
+      "learning_rate": 1.7924247514653846e-05,
+      "loss": 0.1067,
+      "step": 2200
+    },
+    {
+      "epoch": 1.1524358302776323,
+      "loss_partial": 0.050537109375,
+      "step": 2200
+    },
+    {
+      "epoch": 1.1524358302776323,
+      "loss_partial": 0.051025390625,
+      "step": 2200
+    },
+    {
+      "epoch": 1.1529596647459404,
+      "grad_norm": 1.3909848928451538,
+      "learning_rate": 1.7921458158937435e-05,
+      "loss": 0.1033,
+      "step": 2201
+    },
+    {
+      "epoch": 1.1529596647459404,
+      "loss_partial": 0.052001953125,
+      "step": 2201
+    },
+    {
+      "epoch": 1.1529596647459404,
+      "loss_partial": 0.05224609375,
+      "step": 2201
+    },
+    {
+      "epoch": 1.1534834992142482,
+      "grad_norm": 1.9088151454925537,
+      "learning_rate": 1.791866714768525e-05,
+      "loss": 0.112,
+      "step": 2202
+    },
+    {
+      "epoch": 1.1534834992142482,
+      "loss_partial": 0.059814453125,
+      "step": 2202
+    },
+    {
+      "epoch": 1.1534834992142482,
+      "loss_partial": 0.0625,
+      "step": 2202
+    },
+    {
+      "epoch": 1.1540073336825563,
+      "grad_norm": 1.6230601072311401,
+      "learning_rate": 1.7915874481480603e-05,
+      "loss": 0.1095,
+      "step": 2203
+    },
+    {
+      "epoch": 1.1540073336825563,
+      "loss_partial": 0.048583984375,
+      "step": 2203
+    },
+    {
+      "epoch": 1.1540073336825563,
+      "loss_partial": 0.07470703125,
+      "step": 2203
+    },
+    {
+      "epoch": 1.1545311681508643,
+      "grad_norm": 1.9432939291000366,
+      "learning_rate": 1.791308016090714e-05,
+      "loss": 0.1095,
+      "step": 2204
+    },
+    {
+      "epoch": 1.1545311681508643,
+      "loss_partial": 0.046142578125,
+      "step": 2204
+    },
+    {
+      "epoch": 1.1545311681508643,
+      "loss_partial": 0.05712890625,
+      "step": 2204
+    },
+    {
+      "epoch": 1.1550550026191724,
+      "grad_norm": 2.0192179679870605,
+      "learning_rate": 1.7910284186548856e-05,
+      "loss": 0.1102,
+      "step": 2205
+    },
+    {
+      "epoch": 1.1550550026191724,
+      "loss_partial": 0.048583984375,
+      "step": 2205
+    },
+    {
+      "epoch": 1.1550550026191724,
+      "loss_partial": 0.05029296875,
+      "step": 2205
+    },
+    {
+      "epoch": 1.1555788370874804,
+      "grad_norm": 1.6169761419296265,
+      "learning_rate": 1.790748655899009e-05,
+      "loss": 0.1043,
+      "step": 2206
+    },
+    {
+      "epoch": 1.1555788370874804,
+      "loss_partial": 0.060791015625,
+      "step": 2206
+    },
+    {
+      "epoch": 1.1555788370874804,
+      "loss_partial": 0.0576171875,
+      "step": 2206
+    },
+    {
+      "epoch": 1.1561026715557883,
+      "grad_norm": 2.254611015319824,
+      "learning_rate": 1.790468727881553e-05,
+      "loss": 0.1059,
+      "step": 2207
+    },
+    {
+      "epoch": 1.1561026715557883,
+      "loss_partial": 0.058349609375,
+      "step": 2207
+    },
+    {
+      "epoch": 1.1561026715557883,
+      "loss_partial": 0.058349609375,
+      "step": 2207
+    },
+    {
+      "epoch": 1.1566265060240963,
+      "grad_norm": 1.5392755270004272,
+      "learning_rate": 1.7901886346610216e-05,
+      "loss": 0.1082,
+      "step": 2208
+    },
+    {
+      "epoch": 1.1566265060240963,
+      "loss_partial": 0.04638671875,
+      "step": 2208
+    },
+    {
+      "epoch": 1.1566265060240963,
+      "loss_partial": 0.0419921875,
+      "step": 2208
+    },
+    {
+      "epoch": 1.1571503404924044,
+      "grad_norm": 2.0073626041412354,
+      "learning_rate": 1.7899083762959516e-05,
+      "loss": 0.1068,
+      "step": 2209
+    },
+    {
+      "epoch": 1.1571503404924044,
+      "loss_partial": 0.043701171875,
+      "step": 2209
+    },
+    {
+      "epoch": 1.1571503404924044,
+      "loss_partial": 0.048828125,
+      "step": 2209
+    },
+    {
+      "epoch": 1.1576741749607125,
+      "grad_norm": 1.643103837966919,
+      "learning_rate": 1.7896279528449157e-05,
+      "loss": 0.1045,
+      "step": 2210
+    },
+    {
+      "epoch": 1.1576741749607125,
+      "loss_partial": 0.0458984375,
+      "step": 2210
+    },
+    {
+      "epoch": 1.1576741749607125,
+      "loss_partial": 0.052001953125,
+      "step": 2210
+    },
+    {
+      "epoch": 1.1581980094290205,
+      "grad_norm": 1.409690260887146,
+      "learning_rate": 1.7893473643665206e-05,
+      "loss": 0.1049,
+      "step": 2211
+    },
+    {
+      "epoch": 1.1581980094290205,
+      "loss_partial": 0.045654296875,
+      "step": 2211
+    },
+    {
+      "epoch": 1.1581980094290205,
+      "loss_partial": 0.053955078125,
+      "step": 2211
+    },
+    {
+      "epoch": 1.1587218438973284,
+      "grad_norm": 2.509394407272339,
+      "learning_rate": 1.789066610919407e-05,
+      "loss": 0.1073,
+      "step": 2212
+    },
+    {
+      "epoch": 1.1587218438973284,
+      "loss_partial": 0.060791015625,
+      "step": 2212
+    },
+    {
+      "epoch": 1.1587218438973284,
+      "loss_partial": 0.06640625,
+      "step": 2212
+    },
+    {
+      "epoch": 1.1592456783656364,
+      "grad_norm": 2.4378268718719482,
+      "learning_rate": 1.788785692562252e-05,
+      "loss": 0.1128,
+      "step": 2213
+    },
+    {
+      "epoch": 1.1592456783656364,
+      "loss_partial": 0.055908203125,
+      "step": 2213
+    },
+    {
+      "epoch": 1.1592456783656364,
+      "loss_partial": 0.0556640625,
+      "step": 2213
+    },
+    {
+      "epoch": 1.1597695128339445,
+      "grad_norm": 1.7000328302383423,
+      "learning_rate": 1.788504609353764e-05,
+      "loss": 0.1096,
+      "step": 2214
+    },
+    {
+      "epoch": 1.1597695128339445,
+      "loss_partial": 0.04443359375,
+      "step": 2214
+    },
+    {
+      "epoch": 1.1597695128339445,
+      "loss_partial": 0.048095703125,
+      "step": 2214
+    },
+    {
+      "epoch": 1.1602933473022525,
+      "grad_norm": 2.6112565994262695,
+      "learning_rate": 1.7882233613526896e-05,
+      "loss": 0.1099,
+      "step": 2215
+    },
+    {
+      "epoch": 1.1602933473022525,
+      "loss_partial": 0.05419921875,
+      "step": 2215
+    },
+    {
+      "epoch": 1.1602933473022525,
+      "loss_partial": 0.0576171875,
+      "step": 2215
+    },
+    {
+      "epoch": 1.1608171817705606,
+      "grad_norm": 2.111936569213867,
+      "learning_rate": 1.7879419486178062e-05,
+      "loss": 0.1042,
+      "step": 2216
+    },
+    {
+      "epoch": 1.1608171817705606,
+      "loss_partial": 0.049072265625,
+      "step": 2216
+    },
+    {
+      "epoch": 1.1608171817705606,
+      "loss_partial": 0.055419921875,
+      "step": 2216
+    },
+    {
+      "epoch": 1.1613410162388684,
+      "grad_norm": 2.118408679962158,
+      "learning_rate": 1.7876603712079283e-05,
+      "loss": 0.1112,
+      "step": 2217
+    },
+    {
+      "epoch": 1.1613410162388684,
+      "loss_partial": 0.057373046875,
+      "step": 2217
+    },
+    {
+      "epoch": 1.1613410162388684,
+      "loss_partial": 0.04345703125,
+      "step": 2217
+    },
+    {
+      "epoch": 1.1618648507071765,
+      "grad_norm": 2.114126443862915,
+      "learning_rate": 1.7873786291819035e-05,
+      "loss": 0.1078,
+      "step": 2218
+    },
+    {
+      "epoch": 1.1618648507071765,
+      "loss_partial": 0.04052734375,
+      "step": 2218
+    },
+    {
+      "epoch": 1.1618648507071765,
+      "loss_partial": 0.07568359375,
+      "step": 2218
+    },
+    {
+      "epoch": 1.1623886851754845,
+      "grad_norm": 1.8282406330108643,
+      "learning_rate": 1.7870967225986146e-05,
+      "loss": 0.1059,
+      "step": 2219
+    },
+    {
+      "epoch": 1.1623886851754845,
+      "loss_partial": 0.050537109375,
+      "step": 2219
+    },
+    {
+      "epoch": 1.1623886851754845,
+      "loss_partial": 0.052978515625,
+      "step": 2219
+    },
+    {
+      "epoch": 1.1629125196437926,
+      "grad_norm": 2.281447410583496,
+      "learning_rate": 1.7868146515169774e-05,
+      "loss": 0.1078,
+      "step": 2220
+    },
+    {
+      "epoch": 1.1629125196437926,
+      "loss_partial": 0.050048828125,
+      "step": 2220
+    },
+    {
+      "epoch": 1.1629125196437926,
+      "loss_partial": 0.054931640625,
+      "step": 2220
+    },
+    {
+      "epoch": 1.1634363541121007,
+      "grad_norm": 1.8363308906555176,
+      "learning_rate": 1.7865324159959437e-05,
+      "loss": 0.1069,
+      "step": 2221
+    },
+    {
+      "epoch": 1.1634363541121007,
+      "loss_partial": 0.0556640625,
+      "step": 2221
+    },
+    {
+      "epoch": 1.1634363541121007,
+      "loss_partial": 0.06103515625,
+      "step": 2221
+    },
+    {
+      "epoch": 1.1639601885804085,
+      "grad_norm": 1.7892295122146606,
+      "learning_rate": 1.786250016094499e-05,
+      "loss": 0.1086,
+      "step": 2222
+    },
+    {
+      "epoch": 1.1639601885804085,
+      "loss_partial": 0.056884765625,
+      "step": 2222
+    },
+    {
+      "epoch": 1.1639601885804085,
+      "loss_partial": 0.04931640625,
+      "step": 2222
+    },
+    {
+      "epoch": 1.1644840230487166,
+      "grad_norm": 1.6800442934036255,
+      "learning_rate": 1.7859674518716624e-05,
+      "loss": 0.1097,
+      "step": 2223
+    },
+    {
+      "epoch": 1.1644840230487166,
+      "loss_partial": 0.054931640625,
+      "step": 2223
+    },
+    {
+      "epoch": 1.1644840230487166,
+      "loss_partial": 0.044921875,
+      "step": 2223
+    },
+    {
+      "epoch": 1.1650078575170246,
+      "grad_norm": 2.0552146434783936,
+      "learning_rate": 1.7856847233864892e-05,
+      "loss": 0.1064,
+      "step": 2224
+    },
+    {
+      "epoch": 1.1650078575170246,
+      "loss_partial": 0.059814453125,
+      "step": 2224
+    },
+    {
+      "epoch": 1.1650078575170246,
+      "loss_partial": 0.05078125,
+      "step": 2224
+    },
+    {
+      "epoch": 1.1655316919853327,
+      "grad_norm": 1.7386776208877563,
+      "learning_rate": 1.7854018306980672e-05,
+      "loss": 0.1053,
+      "step": 2225
+    },
+    {
+      "epoch": 1.1655316919853327,
+      "loss_partial": 0.052734375,
+      "step": 2225
+    },
+    {
+      "epoch": 1.1655316919853327,
+      "loss_partial": 0.05078125,
+      "step": 2225
+    },
+    {
+      "epoch": 1.1660555264536407,
+      "grad_norm": 2.1497905254364014,
+      "learning_rate": 1.7851187738655194e-05,
+      "loss": 0.1135,
+      "step": 2226
+    },
+    {
+      "epoch": 1.1660555264536407,
+      "loss_partial": 0.052490234375,
+      "step": 2226
+    },
+    {
+      "epoch": 1.1660555264536407,
+      "loss_partial": 0.052490234375,
+      "step": 2226
+    },
+    {
+      "epoch": 1.1665793609219486,
+      "grad_norm": 1.7205687761306763,
+      "learning_rate": 1.7848355529480028e-05,
+      "loss": 0.1075,
+      "step": 2227
+    },
+    {
+      "epoch": 1.1665793609219486,
+      "loss_partial": 0.046142578125,
+      "step": 2227
+    },
+    {
+      "epoch": 1.1665793609219486,
+      "loss_partial": 0.04541015625,
+      "step": 2227
+    },
+    {
+      "epoch": 1.1671031953902566,
+      "grad_norm": 1.7003082036972046,
+      "learning_rate": 1.784552168004709e-05,
+      "loss": 0.1088,
+      "step": 2228
+    },
+    {
+      "epoch": 1.1671031953902566,
+      "loss_partial": 0.06494140625,
+      "step": 2228
+    },
+    {
+      "epoch": 1.1671031953902566,
+      "loss_partial": 0.05078125,
+      "step": 2228
+    },
+    {
+      "epoch": 1.1676270298585647,
+      "grad_norm": 2.0001821517944336,
+      "learning_rate": 1.7842686190948636e-05,
+      "loss": 0.1057,
+      "step": 2229
+    },
+    {
+      "epoch": 1.1676270298585647,
+      "loss_partial": 0.048828125,
+      "step": 2229
+    },
+    {
+      "epoch": 1.1676270298585647,
+      "loss_partial": 0.0517578125,
+      "step": 2229
+    },
+    {
+      "epoch": 1.1681508643268728,
+      "grad_norm": 1.472770094871521,
+      "learning_rate": 1.7839849062777266e-05,
+      "loss": 0.1062,
+      "step": 2230
+    },
+    {
+      "epoch": 1.1681508643268728,
+      "loss_partial": 0.048583984375,
+      "step": 2230
+    },
+    {
+      "epoch": 1.1681508643268728,
+      "loss_partial": 0.06689453125,
+      "step": 2230
+    },
+    {
+      "epoch": 1.1686746987951806,
+      "grad_norm": 1.9872453212738037,
+      "learning_rate": 1.783701029612592e-05,
+      "loss": 0.1083,
+      "step": 2231
+    },
+    {
+      "epoch": 1.1686746987951806,
+      "loss_partial": 0.0458984375,
+      "step": 2231
+    },
+    {
+      "epoch": 1.1686746987951806,
+      "loss_partial": 0.0478515625,
+      "step": 2231
+    },
+    {
+      "epoch": 1.1691985332634887,
+      "grad_norm": 1.3538129329681396,
+      "learning_rate": 1.7834169891587888e-05,
+      "loss": 0.1047,
+      "step": 2232
+    },
+    {
+      "epoch": 1.1691985332634887,
+      "loss_partial": 0.042724609375,
+      "step": 2232
+    },
+    {
+      "epoch": 1.1691985332634887,
+      "loss_partial": 0.06884765625,
+      "step": 2232
+    },
+    {
+      "epoch": 1.1697223677317967,
+      "grad_norm": 1.9453625679016113,
+      "learning_rate": 1.783132784975679e-05,
+      "loss": 0.1068,
+      "step": 2233
+    },
+    {
+      "epoch": 1.1697223677317967,
+      "loss_partial": 0.058837890625,
+      "step": 2233
+    },
+    {
+      "epoch": 1.1697223677317967,
+      "loss_partial": 0.05029296875,
+      "step": 2233
+    },
+    {
+      "epoch": 1.1702462022001048,
+      "grad_norm": 1.7129507064819336,
+      "learning_rate": 1.7828484171226602e-05,
+      "loss": 0.1066,
+      "step": 2234
+    },
+    {
+      "epoch": 1.1702462022001048,
+      "loss_partial": 0.054443359375,
+      "step": 2234
+    },
+    {
+      "epoch": 1.1702462022001048,
+      "loss_partial": 0.051025390625,
+      "step": 2234
+    },
+    {
+      "epoch": 1.1707700366684128,
+      "grad_norm": 1.4261974096298218,
+      "learning_rate": 1.782563885659163e-05,
+      "loss": 0.1051,
+      "step": 2235
+    },
+    {
+      "epoch": 1.1707700366684128,
+      "loss_partial": 0.0625,
+      "step": 2235
+    },
+    {
+      "epoch": 1.1707700366684128,
+      "loss_partial": 0.05908203125,
+      "step": 2235
+    },
+    {
+      "epoch": 1.171293871136721,
+      "grad_norm": 1.927595615386963,
+      "learning_rate": 1.7822791906446532e-05,
+      "loss": 0.1057,
+      "step": 2236
+    },
+    {
+      "epoch": 1.171293871136721,
+      "loss_partial": 0.0556640625,
+      "step": 2236
+    },
+    {
+      "epoch": 1.171293871136721,
+      "loss_partial": 0.050048828125,
+      "step": 2236
+    },
+    {
+      "epoch": 1.1718177056050287,
+      "grad_norm": 1.7795466184616089,
+      "learning_rate": 1.7819943321386295e-05,
+      "loss": 0.1106,
+      "step": 2237
+    },
+    {
+      "epoch": 1.1718177056050287,
+      "loss_partial": 0.05810546875,
+      "step": 2237
+    },
+    {
+      "epoch": 1.1718177056050287,
+      "loss_partial": 0.0537109375,
+      "step": 2237
+    },
+    {
+      "epoch": 1.1723415400733368,
+      "grad_norm": 2.280431032180786,
+      "learning_rate": 1.7817093102006264e-05,
+      "loss": 0.1064,
+      "step": 2238
+    },
+    {
+      "epoch": 1.1723415400733368,
+      "loss_partial": 0.05908203125,
+      "step": 2238
+    },
+    {
+      "epoch": 1.1723415400733368,
+      "loss_partial": 0.06494140625,
+      "step": 2238
+    },
+    {
+      "epoch": 1.1728653745416449,
+      "grad_norm": 2.41829514503479,
+      "learning_rate": 1.7814241248902114e-05,
+      "loss": 0.1106,
+      "step": 2239
+    },
+    {
+      "epoch": 1.1728653745416449,
+      "loss_partial": 0.052001953125,
+      "step": 2239
+    },
+    {
+      "epoch": 1.1728653745416449,
+      "loss_partial": 0.048095703125,
+      "step": 2239
+    },
+    {
+      "epoch": 1.173389209009953,
+      "grad_norm": 1.6544206142425537,
+      "learning_rate": 1.7811387762669866e-05,
+      "loss": 0.1107,
+      "step": 2240
+    },
+    {
+      "epoch": 1.173389209009953,
+      "loss_partial": 0.05322265625,
+      "step": 2240
+    },
+    {
+      "epoch": 1.173389209009953,
+      "loss_partial": 0.062255859375,
+      "step": 2240
+    },
+    {
+      "epoch": 1.1739130434782608,
+      "grad_norm": 1.938385248184204,
+      "learning_rate": 1.780853264390588e-05,
+      "loss": 0.1044,
+      "step": 2241
+    },
+    {
+      "epoch": 1.1739130434782608,
+      "loss_partial": 0.0517578125,
+      "step": 2241
+    },
+    {
+      "epoch": 1.1739130434782608,
+      "loss_partial": 0.060791015625,
+      "step": 2241
+    },
+    {
+      "epoch": 1.1744368779465688,
+      "grad_norm": 2.894312858581543,
+      "learning_rate": 1.7805675893206857e-05,
+      "loss": 0.109,
+      "step": 2242
+    },
+    {
+      "epoch": 1.1744368779465688,
+      "loss_partial": 0.046630859375,
+      "step": 2242
+    },
+    {
+      "epoch": 1.1744368779465688,
+      "loss_partial": 0.052001953125,
+      "step": 2242
+    },
+    {
+      "epoch": 1.1749607124148769,
+      "grad_norm": 1.6722819805145264,
+      "learning_rate": 1.780281751116984e-05,
+      "loss": 0.1043,
+      "step": 2243
+    },
+    {
+      "epoch": 1.1749607124148769,
+      "loss_partial": 0.06787109375,
+      "step": 2243
+    },
+    {
+      "epoch": 1.1749607124148769,
+      "loss_partial": 0.05126953125,
+      "step": 2243
+    },
+    {
+      "epoch": 1.175484546883185,
+      "grad_norm": 1.686635971069336,
+      "learning_rate": 1.7799957498392216e-05,
+      "loss": 0.1062,
+      "step": 2244
+    },
+    {
+      "epoch": 1.175484546883185,
+      "loss_partial": 0.05078125,
+      "step": 2244
+    },
+    {
+      "epoch": 1.175484546883185,
+      "loss_partial": 0.06201171875,
+      "step": 2244
+    },
+    {
+      "epoch": 1.176008381351493,
+      "grad_norm": 2.206735372543335,
+      "learning_rate": 1.779709585547171e-05,
+      "loss": 0.1046,
+      "step": 2245
+    },
+    {
+      "epoch": 1.176008381351493,
+      "loss_partial": 0.04541015625,
+      "step": 2245
+    },
+    {
+      "epoch": 1.176008381351493,
+      "loss_partial": 0.061279296875,
+      "step": 2245
+    },
+    {
+      "epoch": 1.176532215819801,
+      "grad_norm": 2.6197433471679688,
+      "learning_rate": 1.7794232583006385e-05,
+      "loss": 0.1093,
+      "step": 2246
+    },
+    {
+      "epoch": 1.176532215819801,
+      "loss_partial": 0.054931640625,
+      "step": 2246
+    },
+    {
+      "epoch": 1.176532215819801,
+      "loss_partial": 0.057861328125,
+      "step": 2246
+    },
+    {
+      "epoch": 1.177056050288109,
+      "grad_norm": 1.5924808979034424,
+      "learning_rate": 1.7791367681594643e-05,
+      "loss": 0.1116,
+      "step": 2247
+    },
+    {
+      "epoch": 1.177056050288109,
+      "loss_partial": 0.05078125,
+      "step": 2247
+    },
+    {
+      "epoch": 1.177056050288109,
+      "loss_partial": 0.056396484375,
+      "step": 2247
+    },
+    {
+      "epoch": 1.177579884756417,
+      "grad_norm": 2.6449406147003174,
+      "learning_rate": 1.778850115183524e-05,
+      "loss": 0.1079,
+      "step": 2248
+    },
+    {
+      "epoch": 1.177579884756417,
+      "loss_partial": 0.07080078125,
+      "step": 2248
+    },
+    {
+      "epoch": 1.177579884756417,
+      "loss_partial": 0.04931640625,
+      "step": 2248
+    },
+    {
+      "epoch": 1.178103719224725,
+      "grad_norm": 1.9933537244796753,
+      "learning_rate": 1.778563299432726e-05,
+      "loss": 0.1061,
+      "step": 2249
+    },
+    {
+      "epoch": 1.178103719224725,
+      "loss_partial": 0.052978515625,
+      "step": 2249
+    },
+    {
+      "epoch": 1.178103719224725,
+      "loss_partial": 0.054443359375,
+      "step": 2249
+    },
+    {
+      "epoch": 1.178627553693033,
+      "grad_norm": 2.918323278427124,
+      "learning_rate": 1.7782763209670125e-05,
+      "loss": 0.1108,
+      "step": 2250
+    },
+    {
+      "epoch": 1.178627553693033,
+      "loss_partial": 0.047119140625,
+      "step": 2250
+    },
+    {
+      "epoch": 1.178627553693033,
+      "loss_partial": 0.052734375,
+      "step": 2250
+    },
+    {
+      "epoch": 1.179151388161341,
+      "grad_norm": 2.0788064002990723,
+      "learning_rate": 1.777989179846361e-05,
+      "loss": 0.1063,
+      "step": 2251
+    },
+    {
+      "epoch": 1.179151388161341,
+      "loss_partial": 0.057373046875,
+      "step": 2251
+    },
+    {
+      "epoch": 1.179151388161341,
+      "loss_partial": 0.0537109375,
+      "step": 2251
+    },
+    {
+      "epoch": 1.179675222629649,
+      "grad_norm": 2.293046474456787,
+      "learning_rate": 1.7777018761307818e-05,
+      "loss": 0.1086,
+      "step": 2252
+    },
+    {
+      "epoch": 1.179675222629649,
+      "loss_partial": 0.046630859375,
+      "step": 2252
+    },
+    {
+      "epoch": 1.179675222629649,
+      "loss_partial": 0.06884765625,
+      "step": 2252
+    },
+    {
+      "epoch": 1.180199057097957,
+      "grad_norm": 1.801487684249878,
+      "learning_rate": 1.77741440988032e-05,
+      "loss": 0.1111,
+      "step": 2253
+    },
+    {
+      "epoch": 1.180199057097957,
+      "loss_partial": 0.04296875,
+      "step": 2253
+    },
+    {
+      "epoch": 1.180199057097957,
+      "loss_partial": 0.0576171875,
+      "step": 2253
+    },
+    {
+      "epoch": 1.180722891566265,
+      "grad_norm": 2.3079612255096436,
+      "learning_rate": 1.777126781155053e-05,
+      "loss": 0.1071,
+      "step": 2254
+    },
+    {
+      "epoch": 1.180722891566265,
+      "loss_partial": 0.045166015625,
+      "step": 2254
+    },
+    {
+      "epoch": 1.180722891566265,
+      "loss_partial": 0.054931640625,
+      "step": 2254
+    },
+    {
+      "epoch": 1.1812467260345731,
+      "grad_norm": 1.6397074460983276,
+      "learning_rate": 1.776838990015095e-05,
+      "loss": 0.1061,
+      "step": 2255
+    },
+    {
+      "epoch": 1.1812467260345731,
+      "loss_partial": 0.05712890625,
+      "step": 2255
+    },
+    {
+      "epoch": 1.1812467260345731,
+      "loss_partial": 0.053466796875,
+      "step": 2255
+    },
+    {
+      "epoch": 1.1817705605028812,
+      "grad_norm": 2.2472023963928223,
+      "learning_rate": 1.776551036520592e-05,
+      "loss": 0.1084,
+      "step": 2256
+    },
+    {
+      "epoch": 1.1817705605028812,
+      "loss_partial": 0.044921875,
+      "step": 2256
+    },
+    {
+      "epoch": 1.1817705605028812,
+      "loss_partial": 0.046630859375,
+      "step": 2256
+    },
+    {
+      "epoch": 1.182294394971189,
+      "grad_norm": 1.4854799509048462,
+      "learning_rate": 1.7762629207317244e-05,
+      "loss": 0.1073,
+      "step": 2257
+    },
+    {
+      "epoch": 1.182294394971189,
+      "loss_partial": 0.0537109375,
+      "step": 2257
+    },
+    {
+      "epoch": 1.182294394971189,
+      "loss_partial": 0.0478515625,
+      "step": 2257
+    },
+    {
+      "epoch": 1.182818229439497,
+      "grad_norm": 2.399602174758911,
+      "learning_rate": 1.7759746427087064e-05,
+      "loss": 0.1059,
+      "step": 2258
+    },
+    {
+      "epoch": 1.182818229439497,
+      "loss_partial": 0.041748046875,
+      "step": 2258
+    },
+    {
+      "epoch": 1.182818229439497,
+      "loss_partial": 0.059814453125,
+      "step": 2258
+    },
+    {
+      "epoch": 1.1833420639078052,
+      "grad_norm": 1.7280348539352417,
+      "learning_rate": 1.7756862025117866e-05,
+      "loss": 0.1095,
+      "step": 2259
+    },
+    {
+      "epoch": 1.1833420639078052,
+      "loss_partial": 0.047119140625,
+      "step": 2259
+    },
+    {
+      "epoch": 1.1833420639078052,
+      "loss_partial": 0.052978515625,
+      "step": 2259
+    },
+    {
+      "epoch": 1.1838658983761132,
+      "grad_norm": 1.9814295768737793,
+      "learning_rate": 1.775397600201247e-05,
+      "loss": 0.1096,
+      "step": 2260
+    },
+    {
+      "epoch": 1.1838658983761132,
+      "loss_partial": 0.057373046875,
+      "step": 2260
+    },
+    {
+      "epoch": 1.1838658983761132,
+      "loss_partial": 0.056640625,
+      "step": 2260
+    },
+    {
+      "epoch": 1.184389732844421,
+      "grad_norm": 1.4456642866134644,
+      "learning_rate": 1.7751088358374043e-05,
+      "loss": 0.1062,
+      "step": 2261
+    },
+    {
+      "epoch": 1.184389732844421,
+      "loss_partial": 0.05126953125,
+      "step": 2261
+    },
+    {
+      "epoch": 1.184389732844421,
+      "loss_partial": 0.0546875,
+      "step": 2261
+    },
+    {
+      "epoch": 1.1849135673127291,
+      "grad_norm": 1.8419991731643677,
+      "learning_rate": 1.7748199094806075e-05,
+      "loss": 0.1065,
+      "step": 2262
+    },
+    {
+      "epoch": 1.1849135673127291,
+      "loss_partial": 0.05029296875,
+      "step": 2262
+    },
+    {
+      "epoch": 1.1849135673127291,
+      "loss_partial": 0.0478515625,
+      "step": 2262
+    },
+    {
+      "epoch": 1.1854374017810372,
+      "grad_norm": 1.5829871892929077,
+      "learning_rate": 1.7745308211912413e-05,
+      "loss": 0.1064,
+      "step": 2263
+    },
+    {
+      "epoch": 1.1854374017810372,
+      "loss_partial": 0.05029296875,
+      "step": 2263
+    },
+    {
+      "epoch": 1.1854374017810372,
+      "loss_partial": 0.0546875,
+      "step": 2263
+    },
+    {
+      "epoch": 1.1859612362493452,
+      "grad_norm": 1.541068196296692,
+      "learning_rate": 1.774241571029723e-05,
+      "loss": 0.1075,
+      "step": 2264
+    },
+    {
+      "epoch": 1.1859612362493452,
+      "loss_partial": 0.04443359375,
+      "step": 2264
+    },
+    {
+      "epoch": 1.1859612362493452,
+      "loss_partial": 0.054931640625,
+      "step": 2264
+    },
+    {
+      "epoch": 1.1864850707176533,
+      "grad_norm": 1.7174986600875854,
+      "learning_rate": 1.773952159056504e-05,
+      "loss": 0.1083,
+      "step": 2265
+    },
+    {
+      "epoch": 1.1864850707176533,
+      "loss_partial": 0.059814453125,
+      "step": 2265
+    },
+    {
+      "epoch": 1.1864850707176533,
+      "loss_partial": 0.0615234375,
+      "step": 2265
+    },
+    {
+      "epoch": 1.1870089051859614,
+      "grad_norm": 2.5519378185272217,
+      "learning_rate": 1.77366258533207e-05,
+      "loss": 0.1095,
+      "step": 2266
+    },
+    {
+      "epoch": 1.1870089051859614,
+      "loss_partial": 0.0576171875,
+      "step": 2266
+    },
+    {
+      "epoch": 1.1870089051859614,
+      "loss_partial": 0.051025390625,
+      "step": 2266
+    },
+    {
+      "epoch": 1.1875327396542692,
+      "grad_norm": 2.021193504333496,
+      "learning_rate": 1.7733728499169397e-05,
+      "loss": 0.1082,
+      "step": 2267
+    },
+    {
+      "epoch": 1.1875327396542692,
+      "loss_partial": 0.04833984375,
+      "step": 2267
+    },
+    {
+      "epoch": 1.1875327396542692,
+      "loss_partial": 0.056640625,
+      "step": 2267
+    },
+    {
+      "epoch": 1.1880565741225773,
+      "grad_norm": 1.769669532775879,
+      "learning_rate": 1.7730829528716664e-05,
+      "loss": 0.1092,
+      "step": 2268
+    },
+    {
+      "epoch": 1.1880565741225773,
+      "loss_partial": 0.05712890625,
+      "step": 2268
+    },
+    {
+      "epoch": 1.1880565741225773,
+      "loss_partial": 0.051025390625,
+      "step": 2268
+    },
+    {
+      "epoch": 1.1885804085908853,
+      "grad_norm": 1.949404239654541,
+      "learning_rate": 1.772792894256836e-05,
+      "loss": 0.1014,
+      "step": 2269
+    },
+    {
+      "epoch": 1.1885804085908853,
+      "loss_partial": 0.0673828125,
+      "step": 2269
+    },
+    {
+      "epoch": 1.1885804085908853,
+      "loss_partial": 0.0458984375,
+      "step": 2269
+    },
+    {
+      "epoch": 1.1891042430591934,
+      "grad_norm": 2.1509127616882324,
+      "learning_rate": 1.7725026741330698e-05,
+      "loss": 0.1075,
+      "step": 2270
+    },
+    {
+      "epoch": 1.1891042430591934,
+      "loss_partial": 0.059326171875,
+      "step": 2270
+    },
+    {
+      "epoch": 1.1891042430591934,
+      "loss_partial": 0.048828125,
+      "step": 2270
+    },
+    {
+      "epoch": 1.1896280775275012,
+      "grad_norm": 1.850333333015442,
+      "learning_rate": 1.772212292561022e-05,
+      "loss": 0.1067,
+      "step": 2271
+    },
+    {
+      "epoch": 1.1896280775275012,
+      "loss_partial": 0.05078125,
+      "step": 2271
+    },
+    {
+      "epoch": 1.1896280775275012,
+      "loss_partial": 0.052978515625,
+      "step": 2271
+    },
+    {
+      "epoch": 1.1901519119958093,
+      "grad_norm": 2.189124822616577,
+      "learning_rate": 1.7719217496013798e-05,
+      "loss": 0.1067,
+      "step": 2272
+    },
+    {
+      "epoch": 1.1901519119958093,
+      "loss_partial": 0.04736328125,
+      "step": 2272
+    },
+    {
+      "epoch": 1.1901519119958093,
+      "loss_partial": 0.05224609375,
+      "step": 2272
+    },
+    {
+      "epoch": 1.1906757464641173,
+      "grad_norm": 2.6592519283294678,
+      "learning_rate": 1.7716310453148657e-05,
+      "loss": 0.1064,
+      "step": 2273
+    },
+    {
+      "epoch": 1.1906757464641173,
+      "loss_partial": 0.0615234375,
+      "step": 2273
+    },
+    {
+      "epoch": 1.1906757464641173,
+      "loss_partial": 0.05126953125,
+      "step": 2273
+    },
+    {
+      "epoch": 1.1911995809324254,
+      "grad_norm": 2.1931886672973633,
+      "learning_rate": 1.7713401797622346e-05,
+      "loss": 0.1088,
+      "step": 2274
+    },
+    {
+      "epoch": 1.1911995809324254,
+      "loss_partial": 0.04736328125,
+      "step": 2274
+    },
+    {
+      "epoch": 1.1911995809324254,
+      "loss_partial": 0.052001953125,
+      "step": 2274
+    },
+    {
+      "epoch": 1.1917234154007335,
+      "grad_norm": 2.1364352703094482,
+      "learning_rate": 1.7710491530042754e-05,
+      "loss": 0.1092,
+      "step": 2275
+    },
+    {
+      "epoch": 1.1917234154007335,
+      "loss_partial": 0.051513671875,
+      "step": 2275
+    },
+    {
+      "epoch": 1.1917234154007335,
+      "loss_partial": 0.049560546875,
+      "step": 2275
+    },
+    {
+      "epoch": 1.1922472498690413,
+      "grad_norm": 2.0545995235443115,
+      "learning_rate": 1.7707579651018118e-05,
+      "loss": 0.1093,
+      "step": 2276
+    },
+    {
+      "epoch": 1.1922472498690413,
+      "loss_partial": 0.0478515625,
+      "step": 2276
+    },
+    {
+      "epoch": 1.1922472498690413,
+      "loss_partial": 0.05029296875,
+      "step": 2276
+    },
+    {
+      "epoch": 1.1927710843373494,
+      "grad_norm": 2.337165355682373,
+      "learning_rate": 1.7704666161156994e-05,
+      "loss": 0.1089,
+      "step": 2277
+    },
+    {
+      "epoch": 1.1927710843373494,
+      "loss_partial": 0.0546875,
+      "step": 2277
+    },
+    {
+      "epoch": 1.1927710843373494,
+      "loss_partial": 0.0634765625,
+      "step": 2277
+    },
+    {
+      "epoch": 1.1932949188056574,
+      "grad_norm": 2.7267487049102783,
+      "learning_rate": 1.770175106106829e-05,
+      "loss": 0.1096,
+      "step": 2278
+    },
+    {
+      "epoch": 1.1932949188056574,
+      "loss_partial": 0.0478515625,
+      "step": 2278
+    },
+    {
+      "epoch": 1.1932949188056574,
+      "loss_partial": 0.0498046875,
+      "step": 2278
+    },
+    {
+      "epoch": 1.1938187532739655,
+      "grad_norm": 2.644923686981201,
+      "learning_rate": 1.7698834351361237e-05,
+      "loss": 0.1102,
+      "step": 2279
+    },
+    {
+      "epoch": 1.1938187532739655,
+      "loss_partial": 0.057861328125,
+      "step": 2279
+    },
+    {
+      "epoch": 1.1938187532739655,
+      "loss_partial": 0.05419921875,
+      "step": 2279
+    },
+    {
+      "epoch": 1.1943425877422735,
+      "grad_norm": 2.332526206970215,
+      "learning_rate": 1.7695916032645414e-05,
+      "loss": 0.1104,
+      "step": 2280
+    },
+    {
+      "epoch": 1.1943425877422735,
+      "loss_partial": 0.0517578125,
+      "step": 2280
+    },
+    {
+      "epoch": 1.1943425877422735,
+      "loss_partial": 0.0546875,
+      "step": 2280
+    },
+    {
+      "epoch": 1.1948664222105814,
+      "grad_norm": 1.972867488861084,
+      "learning_rate": 1.7692996105530728e-05,
+      "loss": 0.1062,
+      "step": 2281
+    },
+    {
+      "epoch": 1.1948664222105814,
+      "loss_partial": 0.060302734375,
+      "step": 2281
+    },
+    {
+      "epoch": 1.1948664222105814,
+      "loss_partial": 0.057373046875,
+      "step": 2281
+    },
+    {
+      "epoch": 1.1953902566788894,
+      "grad_norm": 2.6976635456085205,
+      "learning_rate": 1.7690074570627432e-05,
+      "loss": 0.1109,
+      "step": 2282
+    },
+    {
+      "epoch": 1.1953902566788894,
+      "loss_partial": 0.0537109375,
+      "step": 2282
+    },
+    {
+      "epoch": 1.1953902566788894,
+      "loss_partial": 0.044921875,
+      "step": 2282
+    },
+    {
+      "epoch": 1.1959140911471975,
+      "grad_norm": 1.912856936454773,
+      "learning_rate": 1.7687151428546103e-05,
+      "loss": 0.1081,
+      "step": 2283
+    },
+    {
+      "epoch": 1.1959140911471975,
+      "loss_partial": 0.058349609375,
+      "step": 2283
+    },
+    {
+      "epoch": 1.1959140911471975,
+      "loss_partial": 0.06787109375,
+      "step": 2283
+    },
+    {
+      "epoch": 1.1964379256155055,
+      "grad_norm": 1.8669211864471436,
+      "learning_rate": 1.768422667989766e-05,
+      "loss": 0.1079,
+      "step": 2284
+    },
+    {
+      "epoch": 1.1964379256155055,
+      "loss_partial": 0.05029296875,
+      "step": 2284
+    },
+    {
+      "epoch": 1.1964379256155055,
+      "loss_partial": 0.054443359375,
+      "step": 2284
+    },
+    {
+      "epoch": 1.1969617600838136,
+      "grad_norm": 1.3715198040008545,
+      "learning_rate": 1.768130032529336e-05,
+      "loss": 0.109,
+      "step": 2285
+    },
+    {
+      "epoch": 1.1969617600838136,
+      "loss_partial": 0.054443359375,
+      "step": 2285
+    },
+    {
+      "epoch": 1.1969617600838136,
+      "loss_partial": 0.0498046875,
+      "step": 2285
+    },
+    {
+      "epoch": 1.1974855945521214,
+      "grad_norm": 2.1648128032684326,
+      "learning_rate": 1.7678372365344792e-05,
+      "loss": 0.1066,
+      "step": 2286
+    },
+    {
+      "epoch": 1.1974855945521214,
+      "loss_partial": 0.05419921875,
+      "step": 2286
+    },
+    {
+      "epoch": 1.1974855945521214,
+      "loss_partial": 0.05126953125,
+      "step": 2286
+    },
+    {
+      "epoch": 1.1980094290204295,
+      "grad_norm": 2.0568501949310303,
+      "learning_rate": 1.7675442800663882e-05,
+      "loss": 0.1109,
+      "step": 2287
+    },
+    {
+      "epoch": 1.1980094290204295,
+      "loss_partial": 0.06640625,
+      "step": 2287
+    },
+    {
+      "epoch": 1.1980094290204295,
+      "loss_partial": 0.057373046875,
+      "step": 2287
+    },
+    {
+      "epoch": 1.1985332634887376,
+      "grad_norm": 2.0781497955322266,
+      "learning_rate": 1.767251163186289e-05,
+      "loss": 0.1054,
+      "step": 2288
+    },
+    {
+      "epoch": 1.1985332634887376,
+      "loss_partial": 0.0498046875,
+      "step": 2288
+    },
+    {
+      "epoch": 1.1985332634887376,
+      "loss_partial": 0.04931640625,
+      "step": 2288
+    },
+    {
+      "epoch": 1.1990570979570456,
+      "grad_norm": 1.8197554349899292,
+      "learning_rate": 1.7669578859554413e-05,
+      "loss": 0.1037,
+      "step": 2289
+    },
+    {
+      "epoch": 1.1990570979570456,
+      "loss_partial": 0.055419921875,
+      "step": 2289
+    },
+    {
+      "epoch": 1.1990570979570456,
+      "loss_partial": 0.048828125,
+      "step": 2289
+    },
+    {
+      "epoch": 1.1995809324253537,
+      "grad_norm": 1.6309056282043457,
+      "learning_rate": 1.7666644484351378e-05,
+      "loss": 0.1093,
+      "step": 2290
+    },
+    {
+      "epoch": 1.1995809324253537,
+      "loss_partial": 0.05810546875,
+      "step": 2290
+    },
+    {
+      "epoch": 1.1995809324253537,
+      "loss_partial": 0.0478515625,
+      "step": 2290
+    },
+    {
+      "epoch": 1.2001047668936615,
+      "grad_norm": 1.6137115955352783,
+      "learning_rate": 1.7663708506867057e-05,
+      "loss": 0.1092,
+      "step": 2291
+    },
+    {
+      "epoch": 1.2001047668936615,
+      "loss_partial": 0.048095703125,
+      "step": 2291
+    },
+    {
+      "epoch": 1.2001047668936615,
+      "loss_partial": 0.0498046875,
+      "step": 2291
+    },
+    {
+      "epoch": 1.2006286013619696,
+      "grad_norm": 1.960054636001587,
+      "learning_rate": 1.766077092771505e-05,
+      "loss": 0.1067,
+      "step": 2292
+    },
+    {
+      "epoch": 1.2006286013619696,
+      "loss_partial": 0.053955078125,
+      "step": 2292
+    },
+    {
+      "epoch": 1.2006286013619696,
+      "loss_partial": 0.06298828125,
+      "step": 2292
+    },
+    {
+      "epoch": 1.2011524358302776,
+      "grad_norm": 1.7433189153671265,
+      "learning_rate": 1.765783174750929e-05,
+      "loss": 0.1085,
+      "step": 2293
+    },
+    {
+      "epoch": 1.2011524358302776,
+      "loss_partial": 0.05322265625,
+      "step": 2293
+    },
+    {
+      "epoch": 1.2011524358302776,
+      "loss_partial": 0.056884765625,
+      "step": 2293
+    },
+    {
+      "epoch": 1.2016762702985857,
+      "grad_norm": 1.8822896480560303,
+      "learning_rate": 1.765489096686405e-05,
+      "loss": 0.1076,
+      "step": 2294
+    },
+    {
+      "epoch": 1.2016762702985857,
+      "loss_partial": 0.050537109375,
+      "step": 2294
+    },
+    {
+      "epoch": 1.2016762702985857,
+      "loss_partial": 0.0546875,
+      "step": 2294
+    },
+    {
+      "epoch": 1.2022001047668938,
+      "grad_norm": 2.0384342670440674,
+      "learning_rate": 1.7651948586393937e-05,
+      "loss": 0.1032,
+      "step": 2295
+    },
+    {
+      "epoch": 1.2022001047668938,
+      "loss_partial": 0.05419921875,
+      "step": 2295
+    },
+    {
+      "epoch": 1.2022001047668938,
+      "loss_partial": 0.04931640625,
+      "step": 2295
+    },
+    {
+      "epoch": 1.2027239392352016,
+      "grad_norm": 1.8460767269134521,
+      "learning_rate": 1.7649004606713882e-05,
+      "loss": 0.1082,
+      "step": 2296
+    },
+    {
+      "epoch": 1.2027239392352016,
+      "loss_partial": 0.05517578125,
+      "step": 2296
+    },
+    {
+      "epoch": 1.2027239392352016,
+      "loss_partial": 0.046630859375,
+      "step": 2296
+    },
+    {
+      "epoch": 1.2032477737035097,
+      "grad_norm": 3.319392442703247,
+      "learning_rate": 1.764605902843917e-05,
+      "loss": 0.1059,
+      "step": 2297
+    },
+    {
+      "epoch": 1.2032477737035097,
+      "loss_partial": 0.051513671875,
+      "step": 2297
+    },
+    {
+      "epoch": 1.2032477737035097,
+      "loss_partial": 0.0556640625,
+      "step": 2297
+    },
+    {
+      "epoch": 1.2037716081718177,
+      "grad_norm": 3.26777720451355,
+      "learning_rate": 1.76431118521854e-05,
+      "loss": 0.1075,
+      "step": 2298
+    },
+    {
+      "epoch": 1.2037716081718177,
+      "loss_partial": 0.052001953125,
+      "step": 2298
+    },
+    {
+      "epoch": 1.2037716081718177,
+      "loss_partial": 0.052001953125,
+      "step": 2298
+    },
+    {
+      "epoch": 1.2042954426401258,
+      "grad_norm": 2.598848581314087,
+      "learning_rate": 1.7640163078568516e-05,
+      "loss": 0.104,
+      "step": 2299
+    },
+    {
+      "epoch": 1.2042954426401258,
+      "loss_partial": 0.0517578125,
+      "step": 2299
+    },
+    {
+      "epoch": 1.2042954426401258,
+      "loss_partial": 0.048583984375,
+      "step": 2299
+    },
+    {
+      "epoch": 1.2048192771084336,
+      "grad_norm": 3.576598882675171,
+      "learning_rate": 1.7637212708204793e-05,
+      "loss": 0.1099,
+      "step": 2300
+    },
+    {
+      "epoch": 1.2048192771084336,
+      "loss_partial": 0.052001953125,
+      "step": 2300
+    },
+    {
+      "epoch": 1.2048192771084336,
+      "loss_partial": 0.04833984375,
+      "step": 2300
+    },
+    {
+      "epoch": 1.2053431115767417,
+      "grad_norm": 2.367281198501587,
+      "learning_rate": 1.7634260741710843e-05,
+      "loss": 0.1052,
+      "step": 2301
+    },
+    {
+      "epoch": 1.2053431115767417,
+      "loss_partial": 0.050048828125,
+      "step": 2301
+    },
+    {
+      "epoch": 1.2053431115767417,
+      "loss_partial": 0.047119140625,
+      "step": 2301
+    },
+    {
+      "epoch": 1.2058669460450497,
+      "grad_norm": 2.0170371532440186,
+      "learning_rate": 1.763130717970361e-05,
+      "loss": 0.1039,
+      "step": 2302
+    },
+    {
+      "epoch": 1.2058669460450497,
+      "loss_partial": 0.05322265625,
+      "step": 2302
+    },
+    {
+      "epoch": 1.2058669460450497,
+      "loss_partial": 0.054931640625,
+      "step": 2302
+    },
+    {
+      "epoch": 1.2063907805133578,
+      "grad_norm": 1.5450997352600098,
+      "learning_rate": 1.7628352022800363e-05,
+      "loss": 0.1104,
+      "step": 2303
+    },
+    {
+      "epoch": 1.2063907805133578,
+      "loss_partial": 0.0625,
+      "step": 2303
+    },
+    {
+      "epoch": 1.2063907805133578,
+      "loss_partial": 0.0478515625,
+      "step": 2303
+    },
+    {
+      "epoch": 1.2069146149816659,
+      "grad_norm": 2.053037643432617,
+      "learning_rate": 1.7625395271618715e-05,
+      "loss": 0.1045,
+      "step": 2304
+    },
+    {
+      "epoch": 1.2069146149816659,
+      "loss_partial": 0.059326171875,
+      "step": 2304
+    },
+    {
+      "epoch": 1.2069146149816659,
+      "loss_partial": 0.050537109375,
+      "step": 2304
+    },
+    {
+      "epoch": 1.207438449449974,
+      "grad_norm": 1.8468154668807983,
+      "learning_rate": 1.762243692677661e-05,
+      "loss": 0.1056,
+      "step": 2305
+    },
+    {
+      "epoch": 1.207438449449974,
+      "loss_partial": 0.056640625,
+      "step": 2305
+    },
+    {
+      "epoch": 1.207438449449974,
+      "loss_partial": 0.0634765625,
+      "step": 2305
+    },
+    {
+      "epoch": 1.2079622839182818,
+      "grad_norm": 2.904712438583374,
+      "learning_rate": 1.7619476988892325e-05,
+      "loss": 0.1086,
+      "step": 2306
+    },
+    {
+      "epoch": 1.2079622839182818,
+      "loss_partial": 0.05859375,
+      "step": 2306
+    },
+    {
+      "epoch": 1.2079622839182818,
+      "loss_partial": 0.058837890625,
+      "step": 2306
+    },
+    {
+      "epoch": 1.2084861183865898,
+      "grad_norm": 1.7873916625976562,
+      "learning_rate": 1.7616515458584465e-05,
+      "loss": 0.108,
+      "step": 2307
+    },
+    {
+      "epoch": 1.2084861183865898,
+      "loss_partial": 0.048828125,
+      "step": 2307
+    },
+    {
+      "epoch": 1.2084861183865898,
+      "loss_partial": 0.057861328125,
+      "step": 2307
+    },
+    {
+      "epoch": 1.2090099528548979,
+      "grad_norm": 3.0008599758148193,
+      "learning_rate": 1.7613552336471972e-05,
+      "loss": 0.1059,
+      "step": 2308
+    },
+    {
+      "epoch": 1.2090099528548979,
+      "loss_partial": 0.0556640625,
+      "step": 2308
+    },
+    {
+      "epoch": 1.2090099528548979,
+      "loss_partial": 0.0546875,
+      "step": 2308
+    },
+    {
+      "epoch": 1.209533787323206,
+      "grad_norm": 1.8932597637176514,
+      "learning_rate": 1.761058762317412e-05,
+      "loss": 0.1069,
+      "step": 2309
+    },
+    {
+      "epoch": 1.209533787323206,
+      "loss_partial": 0.0556640625,
+      "step": 2309
+    },
+    {
+      "epoch": 1.209533787323206,
+      "loss_partial": 0.0517578125,
+      "step": 2309
+    },
+    {
+      "epoch": 1.2100576217915138,
+      "grad_norm": 2.4881958961486816,
+      "learning_rate": 1.7607621319310525e-05,
+      "loss": 0.1098,
+      "step": 2310
+    },
+    {
+      "epoch": 1.2100576217915138,
+      "loss_partial": 0.05419921875,
+      "step": 2310
+    },
+    {
+      "epoch": 1.2100576217915138,
+      "loss_partial": 0.05419921875,
+      "step": 2310
+    },
+    {
+      "epoch": 1.2105814562598218,
+      "grad_norm": 1.6948766708374023,
+      "learning_rate": 1.760465342550111e-05,
+      "loss": 0.1092,
+      "step": 2311
+    },
+    {
+      "epoch": 1.2105814562598218,
+      "loss_partial": 0.04833984375,
+      "step": 2311
+    },
+    {
+      "epoch": 1.2105814562598218,
+      "loss_partial": 0.046630859375,
+      "step": 2311
+    },
+    {
+      "epoch": 1.21110529072813,
+      "grad_norm": 1.979982614517212,
+      "learning_rate": 1.760168394236616e-05,
+      "loss": 0.1083,
+      "step": 2312
+    },
+    {
+      "epoch": 1.21110529072813,
+      "loss_partial": 0.056884765625,
+      "step": 2312
+    },
+    {
+      "epoch": 1.21110529072813,
+      "loss_partial": 0.054443359375,
+      "step": 2312
+    },
+    {
+      "epoch": 1.211629125196438,
+      "grad_norm": 1.7076646089553833,
+      "learning_rate": 1.7598712870526274e-05,
+      "loss": 0.1061,
+      "step": 2313
+    },
+    {
+      "epoch": 1.211629125196438,
+      "loss_partial": 0.053466796875,
+      "step": 2313
+    },
+    {
+      "epoch": 1.211629125196438,
+      "loss_partial": 0.046875,
+      "step": 2313
+    },
+    {
+      "epoch": 1.212152959664746,
+      "grad_norm": 1.9606776237487793,
+      "learning_rate": 1.7595740210602383e-05,
+      "loss": 0.1069,
+      "step": 2314
+    },
+    {
+      "epoch": 1.212152959664746,
+      "loss_partial": 0.05224609375,
+      "step": 2314
+    },
+    {
+      "epoch": 1.212152959664746,
+      "loss_partial": 0.06005859375,
+      "step": 2314
+    },
+    {
+      "epoch": 1.212676794133054,
+      "grad_norm": 1.8420345783233643,
+      "learning_rate": 1.759276596321576e-05,
+      "loss": 0.1066,
+      "step": 2315
+    },
+    {
+      "epoch": 1.212676794133054,
+      "loss_partial": 0.05419921875,
+      "step": 2315
+    },
+    {
+      "epoch": 1.212676794133054,
+      "loss_partial": 0.04248046875,
+      "step": 2315
+    },
+    {
+      "epoch": 1.213200628601362,
+      "grad_norm": 3.5195159912109375,
+      "learning_rate": 1.7589790128988005e-05,
+      "loss": 0.1043,
+      "step": 2316
+    },
+    {
+      "epoch": 1.213200628601362,
+      "loss_partial": 0.050537109375,
+      "step": 2316
+    },
+    {
+      "epoch": 1.213200628601362,
+      "loss_partial": 0.0634765625,
+      "step": 2316
+    },
+    {
+      "epoch": 1.21372446306967,
+      "grad_norm": 1.4390023946762085,
+      "learning_rate": 1.7586812708541046e-05,
+      "loss": 0.1062,
+      "step": 2317
+    },
+    {
+      "epoch": 1.21372446306967,
+      "loss_partial": 0.0546875,
+      "step": 2317
+    },
+    {
+      "epoch": 1.21372446306967,
+      "loss_partial": 0.055908203125,
+      "step": 2317
+    },
+    {
+      "epoch": 1.214248297537978,
+      "grad_norm": 1.8351249694824219,
+      "learning_rate": 1.7583833702497145e-05,
+      "loss": 0.1089,
+      "step": 2318
+    },
+    {
+      "epoch": 1.214248297537978,
+      "loss_partial": 0.0615234375,
+      "step": 2318
+    },
+    {
+      "epoch": 1.214248297537978,
+      "loss_partial": 0.0537109375,
+      "step": 2318
+    },
+    {
+      "epoch": 1.214772132006286,
+      "grad_norm": 2.095912456512451,
+      "learning_rate": 1.75808531114789e-05,
+      "loss": 0.1071,
+      "step": 2319
+    },
+    {
+      "epoch": 1.214772132006286,
+      "loss_partial": 0.056640625,
+      "step": 2319
+    },
+    {
+      "epoch": 1.214772132006286,
+      "loss_partial": 0.0478515625,
+      "step": 2319
+    },
+    {
+      "epoch": 1.215295966474594,
+      "grad_norm": 1.4381579160690308,
+      "learning_rate": 1.757787093610923e-05,
+      "loss": 0.1062,
+      "step": 2320
+    },
+    {
+      "epoch": 1.215295966474594,
+      "loss_partial": 0.051025390625,
+      "step": 2320
+    },
+    {
+      "epoch": 1.215295966474594,
+      "loss_partial": 0.050048828125,
+      "step": 2320
+    },
+    {
+      "epoch": 1.215819800942902,
+      "grad_norm": 1.6719684600830078,
+      "learning_rate": 1.7574887177011395e-05,
+      "loss": 0.1032,
+      "step": 2321
+    },
+    {
+      "epoch": 1.215819800942902,
+      "loss_partial": 0.053955078125,
+      "step": 2321
+    },
+    {
+      "epoch": 1.215819800942902,
+      "loss_partial": 0.056396484375,
+      "step": 2321
+    },
+    {
+      "epoch": 1.21634363541121,
+      "grad_norm": 1.89189612865448,
+      "learning_rate": 1.757190183480898e-05,
+      "loss": 0.1086,
+      "step": 2322
+    },
+    {
+      "epoch": 1.21634363541121,
+      "loss_partial": 0.051025390625,
+      "step": 2322
+    },
+    {
+      "epoch": 1.21634363541121,
+      "loss_partial": 0.058837890625,
+      "step": 2322
+    },
+    {
+      "epoch": 1.216867469879518,
+      "grad_norm": 1.6034857034683228,
+      "learning_rate": 1.7568914910125913e-05,
+      "loss": 0.1092,
+      "step": 2323
+    },
+    {
+      "epoch": 1.216867469879518,
+      "loss_partial": 0.05078125,
+      "step": 2323
+    },
+    {
+      "epoch": 1.216867469879518,
+      "loss_partial": 0.052734375,
+      "step": 2323
+    },
+    {
+      "epoch": 1.2173913043478262,
+      "grad_norm": 1.6164724826812744,
+      "learning_rate": 1.7565926403586427e-05,
+      "loss": 0.1097,
+      "step": 2324
+    },
+    {
+      "epoch": 1.2173913043478262,
+      "loss_partial": 0.055419921875,
+      "step": 2324
+    },
+    {
+      "epoch": 1.2173913043478262,
+      "loss_partial": 0.04345703125,
+      "step": 2324
+    },
+    {
+      "epoch": 1.2179151388161342,
+      "grad_norm": 1.978447675704956,
+      "learning_rate": 1.7562936315815112e-05,
+      "loss": 0.1075,
+      "step": 2325
+    },
+    {
+      "epoch": 1.2179151388161342,
+      "loss_partial": 0.053955078125,
+      "step": 2325
+    },
+    {
+      "epoch": 1.2179151388161342,
+      "loss_partial": 0.041259765625,
+      "step": 2325
+    },
+    {
+      "epoch": 1.218438973284442,
+      "grad_norm": 1.524668574333191,
+      "learning_rate": 1.7559944647436872e-05,
+      "loss": 0.1074,
+      "step": 2326
+    },
+    {
+      "epoch": 1.218438973284442,
+      "loss_partial": 0.055908203125,
+      "step": 2326
+    },
+    {
+      "epoch": 1.218438973284442,
+      "loss_partial": 0.046630859375,
+      "step": 2326
+    },
+    {
+      "epoch": 1.2189628077527501,
+      "grad_norm": 1.6878584623336792,
+      "learning_rate": 1.755695139907695e-05,
+      "loss": 0.1031,
+      "step": 2327
+    },
+    {
+      "epoch": 1.2189628077527501,
+      "loss_partial": 0.05517578125,
+      "step": 2327
+    },
+    {
+      "epoch": 1.2189628077527501,
+      "loss_partial": 0.053466796875,
+      "step": 2327
+    },
+    {
+      "epoch": 1.2194866422210582,
+      "grad_norm": 1.6202870607376099,
+      "learning_rate": 1.755395657136092e-05,
+      "loss": 0.109,
+      "step": 2328
+    },
+    {
+      "epoch": 1.2194866422210582,
+      "loss_partial": 0.053955078125,
+      "step": 2328
+    },
+    {
+      "epoch": 1.2194866422210582,
+      "loss_partial": 0.048828125,
+      "step": 2328
+    },
+    {
+      "epoch": 1.2200104766893662,
+      "grad_norm": 2.209853172302246,
+      "learning_rate": 1.7550960164914675e-05,
+      "loss": 0.1077,
+      "step": 2329
+    },
+    {
+      "epoch": 1.2200104766893662,
+      "loss_partial": 0.048095703125,
+      "step": 2329
+    },
+    {
+      "epoch": 1.2200104766893662,
+      "loss_partial": 0.0517578125,
+      "step": 2329
+    },
+    {
+      "epoch": 1.220534311157674,
+      "grad_norm": 1.4293665885925293,
+      "learning_rate": 1.754796218036445e-05,
+      "loss": 0.1071,
+      "step": 2330
+    },
+    {
+      "epoch": 1.220534311157674,
+      "loss_partial": 0.052978515625,
+      "step": 2330
+    },
+    {
+      "epoch": 1.220534311157674,
+      "loss_partial": 0.048828125,
+      "step": 2330
+    },
+    {
+      "epoch": 1.2210581456259821,
+      "grad_norm": 2.393134117126465,
+      "learning_rate": 1.7544962618336808e-05,
+      "loss": 0.1081,
+      "step": 2331
+    },
+    {
+      "epoch": 1.2210581456259821,
+      "loss_partial": 0.051025390625,
+      "step": 2331
+    },
+    {
+      "epoch": 1.2210581456259821,
+      "loss_partial": 0.051025390625,
+      "step": 2331
+    },
+    {
+      "epoch": 1.2215819800942902,
+      "grad_norm": 2.144458055496216,
+      "learning_rate": 1.754196147945863e-05,
+      "loss": 0.1064,
+      "step": 2332
+    },
+    {
+      "epoch": 1.2215819800942902,
+      "loss_partial": 0.04833984375,
+      "step": 2332
+    },
+    {
+      "epoch": 1.2215819800942902,
+      "loss_partial": 0.05517578125,
+      "step": 2332
+    },
+    {
+      "epoch": 1.2221058145625983,
+      "grad_norm": 2.473433494567871,
+      "learning_rate": 1.7538958764357142e-05,
+      "loss": 0.1112,
+      "step": 2333
+    },
+    {
+      "epoch": 1.2221058145625983,
+      "loss_partial": 0.046142578125,
+      "step": 2333
+    },
+    {
+      "epoch": 1.2221058145625983,
+      "loss_partial": 0.059326171875,
+      "step": 2333
+    },
+    {
+      "epoch": 1.2226296490309063,
+      "grad_norm": 1.4601160287857056,
+      "learning_rate": 1.7535954473659896e-05,
+      "loss": 0.11,
+      "step": 2334
+    },
+    {
+      "epoch": 1.2226296490309063,
+      "loss_partial": 0.052734375,
+      "step": 2334
+    },
+    {
+      "epoch": 1.2226296490309063,
+      "loss_partial": 0.055908203125,
+      "step": 2334
+    },
+    {
+      "epoch": 1.2231534834992142,
+      "grad_norm": 1.8422839641571045,
+      "learning_rate": 1.7532948607994765e-05,
+      "loss": 0.108,
+      "step": 2335
+    },
+    {
+      "epoch": 1.2231534834992142,
+      "loss_partial": 0.042724609375,
+      "step": 2335
+    },
+    {
+      "epoch": 1.2231534834992142,
+      "loss_partial": 0.056396484375,
+      "step": 2335
+    },
+    {
+      "epoch": 1.2236773179675222,
+      "grad_norm": 1.753292441368103,
+      "learning_rate": 1.7529941167989956e-05,
+      "loss": 0.105,
+      "step": 2336
+    },
+    {
+      "epoch": 1.2236773179675222,
+      "loss_partial": 0.056396484375,
+      "step": 2336
+    },
+    {
+      "epoch": 1.2236773179675222,
+      "loss_partial": 0.060546875,
+      "step": 2336
+    },
+    {
+      "epoch": 1.2242011524358303,
+      "grad_norm": 2.0054478645324707,
+      "learning_rate": 1.7526932154274004e-05,
+      "loss": 0.1108,
+      "step": 2337
+    },
+    {
+      "epoch": 1.2242011524358303,
+      "loss_partial": 0.0546875,
+      "step": 2337
+    },
+    {
+      "epoch": 1.2242011524358303,
+      "loss_partial": 0.05615234375,
+      "step": 2337
+    },
+    {
+      "epoch": 1.2247249869041383,
+      "grad_norm": 1.8600703477859497,
+      "learning_rate": 1.7523921567475782e-05,
+      "loss": 0.107,
+      "step": 2338
+    },
+    {
+      "epoch": 1.2247249869041383,
+      "loss_partial": 0.057861328125,
+      "step": 2338
+    },
+    {
+      "epoch": 1.2247249869041383,
+      "loss_partial": 0.0634765625,
+      "step": 2338
+    },
+    {
+      "epoch": 1.2252488213724464,
+      "grad_norm": 1.6635112762451172,
+      "learning_rate": 1.7520909408224474e-05,
+      "loss": 0.1094,
+      "step": 2339
+    },
+    {
+      "epoch": 1.2252488213724464,
+      "loss_partial": 0.05517578125,
+      "step": 2339
+    },
+    {
+      "epoch": 1.2252488213724464,
+      "loss_partial": 0.060546875,
+      "step": 2339
+    },
+    {
+      "epoch": 1.2257726558407542,
+      "grad_norm": 2.414760112762451,
+      "learning_rate": 1.751789567714961e-05,
+      "loss": 0.1047,
+      "step": 2340
+    },
+    {
+      "epoch": 1.2257726558407542,
+      "loss_partial": 0.059326171875,
+      "step": 2340
+    },
+    {
+      "epoch": 1.2257726558407542,
+      "loss_partial": 0.052734375,
+      "step": 2340
+    },
+    {
+      "epoch": 1.2262964903090623,
+      "grad_norm": 1.9641177654266357,
+      "learning_rate": 1.751488037488104e-05,
+      "loss": 0.1072,
+      "step": 2341
+    },
+    {
+      "epoch": 1.2262964903090623,
+      "loss_partial": 0.050048828125,
+      "step": 2341
+    },
+    {
+      "epoch": 1.2262964903090623,
+      "loss_partial": 0.05078125,
+      "step": 2341
+    },
+    {
+      "epoch": 1.2268203247773704,
+      "grad_norm": 1.7865427732467651,
+      "learning_rate": 1.751186350204894e-05,
+      "loss": 0.1044,
+      "step": 2342
+    },
+    {
+      "epoch": 1.2268203247773704,
+      "loss_partial": 0.0498046875,
+      "step": 2342
+    },
+    {
+      "epoch": 1.2268203247773704,
+      "loss_partial": 0.052734375,
+      "step": 2342
+    },
+    {
+      "epoch": 1.2273441592456784,
+      "grad_norm": 1.667188286781311,
+      "learning_rate": 1.750884505928382e-05,
+      "loss": 0.1071,
+      "step": 2343
+    },
+    {
+      "epoch": 1.2273441592456784,
+      "loss_partial": 0.054931640625,
+      "step": 2343
+    },
+    {
+      "epoch": 1.2273441592456784,
+      "loss_partial": 0.0498046875,
+      "step": 2343
+    },
+    {
+      "epoch": 1.2278679937139865,
+      "grad_norm": 1.6643925905227661,
+      "learning_rate": 1.750582504721652e-05,
+      "loss": 0.1074,
+      "step": 2344
+    },
+    {
+      "epoch": 1.2278679937139865,
+      "loss_partial": 0.06640625,
+      "step": 2344
+    },
+    {
+      "epoch": 1.2278679937139865,
+      "loss_partial": 0.04736328125,
+      "step": 2344
+    },
+    {
+      "epoch": 1.2283918281822943,
+      "grad_norm": 1.7399905920028687,
+      "learning_rate": 1.7502803466478194e-05,
+      "loss": 0.105,
+      "step": 2345
+    },
+    {
+      "epoch": 1.2283918281822943,
+      "loss_partial": 0.05615234375,
+      "step": 2345
+    },
+    {
+      "epoch": 1.2283918281822943,
+      "loss_partial": 0.04638671875,
+      "step": 2345
+    },
+    {
+      "epoch": 1.2289156626506024,
+      "grad_norm": 2.0504841804504395,
+      "learning_rate": 1.749978031770034e-05,
+      "loss": 0.1055,
+      "step": 2346
+    },
+    {
+      "epoch": 1.2289156626506024,
+      "loss_partial": 0.05419921875,
+      "step": 2346
+    },
+    {
+      "epoch": 1.2289156626506024,
+      "loss_partial": 0.0595703125,
+      "step": 2346
+    },
+    {
+      "epoch": 1.2294394971189104,
+      "grad_norm": 2.257741689682007,
+      "learning_rate": 1.749675560151478e-05,
+      "loss": 0.1082,
+      "step": 2347
+    },
+    {
+      "epoch": 1.2294394971189104,
+      "loss_partial": 0.053466796875,
+      "step": 2347
+    },
+    {
+      "epoch": 1.2294394971189104,
+      "loss_partial": 0.05078125,
+      "step": 2347
+    },
+    {
+      "epoch": 1.2299633315872185,
+      "grad_norm": 2.2032761573791504,
+      "learning_rate": 1.7493729318553656e-05,
+      "loss": 0.1063,
+      "step": 2348
+    },
+    {
+      "epoch": 1.2299633315872185,
+      "loss_partial": 0.05126953125,
+      "step": 2348
+    },
+    {
+      "epoch": 1.2299633315872185,
+      "loss_partial": 0.044677734375,
+      "step": 2348
+    },
+    {
+      "epoch": 1.2304871660555265,
+      "grad_norm": 1.817396640777588,
+      "learning_rate": 1.7490701469449445e-05,
+      "loss": 0.1053,
+      "step": 2349
+    },
+    {
+      "epoch": 1.2304871660555265,
+      "loss_partial": 0.05419921875,
+      "step": 2349
+    },
+    {
+      "epoch": 1.2304871660555265,
+      "loss_partial": 0.05615234375,
+      "step": 2349
+    },
+    {
+      "epoch": 1.2310110005238344,
+      "grad_norm": 1.705721139907837,
+      "learning_rate": 1.7487672054834947e-05,
+      "loss": 0.11,
+      "step": 2350
+    },
+    {
+      "epoch": 1.2310110005238344,
+      "loss_partial": 0.048583984375,
+      "step": 2350
+    },
+    {
+      "epoch": 1.2310110005238344,
+      "loss_partial": 0.05322265625,
+      "step": 2350
+    },
+    {
+      "epoch": 1.2315348349921424,
+      "grad_norm": 2.1178669929504395,
+      "learning_rate": 1.7484641075343293e-05,
+      "loss": 0.1066,
+      "step": 2351
+    },
+    {
+      "epoch": 1.2315348349921424,
+      "loss_partial": 0.051513671875,
+      "step": 2351
+    },
+    {
+      "epoch": 1.2315348349921424,
+      "loss_partial": 0.0556640625,
+      "step": 2351
+    },
+    {
+      "epoch": 1.2320586694604505,
+      "grad_norm": 1.9130886793136597,
+      "learning_rate": 1.7481608531607938e-05,
+      "loss": 0.1073,
+      "step": 2352
+    },
+    {
+      "epoch": 1.2320586694604505,
+      "loss_partial": 0.0498046875,
+      "step": 2352
+    },
+    {
+      "epoch": 1.2320586694604505,
+      "loss_partial": 0.0498046875,
+      "step": 2352
+    },
+    {
+      "epoch": 1.2325825039287586,
+      "grad_norm": 1.906533122062683,
+      "learning_rate": 1.7478574424262667e-05,
+      "loss": 0.1071,
+      "step": 2353
+    },
+    {
+      "epoch": 1.2325825039287586,
+      "loss_partial": 0.052734375,
+      "step": 2353
+    },
+    {
+      "epoch": 1.2325825039287586,
+      "loss_partial": 0.04638671875,
+      "step": 2353
+    },
+    {
+      "epoch": 1.2331063383970666,
+      "grad_norm": 2.273561716079712,
+      "learning_rate": 1.7475538753941583e-05,
+      "loss": 0.1076,
+      "step": 2354
+    },
+    {
+      "epoch": 1.2331063383970666,
+      "loss_partial": 0.050048828125,
+      "step": 2354
+    },
+    {
+      "epoch": 1.2331063383970666,
+      "loss_partial": 0.048828125,
+      "step": 2354
+    },
+    {
+      "epoch": 1.2336301728653745,
+      "grad_norm": 2.1980185508728027,
+      "learning_rate": 1.7472501521279128e-05,
+      "loss": 0.1101,
+      "step": 2355
+    },
+    {
+      "epoch": 1.2336301728653745,
+      "loss_partial": 0.06103515625,
+      "step": 2355
+    },
+    {
+      "epoch": 1.2336301728653745,
+      "loss_partial": 0.058349609375,
+      "step": 2355
+    },
+    {
+      "epoch": 1.2341540073336825,
+      "grad_norm": 1.8893439769744873,
+      "learning_rate": 1.7469462726910067e-05,
+      "loss": 0.1071,
+      "step": 2356
+    },
+    {
+      "epoch": 1.2341540073336825,
+      "loss_partial": 0.057861328125,
+      "step": 2356
+    },
+    {
+      "epoch": 1.2341540073336825,
+      "loss_partial": 0.046630859375,
+      "step": 2356
+    },
+    {
+      "epoch": 1.2346778418019906,
+      "grad_norm": 1.4944525957107544,
+      "learning_rate": 1.746642237146948e-05,
+      "loss": 0.1037,
+      "step": 2357
+    },
+    {
+      "epoch": 1.2346778418019906,
+      "loss_partial": 0.052978515625,
+      "step": 2357
+    },
+    {
+      "epoch": 1.2346778418019906,
+      "loss_partial": 0.050048828125,
+      "step": 2357
+    },
+    {
+      "epoch": 1.2352016762702986,
+      "grad_norm": 1.5356884002685547,
+      "learning_rate": 1.7463380455592793e-05,
+      "loss": 0.1071,
+      "step": 2358
+    },
+    {
+      "epoch": 1.2352016762702986,
+      "loss_partial": 0.051513671875,
+      "step": 2358
+    },
+    {
+      "epoch": 1.2352016762702986,
+      "loss_partial": 0.0546875,
+      "step": 2358
+    },
+    {
+      "epoch": 1.2357255107386065,
+      "grad_norm": 1.2907894849777222,
+      "learning_rate": 1.746033697991574e-05,
+      "loss": 0.1115,
+      "step": 2359
+    },
+    {
+      "epoch": 1.2357255107386065,
+      "loss_partial": 0.05517578125,
+      "step": 2359
+    },
+    {
+      "epoch": 1.2357255107386065,
+      "loss_partial": 0.0517578125,
+      "step": 2359
+    },
+    {
+      "epoch": 1.2362493452069145,
+      "grad_norm": 1.6088539361953735,
+      "learning_rate": 1.7457291945074394e-05,
+      "loss": 0.1072,
+      "step": 2360
+    },
+    {
+      "epoch": 1.2362493452069145,
+      "loss_partial": 0.0517578125,
+      "step": 2360
+    },
+    {
+      "epoch": 1.2362493452069145,
+      "loss_partial": 0.0556640625,
+      "step": 2360
+    },
+    {
+      "epoch": 1.2367731796752226,
+      "grad_norm": 1.6251497268676758,
+      "learning_rate": 1.7454245351705145e-05,
+      "loss": 0.1042,
+      "step": 2361
+    },
+    {
+      "epoch": 1.2367731796752226,
+      "loss_partial": 0.05859375,
+      "step": 2361
+    },
+    {
+      "epoch": 1.2367731796752226,
+      "loss_partial": 0.048095703125,
+      "step": 2361
+    },
+    {
+      "epoch": 1.2372970141435307,
+      "grad_norm": 1.6176291704177856,
+      "learning_rate": 1.745119720044471e-05,
+      "loss": 0.1059,
+      "step": 2362
+    },
+    {
+      "epoch": 1.2372970141435307,
+      "loss_partial": 0.05126953125,
+      "step": 2362
+    },
+    {
+      "epoch": 1.2372970141435307,
+      "loss_partial": 0.060791015625,
+      "step": 2362
+    },
+    {
+      "epoch": 1.2378208486118387,
+      "grad_norm": 1.5971535444259644,
+      "learning_rate": 1.7448147491930144e-05,
+      "loss": 0.1089,
+      "step": 2363
+    },
+    {
+      "epoch": 1.2378208486118387,
+      "loss_partial": 0.04833984375,
+      "step": 2363
+    },
+    {
+      "epoch": 1.2378208486118387,
+      "loss_partial": 0.046630859375,
+      "step": 2363
+    },
+    {
+      "epoch": 1.2383446830801468,
+      "grad_norm": 1.7030231952667236,
+      "learning_rate": 1.7445096226798807e-05,
+      "loss": 0.1052,
+      "step": 2364
+    },
+    {
+      "epoch": 1.2383446830801468,
+      "loss_partial": 0.06494140625,
+      "step": 2364
+    },
+    {
+      "epoch": 1.2383446830801468,
+      "loss_partial": 0.057861328125,
+      "step": 2364
+    },
+    {
+      "epoch": 1.2388685175484546,
+      "grad_norm": 1.7126494646072388,
+      "learning_rate": 1.7442043405688393e-05,
+      "loss": 0.1084,
+      "step": 2365
+    },
+    {
+      "epoch": 1.2388685175484546,
+      "loss_partial": 0.06005859375,
+      "step": 2365
+    },
+    {
+      "epoch": 1.2388685175484546,
+      "loss_partial": 0.055419921875,
+      "step": 2365
+    },
+    {
+      "epoch": 1.2393923520167627,
+      "grad_norm": 1.6559396982192993,
+      "learning_rate": 1.7438989029236935e-05,
+      "loss": 0.1047,
+      "step": 2366
+    },
+    {
+      "epoch": 1.2393923520167627,
+      "loss_partial": 0.0517578125,
+      "step": 2366
+    },
+    {
+      "epoch": 1.2393923520167627,
+      "loss_partial": 0.06591796875,
+      "step": 2366
+    },
+    {
+      "epoch": 1.2399161864850707,
+      "grad_norm": 1.8710299730300903,
+      "learning_rate": 1.7435933098082767e-05,
+      "loss": 0.1079,
+      "step": 2367
+    },
+    {
+      "epoch": 1.2399161864850707,
+      "loss_partial": 0.052490234375,
+      "step": 2367
+    },
+    {
+      "epoch": 1.2399161864850707,
+      "loss_partial": 0.0546875,
+      "step": 2367
+    },
+    {
+      "epoch": 1.2404400209533788,
+      "grad_norm": 2.1617627143859863,
+      "learning_rate": 1.7432875612864562e-05,
+      "loss": 0.1122,
+      "step": 2368
+    },
+    {
+      "epoch": 1.2404400209533788,
+      "loss_partial": 0.058349609375,
+      "step": 2368
+    },
+    {
+      "epoch": 1.2404400209533788,
+      "loss_partial": 0.0556640625,
+      "step": 2368
+    },
+    {
+      "epoch": 1.2409638554216866,
+      "grad_norm": 1.6311874389648438,
+      "learning_rate": 1.7429816574221322e-05,
+      "loss": 0.11,
+      "step": 2369
+    },
+    {
+      "epoch": 1.2409638554216866,
+      "loss_partial": 0.045654296875,
+      "step": 2369
+    },
+    {
+      "epoch": 1.2409638554216866,
+      "loss_partial": 0.062255859375,
+      "step": 2369
+    },
+    {
+      "epoch": 1.2414876898899947,
+      "grad_norm": 2.537825345993042,
+      "learning_rate": 1.742675598279236e-05,
+      "loss": 0.1066,
+      "step": 2370
+    },
+    {
+      "epoch": 1.2414876898899947,
+      "loss_partial": 0.044189453125,
+      "step": 2370
+    },
+    {
+      "epoch": 1.2414876898899947,
+      "loss_partial": 0.0615234375,
+      "step": 2370
+    },
+    {
+      "epoch": 1.2420115243583028,
+      "grad_norm": 1.5626002550125122,
+      "learning_rate": 1.7423693839217325e-05,
+      "loss": 0.1042,
+      "step": 2371
+    },
+    {
+      "epoch": 1.2420115243583028,
+      "loss_partial": 0.04931640625,
+      "step": 2371
+    },
+    {
+      "epoch": 1.2420115243583028,
+      "loss_partial": 0.049560546875,
+      "step": 2371
+    },
+    {
+      "epoch": 1.2425353588266108,
+      "grad_norm": 2.5606164932250977,
+      "learning_rate": 1.742063014413618e-05,
+      "loss": 0.1087,
+      "step": 2372
+    },
+    {
+      "epoch": 1.2425353588266108,
+      "loss_partial": 0.044189453125,
+      "step": 2372
+    },
+    {
+      "epoch": 1.2425353588266108,
+      "loss_partial": 0.056396484375,
+      "step": 2372
+    },
+    {
+      "epoch": 1.2430591932949189,
+      "grad_norm": 2.161201000213623,
+      "learning_rate": 1.7417564898189227e-05,
+      "loss": 0.108,
+      "step": 2373
+    },
+    {
+      "epoch": 1.2430591932949189,
+      "loss_partial": 0.043212890625,
+      "step": 2373
+    },
+    {
+      "epoch": 1.2430591932949189,
+      "loss_partial": 0.05908203125,
+      "step": 2373
+    },
+    {
+      "epoch": 1.243583027763227,
+      "grad_norm": 2.274561882019043,
+      "learning_rate": 1.7414498102017077e-05,
+      "loss": 0.1064,
+      "step": 2374
+    },
+    {
+      "epoch": 1.243583027763227,
+      "loss_partial": 0.043701171875,
+      "step": 2374
+    },
+    {
+      "epoch": 1.243583027763227,
+      "loss_partial": 0.0478515625,
+      "step": 2374
+    },
+    {
+      "epoch": 1.2441068622315348,
+      "grad_norm": 1.840720534324646,
+      "learning_rate": 1.7411429756260675e-05,
+      "loss": 0.1078,
+      "step": 2375
+    },
+    {
+      "epoch": 1.2441068622315348,
+      "loss_partial": 0.050048828125,
+      "step": 2375
+    },
+    {
+      "epoch": 1.2441068622315348,
+      "loss_partial": 0.055419921875,
+      "step": 2375
+    },
+    {
+      "epoch": 1.2446306966998428,
+      "grad_norm": 1.9925236701965332,
+      "learning_rate": 1.7408359861561282e-05,
+      "loss": 0.1114,
+      "step": 2376
+    },
+    {
+      "epoch": 1.2446306966998428,
+      "loss_partial": 0.052490234375,
+      "step": 2376
+    },
+    {
+      "epoch": 1.2446306966998428,
+      "loss_partial": 0.0458984375,
+      "step": 2376
+    },
+    {
+      "epoch": 1.245154531168151,
+      "grad_norm": 1.7221970558166504,
+      "learning_rate": 1.7405288418560492e-05,
+      "loss": 0.1087,
+      "step": 2377
+    },
+    {
+      "epoch": 1.245154531168151,
+      "loss_partial": 0.05908203125,
+      "step": 2377
+    },
+    {
+      "epoch": 1.245154531168151,
+      "loss_partial": 0.050537109375,
+      "step": 2377
+    },
+    {
+      "epoch": 1.245678365636459,
+      "grad_norm": 2.4041240215301514,
+      "learning_rate": 1.7402215427900217e-05,
+      "loss": 0.1076,
+      "step": 2378
+    },
+    {
+      "epoch": 1.245678365636459,
+      "loss_partial": 0.052001953125,
+      "step": 2378
+    },
+    {
+      "epoch": 1.245678365636459,
+      "loss_partial": 0.051513671875,
+      "step": 2378
+    },
+    {
+      "epoch": 1.2462022001047668,
+      "grad_norm": 1.4751496315002441,
+      "learning_rate": 1.739914089022269e-05,
+      "loss": 0.1048,
+      "step": 2379
+    },
+    {
+      "epoch": 1.2462022001047668,
+      "loss_partial": 0.054931640625,
+      "step": 2379
+    },
+    {
+      "epoch": 1.2462022001047668,
+      "loss_partial": 0.0546875,
+      "step": 2379
+    },
+    {
+      "epoch": 1.2467260345730748,
+      "grad_norm": 2.2878079414367676,
+      "learning_rate": 1.739606480617047e-05,
+      "loss": 0.1102,
+      "step": 2380
+    },
+    {
+      "epoch": 1.2467260345730748,
+      "loss_partial": 0.0625,
+      "step": 2380
+    },
+    {
+      "epoch": 1.2467260345730748,
+      "loss_partial": 0.05078125,
+      "step": 2380
+    },
+    {
+      "epoch": 1.247249869041383,
+      "grad_norm": 1.7641496658325195,
+      "learning_rate": 1.7392987176386445e-05,
+      "loss": 0.1083,
+      "step": 2381
+    },
+    {
+      "epoch": 1.247249869041383,
+      "loss_partial": 0.05859375,
+      "step": 2381
+    },
+    {
+      "epoch": 1.247249869041383,
+      "loss_partial": 0.05078125,
+      "step": 2381
+    },
+    {
+      "epoch": 1.247773703509691,
+      "grad_norm": 1.875465750694275,
+      "learning_rate": 1.7389908001513817e-05,
+      "loss": 0.1066,
+      "step": 2382
+    },
+    {
+      "epoch": 1.247773703509691,
+      "loss_partial": 0.052978515625,
+      "step": 2382
+    },
+    {
+      "epoch": 1.247773703509691,
+      "loss_partial": 0.04248046875,
+      "step": 2382
+    },
+    {
+      "epoch": 1.248297537977999,
+      "grad_norm": 2.0078659057617188,
+      "learning_rate": 1.738682728219611e-05,
+      "loss": 0.1019,
+      "step": 2383
+    },
+    {
+      "epoch": 1.248297537977999,
+      "loss_partial": 0.0478515625,
+      "step": 2383
+    },
+    {
+      "epoch": 1.248297537977999,
+      "loss_partial": 0.053955078125,
+      "step": 2383
+    },
+    {
+      "epoch": 1.248821372446307,
+      "grad_norm": 1.9613475799560547,
+      "learning_rate": 1.738374501907718e-05,
+      "loss": 0.1056,
+      "step": 2384
+    },
+    {
+      "epoch": 1.248821372446307,
+      "loss_partial": 0.048583984375,
+      "step": 2384
+    },
+    {
+      "epoch": 1.248821372446307,
+      "loss_partial": 0.052001953125,
+      "step": 2384
+    },
+    {
+      "epoch": 1.249345206914615,
+      "grad_norm": 1.3957492113113403,
+      "learning_rate": 1.7380661212801207e-05,
+      "loss": 0.1064,
+      "step": 2385
+    },
+    {
+      "epoch": 1.249345206914615,
+      "loss_partial": 0.06298828125,
+      "step": 2385
+    },
+    {
+      "epoch": 1.249345206914615,
+      "loss_partial": 0.04296875,
+      "step": 2385
+    },
+    {
+      "epoch": 1.249869041382923,
+      "grad_norm": 1.5171129703521729,
+      "learning_rate": 1.7377575864012678e-05,
+      "loss": 0.1059,
+      "step": 2386
+    },
+    {
+      "epoch": 1.249869041382923,
+      "loss_partial": 0.05029296875,
+      "step": 2386
+    },
+    {
+      "epoch": 1.249869041382923,
+      "loss_partial": 0.048095703125,
+      "step": 2386
+    },
+    {
+      "epoch": 1.250392875851231,
+      "grad_norm": 1.5617119073867798,
+      "learning_rate": 1.737448897335641e-05,
+      "loss": 0.1084,
+      "step": 2387
+    },
+    {
+      "epoch": 1.250392875851231,
+      "loss_partial": 0.045166015625,
+      "step": 2387
+    },
+    {
+      "epoch": 1.250392875851231,
+      "loss_partial": 0.0546875,
+      "step": 2387
+    },
+    {
+      "epoch": 1.250916710319539,
+      "grad_norm": 2.1057679653167725,
+      "learning_rate": 1.7371400541477556e-05,
+      "loss": 0.1079,
+      "step": 2388
+    },
+    {
+      "epoch": 1.250916710319539,
+      "loss_partial": 0.059326171875,
+      "step": 2388
+    },
+    {
+      "epoch": 1.250916710319539,
+      "loss_partial": 0.057861328125,
+      "step": 2388
+    },
+    {
+      "epoch": 1.251440544787847,
+      "grad_norm": 1.5732454061508179,
+      "learning_rate": 1.736831056902157e-05,
+      "loss": 0.108,
+      "step": 2389
+    },
+    {
+      "epoch": 1.251440544787847,
+      "loss_partial": 0.05517578125,
+      "step": 2389
+    },
+    {
+      "epoch": 1.251440544787847,
+      "loss_partial": 0.0478515625,
+      "step": 2389
+    },
+    {
+      "epoch": 1.251964379256155,
+      "grad_norm": 1.9490211009979248,
+      "learning_rate": 1.736521905663424e-05,
+      "loss": 0.1093,
+      "step": 2390
+    },
+    {
+      "epoch": 1.251964379256155,
+      "loss_partial": 0.056396484375,
+      "step": 2390
+    },
+    {
+      "epoch": 1.251964379256155,
+      "loss_partial": 0.042236328125,
+      "step": 2390
+    },
+    {
+      "epoch": 1.252488213724463,
+      "grad_norm": 1.8913838863372803,
+      "learning_rate": 1.736212600496167e-05,
+      "loss": 0.1103,
+      "step": 2391
+    },
+    {
+      "epoch": 1.252488213724463,
+      "loss_partial": 0.05078125,
+      "step": 2391
+    },
+    {
+      "epoch": 1.252488213724463,
+      "loss_partial": 0.048095703125,
+      "step": 2391
+    },
+    {
+      "epoch": 1.2530120481927711,
+      "grad_norm": 1.7704545259475708,
+      "learning_rate": 1.735903141465029e-05,
+      "loss": 0.1101,
+      "step": 2392
+    },
+    {
+      "epoch": 1.2530120481927711,
+      "loss_partial": 0.054931640625,
+      "step": 2392
+    },
+    {
+      "epoch": 1.2530120481927711,
+      "loss_partial": 0.0556640625,
+      "step": 2392
+    },
+    {
+      "epoch": 1.2535358826610792,
+      "grad_norm": 1.9987603425979614,
+      "learning_rate": 1.7355935286346857e-05,
+      "loss": 0.1078,
+      "step": 2393
+    },
+    {
+      "epoch": 1.2535358826610792,
+      "loss_partial": 0.04833984375,
+      "step": 2393
+    },
+    {
+      "epoch": 1.2535358826610792,
+      "loss_partial": 0.043701171875,
+      "step": 2393
+    },
+    {
+      "epoch": 1.2540597171293872,
+      "grad_norm": 1.5424355268478394,
+      "learning_rate": 1.7352837620698434e-05,
+      "loss": 0.1085,
+      "step": 2394
+    },
+    {
+      "epoch": 1.2540597171293872,
+      "loss_partial": 0.06298828125,
+      "step": 2394
+    },
+    {
+      "epoch": 1.2540597171293872,
+      "loss_partial": 0.05224609375,
+      "step": 2394
+    },
+    {
+      "epoch": 1.254583551597695,
+      "grad_norm": 1.843084692955017,
+      "learning_rate": 1.7349738418352413e-05,
+      "loss": 0.1041,
+      "step": 2395
+    },
+    {
+      "epoch": 1.254583551597695,
+      "loss_partial": 0.052001953125,
+      "step": 2395
+    },
+    {
+      "epoch": 1.254583551597695,
+      "loss_partial": 0.041748046875,
+      "step": 2395
+    },
+    {
+      "epoch": 1.2551073860660031,
+      "grad_norm": 2.796003818511963,
+      "learning_rate": 1.7346637679956514e-05,
+      "loss": 0.1084,
+      "step": 2396
+    },
+    {
+      "epoch": 1.2551073860660031,
+      "loss_partial": 0.0498046875,
+      "step": 2396
+    },
+    {
+      "epoch": 1.2551073860660031,
+      "loss_partial": 0.054443359375,
+      "step": 2396
+    },
+    {
+      "epoch": 1.2556312205343112,
+      "grad_norm": 1.640023946762085,
+      "learning_rate": 1.7343535406158773e-05,
+      "loss": 0.1074,
+      "step": 2397
+    },
+    {
+      "epoch": 1.2556312205343112,
+      "loss_partial": 0.0625,
+      "step": 2397
+    },
+    {
+      "epoch": 1.2556312205343112,
+      "loss_partial": 0.051025390625,
+      "step": 2397
+    },
+    {
+      "epoch": 1.256155055002619,
+      "grad_norm": 1.8355587720870972,
+      "learning_rate": 1.7340431597607544e-05,
+      "loss": 0.1023,
+      "step": 2398
+    },
+    {
+      "epoch": 1.256155055002619,
+      "loss_partial": 0.0439453125,
+      "step": 2398
+    },
+    {
+      "epoch": 1.256155055002619,
+      "loss_partial": 0.054443359375,
+      "step": 2398
+    },
+    {
+      "epoch": 1.256678889470927,
+      "grad_norm": 2.01053524017334,
+      "learning_rate": 1.7337326254951498e-05,
+      "loss": 0.1083,
+      "step": 2399
+    },
+    {
+      "epoch": 1.256678889470927,
+      "loss_partial": 0.049560546875,
+      "step": 2399
+    },
+    {
+      "epoch": 1.256678889470927,
+      "loss_partial": 0.05126953125,
+      "step": 2399
+    },
+    {
+      "epoch": 1.2572027239392352,
+      "grad_norm": 1.664995789527893,
+      "learning_rate": 1.7334219378839642e-05,
+      "loss": 0.1059,
+      "step": 2400
+    },
+    {
+      "epoch": 1.2572027239392352,
+      "loss_partial": 0.053955078125,
+      "step": 2400
+    },
+    {
+      "epoch": 1.2572027239392352,
+      "loss_partial": 0.05029296875,
+      "step": 2400
+    },
+    {
+      "epoch": 1.2577265584075432,
+      "grad_norm": 2.42087984085083,
+      "learning_rate": 1.7331110969921285e-05,
+      "loss": 0.1041,
+      "step": 2401
+    },
+    {
+      "epoch": 1.2577265584075432,
+      "loss_partial": 0.060302734375,
+      "step": 2401
+    },
+    {
+      "epoch": 1.2577265584075432,
+      "loss_partial": 0.059326171875,
+      "step": 2401
+    },
+    {
+      "epoch": 1.2582503928758513,
+      "grad_norm": 2.1680288314819336,
+      "learning_rate": 1.7328001028846073e-05,
+      "loss": 0.1057,
+      "step": 2402
+    },
+    {
+      "epoch": 1.2582503928758513,
+      "loss_partial": 0.046142578125,
+      "step": 2402
+    },
+    {
+      "epoch": 1.2582503928758513,
+      "loss_partial": 0.0556640625,
+      "step": 2402
+    },
+    {
+      "epoch": 1.2587742273441593,
+      "grad_norm": 2.5054802894592285,
+      "learning_rate": 1.732488955626396e-05,
+      "loss": 0.1052,
+      "step": 2403
+    },
+    {
+      "epoch": 1.2587742273441593,
+      "loss_partial": 0.054443359375,
+      "step": 2403
+    },
+    {
+      "epoch": 1.2587742273441593,
+      "loss_partial": 0.05126953125,
+      "step": 2403
+    },
+    {
+      "epoch": 1.2592980618124674,
+      "grad_norm": 1.6170576810836792,
+      "learning_rate": 1.7321776552825223e-05,
+      "loss": 0.1049,
+      "step": 2404
+    },
+    {
+      "epoch": 1.2592980618124674,
+      "loss_partial": 0.0546875,
+      "step": 2404
+    },
+    {
+      "epoch": 1.2592980618124674,
+      "loss_partial": 0.060302734375,
+      "step": 2404
+    },
+    {
+      "epoch": 1.2598218962807752,
+      "grad_norm": 2.1706206798553467,
+      "learning_rate": 1.7318662019180465e-05,
+      "loss": 0.1039,
+      "step": 2405
+    },
+    {
+      "epoch": 1.2598218962807752,
+      "loss_partial": 0.050537109375,
+      "step": 2405
+    },
+    {
+      "epoch": 1.2598218962807752,
+      "loss_partial": 0.05517578125,
+      "step": 2405
+    },
+    {
+      "epoch": 1.2603457307490833,
+      "grad_norm": 2.0403411388397217,
+      "learning_rate": 1.7315545955980605e-05,
+      "loss": 0.1088,
+      "step": 2406
+    },
+    {
+      "epoch": 1.2603457307490833,
+      "loss_partial": 0.0556640625,
+      "step": 2406
+    },
+    {
+      "epoch": 1.2603457307490833,
+      "loss_partial": 0.054931640625,
+      "step": 2406
+    },
+    {
+      "epoch": 1.2608695652173914,
+      "grad_norm": 1.971869945526123,
+      "learning_rate": 1.731242836387687e-05,
+      "loss": 0.1082,
+      "step": 2407
+    },
+    {
+      "epoch": 1.2608695652173914,
+      "loss_partial": 0.052001953125,
+      "step": 2407
+    },
+    {
+      "epoch": 1.2608695652173914,
+      "loss_partial": 0.052490234375,
+      "step": 2407
+    },
+    {
+      "epoch": 1.2613933996856992,
+      "grad_norm": 1.6500134468078613,
+      "learning_rate": 1.7309309243520832e-05,
+      "loss": 0.1058,
+      "step": 2408
+    },
+    {
+      "epoch": 1.2613933996856992,
+      "loss_partial": 0.0419921875,
+      "step": 2408
+    },
+    {
+      "epoch": 1.2613933996856992,
+      "loss_partial": 0.052734375,
+      "step": 2408
+    },
+    {
+      "epoch": 1.2619172341540073,
+      "grad_norm": 2.078641414642334,
+      "learning_rate": 1.730618859556436e-05,
+      "loss": 0.1044,
+      "step": 2409
+    },
+    {
+      "epoch": 1.2619172341540073,
+      "loss_partial": 0.050537109375,
+      "step": 2409
+    },
+    {
+      "epoch": 1.2619172341540073,
+      "loss_partial": 0.060546875,
+      "step": 2409
+    },
+    {
+      "epoch": 1.2624410686223153,
+      "grad_norm": 1.8596546649932861,
+      "learning_rate": 1.7303066420659656e-05,
+      "loss": 0.1045,
+      "step": 2410
+    },
+    {
+      "epoch": 1.2624410686223153,
+      "loss_partial": 0.04931640625,
+      "step": 2410
+    },
+    {
+      "epoch": 1.2624410686223153,
+      "loss_partial": 0.050048828125,
+      "step": 2410
+    },
+    {
+      "epoch": 1.2629649030906234,
+      "grad_norm": 2.6938838958740234,
+      "learning_rate": 1.7299942719459226e-05,
+      "loss": 0.1083,
+      "step": 2411
+    },
+    {
+      "epoch": 1.2629649030906234,
+      "loss_partial": 0.048095703125,
+      "step": 2411
+    },
+    {
+      "epoch": 1.2629649030906234,
+      "loss_partial": 0.0576171875,
+      "step": 2411
+    },
+    {
+      "epoch": 1.2634887375589314,
+      "grad_norm": 1.870066523551941,
+      "learning_rate": 1.729681749261591e-05,
+      "loss": 0.1046,
+      "step": 2412
+    },
+    {
+      "epoch": 1.2634887375589314,
+      "loss_partial": 0.04541015625,
+      "step": 2412
+    },
+    {
+      "epoch": 1.2634887375589314,
+      "loss_partial": 0.05615234375,
+      "step": 2412
+    },
+    {
+      "epoch": 1.2640125720272395,
+      "grad_norm": 2.0772898197174072,
+      "learning_rate": 1.7293690740782862e-05,
+      "loss": 0.1092,
+      "step": 2413
+    },
+    {
+      "epoch": 1.2640125720272395,
+      "loss_partial": 0.051025390625,
+      "step": 2413
+    },
+    {
+      "epoch": 1.2640125720272395,
+      "loss_partial": 0.050537109375,
+      "step": 2413
+    },
+    {
+      "epoch": 1.2645364064955473,
+      "grad_norm": 1.4462018013000488,
+      "learning_rate": 1.729056246461355e-05,
+      "loss": 0.1056,
+      "step": 2414
+    },
+    {
+      "epoch": 1.2645364064955473,
+      "loss_partial": 0.0439453125,
+      "step": 2414
+    },
+    {
+      "epoch": 1.2645364064955473,
+      "loss_partial": 0.05029296875,
+      "step": 2414
+    },
+    {
+      "epoch": 1.2650602409638554,
+      "grad_norm": 2.2030534744262695,
+      "learning_rate": 1.7287432664761764e-05,
+      "loss": 0.1082,
+      "step": 2415
+    },
+    {
+      "epoch": 1.2650602409638554,
+      "loss_partial": 0.0546875,
+      "step": 2415
+    },
+    {
+      "epoch": 1.2650602409638554,
+      "loss_partial": 0.0546875,
+      "step": 2415
+    },
+    {
+      "epoch": 1.2655840754321634,
+      "grad_norm": 1.8735567331314087,
+      "learning_rate": 1.728430134188162e-05,
+      "loss": 0.1105,
+      "step": 2416
+    },
+    {
+      "epoch": 1.2655840754321634,
+      "loss_partial": 0.0439453125,
+      "step": 2416
+    },
+    {
+      "epoch": 1.2655840754321634,
+      "loss_partial": 0.045654296875,
+      "step": 2416
+    },
+    {
+      "epoch": 1.2661079099004715,
+      "grad_norm": 1.709725260734558,
+      "learning_rate": 1.728116849662754e-05,
+      "loss": 0.1045,
+      "step": 2417
+    },
+    {
+      "epoch": 1.2661079099004715,
+      "loss_partial": 0.057373046875,
+      "step": 2417
+    },
+    {
+      "epoch": 1.2661079099004715,
+      "loss_partial": 0.054931640625,
+      "step": 2417
+    },
+    {
+      "epoch": 1.2666317443687793,
+      "grad_norm": 1.4491032361984253,
+      "learning_rate": 1.7278034129654265e-05,
+      "loss": 0.1073,
+      "step": 2418
+    },
+    {
+      "epoch": 1.2666317443687793,
+      "loss_partial": 0.0595703125,
+      "step": 2418
+    },
+    {
+      "epoch": 1.2666317443687793,
+      "loss_partial": 0.05517578125,
+      "step": 2418
+    },
+    {
+      "epoch": 1.2671555788370874,
+      "grad_norm": 1.9610782861709595,
+      "learning_rate": 1.7274898241616867e-05,
+      "loss": 0.1081,
+      "step": 2419
+    },
+    {
+      "epoch": 1.2671555788370874,
+      "loss_partial": 0.05810546875,
+      "step": 2419
+    },
+    {
+      "epoch": 1.2671555788370874,
+      "loss_partial": 0.04638671875,
+      "step": 2419
+    },
+    {
+      "epoch": 1.2676794133053955,
+      "grad_norm": 1.5237140655517578,
+      "learning_rate": 1.727176083317072e-05,
+      "loss": 0.1093,
+      "step": 2420
+    },
+    {
+      "epoch": 1.2676794133053955,
+      "loss_partial": 0.052734375,
+      "step": 2420
+    },
+    {
+      "epoch": 1.2676794133053955,
+      "loss_partial": 0.06005859375,
+      "step": 2420
+    },
+    {
+      "epoch": 1.2682032477737035,
+      "grad_norm": 2.3252930641174316,
+      "learning_rate": 1.726862190497153e-05,
+      "loss": 0.1054,
+      "step": 2421
+    },
+    {
+      "epoch": 1.2682032477737035,
+      "loss_partial": 0.058349609375,
+      "step": 2421
+    },
+    {
+      "epoch": 1.2682032477737035,
+      "loss_partial": 0.057861328125,
+      "step": 2421
+    },
+    {
+      "epoch": 1.2687270822420116,
+      "grad_norm": 2.5265793800354004,
+      "learning_rate": 1.7265481457675307e-05,
+      "loss": 0.1099,
+      "step": 2422
+    },
+    {
+      "epoch": 1.2687270822420116,
+      "loss_partial": 0.057373046875,
+      "step": 2422
+    },
+    {
+      "epoch": 1.2687270822420116,
+      "loss_partial": 0.05126953125,
+      "step": 2422
+    },
+    {
+      "epoch": 1.2692509167103196,
+      "grad_norm": 2.9764373302459717,
+      "learning_rate": 1.7262339491938388e-05,
+      "loss": 0.106,
+      "step": 2423
+    },
+    {
+      "epoch": 1.2692509167103196,
+      "loss_partial": 0.057373046875,
+      "step": 2423
+    },
+    {
+      "epoch": 1.2692509167103196,
+      "loss_partial": 0.052001953125,
+      "step": 2423
+    },
+    {
+      "epoch": 1.2697747511786275,
+      "grad_norm": 1.5970494747161865,
+      "learning_rate": 1.7259196008417425e-05,
+      "loss": 0.1046,
+      "step": 2424
+    },
+    {
+      "epoch": 1.2697747511786275,
+      "loss_partial": 0.05712890625,
+      "step": 2424
+    },
+    {
+      "epoch": 1.2697747511786275,
+      "loss_partial": 0.046875,
+      "step": 2424
+    },
+    {
+      "epoch": 1.2702985856469355,
+      "grad_norm": 2.4368228912353516,
+      "learning_rate": 1.7256051007769384e-05,
+      "loss": 0.1048,
+      "step": 2425
+    },
+    {
+      "epoch": 1.2702985856469355,
+      "loss_partial": 0.0615234375,
+      "step": 2425
+    },
+    {
+      "epoch": 1.2702985856469355,
+      "loss_partial": 0.05517578125,
+      "step": 2425
+    },
+    {
+      "epoch": 1.2708224201152436,
+      "grad_norm": 1.6079792976379395,
+      "learning_rate": 1.7252904490651556e-05,
+      "loss": 0.1096,
+      "step": 2426
+    },
+    {
+      "epoch": 1.2708224201152436,
+      "loss_partial": 0.0615234375,
+      "step": 2426
+    },
+    {
+      "epoch": 1.2708224201152436,
+      "loss_partial": 0.052978515625,
+      "step": 2426
+    },
+    {
+      "epoch": 1.2713462545835517,
+      "grad_norm": 1.934826135635376,
+      "learning_rate": 1.724975645772154e-05,
+      "loss": 0.1106,
+      "step": 2427
+    },
+    {
+      "epoch": 1.2713462545835517,
+      "loss_partial": 0.0556640625,
+      "step": 2427
+    },
+    {
+      "epoch": 1.2713462545835517,
+      "loss_partial": 0.052734375,
+      "step": 2427
+    },
+    {
+      "epoch": 1.2718700890518595,
+      "grad_norm": 1.9331886768341064,
+      "learning_rate": 1.7246606909637252e-05,
+      "loss": 0.1054,
+      "step": 2428
+    },
+    {
+      "epoch": 1.2718700890518595,
+      "loss_partial": 0.04296875,
+      "step": 2428
+    },
+    {
+      "epoch": 1.2718700890518595,
+      "loss_partial": 0.059326171875,
+      "step": 2428
+    },
+    {
+      "epoch": 1.2723939235201676,
+      "grad_norm": 2.1888225078582764,
+      "learning_rate": 1.724345584705693e-05,
+      "loss": 0.1096,
+      "step": 2429
+    },
+    {
+      "epoch": 1.2723939235201676,
+      "loss_partial": 0.05810546875,
+      "step": 2429
+    },
+    {
+      "epoch": 1.2723939235201676,
+      "loss_partial": 0.05224609375,
+      "step": 2429
+    },
+    {
+      "epoch": 1.2729177579884756,
+      "grad_norm": 2.4058454036712646,
+      "learning_rate": 1.7240303270639132e-05,
+      "loss": 0.1077,
+      "step": 2430
+    },
+    {
+      "epoch": 1.2729177579884756,
+      "loss_partial": 0.047119140625,
+      "step": 2430
+    },
+    {
+      "epoch": 1.2729177579884756,
+      "loss_partial": 0.052978515625,
+      "step": 2430
+    },
+    {
+      "epoch": 1.2734415924567837,
+      "grad_norm": 1.5090633630752563,
+      "learning_rate": 1.723714918104272e-05,
+      "loss": 0.1058,
+      "step": 2431
+    },
+    {
+      "epoch": 1.2734415924567837,
+      "loss_partial": 0.050048828125,
+      "step": 2431
+    },
+    {
+      "epoch": 1.2734415924567837,
+      "loss_partial": 0.046630859375,
+      "step": 2431
+    },
+    {
+      "epoch": 1.2739654269250917,
+      "grad_norm": 2.0541272163391113,
+      "learning_rate": 1.7233993578926886e-05,
+      "loss": 0.1053,
+      "step": 2432
+    },
+    {
+      "epoch": 1.2739654269250917,
+      "loss_partial": 0.040771484375,
+      "step": 2432
+    },
+    {
+      "epoch": 1.2739654269250917,
+      "loss_partial": 0.050537109375,
+      "step": 2432
+    },
+    {
+      "epoch": 1.2744892613933998,
+      "grad_norm": 1.9610031843185425,
+      "learning_rate": 1.7230836464951126e-05,
+      "loss": 0.103,
+      "step": 2433
+    },
+    {
+      "epoch": 1.2744892613933998,
+      "loss_partial": 0.055419921875,
+      "step": 2433
+    },
+    {
+      "epoch": 1.2744892613933998,
+      "loss_partial": 0.057373046875,
+      "step": 2433
+    },
+    {
+      "epoch": 1.2750130958617076,
+      "grad_norm": 2.258450508117676,
+      "learning_rate": 1.722767783977526e-05,
+      "loss": 0.1057,
+      "step": 2434
+    },
+    {
+      "epoch": 1.2750130958617076,
+      "loss_partial": 0.05712890625,
+      "step": 2434
+    },
+    {
+      "epoch": 1.2750130958617076,
+      "loss_partial": 0.046630859375,
+      "step": 2434
+    },
+    {
+      "epoch": 1.2755369303300157,
+      "grad_norm": 1.8151472806930542,
+      "learning_rate": 1.7224517704059418e-05,
+      "loss": 0.1054,
+      "step": 2435
+    },
+    {
+      "epoch": 1.2755369303300157,
+      "loss_partial": 0.058349609375,
+      "step": 2435
+    },
+    {
+      "epoch": 1.2755369303300157,
+      "loss_partial": 0.0517578125,
+      "step": 2435
+    },
+    {
+      "epoch": 1.2760607647983238,
+      "grad_norm": 1.9549849033355713,
+      "learning_rate": 1.7221356058464054e-05,
+      "loss": 0.1082,
+      "step": 2436
+    },
+    {
+      "epoch": 1.2760607647983238,
+      "loss_partial": 0.056640625,
+      "step": 2436
+    },
+    {
+      "epoch": 1.2760607647983238,
+      "loss_partial": 0.0498046875,
+      "step": 2436
+    },
+    {
+      "epoch": 1.2765845992666318,
+      "grad_norm": 1.847550392150879,
+      "learning_rate": 1.7218192903649926e-05,
+      "loss": 0.1065,
+      "step": 2437
+    },
+    {
+      "epoch": 1.2765845992666318,
+      "loss_partial": 0.05029296875,
+      "step": 2437
+    },
+    {
+      "epoch": 1.2765845992666318,
+      "loss_partial": 0.048583984375,
+      "step": 2437
+    },
+    {
+      "epoch": 1.2771084337349397,
+      "grad_norm": 2.2522311210632324,
+      "learning_rate": 1.721502824027812e-05,
+      "loss": 0.1095,
+      "step": 2438
+    },
+    {
+      "epoch": 1.2771084337349397,
+      "loss_partial": 0.0498046875,
+      "step": 2438
+    },
+    {
+      "epoch": 1.2771084337349397,
+      "loss_partial": 0.05810546875,
+      "step": 2438
+    },
+    {
+      "epoch": 1.2776322682032477,
+      "grad_norm": 2.007896661758423,
+      "learning_rate": 1.7211862069010027e-05,
+      "loss": 0.1073,
+      "step": 2439
+    },
+    {
+      "epoch": 1.2776322682032477,
+      "loss_partial": 0.04931640625,
+      "step": 2439
+    },
+    {
+      "epoch": 1.2776322682032477,
+      "loss_partial": 0.048828125,
+      "step": 2439
+    },
+    {
+      "epoch": 1.2781561026715558,
+      "grad_norm": 1.9542776346206665,
+      "learning_rate": 1.7208694390507356e-05,
+      "loss": 0.1084,
+      "step": 2440
+    },
+    {
+      "epoch": 1.2781561026715558,
+      "loss_partial": 0.05712890625,
+      "step": 2440
+    },
+    {
+      "epoch": 1.2781561026715558,
+      "loss_partial": 0.055419921875,
+      "step": 2440
+    },
+    {
+      "epoch": 1.2786799371398638,
+      "grad_norm": 1.6021829843521118,
+      "learning_rate": 1.7205525205432136e-05,
+      "loss": 0.1076,
+      "step": 2441
+    },
+    {
+      "epoch": 1.2786799371398638,
+      "loss_partial": 0.04833984375,
+      "step": 2441
+    },
+    {
+      "epoch": 1.2786799371398638,
+      "loss_partial": 0.049072265625,
+      "step": 2441
+    },
+    {
+      "epoch": 1.279203771608172,
+      "grad_norm": 1.4790079593658447,
+      "learning_rate": 1.7202354514446708e-05,
+      "loss": 0.1034,
+      "step": 2442
+    },
+    {
+      "epoch": 1.279203771608172,
+      "loss_partial": 0.052734375,
+      "step": 2442
+    },
+    {
+      "epoch": 1.279203771608172,
+      "loss_partial": 0.053955078125,
+      "step": 2442
+    },
+    {
+      "epoch": 1.27972760607648,
+      "grad_norm": 2.117745876312256,
+      "learning_rate": 1.7199182318213724e-05,
+      "loss": 0.1099,
+      "step": 2443
+    },
+    {
+      "epoch": 1.27972760607648,
+      "loss_partial": 0.046875,
+      "step": 2443
+    },
+    {
+      "epoch": 1.27972760607648,
+      "loss_partial": 0.05322265625,
+      "step": 2443
+    },
+    {
+      "epoch": 1.2802514405447878,
+      "grad_norm": 2.1710453033447266,
+      "learning_rate": 1.719600861739615e-05,
+      "loss": 0.1061,
+      "step": 2444
+    },
+    {
+      "epoch": 1.2802514405447878,
+      "loss_partial": 0.051513671875,
+      "step": 2444
+    },
+    {
+      "epoch": 1.2802514405447878,
+      "loss_partial": 0.054931640625,
+      "step": 2444
+    },
+    {
+      "epoch": 1.2807752750130958,
+      "grad_norm": 1.681764841079712,
+      "learning_rate": 1.7192833412657277e-05,
+      "loss": 0.1063,
+      "step": 2445
+    },
+    {
+      "epoch": 1.2807752750130958,
+      "loss_partial": 0.05419921875,
+      "step": 2445
+    },
+    {
+      "epoch": 1.2807752750130958,
+      "loss_partial": 0.044921875,
+      "step": 2445
+    },
+    {
+      "epoch": 1.281299109481404,
+      "grad_norm": 1.9181956052780151,
+      "learning_rate": 1.71896567046607e-05,
+      "loss": 0.1046,
+      "step": 2446
+    },
+    {
+      "epoch": 1.281299109481404,
+      "loss_partial": 0.055908203125,
+      "step": 2446
+    },
+    {
+      "epoch": 1.281299109481404,
+      "loss_partial": 0.05029296875,
+      "step": 2446
+    },
+    {
+      "epoch": 1.281822943949712,
+      "grad_norm": 2.265064001083374,
+      "learning_rate": 1.7186478494070332e-05,
+      "loss": 0.1125,
+      "step": 2447
+    },
+    {
+      "epoch": 1.281822943949712,
+      "loss_partial": 0.057373046875,
+      "step": 2447
+    },
+    {
+      "epoch": 1.281822943949712,
+      "loss_partial": 0.0595703125,
+      "step": 2447
+    },
+    {
+      "epoch": 1.2823467784180198,
+      "grad_norm": 1.4983469247817993,
+      "learning_rate": 1.71832987815504e-05,
+      "loss": 0.1059,
+      "step": 2448
+    },
+    {
+      "epoch": 1.2823467784180198,
+      "loss_partial": 0.05517578125,
+      "step": 2448
+    },
+    {
+      "epoch": 1.2823467784180198,
+      "loss_partial": 0.04443359375,
+      "step": 2448
+    },
+    {
+      "epoch": 1.2828706128863279,
+      "grad_norm": 2.195044994354248,
+      "learning_rate": 1.7180117567765434e-05,
+      "loss": 0.1048,
+      "step": 2449
+    },
+    {
+      "epoch": 1.2828706128863279,
+      "loss_partial": 0.049560546875,
+      "step": 2449
+    },
+    {
+      "epoch": 1.2828706128863279,
+      "loss_partial": 0.051025390625,
+      "step": 2449
+    },
+    {
+      "epoch": 1.283394447354636,
+      "grad_norm": 2.0027363300323486,
+      "learning_rate": 1.7176934853380305e-05,
+      "loss": 0.1066,
+      "step": 2450
+    },
+    {
+      "epoch": 1.283394447354636,
+      "loss_partial": 0.044677734375,
+      "step": 2450
+    },
+    {
+      "epoch": 1.283394447354636,
+      "loss_partial": 0.06005859375,
+      "step": 2450
+    },
+    {
+      "epoch": 1.283918281822944,
+      "grad_norm": 3.362736225128174,
+      "learning_rate": 1.7173750639060173e-05,
+      "loss": 0.1091,
+      "step": 2451
+    },
+    {
+      "epoch": 1.283918281822944,
+      "loss_partial": 0.0517578125,
+      "step": 2451
+    },
+    {
+      "epoch": 1.283918281822944,
+      "loss_partial": 0.049560546875,
+      "step": 2451
+    },
+    {
+      "epoch": 1.284442116291252,
+      "grad_norm": 3.2353312969207764,
+      "learning_rate": 1.717056492547051e-05,
+      "loss": 0.1132,
+      "step": 2452
+    },
+    {
+      "epoch": 1.284442116291252,
+      "loss_partial": 0.060302734375,
+      "step": 2452
+    },
+    {
+      "epoch": 1.284442116291252,
+      "loss_partial": 0.06591796875,
+      "step": 2452
+    },
+    {
+      "epoch": 1.28496595075956,
+      "grad_norm": 1.7264126539230347,
+      "learning_rate": 1.7167377713277126e-05,
+      "loss": 0.1094,
+      "step": 2453
+    },
+    {
+      "epoch": 1.28496595075956,
+      "loss_partial": 0.061767578125,
+      "step": 2453
+    },
+    {
+      "epoch": 1.28496595075956,
+      "loss_partial": 0.048828125,
+      "step": 2453
+    },
+    {
+      "epoch": 1.285489785227868,
+      "grad_norm": 1.683086633682251,
+      "learning_rate": 1.7164189003146125e-05,
+      "loss": 0.1103,
+      "step": 2454
+    },
+    {
+      "epoch": 1.285489785227868,
+      "loss_partial": 0.047607421875,
+      "step": 2454
+    },
+    {
+      "epoch": 1.285489785227868,
+      "loss_partial": 0.051025390625,
+      "step": 2454
+    },
+    {
+      "epoch": 1.286013619696176,
+      "grad_norm": 2.199923515319824,
+      "learning_rate": 1.7160998795743918e-05,
+      "loss": 0.107,
+      "step": 2455
+    },
+    {
+      "epoch": 1.286013619696176,
+      "loss_partial": 0.0625,
+      "step": 2455
+    },
+    {
+      "epoch": 1.286013619696176,
+      "loss_partial": 0.053466796875,
+      "step": 2455
+    },
+    {
+      "epoch": 1.286537454164484,
+      "grad_norm": 1.8684533834457397,
+      "learning_rate": 1.715780709173725e-05,
+      "loss": 0.1077,
+      "step": 2456
+    },
+    {
+      "epoch": 1.286537454164484,
+      "loss_partial": 0.052734375,
+      "step": 2456
+    },
+    {
+      "epoch": 1.286537454164484,
+      "loss_partial": 0.0498046875,
+      "step": 2456
+    },
+    {
+      "epoch": 1.287061288632792,
+      "grad_norm": 1.8349395990371704,
+      "learning_rate": 1.7154613891793155e-05,
+      "loss": 0.1083,
+      "step": 2457
+    },
+    {
+      "epoch": 1.287061288632792,
+      "loss_partial": 0.0546875,
+      "step": 2457
+    },
+    {
+      "epoch": 1.287061288632792,
+      "loss_partial": 0.044189453125,
+      "step": 2457
+    },
+    {
+      "epoch": 1.2875851231011,
+      "grad_norm": 1.3041032552719116,
+      "learning_rate": 1.7151419196579005e-05,
+      "loss": 0.1033,
+      "step": 2458
+    },
+    {
+      "epoch": 1.2875851231011,
+      "loss_partial": 0.0625,
+      "step": 2458
+    },
+    {
+      "epoch": 1.2875851231011,
+      "loss_partial": 0.055908203125,
+      "step": 2458
+    },
+    {
+      "epoch": 1.288108957569408,
+      "grad_norm": 2.1932971477508545,
+      "learning_rate": 1.7148223006762466e-05,
+      "loss": 0.1079,
+      "step": 2459
+    },
+    {
+      "epoch": 1.288108957569408,
+      "loss_partial": 0.0556640625,
+      "step": 2459
+    },
+    {
+      "epoch": 1.288108957569408,
+      "loss_partial": 0.057861328125,
+      "step": 2459
+    },
+    {
+      "epoch": 1.288632792037716,
+      "grad_norm": 1.7836629152297974,
+      "learning_rate": 1.714502532301152e-05,
+      "loss": 0.1066,
+      "step": 2460
+    },
+    {
+      "epoch": 1.288632792037716,
+      "loss_partial": 0.049560546875,
+      "step": 2460
+    },
+    {
+      "epoch": 1.288632792037716,
+      "loss_partial": 0.053955078125,
+      "step": 2460
+    },
+    {
+      "epoch": 1.2891566265060241,
+      "grad_norm": 2.0578224658966064,
+      "learning_rate": 1.7141826145994467e-05,
+      "loss": 0.107,
+      "step": 2461
+    },
+    {
+      "epoch": 1.2891566265060241,
+      "loss_partial": 0.048583984375,
+      "step": 2461
+    },
+    {
+      "epoch": 1.2891566265060241,
+      "loss_partial": 0.0546875,
+      "step": 2461
+    },
+    {
+      "epoch": 1.2896804609743322,
+      "grad_norm": 1.7020906209945679,
+      "learning_rate": 1.713862547637991e-05,
+      "loss": 0.1065,
+      "step": 2462
+    },
+    {
+      "epoch": 1.2896804609743322,
+      "loss_partial": 0.056640625,
+      "step": 2462
+    },
+    {
+      "epoch": 1.2896804609743322,
+      "loss_partial": 0.0576171875,
+      "step": 2462
+    },
+    {
+      "epoch": 1.2902042954426403,
+      "grad_norm": 2.4571402072906494,
+      "learning_rate": 1.7135423314836774e-05,
+      "loss": 0.1096,
+      "step": 2463
+    },
+    {
+      "epoch": 1.2902042954426403,
+      "loss_partial": 0.048583984375,
+      "step": 2463
+    },
+    {
+      "epoch": 1.2902042954426403,
+      "loss_partial": 0.05078125,
+      "step": 2463
+    },
+    {
+      "epoch": 1.290728129910948,
+      "grad_norm": 2.0472559928894043,
+      "learning_rate": 1.713221966203429e-05,
+      "loss": 0.1062,
+      "step": 2464
+    },
+    {
+      "epoch": 1.290728129910948,
+      "loss_partial": 0.05078125,
+      "step": 2464
+    },
+    {
+      "epoch": 1.290728129910948,
+      "loss_partial": 0.0498046875,
+      "step": 2464
+    },
+    {
+      "epoch": 1.2912519643792562,
+      "grad_norm": 1.8440275192260742,
+      "learning_rate": 1.7129014518642e-05,
+      "loss": 0.1057,
+      "step": 2465
+    },
+    {
+      "epoch": 1.2912519643792562,
+      "loss_partial": 0.05517578125,
+      "step": 2465
+    },
+    {
+      "epoch": 1.2912519643792562,
+      "loss_partial": 0.051513671875,
+      "step": 2465
+    },
+    {
+      "epoch": 1.2917757988475642,
+      "grad_norm": 1.5193103551864624,
+      "learning_rate": 1.712580788532976e-05,
+      "loss": 0.105,
+      "step": 2466
+    },
+    {
+      "epoch": 1.2917757988475642,
+      "loss_partial": 0.054931640625,
+      "step": 2466
+    },
+    {
+      "epoch": 1.2917757988475642,
+      "loss_partial": 0.04443359375,
+      "step": 2466
+    },
+    {
+      "epoch": 1.292299633315872,
+      "grad_norm": 1.6439557075500488,
+      "learning_rate": 1.7122599762767736e-05,
+      "loss": 0.103,
+      "step": 2467
+    },
+    {
+      "epoch": 1.292299633315872,
+      "loss_partial": 0.046142578125,
+      "step": 2467
+    },
+    {
+      "epoch": 1.292299633315872,
+      "loss_partial": 0.056396484375,
+      "step": 2467
+    },
+    {
+      "epoch": 1.2928234677841801,
+      "grad_norm": 1.8413804769515991,
+      "learning_rate": 1.711939015162641e-05,
+      "loss": 0.1046,
+      "step": 2468
+    },
+    {
+      "epoch": 1.2928234677841801,
+      "loss_partial": 0.05322265625,
+      "step": 2468
+    },
+    {
+      "epoch": 1.2928234677841801,
+      "loss_partial": 0.053955078125,
+      "step": 2468
+    },
+    {
+      "epoch": 1.2933473022524882,
+      "grad_norm": 1.3116263151168823,
+      "learning_rate": 1.7116179052576565e-05,
+      "loss": 0.1036,
+      "step": 2469
+    },
+    {
+      "epoch": 1.2933473022524882,
+      "loss_partial": 0.047119140625,
+      "step": 2469
+    },
+    {
+      "epoch": 1.2933473022524882,
+      "loss_partial": 0.052001953125,
+      "step": 2469
+    },
+    {
+      "epoch": 1.2938711367207962,
+      "grad_norm": 3.1703720092773438,
+      "learning_rate": 1.7112966466289304e-05,
+      "loss": 0.1047,
+      "step": 2470
+    },
+    {
+      "epoch": 1.2938711367207962,
+      "loss_partial": 0.050537109375,
+      "step": 2470
+    },
+    {
+      "epoch": 1.2938711367207962,
+      "loss_partial": 0.054443359375,
+      "step": 2470
+    },
+    {
+      "epoch": 1.2943949711891043,
+      "grad_norm": 1.5497941970825195,
+      "learning_rate": 1.7109752393436035e-05,
+      "loss": 0.1103,
+      "step": 2471
+    },
+    {
+      "epoch": 1.2943949711891043,
+      "loss_partial": 0.040771484375,
+      "step": 2471
+    },
+    {
+      "epoch": 1.2943949711891043,
+      "loss_partial": 0.061767578125,
+      "step": 2471
+    },
+    {
+      "epoch": 1.2949188056574124,
+      "grad_norm": 2.5998013019561768,
+      "learning_rate": 1.710653683468848e-05,
+      "loss": 0.107,
+      "step": 2472
+    },
+    {
+      "epoch": 1.2949188056574124,
+      "loss_partial": 0.05615234375,
+      "step": 2472
+    },
+    {
+      "epoch": 1.2949188056574124,
+      "loss_partial": 0.049560546875,
+      "step": 2472
+    },
+    {
+      "epoch": 1.2954426401257202,
+      "grad_norm": 1.9776848554611206,
+      "learning_rate": 1.7103319790718675e-05,
+      "loss": 0.1058,
+      "step": 2473
+    },
+    {
+      "epoch": 1.2954426401257202,
+      "loss_partial": 0.05859375,
+      "step": 2473
+    },
+    {
+      "epoch": 1.2954426401257202,
+      "loss_partial": 0.054443359375,
+      "step": 2473
+    },
+    {
+      "epoch": 1.2959664745940283,
+      "grad_norm": 1.493775486946106,
+      "learning_rate": 1.710010126219896e-05,
+      "loss": 0.1082,
+      "step": 2474
+    },
+    {
+      "epoch": 1.2959664745940283,
+      "loss_partial": 0.052978515625,
+      "step": 2474
+    },
+    {
+      "epoch": 1.2959664745940283,
+      "loss_partial": 0.0458984375,
+      "step": 2474
+    },
+    {
+      "epoch": 1.2964903090623363,
+      "grad_norm": 1.684777855873108,
+      "learning_rate": 1.7096881249801983e-05,
+      "loss": 0.1052,
+      "step": 2475
+    },
+    {
+      "epoch": 1.2964903090623363,
+      "loss_partial": 0.051513671875,
+      "step": 2475
+    },
+    {
+      "epoch": 1.2964903090623363,
+      "loss_partial": 0.051025390625,
+      "step": 2475
+    },
+    {
+      "epoch": 1.2970141435306444,
+      "grad_norm": 2.219202756881714,
+      "learning_rate": 1.7093659754200714e-05,
+      "loss": 0.1062,
+      "step": 2476
+    },
+    {
+      "epoch": 1.2970141435306444,
+      "loss_partial": 0.04345703125,
+      "step": 2476
+    },
+    {
+      "epoch": 1.2970141435306444,
+      "loss_partial": 0.058837890625,
+      "step": 2476
+    },
+    {
+      "epoch": 1.2975379779989522,
+      "grad_norm": 2.07788348197937,
+      "learning_rate": 1.7090436776068422e-05,
+      "loss": 0.1063,
+      "step": 2477
+    },
+    {
+      "epoch": 1.2975379779989522,
+      "loss_partial": 0.052734375,
+      "step": 2477
+    },
+    {
+      "epoch": 1.2975379779989522,
+      "loss_partial": 0.056396484375,
+      "step": 2477
+    },
+    {
+      "epoch": 1.2980618124672603,
+      "grad_norm": 2.3151140213012695,
+      "learning_rate": 1.7087212316078687e-05,
+      "loss": 0.1105,
+      "step": 2478
+    },
+    {
+      "epoch": 1.2980618124672603,
+      "loss_partial": 0.046630859375,
+      "step": 2478
+    },
+    {
+      "epoch": 1.2980618124672603,
+      "loss_partial": 0.057861328125,
+      "step": 2478
+    },
+    {
+      "epoch": 1.2985856469355683,
+      "grad_norm": 2.0728745460510254,
+      "learning_rate": 1.7083986374905408e-05,
+      "loss": 0.106,
+      "step": 2479
+    },
+    {
+      "epoch": 1.2985856469355683,
+      "loss_partial": 0.044677734375,
+      "step": 2479
+    },
+    {
+      "epoch": 1.2985856469355683,
+      "loss_partial": 0.04833984375,
+      "step": 2479
+    },
+    {
+      "epoch": 1.2991094814038764,
+      "grad_norm": 1.9099867343902588,
+      "learning_rate": 1.7080758953222777e-05,
+      "loss": 0.1039,
+      "step": 2480
+    },
+    {
+      "epoch": 1.2991094814038764,
+      "loss_partial": 0.05908203125,
+      "step": 2480
+    },
+    {
+      "epoch": 1.2991094814038764,
+      "loss_partial": 0.0498046875,
+      "step": 2480
+    },
+    {
+      "epoch": 1.2996333158721844,
+      "grad_norm": 1.4293816089630127,
+      "learning_rate": 1.7077530051705316e-05,
+      "loss": 0.106,
+      "step": 2481
+    },
+    {
+      "epoch": 1.2996333158721844,
+      "loss_partial": 0.0693359375,
+      "step": 2481
+    },
+    {
+      "epoch": 1.2996333158721844,
+      "loss_partial": 0.061767578125,
+      "step": 2481
+    },
+    {
+      "epoch": 1.3001571503404925,
+      "grad_norm": 1.7678766250610352,
+      "learning_rate": 1.707429967102784e-05,
+      "loss": 0.1084,
+      "step": 2482
+    },
+    {
+      "epoch": 1.3001571503404925,
+      "loss_partial": 0.060791015625,
+      "step": 2482
+    },
+    {
+      "epoch": 1.3001571503404925,
+      "loss_partial": 0.043212890625,
+      "step": 2482
+    },
+    {
+      "epoch": 1.3006809848088003,
+      "grad_norm": 2.0998642444610596,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 0.1085,
+      "step": 2483
+    },
+    {
+      "epoch": 1.3006809848088003,
+      "loss_partial": 0.048828125,
+      "step": 2483
+    },
+    {
+      "epoch": 1.3006809848088003,
+      "loss_partial": 0.051025390625,
+      "step": 2483
+    },
+    {
+      "epoch": 1.3012048192771084,
+      "grad_norm": 1.6753689050674438,
+      "learning_rate": 1.7067834474893666e-05,
+      "loss": 0.1048,
+      "step": 2484
+    },
+    {
+      "epoch": 1.3012048192771084,
+      "loss_partial": 0.047607421875,
+      "step": 2484
+    },
+    {
+      "epoch": 1.3012048192771084,
+      "loss_partial": 0.0537109375,
+      "step": 2484
+    },
+    {
+      "epoch": 1.3017286537454165,
+      "grad_norm": 2.1330151557922363,
+      "learning_rate": 1.706459966078816e-05,
+      "loss": 0.1063,
+      "step": 2485
+    },
+    {
+      "epoch": 1.3017286537454165,
+      "loss_partial": 0.05859375,
+      "step": 2485
+    },
+    {
+      "epoch": 1.3017286537454165,
+      "loss_partial": 0.04931640625,
+      "step": 2485
+    },
+    {
+      "epoch": 1.3022524882137245,
+      "grad_norm": 1.8193528652191162,
+      "learning_rate": 1.706136337022501e-05,
+      "loss": 0.1104,
+      "step": 2486
+    },
+    {
+      "epoch": 1.3022524882137245,
+      "loss_partial": 0.05126953125,
+      "step": 2486
+    },
+    {
+      "epoch": 1.3022524882137245,
+      "loss_partial": 0.04638671875,
+      "step": 2486
+    },
+    {
+      "epoch": 1.3027763226820324,
+      "grad_norm": 1.9610450267791748,
+      "learning_rate": 1.705812560388058e-05,
+      "loss": 0.1028,
+      "step": 2487
+    },
+    {
+      "epoch": 1.3027763226820324,
+      "loss_partial": 0.052001953125,
+      "step": 2487
+    },
+    {
+      "epoch": 1.3027763226820324,
+      "loss_partial": 0.055419921875,
+      "step": 2487
+    },
+    {
+      "epoch": 1.3033001571503404,
+      "grad_norm": 2.1237974166870117,
+      "learning_rate": 1.7054886362431547e-05,
+      "loss": 0.1101,
+      "step": 2488
+    },
+    {
+      "epoch": 1.3033001571503404,
+      "loss_partial": 0.04833984375,
+      "step": 2488
+    },
+    {
+      "epoch": 1.3033001571503404,
+      "loss_partial": 0.05029296875,
+      "step": 2488
+    },
+    {
+      "epoch": 1.3038239916186485,
+      "grad_norm": 2.726062774658203,
+      "learning_rate": 1.7051645646554886e-05,
+      "loss": 0.1107,
+      "step": 2489
+    },
+    {
+      "epoch": 1.3038239916186485,
+      "loss_partial": 0.048095703125,
+      "step": 2489
+    },
+    {
+      "epoch": 1.3038239916186485,
+      "loss_partial": 0.058837890625,
+      "step": 2489
+    },
+    {
+      "epoch": 1.3043478260869565,
+      "grad_norm": 3.1448233127593994,
+      "learning_rate": 1.7048403456927894e-05,
+      "loss": 0.1109,
+      "step": 2490
+    },
+    {
+      "epoch": 1.3043478260869565,
+      "loss_partial": 0.0673828125,
+      "step": 2490
+    },
+    {
+      "epoch": 1.3043478260869565,
+      "loss_partial": 0.059326171875,
+      "step": 2490
+    },
+    {
+      "epoch": 1.3048716605552646,
+      "grad_norm": 1.8101024627685547,
+      "learning_rate": 1.7045159794228166e-05,
+      "loss": 0.11,
+      "step": 2491
+    },
+    {
+      "epoch": 1.3048716605552646,
+      "loss_partial": 0.05859375,
+      "step": 2491
+    },
+    {
+      "epoch": 1.3048716605552646,
+      "loss_partial": 0.049072265625,
+      "step": 2491
+    },
+    {
+      "epoch": 1.3053954950235727,
+      "grad_norm": 2.3391761779785156,
+      "learning_rate": 1.7041914659133608e-05,
+      "loss": 0.1061,
+      "step": 2492
+    },
+    {
+      "epoch": 1.3053954950235727,
+      "loss_partial": 0.0498046875,
+      "step": 2492
+    },
+    {
+      "epoch": 1.3053954950235727,
+      "loss_partial": 0.057861328125,
+      "step": 2492
+    },
+    {
+      "epoch": 1.3059193294918805,
+      "grad_norm": 2.5070085525512695,
+      "learning_rate": 1.703866805232243e-05,
+      "loss": 0.1064,
+      "step": 2493
+    },
+    {
+      "epoch": 1.3059193294918805,
+      "loss_partial": 0.048583984375,
+      "step": 2493
+    },
+    {
+      "epoch": 1.3059193294918805,
+      "loss_partial": 0.052978515625,
+      "step": 2493
+    },
+    {
+      "epoch": 1.3064431639601886,
+      "grad_norm": 1.7604641914367676,
+      "learning_rate": 1.7035419974473155e-05,
+      "loss": 0.1056,
+      "step": 2494
+    },
+    {
+      "epoch": 1.3064431639601886,
+      "loss_partial": 0.05419921875,
+      "step": 2494
+    },
+    {
+      "epoch": 1.3064431639601886,
+      "loss_partial": 0.05322265625,
+      "step": 2494
+    },
+    {
+      "epoch": 1.3069669984284966,
+      "grad_norm": 2.3360683917999268,
+      "learning_rate": 1.703217042626461e-05,
+      "loss": 0.1078,
+      "step": 2495
+    },
+    {
+      "epoch": 1.3069669984284966,
+      "loss_partial": 0.047119140625,
+      "step": 2495
+    },
+    {
+      "epoch": 1.3069669984284966,
+      "loss_partial": 0.059326171875,
+      "step": 2495
+    },
+    {
+      "epoch": 1.3074908328968047,
+      "grad_norm": 1.7378966808319092,
+      "learning_rate": 1.7028919408375937e-05,
+      "loss": 0.1068,
+      "step": 2496
+    },
+    {
+      "epoch": 1.3074908328968047,
+      "loss_partial": 0.0576171875,
+      "step": 2496
+    },
+    {
+      "epoch": 1.3074908328968047,
+      "loss_partial": 0.048095703125,
+      "step": 2496
+    },
+    {
+      "epoch": 1.3080146673651125,
+      "grad_norm": 1.9857231378555298,
+      "learning_rate": 1.7025666921486566e-05,
+      "loss": 0.1056,
+      "step": 2497
+    },
+    {
+      "epoch": 1.3080146673651125,
+      "loss_partial": 0.045654296875,
+      "step": 2497
+    },
+    {
+      "epoch": 1.3080146673651125,
+      "loss_partial": 0.05810546875,
+      "step": 2497
+    },
+    {
+      "epoch": 1.3085385018334206,
+      "grad_norm": 2.0125534534454346,
+      "learning_rate": 1.702241296627626e-05,
+      "loss": 0.1042,
+      "step": 2498
+    },
+    {
+      "epoch": 1.3085385018334206,
+      "loss_partial": 0.0615234375,
+      "step": 2498
+    },
+    {
+      "epoch": 1.3085385018334206,
+      "loss_partial": 0.0537109375,
+      "step": 2498
+    },
+    {
+      "epoch": 1.3090623363017286,
+      "grad_norm": 2.0567286014556885,
+      "learning_rate": 1.7019157543425063e-05,
+      "loss": 0.1082,
+      "step": 2499
+    },
+    {
+      "epoch": 1.3090623363017286,
+      "loss_partial": 0.051513671875,
+      "step": 2499
+    },
+    {
+      "epoch": 1.3090623363017286,
+      "loss_partial": 0.058837890625,
+      "step": 2499
+    },
+    {
+      "epoch": 1.3095861707700367,
+      "grad_norm": 1.7580924034118652,
+      "learning_rate": 1.701590065361335e-05,
+      "loss": 0.1089,
+      "step": 2500
+    },
+    {
+      "epoch": 1.3095861707700367,
+      "loss_partial": 0.044921875,
+      "step": 2500
+    },
+    {
+      "epoch": 1.3095861707700367,
+      "loss_partial": 0.0546875,
+      "step": 2500
+    },
+    {
+      "epoch": 1.3101100052383448,
+      "grad_norm": 1.897868275642395,
+      "learning_rate": 1.7012642297521785e-05,
+      "loss": 0.1064,
+      "step": 2501
+    },
+    {
+      "epoch": 1.3101100052383448,
+      "loss_partial": 0.048583984375,
+      "step": 2501
+    },
+    {
+      "epoch": 1.3101100052383448,
+      "loss_partial": 0.055419921875,
+      "step": 2501
+    },
+    {
+      "epoch": 1.3106338397066528,
+      "grad_norm": 1.673621416091919,
+      "learning_rate": 1.7009382475831344e-05,
+      "loss": 0.1053,
+      "step": 2502
+    },
+    {
+      "epoch": 1.3106338397066528,
+      "loss_partial": 0.05322265625,
+      "step": 2502
+    },
+    {
+      "epoch": 1.3106338397066528,
+      "loss_partial": 0.05029296875,
+      "step": 2502
+    },
+    {
+      "epoch": 1.3111576741749607,
+      "grad_norm": 1.7159337997436523,
+      "learning_rate": 1.700612118922331e-05,
+      "loss": 0.1049,
+      "step": 2503
+    },
+    {
+      "epoch": 1.3111576741749607,
+      "loss_partial": 0.049072265625,
+      "step": 2503
+    },
+    {
+      "epoch": 1.3111576741749607,
+      "loss_partial": 0.0498046875,
+      "step": 2503
+    },
+    {
+      "epoch": 1.3116815086432687,
+      "grad_norm": 1.902611494064331,
+      "learning_rate": 1.700285843837927e-05,
+      "loss": 0.1066,
+      "step": 2504
+    },
+    {
+      "epoch": 1.3116815086432687,
+      "loss_partial": 0.048828125,
+      "step": 2504
+    },
+    {
+      "epoch": 1.3116815086432687,
+      "loss_partial": 0.05419921875,
+      "step": 2504
+    },
+    {
+      "epoch": 1.3122053431115768,
+      "grad_norm": 1.256926417350769,
+      "learning_rate": 1.699959422398112e-05,
+      "loss": 0.1073,
+      "step": 2505
+    },
+    {
+      "epoch": 1.3122053431115768,
+      "loss_partial": 0.0478515625,
+      "step": 2505
+    },
+    {
+      "epoch": 1.3122053431115768,
+      "loss_partial": 0.04736328125,
+      "step": 2505
+    },
+    {
+      "epoch": 1.3127291775798848,
+      "grad_norm": 1.5790635347366333,
+      "learning_rate": 1.6996328546711068e-05,
+      "loss": 0.1055,
+      "step": 2506
+    },
+    {
+      "epoch": 1.3127291775798848,
+      "loss_partial": 0.041015625,
+      "step": 2506
+    },
+    {
+      "epoch": 1.3127291775798848,
+      "loss_partial": 0.055419921875,
+      "step": 2506
+    },
+    {
+      "epoch": 1.3132530120481927,
+      "grad_norm": 1.6210788488388062,
+      "learning_rate": 1.6993061407251606e-05,
+      "loss": 0.1037,
+      "step": 2507
+    },
+    {
+      "epoch": 1.3132530120481927,
+      "loss_partial": 0.048583984375,
+      "step": 2507
+    },
+    {
+      "epoch": 1.3132530120481927,
+      "loss_partial": 0.045166015625,
+      "step": 2507
+    },
+    {
+      "epoch": 1.3137768465165007,
+      "grad_norm": 1.9911184310913086,
+      "learning_rate": 1.698979280628556e-05,
+      "loss": 0.1078,
+      "step": 2508
+    },
+    {
+      "epoch": 1.3137768465165007,
+      "loss_partial": 0.05517578125,
+      "step": 2508
+    },
+    {
+      "epoch": 1.3137768465165007,
+      "loss_partial": 0.057373046875,
+      "step": 2508
+    },
+    {
+      "epoch": 1.3143006809848088,
+      "grad_norm": 1.9967161417007446,
+      "learning_rate": 1.6986522744496035e-05,
+      "loss": 0.1078,
+      "step": 2509
+    },
+    {
+      "epoch": 1.3143006809848088,
+      "loss_partial": 0.04736328125,
+      "step": 2509
+    },
+    {
+      "epoch": 1.3143006809848088,
+      "loss_partial": 0.05126953125,
+      "step": 2509
+    },
+    {
+      "epoch": 1.3148245154531168,
+      "grad_norm": 1.5034115314483643,
+      "learning_rate": 1.6983251222566465e-05,
+      "loss": 0.1019,
+      "step": 2510
+    },
+    {
+      "epoch": 1.3148245154531168,
+      "loss_partial": 0.052001953125,
+      "step": 2510
+    },
+    {
+      "epoch": 1.3148245154531168,
+      "loss_partial": 0.053955078125,
+      "step": 2510
+    },
+    {
+      "epoch": 1.315348349921425,
+      "grad_norm": 1.5173554420471191,
+      "learning_rate": 1.697997824118057e-05,
+      "loss": 0.1092,
+      "step": 2511
+    },
+    {
+      "epoch": 1.315348349921425,
+      "loss_partial": 0.057861328125,
+      "step": 2511
+    },
+    {
+      "epoch": 1.315348349921425,
+      "loss_partial": 0.05908203125,
+      "step": 2511
+    },
+    {
+      "epoch": 1.315872184389733,
+      "grad_norm": 1.5773869752883911,
+      "learning_rate": 1.6976703801022386e-05,
+      "loss": 0.1073,
+      "step": 2512
+    },
+    {
+      "epoch": 1.315872184389733,
+      "loss_partial": 0.048583984375,
+      "step": 2512
+    },
+    {
+      "epoch": 1.315872184389733,
+      "loss_partial": 0.05859375,
+      "step": 2512
+    },
+    {
+      "epoch": 1.3163960188580408,
+      "grad_norm": 1.6395657062530518,
+      "learning_rate": 1.697342790277625e-05,
+      "loss": 0.1035,
+      "step": 2513
+    },
+    {
+      "epoch": 1.3163960188580408,
+      "loss_partial": 0.052001953125,
+      "step": 2513
+    },
+    {
+      "epoch": 1.3163960188580408,
+      "loss_partial": 0.054443359375,
+      "step": 2513
+    },
+    {
+      "epoch": 1.3169198533263489,
+      "grad_norm": 1.6242003440856934,
+      "learning_rate": 1.6970150547126806e-05,
+      "loss": 0.1065,
+      "step": 2514
+    },
+    {
+      "epoch": 1.3169198533263489,
+      "loss_partial": 0.059814453125,
+      "step": 2514
+    },
+    {
+      "epoch": 1.3169198533263489,
+      "loss_partial": 0.047119140625,
+      "step": 2514
+    },
+    {
+      "epoch": 1.317443687794657,
+      "grad_norm": 1.5596643686294556,
+      "learning_rate": 1.6966871734758997e-05,
+      "loss": 0.1059,
+      "step": 2515
+    },
+    {
+      "epoch": 1.317443687794657,
+      "loss_partial": 0.054931640625,
+      "step": 2515
+    },
+    {
+      "epoch": 1.317443687794657,
+      "loss_partial": 0.04541015625,
+      "step": 2515
+    },
+    {
+      "epoch": 1.317967522262965,
+      "grad_norm": 1.3608126640319824,
+      "learning_rate": 1.696359146635808e-05,
+      "loss": 0.1066,
+      "step": 2516
+    },
+    {
+      "epoch": 1.317967522262965,
+      "loss_partial": 0.0634765625,
+      "step": 2516
+    },
+    {
+      "epoch": 1.317967522262965,
+      "loss_partial": 0.0576171875,
+      "step": 2516
+    },
+    {
+      "epoch": 1.3184913567312728,
+      "grad_norm": 1.4934333562850952,
+      "learning_rate": 1.6960309742609603e-05,
+      "loss": 0.1074,
+      "step": 2517
+    },
+    {
+      "epoch": 1.3184913567312728,
+      "loss_partial": 0.054443359375,
+      "step": 2517
+    },
+    {
+      "epoch": 1.3184913567312728,
+      "loss_partial": 0.054931640625,
+      "step": 2517
+    },
+    {
+      "epoch": 1.3190151911995809,
+      "grad_norm": 1.5376369953155518,
+      "learning_rate": 1.6957026564199435e-05,
+      "loss": 0.1087,
+      "step": 2518
+    },
+    {
+      "epoch": 1.3190151911995809,
+      "loss_partial": 0.0546875,
+      "step": 2518
+    },
+    {
+      "epoch": 1.3190151911995809,
+      "loss_partial": 0.050537109375,
+      "step": 2518
+    },
+    {
+      "epoch": 1.319539025667889,
+      "grad_norm": 1.5920052528381348,
+      "learning_rate": 1.6953741931813738e-05,
+      "loss": 0.106,
+      "step": 2519
+    },
+    {
+      "epoch": 1.319539025667889,
+      "loss_partial": 0.0625,
+      "step": 2519
+    },
+    {
+      "epoch": 1.319539025667889,
+      "loss_partial": 0.052978515625,
+      "step": 2519
+    },
+    {
+      "epoch": 1.320062860136197,
+      "grad_norm": 1.801247000694275,
+      "learning_rate": 1.695045584613898e-05,
+      "loss": 0.106,
+      "step": 2520
+    },
+    {
+      "epoch": 1.320062860136197,
+      "loss_partial": 0.047119140625,
+      "step": 2520
+    },
+    {
+      "epoch": 1.320062860136197,
+      "loss_partial": 0.05224609375,
+      "step": 2520
+    },
+    {
+      "epoch": 1.320586694604505,
+      "grad_norm": 1.3762463331222534,
+      "learning_rate": 1.6947168307861923e-05,
+      "loss": 0.1066,
+      "step": 2521
+    },
+    {
+      "epoch": 1.320586694604505,
+      "loss_partial": 0.0576171875,
+      "step": 2521
+    },
+    {
+      "epoch": 1.320586694604505,
+      "loss_partial": 0.064453125,
+      "step": 2521
+    },
+    {
+      "epoch": 1.3211105290728131,
+      "grad_norm": 1.814895510673523,
+      "learning_rate": 1.6943879317669652e-05,
+      "loss": 0.1074,
+      "step": 2522
+    },
+    {
+      "epoch": 1.3211105290728131,
+      "loss_partial": 0.049072265625,
+      "step": 2522
+    },
+    {
+      "epoch": 1.3211105290728131,
+      "loss_partial": 0.068359375,
+      "step": 2522
+    },
+    {
+      "epoch": 1.321634363541121,
+      "grad_norm": 1.8445179462432861,
+      "learning_rate": 1.6940588876249545e-05,
+      "loss": 0.1061,
+      "step": 2523
+    },
+    {
+      "epoch": 1.321634363541121,
+      "loss_partial": 0.052978515625,
+      "step": 2523
+    },
+    {
+      "epoch": 1.321634363541121,
+      "loss_partial": 0.04638671875,
+      "step": 2523
+    },
+    {
+      "epoch": 1.322158198009429,
+      "grad_norm": 2.2000598907470703,
+      "learning_rate": 1.6937296984289283e-05,
+      "loss": 0.1056,
+      "step": 2524
+    },
+    {
+      "epoch": 1.322158198009429,
+      "loss_partial": 0.0537109375,
+      "step": 2524
+    },
+    {
+      "epoch": 1.322158198009429,
+      "loss_partial": 0.04833984375,
+      "step": 2524
+    },
+    {
+      "epoch": 1.322682032477737,
+      "grad_norm": 1.6337918043136597,
+      "learning_rate": 1.6934003642476848e-05,
+      "loss": 0.1034,
+      "step": 2525
+    },
+    {
+      "epoch": 1.322682032477737,
+      "loss_partial": 0.0537109375,
+      "step": 2525
+    },
+    {
+      "epoch": 1.322682032477737,
+      "loss_partial": 0.051025390625,
+      "step": 2525
+    },
+    {
+      "epoch": 1.323205866946045,
+      "grad_norm": 2.706295967102051,
+      "learning_rate": 1.693070885150053e-05,
+      "loss": 0.107,
+      "step": 2526
+    },
+    {
+      "epoch": 1.323205866946045,
+      "loss_partial": 0.048583984375,
+      "step": 2526
+    },
+    {
+      "epoch": 1.323205866946045,
+      "loss_partial": 0.042724609375,
+      "step": 2526
+    },
+    {
+      "epoch": 1.323729701414353,
+      "grad_norm": 3.9187424182891846,
+      "learning_rate": 1.6927412612048924e-05,
+      "loss": 0.1073,
+      "step": 2527
+    },
+    {
+      "epoch": 1.323729701414353,
+      "loss_partial": 0.052490234375,
+      "step": 2527
+    },
+    {
+      "epoch": 1.323729701414353,
+      "loss_partial": 0.044189453125,
+      "step": 2527
+    },
+    {
+      "epoch": 1.324253535882661,
+      "grad_norm": 2.0816469192504883,
+      "learning_rate": 1.6924114924810915e-05,
+      "loss": 0.1026,
+      "step": 2528
+    },
+    {
+      "epoch": 1.324253535882661,
+      "loss_partial": 0.048828125,
+      "step": 2528
+    },
+    {
+      "epoch": 1.324253535882661,
+      "loss_partial": 0.056884765625,
+      "step": 2528
+    },
+    {
+      "epoch": 1.324777370350969,
+      "grad_norm": 2.105233669281006,
+      "learning_rate": 1.6920815790475707e-05,
+      "loss": 0.1067,
+      "step": 2529
+    },
+    {
+      "epoch": 1.324777370350969,
+      "loss_partial": 0.0498046875,
+      "step": 2529
+    },
+    {
+      "epoch": 1.324777370350969,
+      "loss_partial": 0.052734375,
+      "step": 2529
+    },
+    {
+      "epoch": 1.3253012048192772,
+      "grad_norm": 1.5565385818481445,
+      "learning_rate": 1.6917515209732794e-05,
+      "loss": 0.1058,
+      "step": 2530
+    },
+    {
+      "epoch": 1.3253012048192772,
+      "loss_partial": 0.053466796875,
+      "step": 2530
+    },
+    {
+      "epoch": 1.3253012048192772,
+      "loss_partial": 0.050537109375,
+      "step": 2530
+    },
+    {
+      "epoch": 1.3258250392875852,
+      "grad_norm": 2.26682186126709,
+      "learning_rate": 1.691421318327198e-05,
+      "loss": 0.1074,
+      "step": 2531
+    },
+    {
+      "epoch": 1.3258250392875852,
+      "loss_partial": 0.04638671875,
+      "step": 2531
+    },
+    {
+      "epoch": 1.3258250392875852,
+      "loss_partial": 0.05078125,
+      "step": 2531
+    },
+    {
+      "epoch": 1.3263488737558933,
+      "grad_norm": 1.8382697105407715,
+      "learning_rate": 1.691090971178337e-05,
+      "loss": 0.103,
+      "step": 2532
+    },
+    {
+      "epoch": 1.3263488737558933,
+      "loss_partial": 0.05712890625,
+      "step": 2532
+    },
+    {
+      "epoch": 1.3263488737558933,
+      "loss_partial": 0.06787109375,
+      "step": 2532
+    },
+    {
+      "epoch": 1.3268727082242011,
+      "grad_norm": 2.530897617340088,
+      "learning_rate": 1.6907604795957363e-05,
+      "loss": 0.1088,
+      "step": 2533
+    },
+    {
+      "epoch": 1.3268727082242011,
+      "loss_partial": 0.05078125,
+      "step": 2533
+    },
+    {
+      "epoch": 1.3268727082242011,
+      "loss_partial": 0.0498046875,
+      "step": 2533
+    },
+    {
+      "epoch": 1.3273965426925092,
+      "grad_norm": 2.225139856338501,
+      "learning_rate": 1.6904298436484668e-05,
+      "loss": 0.1024,
+      "step": 2534
+    },
+    {
+      "epoch": 1.3273965426925092,
+      "loss_partial": 0.05224609375,
+      "step": 2534
+    },
+    {
+      "epoch": 1.3273965426925092,
+      "loss_partial": 0.055908203125,
+      "step": 2534
+    },
+    {
+      "epoch": 1.3279203771608172,
+      "grad_norm": 1.6775131225585938,
+      "learning_rate": 1.6900990634056293e-05,
+      "loss": 0.1007,
+      "step": 2535
+    },
+    {
+      "epoch": 1.3279203771608172,
+      "loss_partial": 0.06591796875,
+      "step": 2535
+    },
+    {
+      "epoch": 1.3279203771608172,
+      "loss_partial": 0.052734375,
+      "step": 2535
+    },
+    {
+      "epoch": 1.328444211629125,
+      "grad_norm": 2.6129956245422363,
+      "learning_rate": 1.6897681389363553e-05,
+      "loss": 0.109,
+      "step": 2536
+    },
+    {
+      "epoch": 1.328444211629125,
+      "loss_partial": 0.0517578125,
+      "step": 2536
+    },
+    {
+      "epoch": 1.328444211629125,
+      "loss_partial": 0.064453125,
+      "step": 2536
+    },
+    {
+      "epoch": 1.3289680460974331,
+      "grad_norm": 1.8077596426010132,
+      "learning_rate": 1.6894370703098058e-05,
+      "loss": 0.1064,
+      "step": 2537
+    },
+    {
+      "epoch": 1.3289680460974331,
+      "loss_partial": 0.05078125,
+      "step": 2537
+    },
+    {
+      "epoch": 1.3289680460974331,
+      "loss_partial": 0.050537109375,
+      "step": 2537
+    },
+    {
+      "epoch": 1.3294918805657412,
+      "grad_norm": 1.4247947931289673,
+      "learning_rate": 1.6891058575951718e-05,
+      "loss": 0.1039,
+      "step": 2538
+    },
+    {
+      "epoch": 1.3294918805657412,
+      "loss_partial": 0.052978515625,
+      "step": 2538
+    },
+    {
+      "epoch": 1.3294918805657412,
+      "loss_partial": 0.06640625,
+      "step": 2538
+    },
+    {
+      "epoch": 1.3300157150340493,
+      "grad_norm": 1.7924370765686035,
+      "learning_rate": 1.6887745008616745e-05,
+      "loss": 0.1078,
+      "step": 2539
+    },
+    {
+      "epoch": 1.3300157150340493,
+      "loss_partial": 0.049560546875,
+      "step": 2539
+    },
+    {
+      "epoch": 1.3300157150340493,
+      "loss_partial": 0.056640625,
+      "step": 2539
+    },
+    {
+      "epoch": 1.3305395495023573,
+      "grad_norm": 1.721665620803833,
+      "learning_rate": 1.6884430001785663e-05,
+      "loss": 0.1084,
+      "step": 2540
+    },
+    {
+      "epoch": 1.3305395495023573,
+      "loss_partial": 0.05126953125,
+      "step": 2540
+    },
+    {
+      "epoch": 1.3305395495023573,
+      "loss_partial": 0.056884765625,
+      "step": 2540
+    },
+    {
+      "epoch": 1.3310633839706654,
+      "grad_norm": 2.9097461700439453,
+      "learning_rate": 1.688111355615128e-05,
+      "loss": 0.1078,
+      "step": 2541
+    },
+    {
+      "epoch": 1.3310633839706654,
+      "loss_partial": 0.04296875,
+      "step": 2541
+    },
+    {
+      "epoch": 1.3310633839706654,
+      "loss_partial": 0.05029296875,
+      "step": 2541
+    },
+    {
+      "epoch": 1.3315872184389732,
+      "grad_norm": 2.1348021030426025,
+      "learning_rate": 1.687779567240672e-05,
+      "loss": 0.1046,
+      "step": 2542
+    },
+    {
+      "epoch": 1.3315872184389732,
+      "loss_partial": 0.05419921875,
+      "step": 2542
+    },
+    {
+      "epoch": 1.3315872184389732,
+      "loss_partial": 0.051025390625,
+      "step": 2542
+    },
+    {
+      "epoch": 1.3321110529072813,
+      "grad_norm": 1.2556531429290771,
+      "learning_rate": 1.687447635124539e-05,
+      "loss": 0.104,
+      "step": 2543
+    },
+    {
+      "epoch": 1.3321110529072813,
+      "loss_partial": 0.05615234375,
+      "step": 2543
+    },
+    {
+      "epoch": 1.3321110529072813,
+      "loss_partial": 0.060791015625,
+      "step": 2543
+    },
+    {
+      "epoch": 1.3326348873755893,
+      "grad_norm": 1.4898686408996582,
+      "learning_rate": 1.6871155593361012e-05,
+      "loss": 0.1082,
+      "step": 2544
+    },
+    {
+      "epoch": 1.3326348873755893,
+      "loss_partial": 0.054931640625,
+      "step": 2544
+    },
+    {
+      "epoch": 1.3326348873755893,
+      "loss_partial": 0.0478515625,
+      "step": 2544
+    },
+    {
+      "epoch": 1.3331587218438974,
+      "grad_norm": 1.9333842992782593,
+      "learning_rate": 1.686783339944761e-05,
+      "loss": 0.1092,
+      "step": 2545
+    },
+    {
+      "epoch": 1.3331587218438974,
+      "loss_partial": 0.045654296875,
+      "step": 2545
+    },
+    {
+      "epoch": 1.3331587218438974,
+      "loss_partial": 0.0498046875,
+      "step": 2545
+    },
+    {
+      "epoch": 1.3336825563122052,
+      "grad_norm": 1.5476531982421875,
+      "learning_rate": 1.6864509770199492e-05,
+      "loss": 0.1075,
+      "step": 2546
+    },
+    {
+      "epoch": 1.3336825563122052,
+      "loss_partial": 0.04296875,
+      "step": 2546
+    },
+    {
+      "epoch": 1.3336825563122052,
+      "loss_partial": 0.0537109375,
+      "step": 2546
+    },
+    {
+      "epoch": 1.3342063907805133,
+      "grad_norm": 1.600963830947876,
+      "learning_rate": 1.6861184706311284e-05,
+      "loss": 0.1085,
+      "step": 2547
+    },
+    {
+      "epoch": 1.3342063907805133,
+      "loss_partial": 0.04736328125,
+      "step": 2547
+    },
+    {
+      "epoch": 1.3342063907805133,
+      "loss_partial": 0.0654296875,
+      "step": 2547
+    },
+    {
+      "epoch": 1.3347302252488213,
+      "grad_norm": 1.8654841184616089,
+      "learning_rate": 1.68578582084779e-05,
+      "loss": 0.1077,
+      "step": 2548
+    },
+    {
+      "epoch": 1.3347302252488213,
+      "loss_partial": 0.05517578125,
+      "step": 2548
+    },
+    {
+      "epoch": 1.3347302252488213,
+      "loss_partial": 0.056396484375,
+      "step": 2548
+    },
+    {
+      "epoch": 1.3352540597171294,
+      "grad_norm": 1.6855294704437256,
+      "learning_rate": 1.6854530277394557e-05,
+      "loss": 0.1086,
+      "step": 2549
+    },
+    {
+      "epoch": 1.3352540597171294,
+      "loss_partial": 0.053955078125,
+      "step": 2549
+    },
+    {
+      "epoch": 1.3352540597171294,
+      "loss_partial": 0.0537109375,
+      "step": 2549
+    },
+    {
+      "epoch": 1.3357778941854375,
+      "grad_norm": 1.474547028541565,
+      "learning_rate": 1.6851200913756774e-05,
+      "loss": 0.108,
+      "step": 2550
+    },
+    {
+      "epoch": 1.3357778941854375,
+      "loss_partial": 0.0498046875,
+      "step": 2550
+    },
+    {
+      "epoch": 1.3357778941854375,
+      "loss_partial": 0.056884765625,
+      "step": 2550
+    },
+    {
+      "epoch": 1.3363017286537455,
+      "grad_norm": 2.5148749351501465,
+      "learning_rate": 1.6847870118260368e-05,
+      "loss": 0.1085,
+      "step": 2551
+    },
+    {
+      "epoch": 1.3363017286537455,
+      "loss_partial": 0.051513671875,
+      "step": 2551
+    },
+    {
+      "epoch": 1.3363017286537455,
+      "loss_partial": 0.06201171875,
+      "step": 2551
+    },
+    {
+      "epoch": 1.3368255631220534,
+      "grad_norm": 2.2981367111206055,
+      "learning_rate": 1.6844537891601455e-05,
+      "loss": 0.1074,
+      "step": 2552
+    },
+    {
+      "epoch": 1.3368255631220534,
+      "loss_partial": 0.06201171875,
+      "step": 2552
+    },
+    {
+      "epoch": 1.3368255631220534,
+      "loss_partial": 0.057373046875,
+      "step": 2552
+    },
+    {
+      "epoch": 1.3373493975903614,
+      "grad_norm": 2.4515957832336426,
+      "learning_rate": 1.684120423447644e-05,
+      "loss": 0.1115,
+      "step": 2553
+    },
+    {
+      "epoch": 1.3373493975903614,
+      "loss_partial": 0.050048828125,
+      "step": 2553
+    },
+    {
+      "epoch": 1.3373493975903614,
+      "loss_partial": 0.050537109375,
+      "step": 2553
+    },
+    {
+      "epoch": 1.3378732320586695,
+      "grad_norm": 1.8667863607406616,
+      "learning_rate": 1.6837869147582052e-05,
+      "loss": 0.1089,
+      "step": 2554
+    },
+    {
+      "epoch": 1.3378732320586695,
+      "loss_partial": 0.0458984375,
+      "step": 2554
+    },
+    {
+      "epoch": 1.3378732320586695,
+      "loss_partial": 0.05615234375,
+      "step": 2554
+    },
+    {
+      "epoch": 1.3383970665269775,
+      "grad_norm": 1.3677138090133667,
+      "learning_rate": 1.6834532631615295e-05,
+      "loss": 0.106,
+      "step": 2555
+    },
+    {
+      "epoch": 1.3383970665269775,
+      "loss_partial": 0.048095703125,
+      "step": 2555
+    },
+    {
+      "epoch": 1.3383970665269775,
+      "loss_partial": 0.060546875,
+      "step": 2555
+    },
+    {
+      "epoch": 1.3389209009952854,
+      "grad_norm": 1.9567636251449585,
+      "learning_rate": 1.683119468727348e-05,
+      "loss": 0.1084,
+      "step": 2556
+    },
+    {
+      "epoch": 1.3389209009952854,
+      "loss_partial": 0.048828125,
+      "step": 2556
+    },
+    {
+      "epoch": 1.3389209009952854,
+      "loss_partial": 0.05419921875,
+      "step": 2556
+    },
+    {
+      "epoch": 1.3394447354635934,
+      "grad_norm": 1.4876645803451538,
+      "learning_rate": 1.682785531525422e-05,
+      "loss": 0.1042,
+      "step": 2557
+    },
+    {
+      "epoch": 1.3394447354635934,
+      "loss_partial": 0.05078125,
+      "step": 2557
+    },
+    {
+      "epoch": 1.3394447354635934,
+      "loss_partial": 0.048828125,
+      "step": 2557
+    },
+    {
+      "epoch": 1.3399685699319015,
+      "grad_norm": 1.8194169998168945,
+      "learning_rate": 1.682451451625542e-05,
+      "loss": 0.1055,
+      "step": 2558
+    },
+    {
+      "epoch": 1.3399685699319015,
+      "loss_partial": 0.047119140625,
+      "step": 2558
+    },
+    {
+      "epoch": 1.3399685699319015,
+      "loss_partial": 0.05224609375,
+      "step": 2558
+    },
+    {
+      "epoch": 1.3404924044002096,
+      "grad_norm": 1.9920153617858887,
+      "learning_rate": 1.6821172290975286e-05,
+      "loss": 0.1065,
+      "step": 2559
+    },
+    {
+      "epoch": 1.3404924044002096,
+      "loss_partial": 0.0517578125,
+      "step": 2559
+    },
+    {
+      "epoch": 1.3404924044002096,
+      "loss_partial": 0.050048828125,
+      "step": 2559
+    },
+    {
+      "epoch": 1.3410162388685176,
+      "grad_norm": 1.8870298862457275,
+      "learning_rate": 1.6817828640112326e-05,
+      "loss": 0.1099,
+      "step": 2560
+    },
+    {
+      "epoch": 1.3410162388685176,
+      "loss_partial": 0.0654296875,
+      "step": 2560
+    },
+    {
+      "epoch": 1.3410162388685176,
+      "loss_partial": 0.04833984375,
+      "step": 2560
+    },
+    {
+      "epoch": 1.3415400733368257,
+      "grad_norm": 1.6746149063110352,
+      "learning_rate": 1.681448356436534e-05,
+      "loss": 0.1054,
+      "step": 2561
+    },
+    {
+      "epoch": 1.3415400733368257,
+      "loss_partial": 0.0458984375,
+      "step": 2561
+    },
+    {
+      "epoch": 1.3415400733368257,
+      "loss_partial": 0.048828125,
+      "step": 2561
+    },
+    {
+      "epoch": 1.3420639078051335,
+      "grad_norm": 1.644716501235962,
+      "learning_rate": 1.681113706443343e-05,
+      "loss": 0.1048,
+      "step": 2562
+    },
+    {
+      "epoch": 1.3420639078051335,
+      "loss_partial": 0.0634765625,
+      "step": 2562
+    },
+    {
+      "epoch": 1.3420639078051335,
+      "loss_partial": 0.050537109375,
+      "step": 2562
+    },
+    {
+      "epoch": 1.3425877422734416,
+      "grad_norm": 2.3638720512390137,
+      "learning_rate": 1.6807789141015992e-05,
+      "loss": 0.1081,
+      "step": 2563
+    },
+    {
+      "epoch": 1.3425877422734416,
+      "loss_partial": 0.05419921875,
+      "step": 2563
+    },
+    {
+      "epoch": 1.3425877422734416,
+      "loss_partial": 0.05810546875,
+      "step": 2563
+    },
+    {
+      "epoch": 1.3431115767417496,
+      "grad_norm": 1.4860305786132812,
+      "learning_rate": 1.6804439794812717e-05,
+      "loss": 0.1048,
+      "step": 2564
+    },
+    {
+      "epoch": 1.3431115767417496,
+      "loss_partial": 0.0546875,
+      "step": 2564
+    },
+    {
+      "epoch": 1.3431115767417496,
+      "loss_partial": 0.062255859375,
+      "step": 2564
+    },
+    {
+      "epoch": 1.3436354112100577,
+      "grad_norm": 1.825968861579895,
+      "learning_rate": 1.6801089026523608e-05,
+      "loss": 0.1075,
+      "step": 2565
+    },
+    {
+      "epoch": 1.3436354112100577,
+      "loss_partial": 0.055908203125,
+      "step": 2565
+    },
+    {
+      "epoch": 1.3436354112100577,
+      "loss_partial": 0.0439453125,
+      "step": 2565
+    },
+    {
+      "epoch": 1.3441592456783655,
+      "grad_norm": 1.9290879964828491,
+      "learning_rate": 1.6797736836848947e-05,
+      "loss": 0.1082,
+      "step": 2566
+    },
+    {
+      "epoch": 1.3441592456783655,
+      "loss_partial": 0.050048828125,
+      "step": 2566
+    },
+    {
+      "epoch": 1.3441592456783655,
+      "loss_partial": 0.05810546875,
+      "step": 2566
+    },
+    {
+      "epoch": 1.3446830801466736,
+      "grad_norm": 1.7689162492752075,
+      "learning_rate": 1.6794383226489322e-05,
+      "loss": 0.1086,
+      "step": 2567
+    },
+    {
+      "epoch": 1.3446830801466736,
+      "loss_partial": 0.046630859375,
+      "step": 2567
+    },
+    {
+      "epoch": 1.3446830801466736,
+      "loss_partial": 0.05029296875,
+      "step": 2567
+    },
+    {
+      "epoch": 1.3452069146149817,
+      "grad_norm": 1.7440526485443115,
+      "learning_rate": 1.679102819614562e-05,
+      "loss": 0.1027,
+      "step": 2568
+    },
+    {
+      "epoch": 1.3452069146149817,
+      "loss_partial": 0.055908203125,
+      "step": 2568
+    },
+    {
+      "epoch": 1.3452069146149817,
+      "loss_partial": 0.06689453125,
+      "step": 2568
+    },
+    {
+      "epoch": 1.3457307490832897,
+      "grad_norm": 1.5327471494674683,
+      "learning_rate": 1.678767174651902e-05,
+      "loss": 0.1031,
+      "step": 2569
+    },
+    {
+      "epoch": 1.3457307490832897,
+      "loss_partial": 0.06201171875,
+      "step": 2569
+    },
+    {
+      "epoch": 1.3457307490832897,
+      "loss_partial": 0.0478515625,
+      "step": 2569
+    },
+    {
+      "epoch": 1.3462545835515978,
+      "grad_norm": 2.004622220993042,
+      "learning_rate": 1.6784313878310998e-05,
+      "loss": 0.1018,
+      "step": 2570
+    },
+    {
+      "epoch": 1.3462545835515978,
+      "loss_partial": 0.04736328125,
+      "step": 2570
+    },
+    {
+      "epoch": 1.3462545835515978,
+      "loss_partial": 0.046142578125,
+      "step": 2570
+    },
+    {
+      "epoch": 1.3467784180199058,
+      "grad_norm": 1.4656383991241455,
+      "learning_rate": 1.6780954592223327e-05,
+      "loss": 0.109,
+      "step": 2571
+    },
+    {
+      "epoch": 1.3467784180199058,
+      "loss_partial": 0.051025390625,
+      "step": 2571
+    },
+    {
+      "epoch": 1.3467784180199058,
+      "loss_partial": 0.049072265625,
+      "step": 2571
+    },
+    {
+      "epoch": 1.3473022524882137,
+      "grad_norm": 2.0977134704589844,
+      "learning_rate": 1.6777593888958077e-05,
+      "loss": 0.1066,
+      "step": 2572
+    },
+    {
+      "epoch": 1.3473022524882137,
+      "loss_partial": 0.053955078125,
+      "step": 2572
+    },
+    {
+      "epoch": 1.3473022524882137,
+      "loss_partial": 0.054931640625,
+      "step": 2572
+    },
+    {
+      "epoch": 1.3478260869565217,
+      "grad_norm": 2.416719913482666,
+      "learning_rate": 1.6774231769217617e-05,
+      "loss": 0.1109,
+      "step": 2573
+    },
+    {
+      "epoch": 1.3478260869565217,
+      "loss_partial": 0.0439453125,
+      "step": 2573
+    },
+    {
+      "epoch": 1.3478260869565217,
+      "loss_partial": 0.057861328125,
+      "step": 2573
+    },
+    {
+      "epoch": 1.3483499214248298,
+      "grad_norm": 1.7624162435531616,
+      "learning_rate": 1.6770868233704608e-05,
+      "loss": 0.108,
+      "step": 2574
+    },
+    {
+      "epoch": 1.3483499214248298,
+      "loss_partial": 0.06689453125,
+      "step": 2574
+    },
+    {
+      "epoch": 1.3483499214248298,
+      "loss_partial": 0.05126953125,
+      "step": 2574
+    },
+    {
+      "epoch": 1.3488737558931378,
+      "grad_norm": 1.9633654356002808,
+      "learning_rate": 1.676750328312201e-05,
+      "loss": 0.1073,
+      "step": 2575
+    },
+    {
+      "epoch": 1.3488737558931378,
+      "loss_partial": 0.06103515625,
+      "step": 2575
+    },
+    {
+      "epoch": 1.3488737558931378,
+      "loss_partial": 0.060302734375,
+      "step": 2575
+    },
+    {
+      "epoch": 1.3493975903614457,
+      "grad_norm": 2.1855270862579346,
+      "learning_rate": 1.6764136918173072e-05,
+      "loss": 0.108,
+      "step": 2576
+    },
+    {
+      "epoch": 1.3493975903614457,
+      "loss_partial": 0.048583984375,
+      "step": 2576
+    },
+    {
+      "epoch": 1.3493975903614457,
+      "loss_partial": 0.052978515625,
+      "step": 2576
+    },
+    {
+      "epoch": 1.3499214248297537,
+      "grad_norm": 1.884673833847046,
+      "learning_rate": 1.6760769139561345e-05,
+      "loss": 0.1037,
+      "step": 2577
+    },
+    {
+      "epoch": 1.3499214248297537,
+      "loss_partial": 0.052734375,
+      "step": 2577
+    },
+    {
+      "epoch": 1.3499214248297537,
+      "loss_partial": 0.051513671875,
+      "step": 2577
+    },
+    {
+      "epoch": 1.3504452592980618,
+      "grad_norm": 1.7634294033050537,
+      "learning_rate": 1.675739994799068e-05,
+      "loss": 0.1078,
+      "step": 2578
+    },
+    {
+      "epoch": 1.3504452592980618,
+      "loss_partial": 0.0498046875,
+      "step": 2578
+    },
+    {
+      "epoch": 1.3504452592980618,
+      "loss_partial": 0.053955078125,
+      "step": 2578
+    },
+    {
+      "epoch": 1.3509690937663699,
+      "grad_norm": 2.119130849838257,
+      "learning_rate": 1.675402934416521e-05,
+      "loss": 0.1068,
+      "step": 2579
+    },
+    {
+      "epoch": 1.3509690937663699,
+      "loss_partial": 0.053466796875,
+      "step": 2579
+    },
+    {
+      "epoch": 1.3509690937663699,
+      "loss_partial": 0.05322265625,
+      "step": 2579
+    },
+    {
+      "epoch": 1.351492928234678,
+      "grad_norm": 1.7892155647277832,
+      "learning_rate": 1.6750657328789372e-05,
+      "loss": 0.1085,
+      "step": 2580
+    },
+    {
+      "epoch": 1.351492928234678,
+      "loss_partial": 0.052734375,
+      "step": 2580
+    },
+    {
+      "epoch": 1.351492928234678,
+      "loss_partial": 0.04345703125,
+      "step": 2580
+    },
+    {
+      "epoch": 1.352016762702986,
+      "grad_norm": 1.9108575582504272,
+      "learning_rate": 1.67472839025679e-05,
+      "loss": 0.1111,
+      "step": 2581
+    },
+    {
+      "epoch": 1.352016762702986,
+      "loss_partial": 0.043701171875,
+      "step": 2581
+    },
+    {
+      "epoch": 1.352016762702986,
+      "loss_partial": 0.05419921875,
+      "step": 2581
+    },
+    {
+      "epoch": 1.3525405971712938,
+      "grad_norm": 1.6114813089370728,
+      "learning_rate": 1.674390906620581e-05,
+      "loss": 0.1049,
+      "step": 2582
+    },
+    {
+      "epoch": 1.3525405971712938,
+      "loss_partial": 0.05322265625,
+      "step": 2582
+    },
+    {
+      "epoch": 1.3525405971712938,
+      "loss_partial": 0.05517578125,
+      "step": 2582
+    },
+    {
+      "epoch": 1.3530644316396019,
+      "grad_norm": 1.6589665412902832,
+      "learning_rate": 1.6740532820408436e-05,
+      "loss": 0.1048,
+      "step": 2583
+    },
+    {
+      "epoch": 1.3530644316396019,
+      "loss_partial": 0.04248046875,
+      "step": 2583
+    },
+    {
+      "epoch": 1.3530644316396019,
+      "loss_partial": 0.059326171875,
+      "step": 2583
+    },
+    {
+      "epoch": 1.35358826610791,
+      "grad_norm": 1.514016032218933,
+      "learning_rate": 1.6737155165881377e-05,
+      "loss": 0.1094,
+      "step": 2584
+    },
+    {
+      "epoch": 1.35358826610791,
+      "loss_partial": 0.048828125,
+      "step": 2584
+    },
+    {
+      "epoch": 1.35358826610791,
+      "loss_partial": 0.0556640625,
+      "step": 2584
+    },
+    {
+      "epoch": 1.3541121005762178,
+      "grad_norm": 1.9289817810058594,
+      "learning_rate": 1.6733776103330555e-05,
+      "loss": 0.102,
+      "step": 2585
+    },
+    {
+      "epoch": 1.3541121005762178,
+      "loss_partial": 0.051025390625,
+      "step": 2585
+    },
+    {
+      "epoch": 1.3541121005762178,
+      "loss_partial": 0.0615234375,
+      "step": 2585
+    },
+    {
+      "epoch": 1.3546359350445258,
+      "grad_norm": 1.8410050868988037,
+      "learning_rate": 1.6730395633462166e-05,
+      "loss": 0.1124,
+      "step": 2586
+    },
+    {
+      "epoch": 1.3546359350445258,
+      "loss_partial": 0.05224609375,
+      "step": 2586
+    },
+    {
+      "epoch": 1.3546359350445258,
+      "loss_partial": 0.07470703125,
+      "step": 2586
+    },
+    {
+      "epoch": 1.355159769512834,
+      "grad_norm": 2.359964370727539,
+      "learning_rate": 1.672701375698271e-05,
+      "loss": 0.1084,
+      "step": 2587
+    },
+    {
+      "epoch": 1.355159769512834,
+      "loss_partial": 0.053466796875,
+      "step": 2587
+    },
+    {
+      "epoch": 1.355159769512834,
+      "loss_partial": 0.04541015625,
+      "step": 2587
+    },
+    {
+      "epoch": 1.355683603981142,
+      "grad_norm": 1.405699610710144,
+      "learning_rate": 1.6723630474598976e-05,
+      "loss": 0.1017,
+      "step": 2588
+    },
+    {
+      "epoch": 1.355683603981142,
+      "loss_partial": 0.05859375,
+      "step": 2588
+    },
+    {
+      "epoch": 1.355683603981142,
+      "loss_partial": 0.043212890625,
+      "step": 2588
+    },
+    {
+      "epoch": 1.35620743844945,
+      "grad_norm": 2.1427745819091797,
+      "learning_rate": 1.6720245787018053e-05,
+      "loss": 0.1062,
+      "step": 2589
+    },
+    {
+      "epoch": 1.35620743844945,
+      "loss_partial": 0.046142578125,
+      "step": 2589
+    },
+    {
+      "epoch": 1.35620743844945,
+      "loss_partial": 0.05126953125,
+      "step": 2589
+    },
+    {
+      "epoch": 1.356731272917758,
+      "grad_norm": 1.8558450937271118,
+      "learning_rate": 1.6716859694947315e-05,
+      "loss": 0.1056,
+      "step": 2590
+    },
+    {
+      "epoch": 1.356731272917758,
+      "loss_partial": 0.0546875,
+      "step": 2590
+    },
+    {
+      "epoch": 1.356731272917758,
+      "loss_partial": 0.047607421875,
+      "step": 2590
+    },
+    {
+      "epoch": 1.3572551073860661,
+      "grad_norm": 2.2706191539764404,
+      "learning_rate": 1.6713472199094435e-05,
+      "loss": 0.1069,
+      "step": 2591
+    },
+    {
+      "epoch": 1.3572551073860661,
+      "loss_partial": 0.046630859375,
+      "step": 2591
+    },
+    {
+      "epoch": 1.3572551073860661,
+      "loss_partial": 0.04931640625,
+      "step": 2591
+    },
+    {
+      "epoch": 1.357778941854374,
+      "grad_norm": 2.0172533988952637,
+      "learning_rate": 1.6710083300167383e-05,
+      "loss": 0.104,
+      "step": 2592
+    },
+    {
+      "epoch": 1.357778941854374,
+      "loss_partial": 0.052978515625,
+      "step": 2592
+    },
+    {
+      "epoch": 1.357778941854374,
+      "loss_partial": 0.054443359375,
+      "step": 2592
+    },
+    {
+      "epoch": 1.358302776322682,
+      "grad_norm": 2.110934257507324,
+      "learning_rate": 1.6706692998874413e-05,
+      "loss": 0.1028,
+      "step": 2593
+    },
+    {
+      "epoch": 1.358302776322682,
+      "loss_partial": 0.0537109375,
+      "step": 2593
+    },
+    {
+      "epoch": 1.358302776322682,
+      "loss_partial": 0.056396484375,
+      "step": 2593
+    },
+    {
+      "epoch": 1.35882661079099,
+      "grad_norm": 2.796727418899536,
+      "learning_rate": 1.6703301295924085e-05,
+      "loss": 0.1073,
+      "step": 2594
+    },
+    {
+      "epoch": 1.35882661079099,
+      "loss_partial": 0.041015625,
+      "step": 2594
+    },
+    {
+      "epoch": 1.35882661079099,
+      "loss_partial": 0.0693359375,
+      "step": 2594
+    },
+    {
+      "epoch": 1.359350445259298,
+      "grad_norm": 2.0795059204101562,
+      "learning_rate": 1.6699908192025233e-05,
+      "loss": 0.1004,
+      "step": 2595
+    },
+    {
+      "epoch": 1.359350445259298,
+      "loss_partial": 0.046875,
+      "step": 2595
+    },
+    {
+      "epoch": 1.359350445259298,
+      "loss_partial": 0.045654296875,
+      "step": 2595
+    },
+    {
+      "epoch": 1.359874279727606,
+      "grad_norm": 2.198732614517212,
+      "learning_rate": 1.6696513687887e-05,
+      "loss": 0.1047,
+      "step": 2596
+    },
+    {
+      "epoch": 1.359874279727606,
+      "loss_partial": 0.0498046875,
+      "step": 2596
+    },
+    {
+      "epoch": 1.359874279727606,
+      "loss_partial": 0.056884765625,
+      "step": 2596
+    },
+    {
+      "epoch": 1.360398114195914,
+      "grad_norm": 1.9551740884780884,
+      "learning_rate": 1.6693117784218818e-05,
+      "loss": 0.1099,
+      "step": 2597
+    },
+    {
+      "epoch": 1.360398114195914,
+      "loss_partial": 0.049072265625,
+      "step": 2597
+    },
+    {
+      "epoch": 1.360398114195914,
+      "loss_partial": 0.049560546875,
+      "step": 2597
+    },
+    {
+      "epoch": 1.3609219486642221,
+      "grad_norm": 2.6022355556488037,
+      "learning_rate": 1.6689720481730407e-05,
+      "loss": 0.1089,
+      "step": 2598
+    },
+    {
+      "epoch": 1.3609219486642221,
+      "loss_partial": 0.051513671875,
+      "step": 2598
+    },
+    {
+      "epoch": 1.3609219486642221,
+      "loss_partial": 0.05908203125,
+      "step": 2598
+    },
+    {
+      "epoch": 1.3614457831325302,
+      "grad_norm": 2.057737112045288,
+      "learning_rate": 1.6686321781131782e-05,
+      "loss": 0.108,
+      "step": 2599
+    },
+    {
+      "epoch": 1.3614457831325302,
+      "loss_partial": 0.05615234375,
+      "step": 2599
+    },
+    {
+      "epoch": 1.3614457831325302,
+      "loss_partial": 0.0625,
+      "step": 2599
+    },
+    {
+      "epoch": 1.3619696176008382,
+      "grad_norm": 2.5357303619384766,
+      "learning_rate": 1.6682921683133254e-05,
+      "loss": 0.1071,
+      "step": 2600
+    },
+    {
+      "epoch": 1.3619696176008382,
+      "loss_partial": 0.0625,
+      "step": 2600
+    },
+    {
+      "epoch": 1.3619696176008382,
+      "loss_partial": 0.049560546875,
+      "step": 2600
+    },
+    {
+      "epoch": 1.362493452069146,
+      "grad_norm": 1.8605287075042725,
+      "learning_rate": 1.6679520188445417e-05,
+      "loss": 0.1074,
+      "step": 2601
+    },
+    {
+      "epoch": 1.362493452069146,
+      "loss_partial": 0.056396484375,
+      "step": 2601
+    },
+    {
+      "epoch": 1.362493452069146,
+      "loss_partial": 0.051025390625,
+      "step": 2601
+    },
+    {
+      "epoch": 1.3630172865374541,
+      "grad_norm": 1.9693496227264404,
+      "learning_rate": 1.667611729777917e-05,
+      "loss": 0.11,
+      "step": 2602
+    },
+    {
+      "epoch": 1.3630172865374541,
+      "loss_partial": 0.04833984375,
+      "step": 2602
+    },
+    {
+      "epoch": 1.3630172865374541,
+      "loss_partial": 0.05419921875,
+      "step": 2602
+    },
+    {
+      "epoch": 1.3635411210057622,
+      "grad_norm": 1.6343779563903809,
+      "learning_rate": 1.667271301184569e-05,
+      "loss": 0.1034,
+      "step": 2603
+    },
+    {
+      "epoch": 1.3635411210057622,
+      "loss_partial": 0.057373046875,
+      "step": 2603
+    },
+    {
+      "epoch": 1.3635411210057622,
+      "loss_partial": 0.052490234375,
+      "step": 2603
+    },
+    {
+      "epoch": 1.3640649554740703,
+      "grad_norm": 1.9315255880355835,
+      "learning_rate": 1.6669307331356452e-05,
+      "loss": 0.1049,
+      "step": 2604
+    },
+    {
+      "epoch": 1.3640649554740703,
+      "loss_partial": 0.053955078125,
+      "step": 2604
+    },
+    {
+      "epoch": 1.3640649554740703,
+      "loss_partial": 0.0478515625,
+      "step": 2604
+    },
+    {
+      "epoch": 1.364588789942378,
+      "grad_norm": 2.1818645000457764,
+      "learning_rate": 1.6665900257023223e-05,
+      "loss": 0.1055,
+      "step": 2605
+    },
+    {
+      "epoch": 1.364588789942378,
+      "loss_partial": 0.06396484375,
+      "step": 2605
+    },
+    {
+      "epoch": 1.364588789942378,
+      "loss_partial": 0.055419921875,
+      "step": 2605
+    },
+    {
+      "epoch": 1.3651126244106861,
+      "grad_norm": 1.8073878288269043,
+      "learning_rate": 1.6662491789558062e-05,
+      "loss": 0.1064,
+      "step": 2606
+    },
+    {
+      "epoch": 1.3651126244106861,
+      "loss_partial": 0.049560546875,
+      "step": 2606
+    },
+    {
+      "epoch": 1.3651126244106861,
+      "loss_partial": 0.052001953125,
+      "step": 2606
+    },
+    {
+      "epoch": 1.3656364588789942,
+      "grad_norm": 2.399332046508789,
+      "learning_rate": 1.6659081929673316e-05,
+      "loss": 0.1077,
+      "step": 2607
+    },
+    {
+      "epoch": 1.3656364588789942,
+      "loss_partial": 0.05419921875,
+      "step": 2607
+    },
+    {
+      "epoch": 1.3656364588789942,
+      "loss_partial": 0.052001953125,
+      "step": 2607
+    },
+    {
+      "epoch": 1.3661602933473023,
+      "grad_norm": 1.7398499250411987,
+      "learning_rate": 1.665567067808163e-05,
+      "loss": 0.1059,
+      "step": 2608
+    },
+    {
+      "epoch": 1.3661602933473023,
+      "loss_partial": 0.048095703125,
+      "step": 2608
+    },
+    {
+      "epoch": 1.3661602933473023,
+      "loss_partial": 0.061279296875,
+      "step": 2608
+    },
+    {
+      "epoch": 1.3666841278156103,
+      "grad_norm": 1.8592264652252197,
+      "learning_rate": 1.6652258035495924e-05,
+      "loss": 0.1057,
+      "step": 2609
+    },
+    {
+      "epoch": 1.3666841278156103,
+      "loss_partial": 0.059326171875,
+      "step": 2609
+    },
+    {
+      "epoch": 1.3666841278156103,
+      "loss_partial": 0.0458984375,
+      "step": 2609
+    },
+    {
+      "epoch": 1.3672079622839184,
+      "grad_norm": 1.9104726314544678,
+      "learning_rate": 1.6648844002629427e-05,
+      "loss": 0.105,
+      "step": 2610
+    },
+    {
+      "epoch": 1.3672079622839184,
+      "loss_partial": 0.056640625,
+      "step": 2610
+    },
+    {
+      "epoch": 1.3672079622839184,
+      "loss_partial": 0.04931640625,
+      "step": 2610
+    },
+    {
+      "epoch": 1.3677317967522262,
+      "grad_norm": 2.0557048320770264,
+      "learning_rate": 1.664542858019565e-05,
+      "loss": 0.1036,
+      "step": 2611
+    },
+    {
+      "epoch": 1.3677317967522262,
+      "loss_partial": 0.04833984375,
+      "step": 2611
+    },
+    {
+      "epoch": 1.3677317967522262,
+      "loss_partial": 0.048828125,
+      "step": 2611
+    },
+    {
+      "epoch": 1.3682556312205343,
+      "grad_norm": 1.6146889925003052,
+      "learning_rate": 1.664201176890839e-05,
+      "loss": 0.1051,
+      "step": 2612
+    },
+    {
+      "epoch": 1.3682556312205343,
+      "loss_partial": 0.06201171875,
+      "step": 2612
+    },
+    {
+      "epoch": 1.3682556312205343,
+      "loss_partial": 0.051025390625,
+      "step": 2612
+    },
+    {
+      "epoch": 1.3687794656888423,
+      "grad_norm": 1.530405044555664,
+      "learning_rate": 1.6638593569481746e-05,
+      "loss": 0.1055,
+      "step": 2613
+    },
+    {
+      "epoch": 1.3687794656888423,
+      "loss_partial": 0.0537109375,
+      "step": 2613
+    },
+    {
+      "epoch": 1.3687794656888423,
+      "loss_partial": 0.042236328125,
+      "step": 2613
+    },
+    {
+      "epoch": 1.3693033001571504,
+      "grad_norm": 2.1039977073669434,
+      "learning_rate": 1.6635173982630097e-05,
+      "loss": 0.1061,
+      "step": 2614
+    },
+    {
+      "epoch": 1.3693033001571504,
+      "loss_partial": 0.05908203125,
+      "step": 2614
+    },
+    {
+      "epoch": 1.3693033001571504,
+      "loss_partial": 0.04638671875,
+      "step": 2614
+    },
+    {
+      "epoch": 1.3698271346254582,
+      "grad_norm": 1.4233124256134033,
+      "learning_rate": 1.6631753009068118e-05,
+      "loss": 0.1088,
+      "step": 2615
+    },
+    {
+      "epoch": 1.3698271346254582,
+      "loss_partial": 0.052734375,
+      "step": 2615
+    },
+    {
+      "epoch": 1.3698271346254582,
+      "loss_partial": 0.06103515625,
+      "step": 2615
+    },
+    {
+      "epoch": 1.3703509690937663,
+      "grad_norm": 1.7800705432891846,
+      "learning_rate": 1.6628330649510766e-05,
+      "loss": 0.1059,
+      "step": 2616
+    },
+    {
+      "epoch": 1.3703509690937663,
+      "loss_partial": 0.051513671875,
+      "step": 2616
+    },
+    {
+      "epoch": 1.3703509690937663,
+      "loss_partial": 0.0478515625,
+      "step": 2616
+    },
+    {
+      "epoch": 1.3708748035620744,
+      "grad_norm": 1.7824389934539795,
+      "learning_rate": 1.66249069046733e-05,
+      "loss": 0.1033,
+      "step": 2617
+    },
+    {
+      "epoch": 1.3708748035620744,
+      "loss_partial": 0.048095703125,
+      "step": 2617
+    },
+    {
+      "epoch": 1.3708748035620744,
+      "loss_partial": 0.05810546875,
+      "step": 2617
+    },
+    {
+      "epoch": 1.3713986380303824,
+      "grad_norm": 1.800998330116272,
+      "learning_rate": 1.6621481775271257e-05,
+      "loss": 0.1043,
+      "step": 2618
+    },
+    {
+      "epoch": 1.3713986380303824,
+      "loss_partial": 0.060546875,
+      "step": 2618
+    },
+    {
+      "epoch": 1.3713986380303824,
+      "loss_partial": 0.0556640625,
+      "step": 2618
+    },
+    {
+      "epoch": 1.3719224724986905,
+      "grad_norm": 2.06048846244812,
+      "learning_rate": 1.6618055262020467e-05,
+      "loss": 0.1089,
+      "step": 2619
+    },
+    {
+      "epoch": 1.3719224724986905,
+      "loss_partial": 0.0458984375,
+      "step": 2619
+    },
+    {
+      "epoch": 1.3719224724986905,
+      "loss_partial": 0.054931640625,
+      "step": 2619
+    },
+    {
+      "epoch": 1.3724463069669985,
+      "grad_norm": 2.591367721557617,
+      "learning_rate": 1.6614627365637055e-05,
+      "loss": 0.1079,
+      "step": 2620
+    },
+    {
+      "epoch": 1.3724463069669985,
+      "loss_partial": 0.051513671875,
+      "step": 2620
+    },
+    {
+      "epoch": 1.3724463069669985,
+      "loss_partial": 0.056640625,
+      "step": 2620
+    },
+    {
+      "epoch": 1.3729701414353064,
+      "grad_norm": 2.6660497188568115,
+      "learning_rate": 1.6611198086837427e-05,
+      "loss": 0.1041,
+      "step": 2621
+    },
+    {
+      "epoch": 1.3729701414353064,
+      "loss_partial": 0.052490234375,
+      "step": 2621
+    },
+    {
+      "epoch": 1.3729701414353064,
+      "loss_partial": 0.0537109375,
+      "step": 2621
+    },
+    {
+      "epoch": 1.3734939759036144,
+      "grad_norm": 2.1806578636169434,
+      "learning_rate": 1.6607767426338283e-05,
+      "loss": 0.1067,
+      "step": 2622
+    },
+    {
+      "epoch": 1.3734939759036144,
+      "loss_partial": 0.05029296875,
+      "step": 2622
+    },
+    {
+      "epoch": 1.3734939759036144,
+      "loss_partial": 0.056884765625,
+      "step": 2622
+    },
+    {
+      "epoch": 1.3740178103719225,
+      "grad_norm": 1.8231455087661743,
+      "learning_rate": 1.66043353848566e-05,
+      "loss": 0.1016,
+      "step": 2623
+    },
+    {
+      "epoch": 1.3740178103719225,
+      "loss_partial": 0.052978515625,
+      "step": 2623
+    },
+    {
+      "epoch": 1.3740178103719225,
+      "loss_partial": 0.05419921875,
+      "step": 2623
+    },
+    {
+      "epoch": 1.3745416448402306,
+      "grad_norm": 2.4937222003936768,
+      "learning_rate": 1.660090196310967e-05,
+      "loss": 0.1072,
+      "step": 2624
+    },
+    {
+      "epoch": 1.3745416448402306,
+      "loss_partial": 0.05078125,
+      "step": 2624
+    },
+    {
+      "epoch": 1.3745416448402306,
+      "loss_partial": 0.0576171875,
+      "step": 2624
+    },
+    {
+      "epoch": 1.3750654793085384,
+      "grad_norm": 1.605314016342163,
+      "learning_rate": 1.6597467161815044e-05,
+      "loss": 0.1054,
+      "step": 2625
+    },
+    {
+      "epoch": 1.3750654793085384,
+      "loss_partial": 0.056884765625,
+      "step": 2625
+    },
+    {
+      "epoch": 1.3750654793085384,
+      "loss_partial": 0.04296875,
+      "step": 2625
+    },
+    {
+      "epoch": 1.3755893137768465,
+      "grad_norm": 2.289734363555908,
+      "learning_rate": 1.6594030981690582e-05,
+      "loss": 0.1048,
+      "step": 2626
+    },
+    {
+      "epoch": 1.3755893137768465,
+      "loss_partial": 0.0556640625,
+      "step": 2626
+    },
+    {
+      "epoch": 1.3755893137768465,
+      "loss_partial": 0.0478515625,
+      "step": 2626
+    },
+    {
+      "epoch": 1.3761131482451545,
+      "grad_norm": 1.7012845277786255,
+      "learning_rate": 1.659059342345442e-05,
+      "loss": 0.1009,
+      "step": 2627
+    },
+    {
+      "epoch": 1.3761131482451545,
+      "loss_partial": 0.060302734375,
+      "step": 2627
+    },
+    {
+      "epoch": 1.3761131482451545,
+      "loss_partial": 0.044921875,
+      "step": 2627
+    },
+    {
+      "epoch": 1.3766369827134626,
+      "grad_norm": 2.11164927482605,
+      "learning_rate": 1.6587154487824987e-05,
+      "loss": 0.1057,
+      "step": 2628
+    },
+    {
+      "epoch": 1.3766369827134626,
+      "loss_partial": 0.060302734375,
+      "step": 2628
+    },
+    {
+      "epoch": 1.3766369827134626,
+      "loss_partial": 0.049072265625,
+      "step": 2628
+    },
+    {
+      "epoch": 1.3771608171817706,
+      "grad_norm": 2.1297664642333984,
+      "learning_rate": 1.6583714175521e-05,
+      "loss": 0.1063,
+      "step": 2629
+    },
+    {
+      "epoch": 1.3771608171817706,
+      "loss_partial": 0.064453125,
+      "step": 2629
+    },
+    {
+      "epoch": 1.3771608171817706,
+      "loss_partial": 0.049072265625,
+      "step": 2629
+    },
+    {
+      "epoch": 1.3776846516500787,
+      "grad_norm": 1.6213065385818481,
+      "learning_rate": 1.658027248726146e-05,
+      "loss": 0.1066,
+      "step": 2630
+    },
+    {
+      "epoch": 1.3776846516500787,
+      "loss_partial": 0.055908203125,
+      "step": 2630
+    },
+    {
+      "epoch": 1.3776846516500787,
+      "loss_partial": 0.057373046875,
+      "step": 2630
+    },
+    {
+      "epoch": 1.3782084861183865,
+      "grad_norm": 2.2127878665924072,
+      "learning_rate": 1.6576829423765664e-05,
+      "loss": 0.1052,
+      "step": 2631
+    },
+    {
+      "epoch": 1.3782084861183865,
+      "loss_partial": 0.047607421875,
+      "step": 2631
+    },
+    {
+      "epoch": 1.3782084861183865,
+      "loss_partial": 0.05224609375,
+      "step": 2631
+    },
+    {
+      "epoch": 1.3787323205866946,
+      "grad_norm": 1.743671178817749,
+      "learning_rate": 1.6573384985753186e-05,
+      "loss": 0.1057,
+      "step": 2632
+    },
+    {
+      "epoch": 1.3787323205866946,
+      "loss_partial": 0.04931640625,
+      "step": 2632
+    },
+    {
+      "epoch": 1.3787323205866946,
+      "loss_partial": 0.059326171875,
+      "step": 2632
+    },
+    {
+      "epoch": 1.3792561550550027,
+      "grad_norm": 2.342196464538574,
+      "learning_rate": 1.65699391739439e-05,
+      "loss": 0.1081,
+      "step": 2633
+    },
+    {
+      "epoch": 1.3792561550550027,
+      "loss_partial": 0.04931640625,
+      "step": 2633
+    },
+    {
+      "epoch": 1.3792561550550027,
+      "loss_partial": 0.045654296875,
+      "step": 2633
+    },
+    {
+      "epoch": 1.3797799895233107,
+      "grad_norm": 1.3113583326339722,
+      "learning_rate": 1.656649198905795e-05,
+      "loss": 0.1047,
+      "step": 2634
+    },
+    {
+      "epoch": 1.3797799895233107,
+      "loss_partial": 0.04638671875,
+      "step": 2634
+    },
+    {
+      "epoch": 1.3797799895233107,
+      "loss_partial": 0.0498046875,
+      "step": 2634
+    },
+    {
+      "epoch": 1.3803038239916186,
+      "grad_norm": 2.654012680053711,
+      "learning_rate": 1.656304343181578e-05,
+      "loss": 0.1024,
+      "step": 2635
+    },
+    {
+      "epoch": 1.3803038239916186,
+      "loss_partial": 0.052490234375,
+      "step": 2635
+    },
+    {
+      "epoch": 1.3803038239916186,
+      "loss_partial": 0.04833984375,
+      "step": 2635
+    },
+    {
+      "epoch": 1.3808276584599266,
+      "grad_norm": 1.7663350105285645,
+      "learning_rate": 1.655959350293812e-05,
+      "loss": 0.111,
+      "step": 2636
+    },
+    {
+      "epoch": 1.3808276584599266,
+      "loss_partial": 0.052734375,
+      "step": 2636
+    },
+    {
+      "epoch": 1.3808276584599266,
+      "loss_partial": 0.052001953125,
+      "step": 2636
+    },
+    {
+      "epoch": 1.3813514929282347,
+      "grad_norm": 2.151789426803589,
+      "learning_rate": 1.655614220314598e-05,
+      "loss": 0.1045,
+      "step": 2637
+    },
+    {
+      "epoch": 1.3813514929282347,
+      "loss_partial": 0.05126953125,
+      "step": 2637
+    },
+    {
+      "epoch": 1.3813514929282347,
+      "loss_partial": 0.058349609375,
+      "step": 2637
+    },
+    {
+      "epoch": 1.3818753273965427,
+      "grad_norm": 1.5435978174209595,
+      "learning_rate": 1.655268953316066e-05,
+      "loss": 0.1059,
+      "step": 2638
+    },
+    {
+      "epoch": 1.3818753273965427,
+      "loss_partial": 0.054931640625,
+      "step": 2638
+    },
+    {
+      "epoch": 1.3818753273965427,
+      "loss_partial": 0.04931640625,
+      "step": 2638
+    },
+    {
+      "epoch": 1.3823991618648508,
+      "grad_norm": 3.0968611240386963,
+      "learning_rate": 1.6549235493703748e-05,
+      "loss": 0.1068,
+      "step": 2639
+    },
+    {
+      "epoch": 1.3823991618648508,
+      "loss_partial": 0.048583984375,
+      "step": 2639
+    },
+    {
+      "epoch": 1.3823991618648508,
+      "loss_partial": 0.05078125,
+      "step": 2639
+    },
+    {
+      "epoch": 1.3829229963331588,
+      "grad_norm": 1.6360676288604736,
+      "learning_rate": 1.6545780085497116e-05,
+      "loss": 0.1052,
+      "step": 2640
+    },
+    {
+      "epoch": 1.3829229963331588,
+      "loss_partial": 0.052978515625,
+      "step": 2640
+    },
+    {
+      "epoch": 1.3829229963331588,
+      "loss_partial": 0.05126953125,
+      "step": 2640
+    },
+    {
+      "epoch": 1.3834468308014667,
+      "grad_norm": 2.2131872177124023,
+      "learning_rate": 1.6542323309262926e-05,
+      "loss": 0.1003,
+      "step": 2641
+    },
+    {
+      "epoch": 1.3834468308014667,
+      "loss_partial": 0.06494140625,
+      "step": 2641
+    },
+    {
+      "epoch": 1.3834468308014667,
+      "loss_partial": 0.046875,
+      "step": 2641
+    },
+    {
+      "epoch": 1.3839706652697747,
+      "grad_norm": 1.6055759191513062,
+      "learning_rate": 1.6538865165723622e-05,
+      "loss": 0.1103,
+      "step": 2642
+    },
+    {
+      "epoch": 1.3839706652697747,
+      "loss_partial": 0.046142578125,
+      "step": 2642
+    },
+    {
+      "epoch": 1.3839706652697747,
+      "loss_partial": 0.049560546875,
+      "step": 2642
+    },
+    {
+      "epoch": 1.3844944997380828,
+      "grad_norm": 2.187107563018799,
+      "learning_rate": 1.6535405655601934e-05,
+      "loss": 0.1023,
+      "step": 2643
+    },
+    {
+      "epoch": 1.3844944997380828,
+      "loss_partial": 0.05615234375,
+      "step": 2643
+    },
+    {
+      "epoch": 1.3844944997380828,
+      "loss_partial": 0.05419921875,
+      "step": 2643
+    },
+    {
+      "epoch": 1.3850183342063906,
+      "grad_norm": 1.63357675075531,
+      "learning_rate": 1.6531944779620875e-05,
+      "loss": 0.1114,
+      "step": 2644
+    },
+    {
+      "epoch": 1.3850183342063906,
+      "loss_partial": 0.051025390625,
+      "step": 2644
+    },
+    {
+      "epoch": 1.3850183342063906,
+      "loss_partial": 0.058837890625,
+      "step": 2644
+    },
+    {
+      "epoch": 1.3855421686746987,
+      "grad_norm": 2.014730930328369,
+      "learning_rate": 1.652848253850375e-05,
+      "loss": 0.1057,
+      "step": 2645
+    },
+    {
+      "epoch": 1.3855421686746987,
+      "loss_partial": 0.046142578125,
+      "step": 2645
+    },
+    {
+      "epoch": 1.3855421686746987,
+      "loss_partial": 0.046630859375,
+      "step": 2645
+    },
+    {
+      "epoch": 1.3860660031430068,
+      "grad_norm": 1.8500837087631226,
+      "learning_rate": 1.6525018932974148e-05,
+      "loss": 0.1046,
+      "step": 2646
+    },
+    {
+      "epoch": 1.3860660031430068,
+      "loss_partial": 0.04833984375,
+      "step": 2646
+    },
+    {
+      "epoch": 1.3860660031430068,
+      "loss_partial": 0.06201171875,
+      "step": 2646
+    },
+    {
+      "epoch": 1.3865898376113148,
+      "grad_norm": 1.9444588422775269,
+      "learning_rate": 1.6521553963755933e-05,
+      "loss": 0.1042,
+      "step": 2647
+    },
+    {
+      "epoch": 1.3865898376113148,
+      "loss_partial": 0.057373046875,
+      "step": 2647
+    },
+    {
+      "epoch": 1.3865898376113148,
+      "loss_partial": 0.048828125,
+      "step": 2647
+    },
+    {
+      "epoch": 1.3871136720796229,
+      "grad_norm": 1.8956007957458496,
+      "learning_rate": 1.6518087631573273e-05,
+      "loss": 0.1078,
+      "step": 2648
+    },
+    {
+      "epoch": 1.3871136720796229,
+      "loss_partial": 0.048828125,
+      "step": 2648
+    },
+    {
+      "epoch": 1.3871136720796229,
+      "loss_partial": 0.046142578125,
+      "step": 2648
+    },
+    {
+      "epoch": 1.387637506547931,
+      "grad_norm": 1.569353699684143,
+      "learning_rate": 1.6514619937150603e-05,
+      "loss": 0.1071,
+      "step": 2649
+    },
+    {
+      "epoch": 1.387637506547931,
+      "loss_partial": 0.0478515625,
+      "step": 2649
+    },
+    {
+      "epoch": 1.387637506547931,
+      "loss_partial": 0.051025390625,
+      "step": 2649
+    },
+    {
+      "epoch": 1.388161341016239,
+      "grad_norm": 2.875413656234741,
+      "learning_rate": 1.651115088121265e-05,
+      "loss": 0.1033,
+      "step": 2650
+    },
+    {
+      "epoch": 1.388161341016239,
+      "loss_partial": 0.046875,
+      "step": 2650
+    },
+    {
+      "epoch": 1.388161341016239,
+      "loss_partial": 0.06591796875,
+      "step": 2650
+    },
+    {
+      "epoch": 1.3886851754845468,
+      "grad_norm": 2.006176233291626,
+      "learning_rate": 1.6507680464484428e-05,
+      "loss": 0.1089,
+      "step": 2651
+    },
+    {
+      "epoch": 1.3886851754845468,
+      "loss_partial": 0.05859375,
+      "step": 2651
+    },
+    {
+      "epoch": 1.3886851754845468,
+      "loss_partial": 0.050048828125,
+      "step": 2651
+    },
+    {
+      "epoch": 1.389209009952855,
+      "grad_norm": 2.0382401943206787,
+      "learning_rate": 1.6504208687691232e-05,
+      "loss": 0.1083,
+      "step": 2652
+    },
+    {
+      "epoch": 1.389209009952855,
+      "loss_partial": 0.0625,
+      "step": 2652
+    },
+    {
+      "epoch": 1.389209009952855,
+      "loss_partial": 0.051513671875,
+      "step": 2652
+    },
+    {
+      "epoch": 1.389732844421163,
+      "grad_norm": 2.2320353984832764,
+      "learning_rate": 1.650073555155864e-05,
+      "loss": 0.1104,
+      "step": 2653
+    },
+    {
+      "epoch": 1.389732844421163,
+      "loss_partial": 0.04833984375,
+      "step": 2653
+    },
+    {
+      "epoch": 1.389732844421163,
+      "loss_partial": 0.052978515625,
+      "step": 2653
+    },
+    {
+      "epoch": 1.3902566788894708,
+      "grad_norm": 1.8515186309814453,
+      "learning_rate": 1.6497261056812518e-05,
+      "loss": 0.1039,
+      "step": 2654
+    },
+    {
+      "epoch": 1.3902566788894708,
+      "loss_partial": 0.05224609375,
+      "step": 2654
+    },
+    {
+      "epoch": 1.3902566788894708,
+      "loss_partial": 0.061279296875,
+      "step": 2654
+    },
+    {
+      "epoch": 1.3907805133577789,
+      "grad_norm": 1.4896795749664307,
+      "learning_rate": 1.649378520417901e-05,
+      "loss": 0.1057,
+      "step": 2655
+    },
+    {
+      "epoch": 1.3907805133577789,
+      "loss_partial": 0.049072265625,
+      "step": 2655
+    },
+    {
+      "epoch": 1.3907805133577789,
+      "loss_partial": 0.046875,
+      "step": 2655
+    },
+    {
+      "epoch": 1.391304347826087,
+      "grad_norm": 2.1742210388183594,
+      "learning_rate": 1.6490307994384553e-05,
+      "loss": 0.1046,
+      "step": 2656
+    },
+    {
+      "epoch": 1.391304347826087,
+      "loss_partial": 0.05126953125,
+      "step": 2656
+    },
+    {
+      "epoch": 1.391304347826087,
+      "loss_partial": 0.0576171875,
+      "step": 2656
+    },
+    {
+      "epoch": 1.391828182294395,
+      "grad_norm": 1.600344181060791,
+      "learning_rate": 1.648682942815586e-05,
+      "loss": 0.1079,
+      "step": 2657
+    },
+    {
+      "epoch": 1.391828182294395,
+      "loss_partial": 0.052734375,
+      "step": 2657
+    },
+    {
+      "epoch": 1.391828182294395,
+      "loss_partial": 0.047607421875,
+      "step": 2657
+    },
+    {
+      "epoch": 1.392352016762703,
+      "grad_norm": 1.4313991069793701,
+      "learning_rate": 1.648334950621993e-05,
+      "loss": 0.1043,
+      "step": 2658
+    },
+    {
+      "epoch": 1.392352016762703,
+      "loss_partial": 0.054443359375,
+      "step": 2658
+    },
+    {
+      "epoch": 1.392352016762703,
+      "loss_partial": 0.06640625,
+      "step": 2658
+    },
+    {
+      "epoch": 1.392875851231011,
+      "grad_norm": 1.9930909872055054,
+      "learning_rate": 1.6479868229304044e-05,
+      "loss": 0.1106,
+      "step": 2659
+    },
+    {
+      "epoch": 1.392875851231011,
+      "loss_partial": 0.04931640625,
+      "step": 2659
+    },
+    {
+      "epoch": 1.392875851231011,
+      "loss_partial": 0.0595703125,
+      "step": 2659
+    },
+    {
+      "epoch": 1.393399685699319,
+      "grad_norm": 1.3586688041687012,
+      "learning_rate": 1.647638559813577e-05,
+      "loss": 0.1066,
+      "step": 2660
+    },
+    {
+      "epoch": 1.393399685699319,
+      "loss_partial": 0.052001953125,
+      "step": 2660
+    },
+    {
+      "epoch": 1.393399685699319,
+      "loss_partial": 0.052001953125,
+      "step": 2660
+    },
+    {
+      "epoch": 1.393923520167627,
+      "grad_norm": 1.8975526094436646,
+      "learning_rate": 1.6472901613442953e-05,
+      "loss": 0.1075,
+      "step": 2661
+    },
+    {
+      "epoch": 1.393923520167627,
+      "loss_partial": 0.0537109375,
+      "step": 2661
+    },
+    {
+      "epoch": 1.393923520167627,
+      "loss_partial": 0.048828125,
+      "step": 2661
+    },
+    {
+      "epoch": 1.394447354635935,
+      "grad_norm": 2.1543445587158203,
+      "learning_rate": 1.6469416275953728e-05,
+      "loss": 0.1066,
+      "step": 2662
+    },
+    {
+      "epoch": 1.394447354635935,
+      "loss_partial": 0.0546875,
+      "step": 2662
+    },
+    {
+      "epoch": 1.394447354635935,
+      "loss_partial": 0.055419921875,
+      "step": 2662
+    },
+    {
+      "epoch": 1.3949711891042431,
+      "grad_norm": 1.5739915370941162,
+      "learning_rate": 1.6465929586396505e-05,
+      "loss": 0.1046,
+      "step": 2663
+    },
+    {
+      "epoch": 1.3949711891042431,
+      "loss_partial": 0.057373046875,
+      "step": 2663
+    },
+    {
+      "epoch": 1.3949711891042431,
+      "loss_partial": 0.0615234375,
+      "step": 2663
+    },
+    {
+      "epoch": 1.395495023572551,
+      "grad_norm": 1.6778264045715332,
+      "learning_rate": 1.6462441545499984e-05,
+      "loss": 0.1102,
+      "step": 2664
+    },
+    {
+      "epoch": 1.395495023572551,
+      "loss_partial": 0.045654296875,
+      "step": 2664
+    },
+    {
+      "epoch": 1.395495023572551,
+      "loss_partial": 0.05615234375,
+      "step": 2664
+    },
+    {
+      "epoch": 1.396018858040859,
+      "grad_norm": 1.6990107297897339,
+      "learning_rate": 1.6458952153993143e-05,
+      "loss": 0.1043,
+      "step": 2665
+    },
+    {
+      "epoch": 1.396018858040859,
+      "loss_partial": 0.041015625,
+      "step": 2665
+    },
+    {
+      "epoch": 1.396018858040859,
+      "loss_partial": 0.059326171875,
+      "step": 2665
+    },
+    {
+      "epoch": 1.396542692509167,
+      "grad_norm": 1.6248732805252075,
+      "learning_rate": 1.6455461412605238e-05,
+      "loss": 0.1061,
+      "step": 2666
+    },
+    {
+      "epoch": 1.396542692509167,
+      "loss_partial": 0.049560546875,
+      "step": 2666
+    },
+    {
+      "epoch": 1.396542692509167,
+      "loss_partial": 0.0419921875,
+      "step": 2666
+    },
+    {
+      "epoch": 1.3970665269774751,
+      "grad_norm": 1.903109073638916,
+      "learning_rate": 1.6451969322065823e-05,
+      "loss": 0.1051,
+      "step": 2667
+    },
+    {
+      "epoch": 1.3970665269774751,
+      "loss_partial": 0.048828125,
+      "step": 2667
+    },
+    {
+      "epoch": 1.3970665269774751,
+      "loss_partial": 0.059814453125,
+      "step": 2667
+    },
+    {
+      "epoch": 1.3975903614457832,
+      "grad_norm": 1.9787544012069702,
+      "learning_rate": 1.6448475883104714e-05,
+      "loss": 0.1029,
+      "step": 2668
+    },
+    {
+      "epoch": 1.3975903614457832,
+      "loss_partial": 0.0556640625,
+      "step": 2668
+    },
+    {
+      "epoch": 1.3975903614457832,
+      "loss_partial": 0.043212890625,
+      "step": 2668
+    },
+    {
+      "epoch": 1.3981141959140913,
+      "grad_norm": 1.9250776767730713,
+      "learning_rate": 1.6444981096452025e-05,
+      "loss": 0.1088,
+      "step": 2669
+    },
+    {
+      "epoch": 1.3981141959140913,
+      "loss_partial": 0.0478515625,
+      "step": 2669
+    },
+    {
+      "epoch": 1.3981141959140913,
+      "loss_partial": 0.05712890625,
+      "step": 2669
+    },
+    {
+      "epoch": 1.398638030382399,
+      "grad_norm": 1.948352575302124,
+      "learning_rate": 1.6441484962838138e-05,
+      "loss": 0.1,
+      "step": 2670
+    },
+    {
+      "epoch": 1.398638030382399,
+      "loss_partial": 0.052978515625,
+      "step": 2670
+    },
+    {
+      "epoch": 1.398638030382399,
+      "loss_partial": 0.051025390625,
+      "step": 2670
+    },
+    {
+      "epoch": 1.3991618648507071,
+      "grad_norm": 1.1603249311447144,
+      "learning_rate": 1.6437987482993726e-05,
+      "loss": 0.1061,
+      "step": 2671
+    },
+    {
+      "epoch": 1.3991618648507071,
+      "loss_partial": 0.04345703125,
+      "step": 2671
+    },
+    {
+      "epoch": 1.3991618648507071,
+      "loss_partial": 0.07373046875,
+      "step": 2671
+    },
+    {
+      "epoch": 1.3996856993190152,
+      "grad_norm": 1.7563581466674805,
+      "learning_rate": 1.6434488657649747e-05,
+      "loss": 0.1104,
+      "step": 2672
+    },
+    {
+      "epoch": 1.3996856993190152,
+      "loss_partial": 0.05712890625,
+      "step": 2672
+    },
+    {
+      "epoch": 1.3996856993190152,
+      "loss_partial": 0.046875,
+      "step": 2672
+    },
+    {
+      "epoch": 1.4002095337873233,
+      "grad_norm": 1.7878752946853638,
+      "learning_rate": 1.6430988487537422e-05,
+      "loss": 0.1085,
+      "step": 2673
+    },
+    {
+      "epoch": 1.4002095337873233,
+      "loss_partial": 0.048095703125,
+      "step": 2673
+    },
+    {
+      "epoch": 1.4002095337873233,
+      "loss_partial": 0.0703125,
+      "step": 2673
+    },
+    {
+      "epoch": 1.400733368255631,
+      "grad_norm": 2.0322349071502686,
+      "learning_rate": 1.6427486973388274e-05,
+      "loss": 0.108,
+      "step": 2674
+    },
+    {
+      "epoch": 1.400733368255631,
+      "loss_partial": 0.05126953125,
+      "step": 2674
+    },
+    {
+      "epoch": 1.400733368255631,
+      "loss_partial": 0.045654296875,
+      "step": 2674
+    },
+    {
+      "epoch": 1.4012572027239392,
+      "grad_norm": 1.8649836778640747,
+      "learning_rate": 1.6423984115934094e-05,
+      "loss": 0.1044,
+      "step": 2675
+    },
+    {
+      "epoch": 1.4012572027239392,
+      "loss_partial": 0.052978515625,
+      "step": 2675
+    },
+    {
+      "epoch": 1.4012572027239392,
+      "loss_partial": 0.05615234375,
+      "step": 2675
+    },
+    {
+      "epoch": 1.4017810371922472,
+      "grad_norm": 1.4643440246582031,
+      "learning_rate": 1.642047991590696e-05,
+      "loss": 0.1026,
+      "step": 2676
+    },
+    {
+      "epoch": 1.4017810371922472,
+      "loss_partial": 0.057861328125,
+      "step": 2676
+    },
+    {
+      "epoch": 1.4017810371922472,
+      "loss_partial": 0.045654296875,
+      "step": 2676
+    },
+    {
+      "epoch": 1.4023048716605553,
+      "grad_norm": 1.5632179975509644,
+      "learning_rate": 1.6416974374039227e-05,
+      "loss": 0.1118,
+      "step": 2677
+    },
+    {
+      "epoch": 1.4023048716605553,
+      "loss_partial": 0.05078125,
+      "step": 2677
+    },
+    {
+      "epoch": 1.4023048716605553,
+      "loss_partial": 0.045654296875,
+      "step": 2677
+    },
+    {
+      "epoch": 1.4028287061288633,
+      "grad_norm": 2.015209674835205,
+      "learning_rate": 1.6413467491063527e-05,
+      "loss": 0.1061,
+      "step": 2678
+    },
+    {
+      "epoch": 1.4028287061288633,
+      "loss_partial": 0.05126953125,
+      "step": 2678
+    },
+    {
+      "epoch": 1.4028287061288633,
+      "loss_partial": 0.042724609375,
+      "step": 2678
+    },
+    {
+      "epoch": 1.4033525405971714,
+      "grad_norm": 1.8062039613723755,
+      "learning_rate": 1.6409959267712788e-05,
+      "loss": 0.1047,
+      "step": 2679
+    },
+    {
+      "epoch": 1.4033525405971714,
+      "loss_partial": 0.060302734375,
+      "step": 2679
+    },
+    {
+      "epoch": 1.4033525405971714,
+      "loss_partial": 0.048583984375,
+      "step": 2679
+    },
+    {
+      "epoch": 1.4038763750654792,
+      "grad_norm": 2.2721617221832275,
+      "learning_rate": 1.6406449704720194e-05,
+      "loss": 0.111,
+      "step": 2680
+    },
+    {
+      "epoch": 1.4038763750654792,
+      "loss_partial": 0.05419921875,
+      "step": 2680
+    },
+    {
+      "epoch": 1.4038763750654792,
+      "loss_partial": 0.06005859375,
+      "step": 2680
+    },
+    {
+      "epoch": 1.4044002095337873,
+      "grad_norm": 2.061075210571289,
+      "learning_rate": 1.640293880281923e-05,
+      "loss": 0.1046,
+      "step": 2681
+    },
+    {
+      "epoch": 1.4044002095337873,
+      "loss_partial": 0.05224609375,
+      "step": 2681
+    },
+    {
+      "epoch": 1.4044002095337873,
+      "loss_partial": 0.0625,
+      "step": 2681
+    },
+    {
+      "epoch": 1.4049240440020954,
+      "grad_norm": 2.0294575691223145,
+      "learning_rate": 1.639942656274365e-05,
+      "loss": 0.1065,
+      "step": 2682
+    },
+    {
+      "epoch": 1.4049240440020954,
+      "loss_partial": 0.051025390625,
+      "step": 2682
+    },
+    {
+      "epoch": 1.4049240440020954,
+      "loss_partial": 0.036376953125,
+      "step": 2682
+    },
+    {
+      "epoch": 1.4054478784704034,
+      "grad_norm": 1.3090449571609497,
+      "learning_rate": 1.639591298522749e-05,
+      "loss": 0.1099,
+      "step": 2683
+    },
+    {
+      "epoch": 1.4054478784704034,
+      "loss_partial": 0.047119140625,
+      "step": 2683
+    },
+    {
+      "epoch": 1.4054478784704034,
+      "loss_partial": 0.052490234375,
+      "step": 2683
+    },
+    {
+      "epoch": 1.4059717129387113,
+      "grad_norm": 2.2056431770324707,
+      "learning_rate": 1.639239807100507e-05,
+      "loss": 0.1056,
+      "step": 2684
+    },
+    {
+      "epoch": 1.4059717129387113,
+      "loss_partial": 0.050048828125,
+      "step": 2684
+    },
+    {
+      "epoch": 1.4059717129387113,
+      "loss_partial": 0.045654296875,
+      "step": 2684
+    },
+    {
+      "epoch": 1.4064955474070193,
+      "grad_norm": 1.5968899726867676,
+      "learning_rate": 1.6388881820810985e-05,
+      "loss": 0.1039,
+      "step": 2685
+    },
+    {
+      "epoch": 1.4064955474070193,
+      "loss_partial": 0.048828125,
+      "step": 2685
+    },
+    {
+      "epoch": 1.4064955474070193,
+      "loss_partial": 0.0576171875,
+      "step": 2685
+    },
+    {
+      "epoch": 1.4070193818753274,
+      "grad_norm": 1.9786771535873413,
+      "learning_rate": 1.63853642353801e-05,
+      "loss": 0.1033,
+      "step": 2686
+    },
+    {
+      "epoch": 1.4070193818753274,
+      "loss_partial": 0.0517578125,
+      "step": 2686
+    },
+    {
+      "epoch": 1.4070193818753274,
+      "loss_partial": 0.048095703125,
+      "step": 2686
+    },
+    {
+      "epoch": 1.4075432163436354,
+      "grad_norm": 1.6640609502792358,
+      "learning_rate": 1.638184531544758e-05,
+      "loss": 0.1029,
+      "step": 2687
+    },
+    {
+      "epoch": 1.4075432163436354,
+      "loss_partial": 0.061279296875,
+      "step": 2687
+    },
+    {
+      "epoch": 1.4075432163436354,
+      "loss_partial": 0.052490234375,
+      "step": 2687
+    },
+    {
+      "epoch": 1.4080670508119435,
+      "grad_norm": 1.538978099822998,
+      "learning_rate": 1.6378325061748855e-05,
+      "loss": 0.1041,
+      "step": 2688
+    },
+    {
+      "epoch": 1.4080670508119435,
+      "loss_partial": 0.050537109375,
+      "step": 2688
+    },
+    {
+      "epoch": 1.4080670508119435,
+      "loss_partial": 0.054931640625,
+      "step": 2688
+    },
+    {
+      "epoch": 1.4085908852802516,
+      "grad_norm": 1.8756133317947388,
+      "learning_rate": 1.6374803475019634e-05,
+      "loss": 0.1071,
+      "step": 2689
+    },
+    {
+      "epoch": 1.4085908852802516,
+      "loss_partial": 0.04736328125,
+      "step": 2689
+    },
+    {
+      "epoch": 1.4085908852802516,
+      "loss_partial": 0.04443359375,
+      "step": 2689
+    },
+    {
+      "epoch": 1.4091147197485594,
+      "grad_norm": 1.5843660831451416,
+      "learning_rate": 1.6371280555995905e-05,
+      "loss": 0.1027,
+      "step": 2690
+    },
+    {
+      "epoch": 1.4091147197485594,
+      "loss_partial": 0.055908203125,
+      "step": 2690
+    },
+    {
+      "epoch": 1.4091147197485594,
+      "loss_partial": 0.0517578125,
+      "step": 2690
+    },
+    {
+      "epoch": 1.4096385542168675,
+      "grad_norm": 1.5694762468338013,
+      "learning_rate": 1.636775630541394e-05,
+      "loss": 0.1067,
+      "step": 2691
+    },
+    {
+      "epoch": 1.4096385542168675,
+      "loss_partial": 0.0517578125,
+      "step": 2691
+    },
+    {
+      "epoch": 1.4096385542168675,
+      "loss_partial": 0.049560546875,
+      "step": 2691
+    },
+    {
+      "epoch": 1.4101623886851755,
+      "grad_norm": 1.5777950286865234,
+      "learning_rate": 1.6364230724010282e-05,
+      "loss": 0.1071,
+      "step": 2692
+    },
+    {
+      "epoch": 1.4101623886851755,
+      "loss_partial": 0.046142578125,
+      "step": 2692
+    },
+    {
+      "epoch": 1.4101623886851755,
+      "loss_partial": 0.044677734375,
+      "step": 2692
+    },
+    {
+      "epoch": 1.4106862231534836,
+      "grad_norm": 1.882851004600525,
+      "learning_rate": 1.6360703812521758e-05,
+      "loss": 0.107,
+      "step": 2693
+    },
+    {
+      "epoch": 1.4106862231534836,
+      "loss_partial": 0.04443359375,
+      "step": 2693
+    },
+    {
+      "epoch": 1.4106862231534836,
+      "loss_partial": 0.05712890625,
+      "step": 2693
+    },
+    {
+      "epoch": 1.4112100576217914,
+      "grad_norm": 1.2879184484481812,
+      "learning_rate": 1.6357175571685472e-05,
+      "loss": 0.1045,
+      "step": 2694
+    },
+    {
+      "epoch": 1.4112100576217914,
+      "loss_partial": 0.05859375,
+      "step": 2694
+    },
+    {
+      "epoch": 1.4112100576217914,
+      "loss_partial": 0.05322265625,
+      "step": 2694
+    },
+    {
+      "epoch": 1.4117338920900995,
+      "grad_norm": 1.7097656726837158,
+      "learning_rate": 1.6353646002238804e-05,
+      "loss": 0.1096,
+      "step": 2695
+    },
+    {
+      "epoch": 1.4117338920900995,
+      "loss_partial": 0.050537109375,
+      "step": 2695
+    },
+    {
+      "epoch": 1.4117338920900995,
+      "loss_partial": 0.053466796875,
+      "step": 2695
+    },
+    {
+      "epoch": 1.4122577265584075,
+      "grad_norm": 1.5604887008666992,
+      "learning_rate": 1.6350115104919412e-05,
+      "loss": 0.1069,
+      "step": 2696
+    },
+    {
+      "epoch": 1.4122577265584075,
+      "loss_partial": 0.057861328125,
+      "step": 2696
+    },
+    {
+      "epoch": 1.4122577265584075,
+      "loss_partial": 0.05419921875,
+      "step": 2696
+    },
+    {
+      "epoch": 1.4127815610267156,
+      "grad_norm": 1.4179946184158325,
+      "learning_rate": 1.6346582880465227e-05,
+      "loss": 0.1109,
+      "step": 2697
+    },
+    {
+      "epoch": 1.4127815610267156,
+      "loss_partial": 0.042724609375,
+      "step": 2697
+    },
+    {
+      "epoch": 1.4127815610267156,
+      "loss_partial": 0.056640625,
+      "step": 2697
+    },
+    {
+      "epoch": 1.4133053954950237,
+      "grad_norm": 1.6859416961669922,
+      "learning_rate": 1.634304932961447e-05,
+      "loss": 0.1024,
+      "step": 2698
+    },
+    {
+      "epoch": 1.4133053954950237,
+      "loss_partial": 0.051025390625,
+      "step": 2698
+    },
+    {
+      "epoch": 1.4133053954950237,
+      "loss_partial": 0.05224609375,
+      "step": 2698
+    },
+    {
+      "epoch": 1.4138292299633317,
+      "grad_norm": 1.7853879928588867,
+      "learning_rate": 1.6339514453105625e-05,
+      "loss": 0.1062,
+      "step": 2699
+    },
+    {
+      "epoch": 1.4138292299633317,
+      "loss_partial": 0.05078125,
+      "step": 2699
+    },
+    {
+      "epoch": 1.4138292299633317,
+      "loss_partial": 0.04541015625,
+      "step": 2699
+    },
+    {
+      "epoch": 1.4143530644316396,
+      "grad_norm": 1.699171543121338,
+      "learning_rate": 1.6335978251677464e-05,
+      "loss": 0.107,
+      "step": 2700
+    },
+    {
+      "epoch": 1.4143530644316396,
+      "loss_partial": 0.053955078125,
+      "step": 2700
+    },
+    {
+      "epoch": 1.4143530644316396,
+      "loss_partial": 0.054931640625,
+      "step": 2700
+    },
+    {
+      "epoch": 1.4148768988999476,
+      "grad_norm": 1.9511959552764893,
+      "learning_rate": 1.633244072606903e-05,
+      "loss": 0.1051,
+      "step": 2701
+    },
+    {
+      "epoch": 1.4148768988999476,
+      "loss_partial": 0.052978515625,
+      "step": 2701
+    },
+    {
+      "epoch": 1.4148768988999476,
+      "loss_partial": 0.04296875,
+      "step": 2701
+    },
+    {
+      "epoch": 1.4154007333682557,
+      "grad_norm": 1.8647345304489136,
+      "learning_rate": 1.6328901877019636e-05,
+      "loss": 0.1074,
+      "step": 2702
+    },
+    {
+      "epoch": 1.4154007333682557,
+      "loss_partial": 0.049560546875,
+      "step": 2702
+    },
+    {
+      "epoch": 1.4154007333682557,
+      "loss_partial": 0.046630859375,
+      "step": 2702
+    },
+    {
+      "epoch": 1.4159245678365635,
+      "grad_norm": 2.7146480083465576,
+      "learning_rate": 1.6325361705268893e-05,
+      "loss": 0.102,
+      "step": 2703
+    },
+    {
+      "epoch": 1.4159245678365635,
+      "loss_partial": 0.051513671875,
+      "step": 2703
+    },
+    {
+      "epoch": 1.4159245678365635,
+      "loss_partial": 0.0625,
+      "step": 2703
+    },
+    {
+      "epoch": 1.4164484023048716,
+      "grad_norm": 3.1347455978393555,
+      "learning_rate": 1.6321820211556665e-05,
+      "loss": 0.1054,
+      "step": 2704
+    },
+    {
+      "epoch": 1.4164484023048716,
+      "loss_partial": 0.054931640625,
+      "step": 2704
+    },
+    {
+      "epoch": 1.4164484023048716,
+      "loss_partial": 0.055419921875,
+      "step": 2704
+    },
+    {
+      "epoch": 1.4169722367731796,
+      "grad_norm": 2.1759910583496094,
+      "learning_rate": 1.6318277396623106e-05,
+      "loss": 0.1072,
+      "step": 2705
+    },
+    {
+      "epoch": 1.4169722367731796,
+      "loss_partial": 0.04931640625,
+      "step": 2705
+    },
+    {
+      "epoch": 1.4169722367731796,
+      "loss_partial": 0.0625,
+      "step": 2705
+    },
+    {
+      "epoch": 1.4174960712414877,
+      "grad_norm": 3.035341739654541,
+      "learning_rate": 1.6314733261208645e-05,
+      "loss": 0.1099,
+      "step": 2706
+    },
+    {
+      "epoch": 1.4174960712414877,
+      "loss_partial": 0.060791015625,
+      "step": 2706
+    },
+    {
+      "epoch": 1.4174960712414877,
+      "loss_partial": 0.046875,
+      "step": 2706
+    },
+    {
+      "epoch": 1.4180199057097957,
+      "grad_norm": 3.371323823928833,
+      "learning_rate": 1.6311187806053976e-05,
+      "loss": 0.1028,
+      "step": 2707
+    },
+    {
+      "epoch": 1.4180199057097957,
+      "loss_partial": 0.057861328125,
+      "step": 2707
+    },
+    {
+      "epoch": 1.4180199057097957,
+      "loss_partial": 0.0498046875,
+      "step": 2707
+    },
+    {
+      "epoch": 1.4185437401781038,
+      "grad_norm": 1.8858076333999634,
+      "learning_rate": 1.630764103190009e-05,
+      "loss": 0.1087,
+      "step": 2708
+    },
+    {
+      "epoch": 1.4185437401781038,
+      "loss_partial": 0.05615234375,
+      "step": 2708
+    },
+    {
+      "epoch": 1.4185437401781038,
+      "loss_partial": 0.056640625,
+      "step": 2708
+    },
+    {
+      "epoch": 1.4190675746464119,
+      "grad_norm": 2.535520076751709,
+      "learning_rate": 1.630409293948823e-05,
+      "loss": 0.1079,
+      "step": 2709
+    },
+    {
+      "epoch": 1.4190675746464119,
+      "loss_partial": 0.056640625,
+      "step": 2709
+    },
+    {
+      "epoch": 1.4190675746464119,
+      "loss_partial": 0.054931640625,
+      "step": 2709
+    },
+    {
+      "epoch": 1.4195914091147197,
+      "grad_norm": 1.915280818939209,
+      "learning_rate": 1.630054352955993e-05,
+      "loss": 0.1069,
+      "step": 2710
+    },
+    {
+      "epoch": 1.4195914091147197,
+      "loss_partial": 0.060546875,
+      "step": 2710
+    },
+    {
+      "epoch": 1.4195914091147197,
+      "loss_partial": 0.0654296875,
+      "step": 2710
+    },
+    {
+      "epoch": 1.4201152435830278,
+      "grad_norm": 2.2387537956237793,
+      "learning_rate": 1.629699280285699e-05,
+      "loss": 0.1023,
+      "step": 2711
+    },
+    {
+      "epoch": 1.4201152435830278,
+      "loss_partial": 0.04638671875,
+      "step": 2711
+    },
+    {
+      "epoch": 1.4201152435830278,
+      "loss_partial": 0.059326171875,
+      "step": 2711
+    },
+    {
+      "epoch": 1.4206390780513358,
+      "grad_norm": 2.797391176223755,
+      "learning_rate": 1.6293440760121497e-05,
+      "loss": 0.1063,
+      "step": 2712
+    },
+    {
+      "epoch": 1.4206390780513358,
+      "loss_partial": 0.05029296875,
+      "step": 2712
+    },
+    {
+      "epoch": 1.4206390780513358,
+      "loss_partial": 0.050048828125,
+      "step": 2712
+    },
+    {
+      "epoch": 1.4211629125196437,
+      "grad_norm": 2.437131643295288,
+      "learning_rate": 1.6289887402095802e-05,
+      "loss": 0.1083,
+      "step": 2713
+    },
+    {
+      "epoch": 1.4211629125196437,
+      "loss_partial": 0.05078125,
+      "step": 2713
+    },
+    {
+      "epoch": 1.4211629125196437,
+      "loss_partial": 0.053955078125,
+      "step": 2713
+    },
+    {
+      "epoch": 1.4216867469879517,
+      "grad_norm": 2.2408084869384766,
+      "learning_rate": 1.6286332729522536e-05,
+      "loss": 0.1046,
+      "step": 2714
+    },
+    {
+      "epoch": 1.4216867469879517,
+      "loss_partial": 0.059326171875,
+      "step": 2714
+    },
+    {
+      "epoch": 1.4216867469879517,
+      "loss_partial": 0.047607421875,
+      "step": 2714
+    },
+    {
+      "epoch": 1.4222105814562598,
+      "grad_norm": 2.170994758605957,
+      "learning_rate": 1.62827767431446e-05,
+      "loss": 0.1076,
+      "step": 2715
+    },
+    {
+      "epoch": 1.4222105814562598,
+      "loss_partial": 0.04296875,
+      "step": 2715
+    },
+    {
+      "epoch": 1.4222105814562598,
+      "loss_partial": 0.045654296875,
+      "step": 2715
+    },
+    {
+      "epoch": 1.4227344159245678,
+      "grad_norm": 1.9374788999557495,
+      "learning_rate": 1.6279219443705182e-05,
+      "loss": 0.1043,
+      "step": 2716
+    },
+    {
+      "epoch": 1.4227344159245678,
+      "loss_partial": 0.050048828125,
+      "step": 2716
+    },
+    {
+      "epoch": 1.4227344159245678,
+      "loss_partial": 0.056396484375,
+      "step": 2716
+    },
+    {
+      "epoch": 1.423258250392876,
+      "grad_norm": 2.000227928161621,
+      "learning_rate": 1.6275660831947725e-05,
+      "loss": 0.1059,
+      "step": 2717
+    },
+    {
+      "epoch": 1.423258250392876,
+      "loss_partial": 0.052001953125,
+      "step": 2717
+    },
+    {
+      "epoch": 1.423258250392876,
+      "loss_partial": 0.06201171875,
+      "step": 2717
+    },
+    {
+      "epoch": 1.423782084861184,
+      "grad_norm": 2.2130353450775146,
+      "learning_rate": 1.6272100908615962e-05,
+      "loss": 0.1071,
+      "step": 2718
+    },
+    {
+      "epoch": 1.423782084861184,
+      "loss_partial": 0.049560546875,
+      "step": 2718
+    },
+    {
+      "epoch": 1.423782084861184,
+      "loss_partial": 0.050048828125,
+      "step": 2718
+    },
+    {
+      "epoch": 1.4243059193294918,
+      "grad_norm": 1.91126549243927,
+      "learning_rate": 1.6268539674453896e-05,
+      "loss": 0.1051,
+      "step": 2719
+    },
+    {
+      "epoch": 1.4243059193294918,
+      "loss_partial": 0.05078125,
+      "step": 2719
+    },
+    {
+      "epoch": 1.4243059193294918,
+      "loss_partial": 0.06005859375,
+      "step": 2719
+    },
+    {
+      "epoch": 1.4248297537977999,
+      "grad_norm": 2.1513171195983887,
+      "learning_rate": 1.62649771302058e-05,
+      "loss": 0.1099,
+      "step": 2720
+    },
+    {
+      "epoch": 1.4248297537977999,
+      "loss_partial": 0.05224609375,
+      "step": 2720
+    },
+    {
+      "epoch": 1.4248297537977999,
+      "loss_partial": 0.054931640625,
+      "step": 2720
+    },
+    {
+      "epoch": 1.425353588266108,
+      "grad_norm": 1.73772132396698,
+      "learning_rate": 1.6261413276616226e-05,
+      "loss": 0.1077,
+      "step": 2721
+    },
+    {
+      "epoch": 1.425353588266108,
+      "loss_partial": 0.05419921875,
+      "step": 2721
+    },
+    {
+      "epoch": 1.425353588266108,
+      "loss_partial": 0.06005859375,
+      "step": 2721
+    },
+    {
+      "epoch": 1.425877422734416,
+      "grad_norm": 1.8920893669128418,
+      "learning_rate": 1.625784811443e-05,
+      "loss": 0.1095,
+      "step": 2722
+    },
+    {
+      "epoch": 1.425877422734416,
+      "loss_partial": 0.05859375,
+      "step": 2722
+    },
+    {
+      "epoch": 1.425877422734416,
+      "loss_partial": 0.04833984375,
+      "step": 2722
+    },
+    {
+      "epoch": 1.4264012572027238,
+      "grad_norm": 1.8747164011001587,
+      "learning_rate": 1.625428164439221e-05,
+      "loss": 0.1076,
+      "step": 2723
+    },
+    {
+      "epoch": 1.4264012572027238,
+      "loss_partial": 0.038818359375,
+      "step": 2723
+    },
+    {
+      "epoch": 1.4264012572027238,
+      "loss_partial": 0.053466796875,
+      "step": 2723
+    },
+    {
+      "epoch": 1.4269250916710319,
+      "grad_norm": 1.8671011924743652,
+      "learning_rate": 1.6250713867248235e-05,
+      "loss": 0.1053,
+      "step": 2724
+    },
+    {
+      "epoch": 1.4269250916710319,
+      "loss_partial": 0.0537109375,
+      "step": 2724
+    },
+    {
+      "epoch": 1.4269250916710319,
+      "loss_partial": 0.05224609375,
+      "step": 2724
+    },
+    {
+      "epoch": 1.42744892613934,
+      "grad_norm": 1.9300923347473145,
+      "learning_rate": 1.6247144783743712e-05,
+      "loss": 0.1042,
+      "step": 2725
+    },
+    {
+      "epoch": 1.42744892613934,
+      "loss_partial": 0.0458984375,
+      "step": 2725
+    },
+    {
+      "epoch": 1.42744892613934,
+      "loss_partial": 0.0498046875,
+      "step": 2725
+    },
+    {
+      "epoch": 1.427972760607648,
+      "grad_norm": 2.7550790309906006,
+      "learning_rate": 1.6243574394624566e-05,
+      "loss": 0.1042,
+      "step": 2726
+    },
+    {
+      "epoch": 1.427972760607648,
+      "loss_partial": 0.064453125,
+      "step": 2726
+    },
+    {
+      "epoch": 1.427972760607648,
+      "loss_partial": 0.0478515625,
+      "step": 2726
+    },
+    {
+      "epoch": 1.428496595075956,
+      "grad_norm": 2.1991429328918457,
+      "learning_rate": 1.6240002700636973e-05,
+      "loss": 0.1093,
+      "step": 2727
+    },
+    {
+      "epoch": 1.428496595075956,
+      "loss_partial": 0.056640625,
+      "step": 2727
+    },
+    {
+      "epoch": 1.428496595075956,
+      "loss_partial": 0.04833984375,
+      "step": 2727
+    },
+    {
+      "epoch": 1.4290204295442641,
+      "grad_norm": 1.9510819911956787,
+      "learning_rate": 1.623642970252741e-05,
+      "loss": 0.1046,
+      "step": 2728
+    },
+    {
+      "epoch": 1.4290204295442641,
+      "loss_partial": 0.046630859375,
+      "step": 2728
+    },
+    {
+      "epoch": 1.4290204295442641,
+      "loss_partial": 0.064453125,
+      "step": 2728
+    },
+    {
+      "epoch": 1.429544264012572,
+      "grad_norm": 1.7601884603500366,
+      "learning_rate": 1.62328554010426e-05,
+      "loss": 0.1046,
+      "step": 2729
+    },
+    {
+      "epoch": 1.429544264012572,
+      "loss_partial": 0.0546875,
+      "step": 2729
+    },
+    {
+      "epoch": 1.429544264012572,
+      "loss_partial": 0.055419921875,
+      "step": 2729
+    },
+    {
+      "epoch": 1.43006809848088,
+      "grad_norm": 1.9357502460479736,
+      "learning_rate": 1.6229279796929556e-05,
+      "loss": 0.1092,
+      "step": 2730
+    },
+    {
+      "epoch": 1.43006809848088,
+      "loss_partial": 0.05322265625,
+      "step": 2730
+    },
+    {
+      "epoch": 1.43006809848088,
+      "loss_partial": 0.062255859375,
+      "step": 2730
+    },
+    {
+      "epoch": 1.430591932949188,
+      "grad_norm": 1.5145844221115112,
+      "learning_rate": 1.622570289093555e-05,
+      "loss": 0.1065,
+      "step": 2731
+    },
+    {
+      "epoch": 1.430591932949188,
+      "loss_partial": 0.0556640625,
+      "step": 2731
+    },
+    {
+      "epoch": 1.430591932949188,
+      "loss_partial": 0.052978515625,
+      "step": 2731
+    },
+    {
+      "epoch": 1.4311157674174961,
+      "grad_norm": 2.1671791076660156,
+      "learning_rate": 1.6222124683808144e-05,
+      "loss": 0.1047,
+      "step": 2732
+    },
+    {
+      "epoch": 1.4311157674174961,
+      "loss_partial": 0.0498046875,
+      "step": 2732
+    },
+    {
+      "epoch": 1.4311157674174961,
+      "loss_partial": 0.049072265625,
+      "step": 2732
+    },
+    {
+      "epoch": 1.431639601885804,
+      "grad_norm": 1.7565250396728516,
+      "learning_rate": 1.6218545176295156e-05,
+      "loss": 0.1059,
+      "step": 2733
+    },
+    {
+      "epoch": 1.431639601885804,
+      "loss_partial": 0.05859375,
+      "step": 2733
+    },
+    {
+      "epoch": 1.431639601885804,
+      "loss_partial": 0.06689453125,
+      "step": 2733
+    },
+    {
+      "epoch": 1.432163436354112,
+      "grad_norm": 1.8875665664672852,
+      "learning_rate": 1.6214964369144678e-05,
+      "loss": 0.1075,
+      "step": 2734
+    },
+    {
+      "epoch": 1.432163436354112,
+      "loss_partial": 0.052978515625,
+      "step": 2734
+    },
+    {
+      "epoch": 1.432163436354112,
+      "loss_partial": 0.06494140625,
+      "step": 2734
+    },
+    {
+      "epoch": 1.43268727082242,
+      "grad_norm": 1.7614755630493164,
+      "learning_rate": 1.621138226310508e-05,
+      "loss": 0.1067,
+      "step": 2735
+    },
+    {
+      "epoch": 1.43268727082242,
+      "loss_partial": 0.061279296875,
+      "step": 2735
+    },
+    {
+      "epoch": 1.43268727082242,
+      "loss_partial": 0.058837890625,
+      "step": 2735
+    },
+    {
+      "epoch": 1.4332111052907281,
+      "grad_norm": 2.1377038955688477,
+      "learning_rate": 1.6207798858924997e-05,
+      "loss": 0.1079,
+      "step": 2736
+    },
+    {
+      "epoch": 1.4332111052907281,
+      "loss_partial": 0.05615234375,
+      "step": 2736
+    },
+    {
+      "epoch": 1.4332111052907281,
+      "loss_partial": 0.0595703125,
+      "step": 2736
+    },
+    {
+      "epoch": 1.4337349397590362,
+      "grad_norm": 1.819609522819519,
+      "learning_rate": 1.6204214157353343e-05,
+      "loss": 0.1049,
+      "step": 2737
+    },
+    {
+      "epoch": 1.4337349397590362,
+      "loss_partial": 0.0615234375,
+      "step": 2737
+    },
+    {
+      "epoch": 1.4337349397590362,
+      "loss_partial": 0.0537109375,
+      "step": 2737
+    },
+    {
+      "epoch": 1.4342587742273443,
+      "grad_norm": 2.0409281253814697,
+      "learning_rate": 1.6200628159139292e-05,
+      "loss": 0.1046,
+      "step": 2738
+    },
+    {
+      "epoch": 1.4342587742273443,
+      "loss_partial": 0.050537109375,
+      "step": 2738
+    },
+    {
+      "epoch": 1.4342587742273443,
+      "loss_partial": 0.04638671875,
+      "step": 2738
+    },
+    {
+      "epoch": 1.434782608695652,
+      "grad_norm": 1.971030592918396,
+      "learning_rate": 1.6197040865032303e-05,
+      "loss": 0.1028,
+      "step": 2739
+    },
+    {
+      "epoch": 1.434782608695652,
+      "loss_partial": 0.0556640625,
+      "step": 2739
+    },
+    {
+      "epoch": 1.434782608695652,
+      "loss_partial": 0.052001953125,
+      "step": 2739
+    },
+    {
+      "epoch": 1.4353064431639602,
+      "grad_norm": 2.047922372817993,
+      "learning_rate": 1.6193452275782095e-05,
+      "loss": 0.1046,
+      "step": 2740
+    },
+    {
+      "epoch": 1.4353064431639602,
+      "loss_partial": 0.060791015625,
+      "step": 2740
+    },
+    {
+      "epoch": 1.4353064431639602,
+      "loss_partial": 0.050048828125,
+      "step": 2740
+    },
+    {
+      "epoch": 1.4358302776322682,
+      "grad_norm": 1.5019525289535522,
+      "learning_rate": 1.6189862392138658e-05,
+      "loss": 0.1059,
+      "step": 2741
+    },
+    {
+      "epoch": 1.4358302776322682,
+      "loss_partial": 0.062255859375,
+      "step": 2741
+    },
+    {
+      "epoch": 1.4358302776322682,
+      "loss_partial": 0.048095703125,
+      "step": 2741
+    },
+    {
+      "epoch": 1.4363541121005763,
+      "grad_norm": 3.155778169631958,
+      "learning_rate": 1.6186271214852257e-05,
+      "loss": 0.1037,
+      "step": 2742
+    },
+    {
+      "epoch": 1.4363541121005763,
+      "loss_partial": 0.0517578125,
+      "step": 2742
+    },
+    {
+      "epoch": 1.4363541121005763,
+      "loss_partial": 0.05224609375,
+      "step": 2742
+    },
+    {
+      "epoch": 1.4368779465688841,
+      "grad_norm": 1.8157745599746704,
+      "learning_rate": 1.618267874467343e-05,
+      "loss": 0.1033,
+      "step": 2743
+    },
+    {
+      "epoch": 1.4368779465688841,
+      "loss_partial": 0.04736328125,
+      "step": 2743
+    },
+    {
+      "epoch": 1.4368779465688841,
+      "loss_partial": 0.05078125,
+      "step": 2743
+    },
+    {
+      "epoch": 1.4374017810371922,
+      "grad_norm": 1.978214144706726,
+      "learning_rate": 1.6179084982352978e-05,
+      "loss": 0.1056,
+      "step": 2744
+    },
+    {
+      "epoch": 1.4374017810371922,
+      "loss_partial": 0.05908203125,
+      "step": 2744
+    },
+    {
+      "epoch": 1.4374017810371922,
+      "loss_partial": 0.046875,
+      "step": 2744
+    },
+    {
+      "epoch": 1.4379256155055002,
+      "grad_norm": 1.349212646484375,
+      "learning_rate": 1.617548992864197e-05,
+      "loss": 0.1099,
+      "step": 2745
+    },
+    {
+      "epoch": 1.4379256155055002,
+      "loss_partial": 0.05615234375,
+      "step": 2745
+    },
+    {
+      "epoch": 1.4379256155055002,
+      "loss_partial": 0.044921875,
+      "step": 2745
+    },
+    {
+      "epoch": 1.4384494499738083,
+      "grad_norm": 1.8673712015151978,
+      "learning_rate": 1.617189358429176e-05,
+      "loss": 0.1052,
+      "step": 2746
+    },
+    {
+      "epoch": 1.4384494499738083,
+      "loss_partial": 0.050048828125,
+      "step": 2746
+    },
+    {
+      "epoch": 1.4384494499738083,
+      "loss_partial": 0.050048828125,
+      "step": 2746
+    },
+    {
+      "epoch": 1.4389732844421164,
+      "grad_norm": 1.5173802375793457,
+      "learning_rate": 1.6168295950053955e-05,
+      "loss": 0.1081,
+      "step": 2747
+    },
+    {
+      "epoch": 1.4389732844421164,
+      "loss_partial": 0.056884765625,
+      "step": 2747
+    },
+    {
+      "epoch": 1.4389732844421164,
+      "loss_partial": 0.04833984375,
+      "step": 2747
+    },
+    {
+      "epoch": 1.4394971189104244,
+      "grad_norm": 1.8856632709503174,
+      "learning_rate": 1.6164697026680443e-05,
+      "loss": 0.1049,
+      "step": 2748
+    },
+    {
+      "epoch": 1.4394971189104244,
+      "loss_partial": 0.050537109375,
+      "step": 2748
+    },
+    {
+      "epoch": 1.4394971189104244,
+      "loss_partial": 0.044189453125,
+      "step": 2748
+    },
+    {
+      "epoch": 1.4400209533787323,
+      "grad_norm": 1.9700082540512085,
+      "learning_rate": 1.616109681492337e-05,
+      "loss": 0.1011,
+      "step": 2749
+    },
+    {
+      "epoch": 1.4400209533787323,
+      "loss_partial": 0.048095703125,
+      "step": 2749
+    },
+    {
+      "epoch": 1.4400209533787323,
+      "loss_partial": 0.06591796875,
+      "step": 2749
+    },
+    {
+      "epoch": 1.4405447878470403,
+      "grad_norm": 1.7205219268798828,
+      "learning_rate": 1.6157495315535164e-05,
+      "loss": 0.1063,
+      "step": 2750
+    },
+    {
+      "epoch": 1.4405447878470403,
+      "loss_partial": 0.05615234375,
+      "step": 2750
+    },
+    {
+      "epoch": 1.4405447878470403,
+      "loss_partial": 0.0556640625,
+      "step": 2750
+    },
+    {
+      "epoch": 1.4410686223153484,
+      "grad_norm": 1.8157057762145996,
+      "learning_rate": 1.6153892529268513e-05,
+      "loss": 0.1093,
+      "step": 2751
+    },
+    {
+      "epoch": 1.4410686223153484,
+      "loss_partial": 0.056884765625,
+      "step": 2751
+    },
+    {
+      "epoch": 1.4410686223153484,
+      "loss_partial": 0.050048828125,
+      "step": 2751
+    },
+    {
+      "epoch": 1.4415924567836564,
+      "grad_norm": 1.6150586605072021,
+      "learning_rate": 1.6150288456876376e-05,
+      "loss": 0.1065,
+      "step": 2752
+    },
+    {
+      "epoch": 1.4415924567836564,
+      "loss_partial": 0.056884765625,
+      "step": 2752
+    },
+    {
+      "epoch": 1.4415924567836564,
+      "loss_partial": 0.046142578125,
+      "step": 2752
+    },
+    {
+      "epoch": 1.4421162912519643,
+      "grad_norm": 1.4641846418380737,
+      "learning_rate": 1.6146683099111985e-05,
+      "loss": 0.1042,
+      "step": 2753
+    },
+    {
+      "epoch": 1.4421162912519643,
+      "loss_partial": 0.04736328125,
+      "step": 2753
+    },
+    {
+      "epoch": 1.4421162912519643,
+      "loss_partial": 0.06103515625,
+      "step": 2753
+    },
+    {
+      "epoch": 1.4426401257202723,
+      "grad_norm": 1.2570242881774902,
+      "learning_rate": 1.6143076456728838e-05,
+      "loss": 0.1064,
+      "step": 2754
+    },
+    {
+      "epoch": 1.4426401257202723,
+      "loss_partial": 0.0654296875,
+      "step": 2754
+    },
+    {
+      "epoch": 1.4426401257202723,
+      "loss_partial": 0.051513671875,
+      "step": 2754
+    },
+    {
+      "epoch": 1.4431639601885804,
+      "grad_norm": 2.2051522731781006,
+      "learning_rate": 1.6139468530480698e-05,
+      "loss": 0.1085,
+      "step": 2755
+    },
+    {
+      "epoch": 1.4431639601885804,
+      "loss_partial": 0.055908203125,
+      "step": 2755
+    },
+    {
+      "epoch": 1.4431639601885804,
+      "loss_partial": 0.059814453125,
+      "step": 2755
+    },
+    {
+      "epoch": 1.4436877946568885,
+      "grad_norm": 1.9054694175720215,
+      "learning_rate": 1.6135859321121602e-05,
+      "loss": 0.1077,
+      "step": 2756
+    },
+    {
+      "epoch": 1.4436877946568885,
+      "loss_partial": 0.060546875,
+      "step": 2756
+    },
+    {
+      "epoch": 1.4436877946568885,
+      "loss_partial": 0.046630859375,
+      "step": 2756
+    },
+    {
+      "epoch": 1.4442116291251965,
+      "grad_norm": 1.7362056970596313,
+      "learning_rate": 1.6132248829405845e-05,
+      "loss": 0.1064,
+      "step": 2757
+    },
+    {
+      "epoch": 1.4442116291251965,
+      "loss_partial": 0.0439453125,
+      "step": 2757
+    },
+    {
+      "epoch": 1.4442116291251965,
+      "loss_partial": 0.047607421875,
+      "step": 2757
+    },
+    {
+      "epoch": 1.4447354635935046,
+      "grad_norm": 1.4246729612350464,
+      "learning_rate": 1.612863705608801e-05,
+      "loss": 0.1039,
+      "step": 2758
+    },
+    {
+      "epoch": 1.4447354635935046,
+      "loss_partial": 0.04638671875,
+      "step": 2758
+    },
+    {
+      "epoch": 1.4447354635935046,
+      "loss_partial": 0.047119140625,
+      "step": 2758
+    },
+    {
+      "epoch": 1.4452592980618124,
+      "grad_norm": 1.4162259101867676,
+      "learning_rate": 1.6125024001922925e-05,
+      "loss": 0.1085,
+      "step": 2759
+    },
+    {
+      "epoch": 1.4452592980618124,
+      "loss_partial": 0.05126953125,
+      "step": 2759
+    },
+    {
+      "epoch": 1.4452592980618124,
+      "loss_partial": 0.05419921875,
+      "step": 2759
+    },
+    {
+      "epoch": 1.4457831325301205,
+      "grad_norm": 1.270880103111267,
+      "learning_rate": 1.61214096676657e-05,
+      "loss": 0.1065,
+      "step": 2760
+    },
+    {
+      "epoch": 1.4457831325301205,
+      "loss_partial": 0.05224609375,
+      "step": 2760
+    },
+    {
+      "epoch": 1.4457831325301205,
+      "loss_partial": 0.0439453125,
+      "step": 2760
+    },
+    {
+      "epoch": 1.4463069669984285,
+      "grad_norm": 1.7133859395980835,
+      "learning_rate": 1.611779405407171e-05,
+      "loss": 0.1086,
+      "step": 2761
+    },
+    {
+      "epoch": 1.4463069669984285,
+      "loss_partial": 0.0537109375,
+      "step": 2761
+    },
+    {
+      "epoch": 1.4463069669984285,
+      "loss_partial": 0.05615234375,
+      "step": 2761
+    },
+    {
+      "epoch": 1.4468308014667366,
+      "grad_norm": 2.1264238357543945,
+      "learning_rate": 1.6114177161896595e-05,
+      "loss": 0.1083,
+      "step": 2762
+    },
+    {
+      "epoch": 1.4468308014667366,
+      "loss_partial": 0.072265625,
+      "step": 2762
+    },
+    {
+      "epoch": 1.4468308014667366,
+      "loss_partial": 0.049072265625,
+      "step": 2762
+    },
+    {
+      "epoch": 1.4473546359350444,
+      "grad_norm": 1.8968329429626465,
+      "learning_rate": 1.611055899189626e-05,
+      "loss": 0.1057,
+      "step": 2763
+    },
+    {
+      "epoch": 1.4473546359350444,
+      "loss_partial": 0.049560546875,
+      "step": 2763
+    },
+    {
+      "epoch": 1.4473546359350444,
+      "loss_partial": 0.051025390625,
+      "step": 2763
+    },
+    {
+      "epoch": 1.4478784704033525,
+      "grad_norm": 1.5370577573776245,
+      "learning_rate": 1.6106939544826882e-05,
+      "loss": 0.1092,
+      "step": 2764
+    },
+    {
+      "epoch": 1.4478784704033525,
+      "loss_partial": 0.04931640625,
+      "step": 2764
+    },
+    {
+      "epoch": 1.4478784704033525,
+      "loss_partial": 0.051513671875,
+      "step": 2764
+    },
+    {
+      "epoch": 1.4484023048716606,
+      "grad_norm": 1.875764012336731,
+      "learning_rate": 1.61033188214449e-05,
+      "loss": 0.1025,
+      "step": 2765
+    },
+    {
+      "epoch": 1.4484023048716606,
+      "loss_partial": 0.058349609375,
+      "step": 2765
+    },
+    {
+      "epoch": 1.4484023048716606,
+      "loss_partial": 0.046875,
+      "step": 2765
+    },
+    {
+      "epoch": 1.4489261393399686,
+      "grad_norm": 1.8485453128814697,
+      "learning_rate": 1.609969682250703e-05,
+      "loss": 0.1054,
+      "step": 2766
+    },
+    {
+      "epoch": 1.4489261393399686,
+      "loss_partial": 0.05126953125,
+      "step": 2766
+    },
+    {
+      "epoch": 1.4489261393399686,
+      "loss_partial": 0.064453125,
+      "step": 2766
+    },
+    {
+      "epoch": 1.4494499738082767,
+      "grad_norm": 2.085080146789551,
+      "learning_rate": 1.6096073548770244e-05,
+      "loss": 0.1088,
+      "step": 2767
+    },
+    {
+      "epoch": 1.4494499738082767,
+      "loss_partial": 0.04736328125,
+      "step": 2767
+    },
+    {
+      "epoch": 1.4494499738082767,
+      "loss_partial": 0.0537109375,
+      "step": 2767
+    },
+    {
+      "epoch": 1.4499738082765847,
+      "grad_norm": 2.093506097793579,
+      "learning_rate": 1.6092449000991783e-05,
+      "loss": 0.1052,
+      "step": 2768
+    },
+    {
+      "epoch": 1.4499738082765847,
+      "loss_partial": 0.052490234375,
+      "step": 2768
+    },
+    {
+      "epoch": 1.4499738082765847,
+      "loss_partial": 0.0537109375,
+      "step": 2768
+    },
+    {
+      "epoch": 1.4504976427448926,
+      "grad_norm": 1.916346788406372,
+      "learning_rate": 1.6088823179929155e-05,
+      "loss": 0.1064,
+      "step": 2769
+    },
+    {
+      "epoch": 1.4504976427448926,
+      "loss_partial": 0.0546875,
+      "step": 2769
+    },
+    {
+      "epoch": 1.4504976427448926,
+      "loss_partial": 0.0478515625,
+      "step": 2769
+    },
+    {
+      "epoch": 1.4510214772132006,
+      "grad_norm": 2.1268043518066406,
+      "learning_rate": 1.6085196086340134e-05,
+      "loss": 0.1106,
+      "step": 2770
+    },
+    {
+      "epoch": 1.4510214772132006,
+      "loss_partial": 0.0478515625,
+      "step": 2770
+    },
+    {
+      "epoch": 1.4510214772132006,
+      "loss_partial": 0.05859375,
+      "step": 2770
+    },
+    {
+      "epoch": 1.4515453116815087,
+      "grad_norm": 1.762880802154541,
+      "learning_rate": 1.6081567720982764e-05,
+      "loss": 0.1021,
+      "step": 2771
+    },
+    {
+      "epoch": 1.4515453116815087,
+      "loss_partial": 0.059814453125,
+      "step": 2771
+    },
+    {
+      "epoch": 1.4515453116815087,
+      "loss_partial": 0.048828125,
+      "step": 2771
+    },
+    {
+      "epoch": 1.4520691461498165,
+      "grad_norm": 2.6798012256622314,
+      "learning_rate": 1.6077938084615344e-05,
+      "loss": 0.1064,
+      "step": 2772
+    },
+    {
+      "epoch": 1.4520691461498165,
+      "loss_partial": 0.05029296875,
+      "step": 2772
+    },
+    {
+      "epoch": 1.4520691461498165,
+      "loss_partial": 0.0556640625,
+      "step": 2772
+    },
+    {
+      "epoch": 1.4525929806181246,
+      "grad_norm": 1.4068498611450195,
+      "learning_rate": 1.6074307177996448e-05,
+      "loss": 0.1038,
+      "step": 2773
+    },
+    {
+      "epoch": 1.4525929806181246,
+      "loss_partial": 0.05615234375,
+      "step": 2773
+    },
+    {
+      "epoch": 1.4525929806181246,
+      "loss_partial": 0.05224609375,
+      "step": 2773
+    },
+    {
+      "epoch": 1.4531168150864326,
+      "grad_norm": 1.8480709791183472,
+      "learning_rate": 1.6070675001884916e-05,
+      "loss": 0.1116,
+      "step": 2774
+    },
+    {
+      "epoch": 1.4531168150864326,
+      "loss_partial": 0.0615234375,
+      "step": 2774
+    },
+    {
+      "epoch": 1.4531168150864326,
+      "loss_partial": 0.052734375,
+      "step": 2774
+    },
+    {
+      "epoch": 1.4536406495547407,
+      "grad_norm": 1.3772658109664917,
+      "learning_rate": 1.606704155703985e-05,
+      "loss": 0.1067,
+      "step": 2775
+    },
+    {
+      "epoch": 1.4536406495547407,
+      "loss_partial": 0.058837890625,
+      "step": 2775
+    },
+    {
+      "epoch": 1.4536406495547407,
+      "loss_partial": 0.05126953125,
+      "step": 2775
+    },
+    {
+      "epoch": 1.4541644840230488,
+      "grad_norm": 1.9178107976913452,
+      "learning_rate": 1.6063406844220614e-05,
+      "loss": 0.1045,
+      "step": 2776
+    },
+    {
+      "epoch": 1.4541644840230488,
+      "loss_partial": 0.06396484375,
+      "step": 2776
+    },
+    {
+      "epoch": 1.4541644840230488,
+      "loss_partial": 0.046630859375,
+      "step": 2776
+    },
+    {
+      "epoch": 1.4546883184913568,
+      "grad_norm": 1.5352592468261719,
+      "learning_rate": 1.6059770864186845e-05,
+      "loss": 0.108,
+      "step": 2777
+    },
+    {
+      "epoch": 1.4546883184913568,
+      "loss_partial": 0.0498046875,
+      "step": 2777
+    },
+    {
+      "epoch": 1.4546883184913568,
+      "loss_partial": 0.04833984375,
+      "step": 2777
+    },
+    {
+      "epoch": 1.4552121529596649,
+      "grad_norm": 1.9351550340652466,
+      "learning_rate": 1.6056133617698432e-05,
+      "loss": 0.1034,
+      "step": 2778
+    },
+    {
+      "epoch": 1.4552121529596649,
+      "loss_partial": 0.060546875,
+      "step": 2778
+    },
+    {
+      "epoch": 1.4552121529596649,
+      "loss_partial": 0.05859375,
+      "step": 2778
+    },
+    {
+      "epoch": 1.4557359874279727,
+      "grad_norm": 1.9188241958618164,
+      "learning_rate": 1.6052495105515547e-05,
+      "loss": 0.1087,
+      "step": 2779
+    },
+    {
+      "epoch": 1.4557359874279727,
+      "loss_partial": 0.05615234375,
+      "step": 2779
+    },
+    {
+      "epoch": 1.4557359874279727,
+      "loss_partial": 0.051513671875,
+      "step": 2779
+    },
+    {
+      "epoch": 1.4562598218962808,
+      "grad_norm": 1.578450083732605,
+      "learning_rate": 1.6048855328398612e-05,
+      "loss": 0.1073,
+      "step": 2780
+    },
+    {
+      "epoch": 1.4562598218962808,
+      "loss_partial": 0.05419921875,
+      "step": 2780
+    },
+    {
+      "epoch": 1.4562598218962808,
+      "loss_partial": 0.0478515625,
+      "step": 2780
+    },
+    {
+      "epoch": 1.4567836563645888,
+      "grad_norm": 2.4706685543060303,
+      "learning_rate": 1.604521428710832e-05,
+      "loss": 0.1036,
+      "step": 2781
+    },
+    {
+      "epoch": 1.4567836563645888,
+      "loss_partial": 0.064453125,
+      "step": 2781
+    },
+    {
+      "epoch": 1.4567836563645888,
+      "loss_partial": 0.052001953125,
+      "step": 2781
+    },
+    {
+      "epoch": 1.4573074908328967,
+      "grad_norm": 1.3168102502822876,
+      "learning_rate": 1.6041571982405624e-05,
+      "loss": 0.1062,
+      "step": 2782
+    },
+    {
+      "epoch": 1.4573074908328967,
+      "loss_partial": 0.043701171875,
+      "step": 2782
+    },
+    {
+      "epoch": 1.4573074908328967,
+      "loss_partial": 0.05322265625,
+      "step": 2782
+    },
+    {
+      "epoch": 1.4578313253012047,
+      "grad_norm": 1.712822675704956,
+      "learning_rate": 1.6037928415051746e-05,
+      "loss": 0.1058,
+      "step": 2783
+    },
+    {
+      "epoch": 1.4578313253012047,
+      "loss_partial": 0.0556640625,
+      "step": 2783
+    },
+    {
+      "epoch": 1.4578313253012047,
+      "loss_partial": 0.053466796875,
+      "step": 2783
+    },
+    {
+      "epoch": 1.4583551597695128,
+      "grad_norm": 1.6045546531677246,
+      "learning_rate": 1.603428358580816e-05,
+      "loss": 0.1074,
+      "step": 2784
+    },
+    {
+      "epoch": 1.4583551597695128,
+      "loss_partial": 0.06103515625,
+      "step": 2784
+    },
+    {
+      "epoch": 1.4583551597695128,
+      "loss_partial": 0.06396484375,
+      "step": 2784
+    },
+    {
+      "epoch": 1.4588789942378209,
+      "grad_norm": 2.117365837097168,
+      "learning_rate": 1.603063749543662e-05,
+      "loss": 0.11,
+      "step": 2785
+    },
+    {
+      "epoch": 1.4588789942378209,
+      "loss_partial": 0.05029296875,
+      "step": 2785
+    },
+    {
+      "epoch": 1.4588789942378209,
+      "loss_partial": 0.062255859375,
+      "step": 2785
+    },
+    {
+      "epoch": 1.459402828706129,
+      "grad_norm": 1.9909448623657227,
+      "learning_rate": 1.602699014469914e-05,
+      "loss": 0.1042,
+      "step": 2786
+    },
+    {
+      "epoch": 1.459402828706129,
+      "loss_partial": 0.05615234375,
+      "step": 2786
+    },
+    {
+      "epoch": 1.459402828706129,
+      "loss_partial": 0.05224609375,
+      "step": 2786
+    },
+    {
+      "epoch": 1.459926663174437,
+      "grad_norm": 1.6260584592819214,
+      "learning_rate": 1.602334153435799e-05,
+      "loss": 0.1106,
+      "step": 2787
+    },
+    {
+      "epoch": 1.459926663174437,
+      "loss_partial": 0.0498046875,
+      "step": 2787
+    },
+    {
+      "epoch": 1.459926663174437,
+      "loss_partial": 0.059814453125,
+      "step": 2787
+    },
+    {
+      "epoch": 1.4604504976427448,
+      "grad_norm": 1.476654052734375,
+      "learning_rate": 1.6019691665175706e-05,
+      "loss": 0.1068,
+      "step": 2788
+    },
+    {
+      "epoch": 1.4604504976427448,
+      "loss_partial": 0.04833984375,
+      "step": 2788
+    },
+    {
+      "epoch": 1.4604504976427448,
+      "loss_partial": 0.0517578125,
+      "step": 2788
+    },
+    {
+      "epoch": 1.4609743321110529,
+      "grad_norm": 1.5282933712005615,
+      "learning_rate": 1.6016040537915087e-05,
+      "loss": 0.1036,
+      "step": 2789
+    },
+    {
+      "epoch": 1.4609743321110529,
+      "loss_partial": 0.0595703125,
+      "step": 2789
+    },
+    {
+      "epoch": 1.4609743321110529,
+      "loss_partial": 0.041748046875,
+      "step": 2789
+    },
+    {
+      "epoch": 1.461498166579361,
+      "grad_norm": 1.354871392250061,
+      "learning_rate": 1.6012388153339194e-05,
+      "loss": 0.1081,
+      "step": 2790
+    },
+    {
+      "epoch": 1.461498166579361,
+      "loss_partial": 0.05078125,
+      "step": 2790
+    },
+    {
+      "epoch": 1.461498166579361,
+      "loss_partial": 0.050537109375,
+      "step": 2790
+    },
+    {
+      "epoch": 1.462022001047669,
+      "grad_norm": 1.9361114501953125,
+      "learning_rate": 1.6008734512211365e-05,
+      "loss": 0.1036,
+      "step": 2791
+    },
+    {
+      "epoch": 1.462022001047669,
+      "loss_partial": 0.0625,
+      "step": 2791
+    },
+    {
+      "epoch": 1.462022001047669,
+      "loss_partial": 0.05126953125,
+      "step": 2791
+    },
+    {
+      "epoch": 1.4625458355159768,
+      "grad_norm": 1.7695542573928833,
+      "learning_rate": 1.6005079615295175e-05,
+      "loss": 0.1088,
+      "step": 2792
+    },
+    {
+      "epoch": 1.4625458355159768,
+      "loss_partial": 0.04931640625,
+      "step": 2792
+    },
+    {
+      "epoch": 1.4625458355159768,
+      "loss_partial": 0.05615234375,
+      "step": 2792
+    },
+    {
+      "epoch": 1.463069669984285,
+      "grad_norm": 1.6564358472824097,
+      "learning_rate": 1.6001423463354478e-05,
+      "loss": 0.1056,
+      "step": 2793
+    },
+    {
+      "epoch": 1.463069669984285,
+      "loss_partial": 0.0771484375,
+      "step": 2793
+    },
+    {
+      "epoch": 1.463069669984285,
+      "loss_partial": 0.04931640625,
+      "step": 2793
+    },
+    {
+      "epoch": 1.463593504452593,
+      "grad_norm": 1.928379774093628,
+      "learning_rate": 1.5997766057153394e-05,
+      "loss": 0.1106,
+      "step": 2794
+    },
+    {
+      "epoch": 1.463593504452593,
+      "loss_partial": 0.0556640625,
+      "step": 2794
+    },
+    {
+      "epoch": 1.463593504452593,
+      "loss_partial": 0.0498046875,
+      "step": 2794
+    },
+    {
+      "epoch": 1.464117338920901,
+      "grad_norm": 2.066202163696289,
+      "learning_rate": 1.5994107397456288e-05,
+      "loss": 0.1066,
+      "step": 2795
+    },
+    {
+      "epoch": 1.464117338920901,
+      "loss_partial": 0.05322265625,
+      "step": 2795
+    },
+    {
+      "epoch": 1.464117338920901,
+      "loss_partial": 0.053466796875,
+      "step": 2795
+    },
+    {
+      "epoch": 1.464641173389209,
+      "grad_norm": 1.5666621923446655,
+      "learning_rate": 1.5990447485027803e-05,
+      "loss": 0.109,
+      "step": 2796
+    },
+    {
+      "epoch": 1.464641173389209,
+      "loss_partial": 0.059326171875,
+      "step": 2796
+    },
+    {
+      "epoch": 1.464641173389209,
+      "loss_partial": 0.0673828125,
+      "step": 2796
+    },
+    {
+      "epoch": 1.4651650078575171,
+      "grad_norm": 1.3984063863754272,
+      "learning_rate": 1.5986786320632842e-05,
+      "loss": 0.1063,
+      "step": 2797
+    },
+    {
+      "epoch": 1.4651650078575171,
+      "loss_partial": 0.05419921875,
+      "step": 2797
+    },
+    {
+      "epoch": 1.4651650078575171,
+      "loss_partial": 0.046630859375,
+      "step": 2797
+    },
+    {
+      "epoch": 1.465688842325825,
+      "grad_norm": 2.2993948459625244,
+      "learning_rate": 1.5983123905036556e-05,
+      "loss": 0.1039,
+      "step": 2798
+    },
+    {
+      "epoch": 1.465688842325825,
+      "loss_partial": 0.056640625,
+      "step": 2798
+    },
+    {
+      "epoch": 1.465688842325825,
+      "loss_partial": 0.0498046875,
+      "step": 2798
+    },
+    {
+      "epoch": 1.466212676794133,
+      "grad_norm": 1.1811585426330566,
+      "learning_rate": 1.5979460239004375e-05,
+      "loss": 0.1057,
+      "step": 2799
+    },
+    {
+      "epoch": 1.466212676794133,
+      "loss_partial": 0.048583984375,
+      "step": 2799
+    },
+    {
+      "epoch": 1.466212676794133,
+      "loss_partial": 0.054931640625,
+      "step": 2799
+    },
+    {
+      "epoch": 1.466736511262441,
+      "grad_norm": 2.1576571464538574,
+      "learning_rate": 1.597579532330198e-05,
+      "loss": 0.1059,
+      "step": 2800
+    },
+    {
+      "epoch": 1.466736511262441,
+      "loss_partial": 0.046875,
+      "step": 2800
+    },
+    {
+      "epoch": 1.466736511262441,
+      "loss_partial": 0.057861328125,
+      "step": 2800
+    },
+    {
+      "epoch": 1.4672603457307491,
+      "grad_norm": 1.592962384223938,
+      "learning_rate": 1.5972129158695314e-05,
+      "loss": 0.1005,
+      "step": 2801
+    },
+    {
+      "epoch": 1.4672603457307491,
+      "loss_partial": 0.0546875,
+      "step": 2801
+    },
+    {
+      "epoch": 1.4672603457307491,
+      "loss_partial": 0.05029296875,
+      "step": 2801
+    },
+    {
+      "epoch": 1.467784180199057,
+      "grad_norm": 1.362370491027832,
+      "learning_rate": 1.5968461745950585e-05,
+      "loss": 0.1072,
+      "step": 2802
+    },
+    {
+      "epoch": 1.467784180199057,
+      "loss_partial": 0.05517578125,
+      "step": 2802
+    },
+    {
+      "epoch": 1.467784180199057,
+      "loss_partial": 0.045654296875,
+      "step": 2802
+    },
+    {
+      "epoch": 1.468308014667365,
+      "grad_norm": 1.7609014511108398,
+      "learning_rate": 1.596479308583426e-05,
+      "loss": 0.1034,
+      "step": 2803
+    },
+    {
+      "epoch": 1.468308014667365,
+      "loss_partial": 0.046875,
+      "step": 2803
+    },
+    {
+      "epoch": 1.468308014667365,
+      "loss_partial": 0.05810546875,
+      "step": 2803
+    },
+    {
+      "epoch": 1.468831849135673,
+      "grad_norm": 1.9004058837890625,
+      "learning_rate": 1.5961123179113064e-05,
+      "loss": 0.1068,
+      "step": 2804
+    },
+    {
+      "epoch": 1.468831849135673,
+      "loss_partial": 0.049560546875,
+      "step": 2804
+    },
+    {
+      "epoch": 1.468831849135673,
+      "loss_partial": 0.059814453125,
+      "step": 2804
+    },
+    {
+      "epoch": 1.4693556836039812,
+      "grad_norm": 1.6897549629211426,
+      "learning_rate": 1.5957452026553988e-05,
+      "loss": 0.1074,
+      "step": 2805
+    },
+    {
+      "epoch": 1.4693556836039812,
+      "loss_partial": 0.06298828125,
+      "step": 2805
+    },
+    {
+      "epoch": 1.4693556836039812,
+      "loss_partial": 0.06298828125,
+      "step": 2805
+    },
+    {
+      "epoch": 1.4698795180722892,
+      "grad_norm": 1.6610709428787231,
+      "learning_rate": 1.5953779628924273e-05,
+      "loss": 0.1062,
+      "step": 2806
+    },
+    {
+      "epoch": 1.4698795180722892,
+      "loss_partial": 0.050537109375,
+      "step": 2806
+    },
+    {
+      "epoch": 1.4698795180722892,
+      "loss_partial": 0.0498046875,
+      "step": 2806
+    },
+    {
+      "epoch": 1.4704033525405973,
+      "grad_norm": 1.7832163572311401,
+      "learning_rate": 1.595010598699144e-05,
+      "loss": 0.1111,
+      "step": 2807
+    },
+    {
+      "epoch": 1.4704033525405973,
+      "loss_partial": 0.064453125,
+      "step": 2807
+    },
+    {
+      "epoch": 1.4704033525405973,
+      "loss_partial": 0.052490234375,
+      "step": 2807
+    },
+    {
+      "epoch": 1.4709271870089051,
+      "grad_norm": 1.7252070903778076,
+      "learning_rate": 1.5946431101523243e-05,
+      "loss": 0.1057,
+      "step": 2808
+    },
+    {
+      "epoch": 1.4709271870089051,
+      "loss_partial": 0.05517578125,
+      "step": 2808
+    },
+    {
+      "epoch": 1.4709271870089051,
+      "loss_partial": 0.046630859375,
+      "step": 2808
+    },
+    {
+      "epoch": 1.4714510214772132,
+      "grad_norm": 2.1118767261505127,
+      "learning_rate": 1.594275497328772e-05,
+      "loss": 0.1002,
+      "step": 2809
+    },
+    {
+      "epoch": 1.4714510214772132,
+      "loss_partial": 0.0517578125,
+      "step": 2809
+    },
+    {
+      "epoch": 1.4714510214772132,
+      "loss_partial": 0.052734375,
+      "step": 2809
+    },
+    {
+      "epoch": 1.4719748559455212,
+      "grad_norm": 1.961474061012268,
+      "learning_rate": 1.593907760305316e-05,
+      "loss": 0.1074,
+      "step": 2810
+    },
+    {
+      "epoch": 1.4719748559455212,
+      "loss_partial": 0.04443359375,
+      "step": 2810
+    },
+    {
+      "epoch": 1.4719748559455212,
+      "loss_partial": 0.051513671875,
+      "step": 2810
+    },
+    {
+      "epoch": 1.4724986904138293,
+      "grad_norm": 2.137991189956665,
+      "learning_rate": 1.5935398991588106e-05,
+      "loss": 0.1047,
+      "step": 2811
+    },
+    {
+      "epoch": 1.4724986904138293,
+      "loss_partial": 0.0615234375,
+      "step": 2811
+    },
+    {
+      "epoch": 1.4724986904138293,
+      "loss_partial": 0.052490234375,
+      "step": 2811
+    },
+    {
+      "epoch": 1.4730225248821371,
+      "grad_norm": 1.961626648902893,
+      "learning_rate": 1.593171913966136e-05,
+      "loss": 0.1031,
+      "step": 2812
+    },
+    {
+      "epoch": 1.4730225248821371,
+      "loss_partial": 0.046142578125,
+      "step": 2812
+    },
+    {
+      "epoch": 1.4730225248821371,
+      "loss_partial": 0.0654296875,
+      "step": 2812
+    },
+    {
+      "epoch": 1.4735463593504452,
+      "grad_norm": 1.8399709463119507,
+      "learning_rate": 1.5928038048042002e-05,
+      "loss": 0.1034,
+      "step": 2813
+    },
+    {
+      "epoch": 1.4735463593504452,
+      "loss_partial": 0.0546875,
+      "step": 2813
+    },
+    {
+      "epoch": 1.4735463593504452,
+      "loss_partial": 0.06298828125,
+      "step": 2813
+    },
+    {
+      "epoch": 1.4740701938187533,
+      "grad_norm": 2.5764272212982178,
+      "learning_rate": 1.5924355717499346e-05,
+      "loss": 0.1068,
+      "step": 2814
+    },
+    {
+      "epoch": 1.4740701938187533,
+      "loss_partial": 0.047119140625,
+      "step": 2814
+    },
+    {
+      "epoch": 1.4740701938187533,
+      "loss_partial": 0.06640625,
+      "step": 2814
+    },
+    {
+      "epoch": 1.4745940282870613,
+      "grad_norm": 2.3762643337249756,
+      "learning_rate": 1.5920672148802982e-05,
+      "loss": 0.1063,
+      "step": 2815
+    },
+    {
+      "epoch": 1.4745940282870613,
+      "loss_partial": 0.051025390625,
+      "step": 2815
+    },
+    {
+      "epoch": 1.4745940282870613,
+      "loss_partial": 0.041748046875,
+      "step": 2815
+    },
+    {
+      "epoch": 1.4751178627553694,
+      "grad_norm": 2.445559024810791,
+      "learning_rate": 1.591698734272275e-05,
+      "loss": 0.1084,
+      "step": 2816
+    },
+    {
+      "epoch": 1.4751178627553694,
+      "loss_partial": 0.049560546875,
+      "step": 2816
+    },
+    {
+      "epoch": 1.4751178627553694,
+      "loss_partial": 0.05419921875,
+      "step": 2816
+    },
+    {
+      "epoch": 1.4756416972236774,
+      "grad_norm": 2.221182346343994,
+      "learning_rate": 1.5913301300028755e-05,
+      "loss": 0.1063,
+      "step": 2817
+    },
+    {
+      "epoch": 1.4756416972236774,
+      "loss_partial": 0.052978515625,
+      "step": 2817
+    },
+    {
+      "epoch": 1.4756416972236774,
+      "loss_partial": 0.052001953125,
+      "step": 2817
+    },
+    {
+      "epoch": 1.4761655316919853,
+      "grad_norm": 1.2733879089355469,
+      "learning_rate": 1.590961402149135e-05,
+      "loss": 0.1034,
+      "step": 2818
+    },
+    {
+      "epoch": 1.4761655316919853,
+      "loss_partial": 0.05908203125,
+      "step": 2818
+    },
+    {
+      "epoch": 1.4761655316919853,
+      "loss_partial": 0.06201171875,
+      "step": 2818
+    },
+    {
+      "epoch": 1.4766893661602933,
+      "grad_norm": 2.1312990188598633,
+      "learning_rate": 1.590592550788116e-05,
+      "loss": 0.1098,
+      "step": 2819
+    },
+    {
+      "epoch": 1.4766893661602933,
+      "loss_partial": 0.06201171875,
+      "step": 2819
+    },
+    {
+      "epoch": 1.4766893661602933,
+      "loss_partial": 0.05078125,
+      "step": 2819
+    },
+    {
+      "epoch": 1.4772132006286014,
+      "grad_norm": 2.03110933303833,
+      "learning_rate": 1.5902235759969064e-05,
+      "loss": 0.106,
+      "step": 2820
+    },
+    {
+      "epoch": 1.4772132006286014,
+      "loss_partial": 0.0458984375,
+      "step": 2820
+    },
+    {
+      "epoch": 1.4772132006286014,
+      "loss_partial": 0.059814453125,
+      "step": 2820
+    },
+    {
+      "epoch": 1.4777370350969095,
+      "grad_norm": 1.729462742805481,
+      "learning_rate": 1.5898544778526194e-05,
+      "loss": 0.106,
+      "step": 2821
+    },
+    {
+      "epoch": 1.4777370350969095,
+      "loss_partial": 0.05517578125,
+      "step": 2821
+    },
+    {
+      "epoch": 1.4777370350969095,
+      "loss_partial": 0.048828125,
+      "step": 2821
+    },
+    {
+      "epoch": 1.4782608695652173,
+      "grad_norm": 1.8551141023635864,
+      "learning_rate": 1.5894852564323935e-05,
+      "loss": 0.1059,
+      "step": 2822
+    },
+    {
+      "epoch": 1.4782608695652173,
+      "loss_partial": 0.050537109375,
+      "step": 2822
+    },
+    {
+      "epoch": 1.4782608695652173,
+      "loss_partial": 0.06591796875,
+      "step": 2822
+    },
+    {
+      "epoch": 1.4787847040335254,
+      "grad_norm": 1.544075608253479,
+      "learning_rate": 1.5891159118133943e-05,
+      "loss": 0.1042,
+      "step": 2823
+    },
+    {
+      "epoch": 1.4787847040335254,
+      "loss_partial": 0.046875,
+      "step": 2823
+    },
+    {
+      "epoch": 1.4787847040335254,
+      "loss_partial": 0.0478515625,
+      "step": 2823
+    },
+    {
+      "epoch": 1.4793085385018334,
+      "grad_norm": 1.7243214845657349,
+      "learning_rate": 1.5887464440728127e-05,
+      "loss": 0.1052,
+      "step": 2824
+    },
+    {
+      "epoch": 1.4793085385018334,
+      "loss_partial": 0.051025390625,
+      "step": 2824
+    },
+    {
+      "epoch": 1.4793085385018334,
+      "loss_partial": 0.0478515625,
+      "step": 2824
+    },
+    {
+      "epoch": 1.4798323729701415,
+      "grad_norm": 2.121458053588867,
+      "learning_rate": 1.588376853287865e-05,
+      "loss": 0.1053,
+      "step": 2825
+    },
+    {
+      "epoch": 1.4798323729701415,
+      "loss_partial": 0.056884765625,
+      "step": 2825
+    },
+    {
+      "epoch": 1.4798323729701415,
+      "loss_partial": 0.0537109375,
+      "step": 2825
+    },
+    {
+      "epoch": 1.4803562074384495,
+      "grad_norm": 1.6859360933303833,
+      "learning_rate": 1.5880071395357935e-05,
+      "loss": 0.1087,
+      "step": 2826
+    },
+    {
+      "epoch": 1.4803562074384495,
+      "loss_partial": 0.0576171875,
+      "step": 2826
+    },
+    {
+      "epoch": 1.4803562074384495,
+      "loss_partial": 0.049072265625,
+      "step": 2826
+    },
+    {
+      "epoch": 1.4808800419067576,
+      "grad_norm": 1.9426050186157227,
+      "learning_rate": 1.5876373028938656e-05,
+      "loss": 0.1073,
+      "step": 2827
+    },
+    {
+      "epoch": 1.4808800419067576,
+      "loss_partial": 0.06005859375,
+      "step": 2827
+    },
+    {
+      "epoch": 1.4808800419067576,
+      "loss_partial": 0.06005859375,
+      "step": 2827
+    },
+    {
+      "epoch": 1.4814038763750654,
+      "grad_norm": 1.494519829750061,
+      "learning_rate": 1.5872673434393756e-05,
+      "loss": 0.1084,
+      "step": 2828
+    },
+    {
+      "epoch": 1.4814038763750654,
+      "loss_partial": 0.06201171875,
+      "step": 2828
+    },
+    {
+      "epoch": 1.4814038763750654,
+      "loss_partial": 0.050048828125,
+      "step": 2828
+    },
+    {
+      "epoch": 1.4819277108433735,
+      "grad_norm": 1.3508449792861938,
+      "learning_rate": 1.586897261249642e-05,
+      "loss": 0.1051,
+      "step": 2829
+    },
+    {
+      "epoch": 1.4819277108433735,
+      "loss_partial": 0.046875,
+      "step": 2829
+    },
+    {
+      "epoch": 1.4819277108433735,
+      "loss_partial": 0.052001953125,
+      "step": 2829
+    },
+    {
+      "epoch": 1.4824515453116816,
+      "grad_norm": 1.4342666864395142,
+      "learning_rate": 1.5865270564020107e-05,
+      "loss": 0.1025,
+      "step": 2830
+    },
+    {
+      "epoch": 1.4824515453116816,
+      "loss_partial": 0.048095703125,
+      "step": 2830
+    },
+    {
+      "epoch": 1.4824515453116816,
+      "loss_partial": 0.052490234375,
+      "step": 2830
+    },
+    {
+      "epoch": 1.4829753797799894,
+      "grad_norm": 1.2182122468948364,
+      "learning_rate": 1.586156728973851e-05,
+      "loss": 0.1067,
+      "step": 2831
+    },
+    {
+      "epoch": 1.4829753797799894,
+      "loss_partial": 0.051513671875,
+      "step": 2831
+    },
+    {
+      "epoch": 1.4829753797799894,
+      "loss_partial": 0.047119140625,
+      "step": 2831
+    },
+    {
+      "epoch": 1.4834992142482974,
+      "grad_norm": 1.6658556461334229,
+      "learning_rate": 1.58578627904256e-05,
+      "loss": 0.1053,
+      "step": 2832
+    },
+    {
+      "epoch": 1.4834992142482974,
+      "loss_partial": 0.044921875,
+      "step": 2832
+    },
+    {
+      "epoch": 1.4834992142482974,
+      "loss_partial": 0.056396484375,
+      "step": 2832
+    },
+    {
+      "epoch": 1.4840230487166055,
+      "grad_norm": 1.6461505889892578,
+      "learning_rate": 1.5854157066855586e-05,
+      "loss": 0.1066,
+      "step": 2833
+    },
+    {
+      "epoch": 1.4840230487166055,
+      "loss_partial": 0.053955078125,
+      "step": 2833
+    },
+    {
+      "epoch": 1.4840230487166055,
+      "loss_partial": 0.055419921875,
+      "step": 2833
+    },
+    {
+      "epoch": 1.4845468831849136,
+      "grad_norm": 2.3424229621887207,
+      "learning_rate": 1.5850450119802948e-05,
+      "loss": 0.1024,
+      "step": 2834
+    },
+    {
+      "epoch": 1.4845468831849136,
+      "loss_partial": 0.059814453125,
+      "step": 2834
+    },
+    {
+      "epoch": 1.4845468831849136,
+      "loss_partial": 0.059814453125,
+      "step": 2834
+    },
+    {
+      "epoch": 1.4850707176532216,
+      "grad_norm": 1.9374055862426758,
+      "learning_rate": 1.5846741950042414e-05,
+      "loss": 0.1104,
+      "step": 2835
+    },
+    {
+      "epoch": 1.4850707176532216,
+      "loss_partial": 0.046630859375,
+      "step": 2835
+    },
+    {
+      "epoch": 1.4850707176532216,
+      "loss_partial": 0.046630859375,
+      "step": 2835
+    },
+    {
+      "epoch": 1.4855945521215297,
+      "grad_norm": 1.8626437187194824,
+      "learning_rate": 1.584303255834897e-05,
+      "loss": 0.1037,
+      "step": 2836
+    },
+    {
+      "epoch": 1.4855945521215297,
+      "loss_partial": 0.05078125,
+      "step": 2836
+    },
+    {
+      "epoch": 1.4855945521215297,
+      "loss_partial": 0.055419921875,
+      "step": 2836
+    },
+    {
+      "epoch": 1.4861183865898377,
+      "grad_norm": 2.1934640407562256,
+      "learning_rate": 1.5839321945497847e-05,
+      "loss": 0.1061,
+      "step": 2837
+    },
+    {
+      "epoch": 1.4861183865898377,
+      "loss_partial": 0.046630859375,
+      "step": 2837
+    },
+    {
+      "epoch": 1.4861183865898377,
+      "loss_partial": 0.05712890625,
+      "step": 2837
+    },
+    {
+      "epoch": 1.4866422210581456,
+      "grad_norm": 1.7107633352279663,
+      "learning_rate": 1.5835610112264553e-05,
+      "loss": 0.1039,
+      "step": 2838
+    },
+    {
+      "epoch": 1.4866422210581456,
+      "loss_partial": 0.0517578125,
+      "step": 2838
+    },
+    {
+      "epoch": 1.4866422210581456,
+      "loss_partial": 0.05126953125,
+      "step": 2838
+    },
+    {
+      "epoch": 1.4871660555264536,
+      "grad_norm": 2.193416118621826,
+      "learning_rate": 1.5831897059424828e-05,
+      "loss": 0.1077,
+      "step": 2839
+    },
+    {
+      "epoch": 1.4871660555264536,
+      "loss_partial": 0.0556640625,
+      "step": 2839
+    },
+    {
+      "epoch": 1.4871660555264536,
+      "loss_partial": 0.057373046875,
+      "step": 2839
+    },
+    {
+      "epoch": 1.4876898899947617,
+      "grad_norm": 2.201401948928833,
+      "learning_rate": 1.582818278775468e-05,
+      "loss": 0.1021,
+      "step": 2840
+    },
+    {
+      "epoch": 1.4876898899947617,
+      "loss_partial": 0.03955078125,
+      "step": 2840
+    },
+    {
+      "epoch": 1.4876898899947617,
+      "loss_partial": 0.052490234375,
+      "step": 2840
+    },
+    {
+      "epoch": 1.4882137244630695,
+      "grad_norm": 2.6966540813446045,
+      "learning_rate": 1.5824467298030376e-05,
+      "loss": 0.1039,
+      "step": 2841
+    },
+    {
+      "epoch": 1.4882137244630695,
+      "loss_partial": 0.044677734375,
+      "step": 2841
+    },
+    {
+      "epoch": 1.4882137244630695,
+      "loss_partial": 0.0634765625,
+      "step": 2841
+    },
+    {
+      "epoch": 1.4887375589313776,
+      "grad_norm": 1.7031128406524658,
+      "learning_rate": 1.582075059102842e-05,
+      "loss": 0.1085,
+      "step": 2842
+    },
+    {
+      "epoch": 1.4887375589313776,
+      "loss_partial": 0.052001953125,
+      "step": 2842
+    },
+    {
+      "epoch": 1.4887375589313776,
+      "loss_partial": 0.05615234375,
+      "step": 2842
+    },
+    {
+      "epoch": 1.4892613933996857,
+      "grad_norm": 1.7023993730545044,
+      "learning_rate": 1.5817032667525583e-05,
+      "loss": 0.1045,
+      "step": 2843
+    },
+    {
+      "epoch": 1.4892613933996857,
+      "loss_partial": 0.0654296875,
+      "step": 2843
+    },
+    {
+      "epoch": 1.4892613933996857,
+      "loss_partial": 0.052490234375,
+      "step": 2843
+    },
+    {
+      "epoch": 1.4897852278679937,
+      "grad_norm": 2.399492025375366,
+      "learning_rate": 1.5813313528298893e-05,
+      "loss": 0.104,
+      "step": 2844
+    },
+    {
+      "epoch": 1.4897852278679937,
+      "loss_partial": 0.055908203125,
+      "step": 2844
+    },
+    {
+      "epoch": 1.4897852278679937,
+      "loss_partial": 0.051025390625,
+      "step": 2844
+    },
+    {
+      "epoch": 1.4903090623363018,
+      "grad_norm": 1.5908406972885132,
+      "learning_rate": 1.5809593174125623e-05,
+      "loss": 0.1046,
+      "step": 2845
+    },
+    {
+      "epoch": 1.4903090623363018,
+      "loss_partial": 0.05126953125,
+      "step": 2845
+    },
+    {
+      "epoch": 1.4903090623363018,
+      "loss_partial": 0.05908203125,
+      "step": 2845
+    },
+    {
+      "epoch": 1.4908328968046098,
+      "grad_norm": 2.402723550796509,
+      "learning_rate": 1.5805871605783308e-05,
+      "loss": 0.1069,
+      "step": 2846
+    },
+    {
+      "epoch": 1.4908328968046098,
+      "loss_partial": 0.057861328125,
+      "step": 2846
+    },
+    {
+      "epoch": 1.4908328968046098,
+      "loss_partial": 0.04638671875,
+      "step": 2846
+    },
+    {
+      "epoch": 1.4913567312729177,
+      "grad_norm": 2.436734199523926,
+      "learning_rate": 1.5802148824049725e-05,
+      "loss": 0.1062,
+      "step": 2847
+    },
+    {
+      "epoch": 1.4913567312729177,
+      "loss_partial": 0.0703125,
+      "step": 2847
+    },
+    {
+      "epoch": 1.4913567312729177,
+      "loss_partial": 0.0537109375,
+      "step": 2847
+    },
+    {
+      "epoch": 1.4918805657412257,
+      "grad_norm": 1.8003733158111572,
+      "learning_rate": 1.579842482970292e-05,
+      "loss": 0.1055,
+      "step": 2848
+    },
+    {
+      "epoch": 1.4918805657412257,
+      "loss_partial": 0.056640625,
+      "step": 2848
+    },
+    {
+      "epoch": 1.4918805657412257,
+      "loss_partial": 0.041748046875,
+      "step": 2848
+    },
+    {
+      "epoch": 1.4924044002095338,
+      "grad_norm": 1.915303111076355,
+      "learning_rate": 1.579469962352118e-05,
+      "loss": 0.1009,
+      "step": 2849
+    },
+    {
+      "epoch": 1.4924044002095338,
+      "loss_partial": 0.04736328125,
+      "step": 2849
+    },
+    {
+      "epoch": 1.4924044002095338,
+      "loss_partial": 0.053466796875,
+      "step": 2849
+    },
+    {
+      "epoch": 1.4929282346778419,
+      "grad_norm": 2.1081244945526123,
+      "learning_rate": 1.5790973206283054e-05,
+      "loss": 0.1028,
+      "step": 2850
+    },
+    {
+      "epoch": 1.4929282346778419,
+      "loss_partial": 0.05078125,
+      "step": 2850
+    },
+    {
+      "epoch": 1.4929282346778419,
+      "loss_partial": 0.05517578125,
+      "step": 2850
+    },
+    {
+      "epoch": 1.4934520691461497,
+      "grad_norm": 1.4611300230026245,
+      "learning_rate": 1.5787245578767337e-05,
+      "loss": 0.1083,
+      "step": 2851
+    },
+    {
+      "epoch": 1.4934520691461497,
+      "loss_partial": 0.056640625,
+      "step": 2851
+    },
+    {
+      "epoch": 1.4934520691461497,
+      "loss_partial": 0.0517578125,
+      "step": 2851
+    },
+    {
+      "epoch": 1.4939759036144578,
+      "grad_norm": 1.578149437904358,
+      "learning_rate": 1.5783516741753084e-05,
+      "loss": 0.1066,
+      "step": 2852
+    },
+    {
+      "epoch": 1.4939759036144578,
+      "loss_partial": 0.046630859375,
+      "step": 2852
+    },
+    {
+      "epoch": 1.4939759036144578,
+      "loss_partial": 0.049072265625,
+      "step": 2852
+    },
+    {
+      "epoch": 1.4944997380827658,
+      "grad_norm": 1.3101202249526978,
+      "learning_rate": 1.5779786696019593e-05,
+      "loss": 0.1031,
+      "step": 2853
+    },
+    {
+      "epoch": 1.4944997380827658,
+      "loss_partial": 0.06640625,
+      "step": 2853
+    },
+    {
+      "epoch": 1.4944997380827658,
+      "loss_partial": 0.044921875,
+      "step": 2853
+    },
+    {
+      "epoch": 1.4950235725510739,
+      "grad_norm": 2.404794692993164,
+      "learning_rate": 1.5776055442346426e-05,
+      "loss": 0.1057,
+      "step": 2854
+    },
+    {
+      "epoch": 1.4950235725510739,
+      "loss_partial": 0.054443359375,
+      "step": 2854
+    },
+    {
+      "epoch": 1.4950235725510739,
+      "loss_partial": 0.046630859375,
+      "step": 2854
+    },
+    {
+      "epoch": 1.495547407019382,
+      "grad_norm": 1.8280065059661865,
+      "learning_rate": 1.5772322981513385e-05,
+      "loss": 0.101,
+      "step": 2855
+    },
+    {
+      "epoch": 1.495547407019382,
+      "loss_partial": 0.05859375,
+      "step": 2855
+    },
+    {
+      "epoch": 1.495547407019382,
+      "loss_partial": 0.06591796875,
+      "step": 2855
+    },
+    {
+      "epoch": 1.49607124148769,
+      "grad_norm": 1.4989367723464966,
+      "learning_rate": 1.576858931430054e-05,
+      "loss": 0.1061,
+      "step": 2856
+    },
+    {
+      "epoch": 1.49607124148769,
+      "loss_partial": 0.057373046875,
+      "step": 2856
+    },
+    {
+      "epoch": 1.49607124148769,
+      "loss_partial": 0.05517578125,
+      "step": 2856
+    },
+    {
+      "epoch": 1.4965950759559978,
+      "grad_norm": 2.404759645462036,
+      "learning_rate": 1.5764854441488198e-05,
+      "loss": 0.1073,
+      "step": 2857
+    },
+    {
+      "epoch": 1.4965950759559978,
+      "loss_partial": 0.0517578125,
+      "step": 2857
+    },
+    {
+      "epoch": 1.4965950759559978,
+      "loss_partial": 0.049560546875,
+      "step": 2857
+    },
+    {
+      "epoch": 1.497118910424306,
+      "grad_norm": 2.1077587604522705,
+      "learning_rate": 1.5761118363856928e-05,
+      "loss": 0.1038,
+      "step": 2858
+    },
+    {
+      "epoch": 1.497118910424306,
+      "loss_partial": 0.056640625,
+      "step": 2858
+    },
+    {
+      "epoch": 1.497118910424306,
+      "loss_partial": 0.0546875,
+      "step": 2858
+    },
+    {
+      "epoch": 1.497642744892614,
+      "grad_norm": 1.600986123085022,
+      "learning_rate": 1.5757381082187547e-05,
+      "loss": 0.1068,
+      "step": 2859
+    },
+    {
+      "epoch": 1.497642744892614,
+      "loss_partial": 0.04443359375,
+      "step": 2859
+    },
+    {
+      "epoch": 1.497642744892614,
+      "loss_partial": 0.049560546875,
+      "step": 2859
+    },
+    {
+      "epoch": 1.498166579360922,
+      "grad_norm": 2.375326633453369,
+      "learning_rate": 1.575364259726112e-05,
+      "loss": 0.1077,
+      "step": 2860
+    },
+    {
+      "epoch": 1.498166579360922,
+      "loss_partial": 0.048095703125,
+      "step": 2860
+    },
+    {
+      "epoch": 1.498166579360922,
+      "loss_partial": 0.05078125,
+      "step": 2860
+    },
+    {
+      "epoch": 1.4986904138292299,
+      "grad_norm": 1.96660578250885,
+      "learning_rate": 1.574990290985897e-05,
+      "loss": 0.1065,
+      "step": 2861
+    },
+    {
+      "epoch": 1.4986904138292299,
+      "loss_partial": 0.059814453125,
+      "step": 2861
+    },
+    {
+      "epoch": 1.4986904138292299,
+      "loss_partial": 0.053466796875,
+      "step": 2861
+    },
+    {
+      "epoch": 1.499214248297538,
+      "grad_norm": 1.9134925603866577,
+      "learning_rate": 1.5746162020762666e-05,
+      "loss": 0.1102,
+      "step": 2862
+    },
+    {
+      "epoch": 1.499214248297538,
+      "loss_partial": 0.055908203125,
+      "step": 2862
+    },
+    {
+      "epoch": 1.499214248297538,
+      "loss_partial": 0.0517578125,
+      "step": 2862
+    },
+    {
+      "epoch": 1.499738082765846,
+      "grad_norm": 2.5109567642211914,
+      "learning_rate": 1.574241993075404e-05,
+      "loss": 0.1085,
+      "step": 2863
+    },
+    {
+      "epoch": 1.499738082765846,
+      "loss_partial": 0.060791015625,
+      "step": 2863
+    },
+    {
+      "epoch": 1.499738082765846,
+      "loss_partial": 0.047119140625,
+      "step": 2863
+    },
+    {
+      "epoch": 1.500261917234154,
+      "grad_norm": 1.349297285079956,
+      "learning_rate": 1.573867664061515e-05,
+      "loss": 0.1052,
+      "step": 2864
+    },
+    {
+      "epoch": 1.500261917234154,
+      "loss_partial": 0.04833984375,
+      "step": 2864
+    },
+    {
+      "epoch": 1.500261917234154,
+      "loss_partial": 0.051025390625,
+      "step": 2864
+    },
+    {
+      "epoch": 1.500785751702462,
+      "grad_norm": 2.2983410358428955,
+      "learning_rate": 1.5734932151128335e-05,
+      "loss": 0.1032,
+      "step": 2865
+    },
+    {
+      "epoch": 1.500785751702462,
+      "loss_partial": 0.05712890625,
+      "step": 2865
+    },
+    {
+      "epoch": 1.500785751702462,
+      "loss_partial": 0.06787109375,
+      "step": 2865
+    },
+    {
+      "epoch": 1.5013095861707701,
+      "grad_norm": 1.7453261613845825,
+      "learning_rate": 1.5731186463076163e-05,
+      "loss": 0.1079,
+      "step": 2866
+    },
+    {
+      "epoch": 1.5013095861707701,
+      "loss_partial": 0.046142578125,
+      "step": 2866
+    },
+    {
+      "epoch": 1.5013095861707701,
+      "loss_partial": 0.06396484375,
+      "step": 2866
+    },
+    {
+      "epoch": 1.5018334206390782,
+      "grad_norm": 1.6986737251281738,
+      "learning_rate": 1.572743957724146e-05,
+      "loss": 0.1069,
+      "step": 2867
+    },
+    {
+      "epoch": 1.5018334206390782,
+      "loss_partial": 0.056640625,
+      "step": 2867
+    },
+    {
+      "epoch": 1.5018334206390782,
+      "loss_partial": 0.047607421875,
+      "step": 2867
+    },
+    {
+      "epoch": 1.502357255107386,
+      "grad_norm": 1.6467151641845703,
+      "learning_rate": 1.5723691494407302e-05,
+      "loss": 0.1039,
+      "step": 2868
+    },
+    {
+      "epoch": 1.502357255107386,
+      "loss_partial": 0.057861328125,
+      "step": 2868
+    },
+    {
+      "epoch": 1.502357255107386,
+      "loss_partial": 0.057861328125,
+      "step": 2868
+    },
+    {
+      "epoch": 1.502881089575694,
+      "grad_norm": 1.8165205717086792,
+      "learning_rate": 1.5719942215357015e-05,
+      "loss": 0.1083,
+      "step": 2869
+    },
+    {
+      "epoch": 1.502881089575694,
+      "loss_partial": 0.052978515625,
+      "step": 2869
+    },
+    {
+      "epoch": 1.502881089575694,
+      "loss_partial": 0.04443359375,
+      "step": 2869
+    },
+    {
+      "epoch": 1.503404924044002,
+      "grad_norm": 1.2821091413497925,
+      "learning_rate": 1.5716191740874182e-05,
+      "loss": 0.106,
+      "step": 2870
+    },
+    {
+      "epoch": 1.503404924044002,
+      "loss_partial": 0.054443359375,
+      "step": 2870
+    },
+    {
+      "epoch": 1.503404924044002,
+      "loss_partial": 0.059326171875,
+      "step": 2870
+    },
+    {
+      "epoch": 1.50392875851231,
+      "grad_norm": 1.7728697061538696,
+      "learning_rate": 1.5712440071742615e-05,
+      "loss": 0.1083,
+      "step": 2871
+    },
+    {
+      "epoch": 1.50392875851231,
+      "loss_partial": 0.04443359375,
+      "step": 2871
+    },
+    {
+      "epoch": 1.50392875851231,
+      "loss_partial": 0.060791015625,
+      "step": 2871
+    },
+    {
+      "epoch": 1.504452592980618,
+      "grad_norm": 1.4947500228881836,
+      "learning_rate": 1.5708687208746403e-05,
+      "loss": 0.1112,
+      "step": 2872
+    },
+    {
+      "epoch": 1.504452592980618,
+      "loss_partial": 0.056396484375,
+      "step": 2872
+    },
+    {
+      "epoch": 1.504452592980618,
+      "loss_partial": 0.05517578125,
+      "step": 2872
+    },
+    {
+      "epoch": 1.5049764274489261,
+      "grad_norm": 1.857672929763794,
+      "learning_rate": 1.5704933152669862e-05,
+      "loss": 0.1073,
+      "step": 2873
+    },
+    {
+      "epoch": 1.5049764274489261,
+      "loss_partial": 0.046875,
+      "step": 2873
+    },
+    {
+      "epoch": 1.5049764274489261,
+      "loss_partial": 0.07763671875,
+      "step": 2873
+    },
+    {
+      "epoch": 1.5055002619172342,
+      "grad_norm": 1.6611169576644897,
+      "learning_rate": 1.5701177904297574e-05,
+      "loss": 0.1083,
+      "step": 2874
+    },
+    {
+      "epoch": 1.5055002619172342,
+      "loss_partial": 0.04736328125,
+      "step": 2874
+    },
+    {
+      "epoch": 1.5055002619172342,
+      "loss_partial": 0.04638671875,
+      "step": 2874
+    },
+    {
+      "epoch": 1.5060240963855422,
+      "grad_norm": 1.3714118003845215,
+      "learning_rate": 1.569742146441436e-05,
+      "loss": 0.1041,
+      "step": 2875
+    },
+    {
+      "epoch": 1.5060240963855422,
+      "loss_partial": 0.057373046875,
+      "step": 2875
+    },
+    {
+      "epoch": 1.5060240963855422,
+      "loss_partial": 0.052978515625,
+      "step": 2875
+    },
+    {
+      "epoch": 1.5065479308538503,
+      "grad_norm": 1.4388598203659058,
+      "learning_rate": 1.569366383380529e-05,
+      "loss": 0.1018,
+      "step": 2876
+    },
+    {
+      "epoch": 1.5065479308538503,
+      "loss_partial": 0.060546875,
+      "step": 2876
+    },
+    {
+      "epoch": 1.5065479308538503,
+      "loss_partial": 0.048828125,
+      "step": 2876
+    },
+    {
+      "epoch": 1.5070717653221584,
+      "grad_norm": 1.8640209436416626,
+      "learning_rate": 1.5689905013255683e-05,
+      "loss": 0.104,
+      "step": 2877
+    },
+    {
+      "epoch": 1.5070717653221584,
+      "loss_partial": 0.047607421875,
+      "step": 2877
+    },
+    {
+      "epoch": 1.5070717653221584,
+      "loss_partial": 0.055908203125,
+      "step": 2877
+    },
+    {
+      "epoch": 1.5075955997904662,
+      "grad_norm": 1.253453016281128,
+      "learning_rate": 1.568614500355112e-05,
+      "loss": 0.1023,
+      "step": 2878
+    },
+    {
+      "epoch": 1.5075955997904662,
+      "loss_partial": 0.04931640625,
+      "step": 2878
+    },
+    {
+      "epoch": 1.5075955997904662,
+      "loss_partial": 0.05322265625,
+      "step": 2878
+    },
+    {
+      "epoch": 1.5081194342587743,
+      "grad_norm": 1.8689688444137573,
+      "learning_rate": 1.5682383805477407e-05,
+      "loss": 0.1061,
+      "step": 2879
+    },
+    {
+      "epoch": 1.5081194342587743,
+      "loss_partial": 0.046630859375,
+      "step": 2879
+    },
+    {
+      "epoch": 1.5081194342587743,
+      "loss_partial": 0.042724609375,
+      "step": 2879
+    },
+    {
+      "epoch": 1.508643268727082,
+      "grad_norm": 1.731392741203308,
+      "learning_rate": 1.567862141982062e-05,
+      "loss": 0.1085,
+      "step": 2880
+    },
+    {
+      "epoch": 1.508643268727082,
+      "loss_partial": 0.0537109375,
+      "step": 2880
+    },
+    {
+      "epoch": 1.508643268727082,
+      "loss_partial": 0.051513671875,
+      "step": 2880
+    },
+    {
+      "epoch": 1.5091671031953902,
+      "grad_norm": 2.017660617828369,
+      "learning_rate": 1.567485784736707e-05,
+      "loss": 0.1078,
+      "step": 2881
+    },
+    {
+      "epoch": 1.5091671031953902,
+      "loss_partial": 0.04931640625,
+      "step": 2881
+    },
+    {
+      "epoch": 1.5091671031953902,
+      "loss_partial": 0.041259765625,
+      "step": 2881
+    },
+    {
+      "epoch": 1.5096909376636982,
+      "grad_norm": 1.5290544033050537,
+      "learning_rate": 1.567109308890333e-05,
+      "loss": 0.1062,
+      "step": 2882
+    },
+    {
+      "epoch": 1.5096909376636982,
+      "loss_partial": 0.06396484375,
+      "step": 2882
+    },
+    {
+      "epoch": 1.5096909376636982,
+      "loss_partial": 0.047607421875,
+      "step": 2882
+    },
+    {
+      "epoch": 1.5102147721320063,
+      "grad_norm": 1.984192967414856,
+      "learning_rate": 1.5667327145216198e-05,
+      "loss": 0.1084,
+      "step": 2883
+    },
+    {
+      "epoch": 1.5102147721320063,
+      "loss_partial": 0.060791015625,
+      "step": 2883
+    },
+    {
+      "epoch": 1.5102147721320063,
+      "loss_partial": 0.05419921875,
+      "step": 2883
+    },
+    {
+      "epoch": 1.5107386066003143,
+      "grad_norm": 1.6246426105499268,
+      "learning_rate": 1.566356001709274e-05,
+      "loss": 0.1107,
+      "step": 2884
+    },
+    {
+      "epoch": 1.5107386066003143,
+      "loss_partial": 0.05322265625,
+      "step": 2884
+    },
+    {
+      "epoch": 1.5107386066003143,
+      "loss_partial": 0.059814453125,
+      "step": 2884
+    },
+    {
+      "epoch": 1.5112624410686224,
+      "grad_norm": 1.4087316989898682,
+      "learning_rate": 1.565979170532026e-05,
+      "loss": 0.1067,
+      "step": 2885
+    },
+    {
+      "epoch": 1.5112624410686224,
+      "loss_partial": 0.050537109375,
+      "step": 2885
+    },
+    {
+      "epoch": 1.5112624410686224,
+      "loss_partial": 0.0537109375,
+      "step": 2885
+    },
+    {
+      "epoch": 1.5117862755369305,
+      "grad_norm": 1.6837798357009888,
+      "learning_rate": 1.5656022210686307e-05,
+      "loss": 0.1063,
+      "step": 2886
+    },
+    {
+      "epoch": 1.5117862755369305,
+      "loss_partial": 0.0498046875,
+      "step": 2886
+    },
+    {
+      "epoch": 1.5117862755369305,
+      "loss_partial": 0.04833984375,
+      "step": 2886
+    },
+    {
+      "epoch": 1.5123101100052383,
+      "grad_norm": 1.9990885257720947,
+      "learning_rate": 1.5652251533978696e-05,
+      "loss": 0.1051,
+      "step": 2887
+    },
+    {
+      "epoch": 1.5123101100052383,
+      "loss_partial": 0.043701171875,
+      "step": 2887
+    },
+    {
+      "epoch": 1.5123101100052383,
+      "loss_partial": 0.05517578125,
+      "step": 2887
+    },
+    {
+      "epoch": 1.5128339444735464,
+      "grad_norm": 1.920280933380127,
+      "learning_rate": 1.5648479675985462e-05,
+      "loss": 0.1037,
+      "step": 2888
+    },
+    {
+      "epoch": 1.5128339444735464,
+      "loss_partial": 0.059326171875,
+      "step": 2888
+    },
+    {
+      "epoch": 1.5128339444735464,
+      "loss_partial": 0.047119140625,
+      "step": 2888
+    },
+    {
+      "epoch": 1.5133577789418544,
+      "grad_norm": 1.8391404151916504,
+      "learning_rate": 1.564470663749491e-05,
+      "loss": 0.1048,
+      "step": 2889
+    },
+    {
+      "epoch": 1.5133577789418544,
+      "loss_partial": 0.053466796875,
+      "step": 2889
+    },
+    {
+      "epoch": 1.5133577789418544,
+      "loss_partial": 0.057373046875,
+      "step": 2889
+    },
+    {
+      "epoch": 1.5138816134101623,
+      "grad_norm": 1.7773481607437134,
+      "learning_rate": 1.5640932419295572e-05,
+      "loss": 0.1045,
+      "step": 2890
+    },
+    {
+      "epoch": 1.5138816134101623,
+      "loss_partial": 0.04345703125,
+      "step": 2890
+    },
+    {
+      "epoch": 1.5138816134101623,
+      "loss_partial": 0.047607421875,
+      "step": 2890
+    },
+    {
+      "epoch": 1.5144054478784703,
+      "grad_norm": 1.6220537424087524,
+      "learning_rate": 1.563715702217624e-05,
+      "loss": 0.1041,
+      "step": 2891
+    },
+    {
+      "epoch": 1.5144054478784703,
+      "loss_partial": 0.052734375,
+      "step": 2891
+    },
+    {
+      "epoch": 1.5144054478784703,
+      "loss_partial": 0.053466796875,
+      "step": 2891
+    },
+    {
+      "epoch": 1.5149292823467784,
+      "grad_norm": 2.3172919750213623,
+      "learning_rate": 1.563338044692595e-05,
+      "loss": 0.1045,
+      "step": 2892
+    },
+    {
+      "epoch": 1.5149292823467784,
+      "loss_partial": 0.053466796875,
+      "step": 2892
+    },
+    {
+      "epoch": 1.5149292823467784,
+      "loss_partial": 0.0458984375,
+      "step": 2892
+    },
+    {
+      "epoch": 1.5154531168150864,
+      "grad_norm": 1.429153561592102,
+      "learning_rate": 1.5629602694333987e-05,
+      "loss": 0.104,
+      "step": 2893
+    },
+    {
+      "epoch": 1.5154531168150864,
+      "loss_partial": 0.047607421875,
+      "step": 2893
+    },
+    {
+      "epoch": 1.5154531168150864,
+      "loss_partial": 0.05224609375,
+      "step": 2893
+    },
+    {
+      "epoch": 1.5159769512833945,
+      "grad_norm": 2.2504336833953857,
+      "learning_rate": 1.5625823765189868e-05,
+      "loss": 0.1036,
+      "step": 2894
+    },
+    {
+      "epoch": 1.5159769512833945,
+      "loss_partial": 0.055908203125,
+      "step": 2894
+    },
+    {
+      "epoch": 1.5159769512833945,
+      "loss_partial": 0.0498046875,
+      "step": 2894
+    },
+    {
+      "epoch": 1.5165007857517026,
+      "grad_norm": 1.4839025735855103,
+      "learning_rate": 1.5622043660283373e-05,
+      "loss": 0.1098,
+      "step": 2895
+    },
+    {
+      "epoch": 1.5165007857517026,
+      "loss_partial": 0.049072265625,
+      "step": 2895
+    },
+    {
+      "epoch": 1.5165007857517026,
+      "loss_partial": 0.059814453125,
+      "step": 2895
+    },
+    {
+      "epoch": 1.5170246202200106,
+      "grad_norm": 2.2656197547912598,
+      "learning_rate": 1.561826238040452e-05,
+      "loss": 0.1034,
+      "step": 2896
+    },
+    {
+      "epoch": 1.5170246202200106,
+      "loss_partial": 0.0478515625,
+      "step": 2896
+    },
+    {
+      "epoch": 1.5170246202200106,
+      "loss_partial": 0.052978515625,
+      "step": 2896
+    },
+    {
+      "epoch": 1.5175484546883184,
+      "grad_norm": 1.9196670055389404,
+      "learning_rate": 1.5614479926343573e-05,
+      "loss": 0.1021,
+      "step": 2897
+    },
+    {
+      "epoch": 1.5175484546883184,
+      "loss_partial": 0.04736328125,
+      "step": 2897
+    },
+    {
+      "epoch": 1.5175484546883184,
+      "loss_partial": 0.058349609375,
+      "step": 2897
+    },
+    {
+      "epoch": 1.5180722891566265,
+      "grad_norm": 1.678430438041687,
+      "learning_rate": 1.5610696298891034e-05,
+      "loss": 0.1058,
+      "step": 2898
+    },
+    {
+      "epoch": 1.5180722891566265,
+      "loss_partial": 0.0498046875,
+      "step": 2898
+    },
+    {
+      "epoch": 1.5180722891566265,
+      "loss_partial": 0.0703125,
+      "step": 2898
+    },
+    {
+      "epoch": 1.5185961236249346,
+      "grad_norm": 2.650829315185547,
+      "learning_rate": 1.560691149883767e-05,
+      "loss": 0.1081,
+      "step": 2899
+    },
+    {
+      "epoch": 1.5185961236249346,
+      "loss_partial": 0.058349609375,
+      "step": 2899
+    },
+    {
+      "epoch": 1.5185961236249346,
+      "loss_partial": 0.0498046875,
+      "step": 2899
+    },
+    {
+      "epoch": 1.5191199580932424,
+      "grad_norm": 1.8772789239883423,
+      "learning_rate": 1.5603125526974472e-05,
+      "loss": 0.1071,
+      "step": 2900
+    },
+    {
+      "epoch": 1.5191199580932424,
+      "loss_partial": 0.0556640625,
+      "step": 2900
+    },
+    {
+      "epoch": 1.5191199580932424,
+      "loss_partial": 0.05029296875,
+      "step": 2900
+    },
+    {
+      "epoch": 1.5196437925615505,
+      "grad_norm": 2.7373101711273193,
+      "learning_rate": 1.559933838409269e-05,
+      "loss": 0.1036,
+      "step": 2901
+    },
+    {
+      "epoch": 1.5196437925615505,
+      "loss_partial": 0.046875,
+      "step": 2901
+    },
+    {
+      "epoch": 1.5196437925615505,
+      "loss_partial": 0.06640625,
+      "step": 2901
+    },
+    {
+      "epoch": 1.5201676270298585,
+      "grad_norm": 2.3051488399505615,
+      "learning_rate": 1.559555007098381e-05,
+      "loss": 0.1095,
+      "step": 2902
+    },
+    {
+      "epoch": 1.5201676270298585,
+      "loss_partial": 0.044677734375,
+      "step": 2902
+    },
+    {
+      "epoch": 1.5201676270298585,
+      "loss_partial": 0.052734375,
+      "step": 2902
+    },
+    {
+      "epoch": 1.5206914614981666,
+      "grad_norm": 2.094113826751709,
+      "learning_rate": 1.5591760588439568e-05,
+      "loss": 0.1035,
+      "step": 2903
+    },
+    {
+      "epoch": 1.5206914614981666,
+      "loss_partial": 0.06396484375,
+      "step": 2903
+    },
+    {
+      "epoch": 1.5206914614981666,
+      "loss_partial": 0.0546875,
+      "step": 2903
+    },
+    {
+      "epoch": 1.5212152959664746,
+      "grad_norm": 1.7703559398651123,
+      "learning_rate": 1.558796993725194e-05,
+      "loss": 0.1071,
+      "step": 2904
+    },
+    {
+      "epoch": 1.5212152959664746,
+      "loss_partial": 0.0517578125,
+      "step": 2904
+    },
+    {
+      "epoch": 1.5212152959664746,
+      "loss_partial": 0.056640625,
+      "step": 2904
+    },
+    {
+      "epoch": 1.5217391304347827,
+      "grad_norm": 1.924858808517456,
+      "learning_rate": 1.5584178118213147e-05,
+      "loss": 0.1046,
+      "step": 2905
+    },
+    {
+      "epoch": 1.5217391304347827,
+      "loss_partial": 0.0556640625,
+      "step": 2905
+    },
+    {
+      "epoch": 1.5217391304347827,
+      "loss_partial": 0.05810546875,
+      "step": 2905
+    },
+    {
+      "epoch": 1.5222629649030908,
+      "grad_norm": 2.402863025665283,
+      "learning_rate": 1.5580385132115662e-05,
+      "loss": 0.1083,
+      "step": 2906
+    },
+    {
+      "epoch": 1.5222629649030908,
+      "loss_partial": 0.051513671875,
+      "step": 2906
+    },
+    {
+      "epoch": 1.5222629649030908,
+      "loss_partial": 0.05517578125,
+      "step": 2906
+    },
+    {
+      "epoch": 1.5227867993713986,
+      "grad_norm": 1.9685168266296387,
+      "learning_rate": 1.5576590979752194e-05,
+      "loss": 0.1015,
+      "step": 2907
+    },
+    {
+      "epoch": 1.5227867993713986,
+      "loss_partial": 0.05078125,
+      "step": 2907
+    },
+    {
+      "epoch": 1.5227867993713986,
+      "loss_partial": 0.05322265625,
+      "step": 2907
+    },
+    {
+      "epoch": 1.5233106338397067,
+      "grad_norm": 1.91448175907135,
+      "learning_rate": 1.5572795661915693e-05,
+      "loss": 0.1018,
+      "step": 2908
+    },
+    {
+      "epoch": 1.5233106338397067,
+      "loss_partial": 0.051513671875,
+      "step": 2908
+    },
+    {
+      "epoch": 1.5233106338397067,
+      "loss_partial": 0.05224609375,
+      "step": 2908
+    },
+    {
+      "epoch": 1.5238344683080147,
+      "grad_norm": 2.2481205463409424,
+      "learning_rate": 1.5568999179399363e-05,
+      "loss": 0.1041,
+      "step": 2909
+    },
+    {
+      "epoch": 1.5238344683080147,
+      "loss_partial": 0.048095703125,
+      "step": 2909
+    },
+    {
+      "epoch": 1.5238344683080147,
+      "loss_partial": 0.059326171875,
+      "step": 2909
+    },
+    {
+      "epoch": 1.5243583027763226,
+      "grad_norm": 1.8512382507324219,
+      "learning_rate": 1.5565201532996638e-05,
+      "loss": 0.1002,
+      "step": 2910
+    },
+    {
+      "epoch": 1.5243583027763226,
+      "loss_partial": 0.055908203125,
+      "step": 2910
+    },
+    {
+      "epoch": 1.5243583027763226,
+      "loss_partial": 0.05224609375,
+      "step": 2910
+    },
+    {
+      "epoch": 1.5248821372446306,
+      "grad_norm": 2.5067145824432373,
+      "learning_rate": 1.556140272350121e-05,
+      "loss": 0.1052,
+      "step": 2911
+    },
+    {
+      "epoch": 1.5248821372446306,
+      "loss_partial": 0.0537109375,
+      "step": 2911
+    },
+    {
+      "epoch": 1.5248821372446306,
+      "loss_partial": 0.0478515625,
+      "step": 2911
+    },
+    {
+      "epoch": 1.5254059717129387,
+      "grad_norm": 2.0607762336730957,
+      "learning_rate": 1.5557602751707e-05,
+      "loss": 0.107,
+      "step": 2912
+    },
+    {
+      "epoch": 1.5254059717129387,
+      "loss_partial": 0.052001953125,
+      "step": 2912
+    },
+    {
+      "epoch": 1.5254059717129387,
+      "loss_partial": 0.049560546875,
+      "step": 2912
+    },
+    {
+      "epoch": 1.5259298061812467,
+      "grad_norm": 2.109957218170166,
+      "learning_rate": 1.5553801618408187e-05,
+      "loss": 0.1063,
+      "step": 2913
+    },
+    {
+      "epoch": 1.5259298061812467,
+      "loss_partial": 0.059326171875,
+      "step": 2913
+    },
+    {
+      "epoch": 1.5259298061812467,
+      "loss_partial": 0.052734375,
+      "step": 2913
+    },
+    {
+      "epoch": 1.5264536406495548,
+      "grad_norm": 1.73371160030365,
+      "learning_rate": 1.5549999324399174e-05,
+      "loss": 0.1075,
+      "step": 2914
+    },
+    {
+      "epoch": 1.5264536406495548,
+      "loss_partial": 0.05126953125,
+      "step": 2914
+    },
+    {
+      "epoch": 1.5264536406495548,
+      "loss_partial": 0.0576171875,
+      "step": 2914
+    },
+    {
+      "epoch": 1.5269774751178629,
+      "grad_norm": 1.8792797327041626,
+      "learning_rate": 1.5546195870474625e-05,
+      "loss": 0.1056,
+      "step": 2915
+    },
+    {
+      "epoch": 1.5269774751178629,
+      "loss_partial": 0.04833984375,
+      "step": 2915
+    },
+    {
+      "epoch": 1.5269774751178629,
+      "loss_partial": 0.06103515625,
+      "step": 2915
+    },
+    {
+      "epoch": 1.527501309586171,
+      "grad_norm": 1.41641104221344,
+      "learning_rate": 1.5542391257429438e-05,
+      "loss": 0.1032,
+      "step": 2916
+    },
+    {
+      "epoch": 1.527501309586171,
+      "loss_partial": 0.04443359375,
+      "step": 2916
+    },
+    {
+      "epoch": 1.527501309586171,
+      "loss_partial": 0.05078125,
+      "step": 2916
+    },
+    {
+      "epoch": 1.5280251440544788,
+      "grad_norm": 1.6860759258270264,
+      "learning_rate": 1.5538585486058747e-05,
+      "loss": 0.1057,
+      "step": 2917
+    },
+    {
+      "epoch": 1.5280251440544788,
+      "loss_partial": 0.055419921875,
+      "step": 2917
+    },
+    {
+      "epoch": 1.5280251440544788,
+      "loss_partial": 0.05224609375,
+      "step": 2917
+    },
+    {
+      "epoch": 1.5285489785227868,
+      "grad_norm": 1.737370252609253,
+      "learning_rate": 1.5534778557157943e-05,
+      "loss": 0.1061,
+      "step": 2918
+    },
+    {
+      "epoch": 1.5285489785227868,
+      "loss_partial": 0.055419921875,
+      "step": 2918
+    },
+    {
+      "epoch": 1.5285489785227868,
+      "loss_partial": 0.051513671875,
+      "step": 2918
+    },
+    {
+      "epoch": 1.5290728129910947,
+      "grad_norm": 1.6334807872772217,
+      "learning_rate": 1.5530970471522644e-05,
+      "loss": 0.1061,
+      "step": 2919
+    },
+    {
+      "epoch": 1.5290728129910947,
+      "loss_partial": 0.060791015625,
+      "step": 2919
+    },
+    {
+      "epoch": 1.5290728129910947,
+      "loss_partial": 0.0576171875,
+      "step": 2919
+    },
+    {
+      "epoch": 1.5295966474594027,
+      "grad_norm": 2.3248369693756104,
+      "learning_rate": 1.552716122994872e-05,
+      "loss": 0.1079,
+      "step": 2920
+    },
+    {
+      "epoch": 1.5295966474594027,
+      "loss_partial": 0.054443359375,
+      "step": 2920
+    },
+    {
+      "epoch": 1.5295966474594027,
+      "loss_partial": 0.06689453125,
+      "step": 2920
+    },
+    {
+      "epoch": 1.5301204819277108,
+      "grad_norm": 1.5151621103286743,
+      "learning_rate": 1.5523350833232278e-05,
+      "loss": 0.1042,
+      "step": 2921
+    },
+    {
+      "epoch": 1.5301204819277108,
+      "loss_partial": 0.0517578125,
+      "step": 2921
+    },
+    {
+      "epoch": 1.5301204819277108,
+      "loss_partial": 0.046875,
+      "step": 2921
+    },
+    {
+      "epoch": 1.5306443163960188,
+      "grad_norm": 2.61857271194458,
+      "learning_rate": 1.5519539282169666e-05,
+      "loss": 0.1053,
+      "step": 2922
+    },
+    {
+      "epoch": 1.5306443163960188,
+      "loss_partial": 0.047119140625,
+      "step": 2922
+    },
+    {
+      "epoch": 1.5306443163960188,
+      "loss_partial": 0.048583984375,
+      "step": 2922
+    },
+    {
+      "epoch": 1.531168150864327,
+      "grad_norm": 2.600139856338501,
+      "learning_rate": 1.5515726577557474e-05,
+      "loss": 0.1052,
+      "step": 2923
+    },
+    {
+      "epoch": 1.531168150864327,
+      "loss_partial": 0.04931640625,
+      "step": 2923
+    },
+    {
+      "epoch": 1.531168150864327,
+      "loss_partial": 0.05419921875,
+      "step": 2923
+    },
+    {
+      "epoch": 1.531691985332635,
+      "grad_norm": 2.0038845539093018,
+      "learning_rate": 1.551191272019254e-05,
+      "loss": 0.1049,
+      "step": 2924
+    },
+    {
+      "epoch": 1.531691985332635,
+      "loss_partial": 0.0478515625,
+      "step": 2924
+    },
+    {
+      "epoch": 1.531691985332635,
+      "loss_partial": 0.0546875,
+      "step": 2924
+    },
+    {
+      "epoch": 1.532215819800943,
+      "grad_norm": 2.0864756107330322,
+      "learning_rate": 1.550809771087193e-05,
+      "loss": 0.1021,
+      "step": 2925
+    },
+    {
+      "epoch": 1.532215819800943,
+      "loss_partial": 0.05224609375,
+      "step": 2925
+    },
+    {
+      "epoch": 1.532215819800943,
+      "loss_partial": 0.05029296875,
+      "step": 2925
+    },
+    {
+      "epoch": 1.532739654269251,
+      "grad_norm": 1.6944499015808105,
+      "learning_rate": 1.550428155039296e-05,
+      "loss": 0.1042,
+      "step": 2926
+    },
+    {
+      "epoch": 1.532739654269251,
+      "loss_partial": 0.04931640625,
+      "step": 2926
+    },
+    {
+      "epoch": 1.532739654269251,
+      "loss_partial": 0.055908203125,
+      "step": 2926
+    },
+    {
+      "epoch": 1.533263488737559,
+      "grad_norm": 1.7469154596328735,
+      "learning_rate": 1.5500464239553186e-05,
+      "loss": 0.1039,
+      "step": 2927
+    },
+    {
+      "epoch": 1.533263488737559,
+      "loss_partial": 0.0654296875,
+      "step": 2927
+    },
+    {
+      "epoch": 1.533263488737559,
+      "loss_partial": 0.053955078125,
+      "step": 2927
+    },
+    {
+      "epoch": 1.533787323205867,
+      "grad_norm": 1.7101576328277588,
+      "learning_rate": 1.5496645779150395e-05,
+      "loss": 0.1077,
+      "step": 2928
+    },
+    {
+      "epoch": 1.533787323205867,
+      "loss_partial": 0.052001953125,
+      "step": 2928
+    },
+    {
+      "epoch": 1.533787323205867,
+      "loss_partial": 0.046875,
+      "step": 2928
+    },
+    {
+      "epoch": 1.5343111576741748,
+      "grad_norm": 2.1079602241516113,
+      "learning_rate": 1.5492826169982626e-05,
+      "loss": 0.1021,
+      "step": 2929
+    },
+    {
+      "epoch": 1.5343111576741748,
+      "loss_partial": 0.0498046875,
+      "step": 2929
+    },
+    {
+      "epoch": 1.5343111576741748,
+      "loss_partial": 0.060546875,
+      "step": 2929
+    },
+    {
+      "epoch": 1.5348349921424829,
+      "grad_norm": 1.5461714267730713,
+      "learning_rate": 1.548900541284816e-05,
+      "loss": 0.1082,
+      "step": 2930
+    },
+    {
+      "epoch": 1.5348349921424829,
+      "loss_partial": 0.055908203125,
+      "step": 2930
+    },
+    {
+      "epoch": 1.5348349921424829,
+      "loss_partial": 0.052978515625,
+      "step": 2930
+    },
+    {
+      "epoch": 1.535358826610791,
+      "grad_norm": 2.2899889945983887,
+      "learning_rate": 1.54851835085455e-05,
+      "loss": 0.1059,
+      "step": 2931
+    },
+    {
+      "epoch": 1.535358826610791,
+      "loss_partial": 0.046630859375,
+      "step": 2931
+    },
+    {
+      "epoch": 1.535358826610791,
+      "loss_partial": 0.057861328125,
+      "step": 2931
+    },
+    {
+      "epoch": 1.535882661079099,
+      "grad_norm": 2.380585193634033,
+      "learning_rate": 1.548136045787341e-05,
+      "loss": 0.1012,
+      "step": 2932
+    },
+    {
+      "epoch": 1.535882661079099,
+      "loss_partial": 0.057861328125,
+      "step": 2932
+    },
+    {
+      "epoch": 1.535882661079099,
+      "loss_partial": 0.059326171875,
+      "step": 2932
+    },
+    {
+      "epoch": 1.536406495547407,
+      "grad_norm": 1.8771768808364868,
+      "learning_rate": 1.547753626163088e-05,
+      "loss": 0.1024,
+      "step": 2933
+    },
+    {
+      "epoch": 1.536406495547407,
+      "loss_partial": 0.0458984375,
+      "step": 2933
+    },
+    {
+      "epoch": 1.536406495547407,
+      "loss_partial": 0.056640625,
+      "step": 2933
+    },
+    {
+      "epoch": 1.536930330015715,
+      "grad_norm": 2.4316565990448,
+      "learning_rate": 1.5473710920617137e-05,
+      "loss": 0.1052,
+      "step": 2934
+    },
+    {
+      "epoch": 1.536930330015715,
+      "loss_partial": 0.0478515625,
+      "step": 2934
+    },
+    {
+      "epoch": 1.536930330015715,
+      "loss_partial": 0.049072265625,
+      "step": 2934
+    },
+    {
+      "epoch": 1.5374541644840232,
+      "grad_norm": 1.854496955871582,
+      "learning_rate": 1.5469884435631666e-05,
+      "loss": 0.1039,
+      "step": 2935
+    },
+    {
+      "epoch": 1.5374541644840232,
+      "loss_partial": 0.048095703125,
+      "step": 2935
+    },
+    {
+      "epoch": 1.5374541644840232,
+      "loss_partial": 0.06396484375,
+      "step": 2935
+    },
+    {
+      "epoch": 1.5379779989523312,
+      "grad_norm": 1.9114913940429688,
+      "learning_rate": 1.546605680747417e-05,
+      "loss": 0.107,
+      "step": 2936
+    },
+    {
+      "epoch": 1.5379779989523312,
+      "loss_partial": 0.05419921875,
+      "step": 2936
+    },
+    {
+      "epoch": 1.5379779989523312,
+      "loss_partial": 0.051513671875,
+      "step": 2936
+    },
+    {
+      "epoch": 1.538501833420639,
+      "grad_norm": 2.692286491394043,
+      "learning_rate": 1.5462228036944608e-05,
+      "loss": 0.1055,
+      "step": 2937
+    },
+    {
+      "epoch": 1.538501833420639,
+      "loss_partial": 0.06884765625,
+      "step": 2937
+    },
+    {
+      "epoch": 1.538501833420639,
+      "loss_partial": 0.04931640625,
+      "step": 2937
+    },
+    {
+      "epoch": 1.5390256678889471,
+      "grad_norm": 1.8939118385314941,
+      "learning_rate": 1.5458398124843155e-05,
+      "loss": 0.1078,
+      "step": 2938
+    },
+    {
+      "epoch": 1.5390256678889471,
+      "loss_partial": 0.05322265625,
+      "step": 2938
+    },
+    {
+      "epoch": 1.5390256678889471,
+      "loss_partial": 0.07763671875,
+      "step": 2938
+    },
+    {
+      "epoch": 1.539549502357255,
+      "grad_norm": 2.108433246612549,
+      "learning_rate": 1.545456707197025e-05,
+      "loss": 0.1035,
+      "step": 2939
+    },
+    {
+      "epoch": 1.539549502357255,
+      "loss_partial": 0.060791015625,
+      "step": 2939
+    },
+    {
+      "epoch": 1.539549502357255,
+      "loss_partial": 0.05126953125,
+      "step": 2939
+    },
+    {
+      "epoch": 1.540073336825563,
+      "grad_norm": 1.7230746746063232,
+      "learning_rate": 1.5450734879126562e-05,
+      "loss": 0.1054,
+      "step": 2940
+    },
+    {
+      "epoch": 1.540073336825563,
+      "loss_partial": 0.058837890625,
+      "step": 2940
+    },
+    {
+      "epoch": 1.540073336825563,
+      "loss_partial": 0.048583984375,
+      "step": 2940
+    },
+    {
+      "epoch": 1.540597171293871,
+      "grad_norm": 1.3404326438903809,
+      "learning_rate": 1.5446901547112986e-05,
+      "loss": 0.104,
+      "step": 2941
+    },
+    {
+      "epoch": 1.540597171293871,
+      "loss_partial": 0.046875,
+      "step": 2941
+    },
+    {
+      "epoch": 1.540597171293871,
+      "loss_partial": 0.05224609375,
+      "step": 2941
+    },
+    {
+      "epoch": 1.5411210057621791,
+      "grad_norm": 1.5903006792068481,
+      "learning_rate": 1.5443067076730673e-05,
+      "loss": 0.1057,
+      "step": 2942
+    },
+    {
+      "epoch": 1.5411210057621791,
+      "loss_partial": 0.0498046875,
+      "step": 2942
+    },
+    {
+      "epoch": 1.5411210057621791,
+      "loss_partial": 0.05224609375,
+      "step": 2942
+    },
+    {
+      "epoch": 1.5416448402304872,
+      "grad_norm": 1.8874188661575317,
+      "learning_rate": 1.5439231468780997e-05,
+      "loss": 0.1021,
+      "step": 2943
+    },
+    {
+      "epoch": 1.5416448402304872,
+      "loss_partial": 0.050048828125,
+      "step": 2943
+    },
+    {
+      "epoch": 1.5416448402304872,
+      "loss_partial": 0.060546875,
+      "step": 2943
+    },
+    {
+      "epoch": 1.5421686746987953,
+      "grad_norm": 1.6640375852584839,
+      "learning_rate": 1.543539472406558e-05,
+      "loss": 0.1005,
+      "step": 2944
+    },
+    {
+      "epoch": 1.5421686746987953,
+      "loss_partial": 0.0615234375,
+      "step": 2944
+    },
+    {
+      "epoch": 1.5421686746987953,
+      "loss_partial": 0.046875,
+      "step": 2944
+    },
+    {
+      "epoch": 1.5426925091671033,
+      "grad_norm": 2.0098888874053955,
+      "learning_rate": 1.5431556843386277e-05,
+      "loss": 0.1047,
+      "step": 2945
+    },
+    {
+      "epoch": 1.5426925091671033,
+      "loss_partial": 0.047607421875,
+      "step": 2945
+    },
+    {
+      "epoch": 1.5426925091671033,
+      "loss_partial": 0.0498046875,
+      "step": 2945
+    },
+    {
+      "epoch": 1.5432163436354112,
+      "grad_norm": 1.649552345275879,
+      "learning_rate": 1.5427717827545184e-05,
+      "loss": 0.103,
+      "step": 2946
+    },
+    {
+      "epoch": 1.5432163436354112,
+      "loss_partial": 0.05810546875,
+      "step": 2946
+    },
+    {
+      "epoch": 1.5432163436354112,
+      "loss_partial": 0.0498046875,
+      "step": 2946
+    },
+    {
+      "epoch": 1.5437401781037192,
+      "grad_norm": 1.7406262159347534,
+      "learning_rate": 1.5423877677344625e-05,
+      "loss": 0.1053,
+      "step": 2947
+    },
+    {
+      "epoch": 1.5437401781037192,
+      "loss_partial": 0.049560546875,
+      "step": 2947
+    },
+    {
+      "epoch": 1.5437401781037192,
+      "loss_partial": 0.049560546875,
+      "step": 2947
+    },
+    {
+      "epoch": 1.5442640125720273,
+      "grad_norm": 1.851324439048767,
+      "learning_rate": 1.542003639358717e-05,
+      "loss": 0.1067,
+      "step": 2948
+    },
+    {
+      "epoch": 1.5442640125720273,
+      "loss_partial": 0.05029296875,
+      "step": 2948
+    },
+    {
+      "epoch": 1.5442640125720273,
+      "loss_partial": 0.05224609375,
+      "step": 2948
+    },
+    {
+      "epoch": 1.5447878470403351,
+      "grad_norm": 1.3547554016113281,
+      "learning_rate": 1.541619397707563e-05,
+      "loss": 0.1068,
+      "step": 2949
+    },
+    {
+      "epoch": 1.5447878470403351,
+      "loss_partial": 0.061279296875,
+      "step": 2949
+    },
+    {
+      "epoch": 1.5447878470403351,
+      "loss_partial": 0.040771484375,
+      "step": 2949
+    },
+    {
+      "epoch": 1.5453116815086432,
+      "grad_norm": 2.6245388984680176,
+      "learning_rate": 1.5412350428613042e-05,
+      "loss": 0.108,
+      "step": 2950
+    },
+    {
+      "epoch": 1.5453116815086432,
+      "loss_partial": 0.04931640625,
+      "step": 2950
+    },
+    {
+      "epoch": 1.5453116815086432,
+      "loss_partial": 0.048583984375,
+      "step": 2950
+    },
+    {
+      "epoch": 1.5458355159769512,
+      "grad_norm": 1.6514451503753662,
+      "learning_rate": 1.540850574900268e-05,
+      "loss": 0.107,
+      "step": 2951
+    },
+    {
+      "epoch": 1.5458355159769512,
+      "loss_partial": 0.054931640625,
+      "step": 2951
+    },
+    {
+      "epoch": 1.5458355159769512,
+      "loss_partial": 0.05712890625,
+      "step": 2951
+    },
+    {
+      "epoch": 1.5463593504452593,
+      "grad_norm": 2.5415048599243164,
+      "learning_rate": 1.5404659939048064e-05,
+      "loss": 0.1059,
+      "step": 2952
+    },
+    {
+      "epoch": 1.5463593504452593,
+      "loss_partial": 0.052978515625,
+      "step": 2952
+    },
+    {
+      "epoch": 1.5463593504452593,
+      "loss_partial": 0.054443359375,
+      "step": 2952
+    },
+    {
+      "epoch": 1.5468831849135674,
+      "grad_norm": 1.959229826927185,
+      "learning_rate": 1.540081299955294e-05,
+      "loss": 0.1079,
+      "step": 2953
+    },
+    {
+      "epoch": 1.5468831849135674,
+      "loss_partial": 0.045654296875,
+      "step": 2953
+    },
+    {
+      "epoch": 1.5468831849135674,
+      "loss_partial": 0.04931640625,
+      "step": 2953
+    },
+    {
+      "epoch": 1.5474070193818754,
+      "grad_norm": 2.179314613342285,
+      "learning_rate": 1.5396964931321297e-05,
+      "loss": 0.1032,
+      "step": 2954
+    },
+    {
+      "epoch": 1.5474070193818754,
+      "loss_partial": 0.049560546875,
+      "step": 2954
+    },
+    {
+      "epoch": 1.5474070193818754,
+      "loss_partial": 0.048583984375,
+      "step": 2954
+    },
+    {
+      "epoch": 1.5479308538501835,
+      "grad_norm": 2.2039008140563965,
+      "learning_rate": 1.539311573515736e-05,
+      "loss": 0.1024,
+      "step": 2955
+    },
+    {
+      "epoch": 1.5479308538501835,
+      "loss_partial": 0.06298828125,
+      "step": 2955
+    },
+    {
+      "epoch": 1.5479308538501835,
+      "loss_partial": 0.043701171875,
+      "step": 2955
+    },
+    {
+      "epoch": 1.5484546883184913,
+      "grad_norm": 2.5562760829925537,
+      "learning_rate": 1.5389265411865582e-05,
+      "loss": 0.1068,
+      "step": 2956
+    },
+    {
+      "epoch": 1.5484546883184913,
+      "loss_partial": 0.052978515625,
+      "step": 2956
+    },
+    {
+      "epoch": 1.5484546883184913,
+      "loss_partial": 0.05419921875,
+      "step": 2956
+    },
+    {
+      "epoch": 1.5489785227867994,
+      "grad_norm": 2.4580495357513428,
+      "learning_rate": 1.5385413962250657e-05,
+      "loss": 0.1053,
+      "step": 2957
+    },
+    {
+      "epoch": 1.5489785227867994,
+      "loss_partial": 0.052001953125,
+      "step": 2957
+    },
+    {
+      "epoch": 1.5489785227867994,
+      "loss_partial": 0.0625,
+      "step": 2957
+    },
+    {
+      "epoch": 1.5495023572551074,
+      "grad_norm": 2.930985450744629,
+      "learning_rate": 1.538156138711752e-05,
+      "loss": 0.1082,
+      "step": 2958
+    },
+    {
+      "epoch": 1.5495023572551074,
+      "loss_partial": 0.043701171875,
+      "step": 2958
+    },
+    {
+      "epoch": 1.5495023572551074,
+      "loss_partial": 0.05224609375,
+      "step": 2958
+    },
+    {
+      "epoch": 1.5500261917234153,
+      "grad_norm": 1.4846974611282349,
+      "learning_rate": 1.537770768727133e-05,
+      "loss": 0.1072,
+      "step": 2959
+    },
+    {
+      "epoch": 1.5500261917234153,
+      "loss_partial": 0.054443359375,
+      "step": 2959
+    },
+    {
+      "epoch": 1.5500261917234153,
+      "loss_partial": 0.068359375,
+      "step": 2959
+    },
+    {
+      "epoch": 1.5505500261917233,
+      "grad_norm": 2.464038848876953,
+      "learning_rate": 1.537385286351749e-05,
+      "loss": 0.11,
+      "step": 2960
+    },
+    {
+      "epoch": 1.5505500261917233,
+      "loss_partial": 0.05224609375,
+      "step": 2960
+    },
+    {
+      "epoch": 1.5505500261917233,
+      "loss_partial": 0.048095703125,
+      "step": 2960
+    },
+    {
+      "epoch": 1.5510738606600314,
+      "grad_norm": 1.5823040008544922,
+      "learning_rate": 1.536999691666163e-05,
+      "loss": 0.1056,
+      "step": 2961
+    },
+    {
+      "epoch": 1.5510738606600314,
+      "loss_partial": 0.05078125,
+      "step": 2961
+    },
+    {
+      "epoch": 1.5510738606600314,
+      "loss_partial": 0.04638671875,
+      "step": 2961
+    },
+    {
+      "epoch": 1.5515976951283394,
+      "grad_norm": 1.4167269468307495,
+      "learning_rate": 1.5366139847509625e-05,
+      "loss": 0.1069,
+      "step": 2962
+    },
+    {
+      "epoch": 1.5515976951283394,
+      "loss_partial": 0.05224609375,
+      "step": 2962
+    },
+    {
+      "epoch": 1.5515976951283394,
+      "loss_partial": 0.060302734375,
+      "step": 2962
+    },
+    {
+      "epoch": 1.5521215295966475,
+      "grad_norm": 2.0336852073669434,
+      "learning_rate": 1.536228165686757e-05,
+      "loss": 0.1049,
+      "step": 2963
+    },
+    {
+      "epoch": 1.5521215295966475,
+      "loss_partial": 0.052978515625,
+      "step": 2963
+    },
+    {
+      "epoch": 1.5521215295966475,
+      "loss_partial": 0.050048828125,
+      "step": 2963
+    },
+    {
+      "epoch": 1.5526453640649556,
+      "grad_norm": 1.6790026426315308,
+      "learning_rate": 1.535842234554181e-05,
+      "loss": 0.1035,
+      "step": 2964
+    },
+    {
+      "epoch": 1.5526453640649556,
+      "loss_partial": 0.046142578125,
+      "step": 2964
+    },
+    {
+      "epoch": 1.5526453640649556,
+      "loss_partial": 0.04736328125,
+      "step": 2964
+    },
+    {
+      "epoch": 1.5531691985332636,
+      "grad_norm": 1.608184814453125,
+      "learning_rate": 1.5354561914338917e-05,
+      "loss": 0.1046,
+      "step": 2965
+    },
+    {
+      "epoch": 1.5531691985332636,
+      "loss_partial": 0.046630859375,
+      "step": 2965
+    },
+    {
+      "epoch": 1.5531691985332636,
+      "loss_partial": 0.0439453125,
+      "step": 2965
+    },
+    {
+      "epoch": 1.5536930330015715,
+      "grad_norm": 2.5588347911834717,
+      "learning_rate": 1.5350700364065696e-05,
+      "loss": 0.1055,
+      "step": 2966
+    },
+    {
+      "epoch": 1.5536930330015715,
+      "loss_partial": 0.04833984375,
+      "step": 2966
+    },
+    {
+      "epoch": 1.5536930330015715,
+      "loss_partial": 0.06787109375,
+      "step": 2966
+    },
+    {
+      "epoch": 1.5542168674698795,
+      "grad_norm": 1.7044183015823364,
+      "learning_rate": 1.5346837695529183e-05,
+      "loss": 0.1076,
+      "step": 2967
+    },
+    {
+      "epoch": 1.5542168674698795,
+      "loss_partial": 0.0556640625,
+      "step": 2967
+    },
+    {
+      "epoch": 1.5542168674698795,
+      "loss_partial": 0.041748046875,
+      "step": 2967
+    },
+    {
+      "epoch": 1.5547407019381876,
+      "grad_norm": 2.044344902038574,
+      "learning_rate": 1.5342973909536662e-05,
+      "loss": 0.1044,
+      "step": 2968
+    },
+    {
+      "epoch": 1.5547407019381876,
+      "loss_partial": 0.056396484375,
+      "step": 2968
+    },
+    {
+      "epoch": 1.5547407019381876,
+      "loss_partial": 0.055908203125,
+      "step": 2968
+    },
+    {
+      "epoch": 1.5552645364064954,
+      "grad_norm": 1.667266845703125,
+      "learning_rate": 1.5339109006895626e-05,
+      "loss": 0.1067,
+      "step": 2969
+    },
+    {
+      "epoch": 1.5552645364064954,
+      "loss_partial": 0.04833984375,
+      "step": 2969
+    },
+    {
+      "epoch": 1.5552645364064954,
+      "loss_partial": 0.056396484375,
+      "step": 2969
+    },
+    {
+      "epoch": 1.5557883708748035,
+      "grad_norm": 2.029794454574585,
+      "learning_rate": 1.533524298841383e-05,
+      "loss": 0.1021,
+      "step": 2970
+    },
+    {
+      "epoch": 1.5557883708748035,
+      "loss_partial": 0.05615234375,
+      "step": 2970
+    },
+    {
+      "epoch": 1.5557883708748035,
+      "loss_partial": 0.052978515625,
+      "step": 2970
+    },
+    {
+      "epoch": 1.5563122053431115,
+      "grad_norm": 1.289414882659912,
+      "learning_rate": 1.533137585489924e-05,
+      "loss": 0.1039,
+      "step": 2971
+    },
+    {
+      "epoch": 1.5563122053431115,
+      "loss_partial": 0.052001953125,
+      "step": 2971
+    },
+    {
+      "epoch": 1.5563122053431115,
+      "loss_partial": 0.062255859375,
+      "step": 2971
+    },
+    {
+      "epoch": 1.5568360398114196,
+      "grad_norm": 1.5065021514892578,
+      "learning_rate": 1.5327507607160058e-05,
+      "loss": 0.1079,
+      "step": 2972
+    },
+    {
+      "epoch": 1.5568360398114196,
+      "loss_partial": 0.054443359375,
+      "step": 2972
+    },
+    {
+      "epoch": 1.5568360398114196,
+      "loss_partial": 0.0517578125,
+      "step": 2972
+    },
+    {
+      "epoch": 1.5573598742797277,
+      "grad_norm": 1.7801318168640137,
+      "learning_rate": 1.5323638246004735e-05,
+      "loss": 0.1068,
+      "step": 2973
+    },
+    {
+      "epoch": 1.5573598742797277,
+      "loss_partial": 0.05810546875,
+      "step": 2973
+    },
+    {
+      "epoch": 1.5573598742797277,
+      "loss_partial": 0.058837890625,
+      "step": 2973
+    },
+    {
+      "epoch": 1.5578837087480357,
+      "grad_norm": 1.8043224811553955,
+      "learning_rate": 1.5319767772241942e-05,
+      "loss": 0.1061,
+      "step": 2974
+    },
+    {
+      "epoch": 1.5578837087480357,
+      "loss_partial": 0.044189453125,
+      "step": 2974
+    },
+    {
+      "epoch": 1.5578837087480357,
+      "loss_partial": 0.048583984375,
+      "step": 2974
+    },
+    {
+      "epoch": 1.5584075432163438,
+      "grad_norm": 1.6318395137786865,
+      "learning_rate": 1.5315896186680575e-05,
+      "loss": 0.0997,
+      "step": 2975
+    },
+    {
+      "epoch": 1.5584075432163438,
+      "loss_partial": 0.056396484375,
+      "step": 2975
+    },
+    {
+      "epoch": 1.5584075432163438,
+      "loss_partial": 0.049072265625,
+      "step": 2975
+    },
+    {
+      "epoch": 1.5589313776846516,
+      "grad_norm": 1.6016838550567627,
+      "learning_rate": 1.5312023490129785e-05,
+      "loss": 0.1035,
+      "step": 2976
+    },
+    {
+      "epoch": 1.5589313776846516,
+      "loss_partial": 0.048583984375,
+      "step": 2976
+    },
+    {
+      "epoch": 1.5589313776846516,
+      "loss_partial": 0.0654296875,
+      "step": 2976
+    },
+    {
+      "epoch": 1.5594552121529597,
+      "grad_norm": 2.5335676670074463,
+      "learning_rate": 1.530814968339893e-05,
+      "loss": 0.1109,
+      "step": 2977
+    },
+    {
+      "epoch": 1.5594552121529597,
+      "loss_partial": 0.06787109375,
+      "step": 2977
+    },
+    {
+      "epoch": 1.5594552121529597,
+      "loss_partial": 0.049072265625,
+      "step": 2977
+    },
+    {
+      "epoch": 1.5599790466212675,
+      "grad_norm": 1.391316533088684,
+      "learning_rate": 1.5304274767297617e-05,
+      "loss": 0.1084,
+      "step": 2978
+    },
+    {
+      "epoch": 1.5599790466212675,
+      "loss_partial": 0.04541015625,
+      "step": 2978
+    },
+    {
+      "epoch": 1.5599790466212675,
+      "loss_partial": 0.05810546875,
+      "step": 2978
+    },
+    {
+      "epoch": 1.5605028810895756,
+      "grad_norm": 2.257941961288452,
+      "learning_rate": 1.5300398742635677e-05,
+      "loss": 0.1056,
+      "step": 2979
+    },
+    {
+      "epoch": 1.5605028810895756,
+      "loss_partial": 0.052490234375,
+      "step": 2979
+    },
+    {
+      "epoch": 1.5605028810895756,
+      "loss_partial": 0.047119140625,
+      "step": 2979
+    },
+    {
+      "epoch": 1.5610267155578836,
+      "grad_norm": 1.6574158668518066,
+      "learning_rate": 1.5296521610223175e-05,
+      "loss": 0.1056,
+      "step": 2980
+    },
+    {
+      "epoch": 1.5610267155578836,
+      "loss_partial": 0.046142578125,
+      "step": 2980
+    },
+    {
+      "epoch": 1.5610267155578836,
+      "loss_partial": 0.04736328125,
+      "step": 2980
+    },
+    {
+      "epoch": 1.5615505500261917,
+      "grad_norm": 2.3131542205810547,
+      "learning_rate": 1.529264337087041e-05,
+      "loss": 0.1046,
+      "step": 2981
+    },
+    {
+      "epoch": 1.5615505500261917,
+      "loss_partial": 0.051025390625,
+      "step": 2981
+    },
+    {
+      "epoch": 1.5615505500261917,
+      "loss_partial": 0.0439453125,
+      "step": 2981
+    },
+    {
+      "epoch": 1.5620743844944998,
+      "grad_norm": 2.1384053230285645,
+      "learning_rate": 1.5288764025387912e-05,
+      "loss": 0.1057,
+      "step": 2982
+    },
+    {
+      "epoch": 1.5620743844944998,
+      "loss_partial": 0.054443359375,
+      "step": 2982
+    },
+    {
+      "epoch": 1.5620743844944998,
+      "loss_partial": 0.059814453125,
+      "step": 2982
+    },
+    {
+      "epoch": 1.5625982189628078,
+      "grad_norm": 2.0659801959991455,
+      "learning_rate": 1.5284883574586436e-05,
+      "loss": 0.1058,
+      "step": 2983
+    },
+    {
+      "epoch": 1.5625982189628078,
+      "loss_partial": 0.059326171875,
+      "step": 2983
+    },
+    {
+      "epoch": 1.5625982189628078,
+      "loss_partial": 0.05224609375,
+      "step": 2983
+    },
+    {
+      "epoch": 1.5631220534311159,
+      "grad_norm": 2.371082067489624,
+      "learning_rate": 1.5281002019276973e-05,
+      "loss": 0.1087,
+      "step": 2984
+    },
+    {
+      "epoch": 1.5631220534311159,
+      "loss_partial": 0.04248046875,
+      "step": 2984
+    },
+    {
+      "epoch": 1.5631220534311159,
+      "loss_partial": 0.057861328125,
+      "step": 2984
+    },
+    {
+      "epoch": 1.563645887899424,
+      "grad_norm": 1.8270812034606934,
+      "learning_rate": 1.5277119360270744e-05,
+      "loss": 0.1049,
+      "step": 2985
+    },
+    {
+      "epoch": 1.563645887899424,
+      "loss_partial": 0.0517578125,
+      "step": 2985
+    },
+    {
+      "epoch": 1.563645887899424,
+      "loss_partial": 0.048583984375,
+      "step": 2985
+    },
+    {
+      "epoch": 1.5641697223677318,
+      "grad_norm": 2.0408644676208496,
+      "learning_rate": 1.5273235598379203e-05,
+      "loss": 0.108,
+      "step": 2986
+    },
+    {
+      "epoch": 1.5641697223677318,
+      "loss_partial": 0.0537109375,
+      "step": 2986
+    },
+    {
+      "epoch": 1.5641697223677318,
+      "loss_partial": 0.05078125,
+      "step": 2986
+    },
+    {
+      "epoch": 1.5646935568360398,
+      "grad_norm": 2.4998841285705566,
+      "learning_rate": 1.5269350734414028e-05,
+      "loss": 0.1062,
+      "step": 2987
+    },
+    {
+      "epoch": 1.5646935568360398,
+      "loss_partial": 0.04736328125,
+      "step": 2987
+    },
+    {
+      "epoch": 1.5646935568360398,
+      "loss_partial": 0.05419921875,
+      "step": 2987
+    },
+    {
+      "epoch": 1.5652173913043477,
+      "grad_norm": 2.021735191345215,
+      "learning_rate": 1.5265464769187136e-05,
+      "loss": 0.1039,
+      "step": 2988
+    },
+    {
+      "epoch": 1.5652173913043477,
+      "loss_partial": 0.053466796875,
+      "step": 2988
+    },
+    {
+      "epoch": 1.5652173913043477,
+      "loss_partial": 0.055419921875,
+      "step": 2988
+    },
+    {
+      "epoch": 1.5657412257726557,
+      "grad_norm": 2.1327035427093506,
+      "learning_rate": 1.5261577703510672e-05,
+      "loss": 0.1083,
+      "step": 2989
+    },
+    {
+      "epoch": 1.5657412257726557,
+      "loss_partial": 0.06201171875,
+      "step": 2989
+    },
+    {
+      "epoch": 1.5657412257726557,
+      "loss_partial": 0.055908203125,
+      "step": 2989
+    },
+    {
+      "epoch": 1.5662650602409638,
+      "grad_norm": 2.5197203159332275,
+      "learning_rate": 1.5257689538197e-05,
+      "loss": 0.1078,
+      "step": 2990
+    },
+    {
+      "epoch": 1.5662650602409638,
+      "loss_partial": 0.06201171875,
+      "step": 2990
+    },
+    {
+      "epoch": 1.5662650602409638,
+      "loss_partial": 0.050048828125,
+      "step": 2990
+    },
+    {
+      "epoch": 1.5667888947092719,
+      "grad_norm": 1.86941397190094,
+      "learning_rate": 1.5253800274058726e-05,
+      "loss": 0.1072,
+      "step": 2991
+    },
+    {
+      "epoch": 1.5667888947092719,
+      "loss_partial": 0.05517578125,
+      "step": 2991
+    },
+    {
+      "epoch": 1.5667888947092719,
+      "loss_partial": 0.06640625,
+      "step": 2991
+    },
+    {
+      "epoch": 1.56731272917758,
+      "grad_norm": 2.183678388595581,
+      "learning_rate": 1.5249909911908685e-05,
+      "loss": 0.1074,
+      "step": 2992
+    },
+    {
+      "epoch": 1.56731272917758,
+      "loss_partial": 0.053955078125,
+      "step": 2992
+    },
+    {
+      "epoch": 1.56731272917758,
+      "loss_partial": 0.05615234375,
+      "step": 2992
+    },
+    {
+      "epoch": 1.567836563645888,
+      "grad_norm": 1.9098516702651978,
+      "learning_rate": 1.5246018452559936e-05,
+      "loss": 0.1058,
+      "step": 2993
+    },
+    {
+      "epoch": 1.567836563645888,
+      "loss_partial": 0.052490234375,
+      "step": 2993
+    },
+    {
+      "epoch": 1.567836563645888,
+      "loss_partial": 0.061279296875,
+      "step": 2993
+    },
+    {
+      "epoch": 1.568360398114196,
+      "grad_norm": 1.7897162437438965,
+      "learning_rate": 1.5242125896825775e-05,
+      "loss": 0.1052,
+      "step": 2994
+    },
+    {
+      "epoch": 1.568360398114196,
+      "loss_partial": 0.05712890625,
+      "step": 2994
+    },
+    {
+      "epoch": 1.568360398114196,
+      "loss_partial": 0.05419921875,
+      "step": 2994
+    },
+    {
+      "epoch": 1.568884232582504,
+      "grad_norm": 2.216892719268799,
+      "learning_rate": 1.5238232245519713e-05,
+      "loss": 0.1028,
+      "step": 2995
+    },
+    {
+      "epoch": 1.568884232582504,
+      "loss_partial": 0.05322265625,
+      "step": 2995
+    },
+    {
+      "epoch": 1.568884232582504,
+      "loss_partial": 0.058349609375,
+      "step": 2995
+    },
+    {
+      "epoch": 1.569408067050812,
+      "grad_norm": 1.4949967861175537,
+      "learning_rate": 1.5234337499455507e-05,
+      "loss": 0.1024,
+      "step": 2996
+    },
+    {
+      "epoch": 1.569408067050812,
+      "loss_partial": 0.055419921875,
+      "step": 2996
+    },
+    {
+      "epoch": 1.569408067050812,
+      "loss_partial": 0.060546875,
+      "step": 2996
+    },
+    {
+      "epoch": 1.56993190151912,
+      "grad_norm": 2.8759686946868896,
+      "learning_rate": 1.5230441659447128e-05,
+      "loss": 0.111,
+      "step": 2997
+    },
+    {
+      "epoch": 1.56993190151912,
+      "loss_partial": 0.0537109375,
+      "step": 2997
+    },
+    {
+      "epoch": 1.56993190151912,
+      "loss_partial": 0.053955078125,
+      "step": 2997
+    },
+    {
+      "epoch": 1.5704557359874278,
+      "grad_norm": 2.1323699951171875,
+      "learning_rate": 1.522654472630879e-05,
+      "loss": 0.1056,
+      "step": 2998
+    },
+    {
+      "epoch": 1.5704557359874278,
+      "loss_partial": 0.0654296875,
+      "step": 2998
+    },
+    {
+      "epoch": 1.5704557359874278,
+      "loss_partial": 0.048828125,
+      "step": 2998
+    },
+    {
+      "epoch": 1.5709795704557359,
+      "grad_norm": 1.8517512083053589,
+      "learning_rate": 1.5222646700854923e-05,
+      "loss": 0.105,
+      "step": 2999
+    },
+    {
+      "epoch": 1.5709795704557359,
+      "loss_partial": 0.048828125,
+      "step": 2999
+    },
+    {
+      "epoch": 1.5709795704557359,
+      "loss_partial": 0.06689453125,
+      "step": 2999
+    },
+    {
+      "epoch": 1.571503404924044,
+      "grad_norm": 2.4281389713287354,
+      "learning_rate": 1.5218747583900191e-05,
+      "loss": 0.1048,
+      "step": 3000
+    },
+    {
+      "epoch": 1.571503404924044,
+      "loss_partial": 0.053466796875,
+      "step": 3000
+    },
+    {
+      "epoch": 1.571503404924044,
+      "loss_partial": 0.056640625,
+      "step": 3000
+    },
+    {
+      "epoch": 1.572027239392352,
+      "grad_norm": 1.5634186267852783,
+      "learning_rate": 1.5214847376259485e-05,
+      "loss": 0.1042,
+      "step": 3001
+    },
+    {
+      "epoch": 1.572027239392352,
+      "loss_partial": 0.0498046875,
+      "step": 3001
+    },
+    {
+      "epoch": 1.572027239392352,
+      "loss_partial": 0.052734375,
+      "step": 3001
+    },
+    {
+      "epoch": 1.57255107386066,
+      "grad_norm": 2.098299503326416,
+      "learning_rate": 1.5210946078747926e-05,
+      "loss": 0.1026,
+      "step": 3002
+    },
+    {
+      "epoch": 1.57255107386066,
+      "loss_partial": 0.052978515625,
+      "step": 3002
+    },
+    {
+      "epoch": 1.57255107386066,
+      "loss_partial": 0.048095703125,
+      "step": 3002
+    },
+    {
+      "epoch": 1.5730749083289681,
+      "grad_norm": 1.9208358526229858,
+      "learning_rate": 1.5207043692180863e-05,
+      "loss": 0.1074,
+      "step": 3003
+    },
+    {
+      "epoch": 1.5730749083289681,
+      "loss_partial": 0.0556640625,
+      "step": 3003
+    },
+    {
+      "epoch": 1.5730749083289681,
+      "loss_partial": 0.053955078125,
+      "step": 3003
+    },
+    {
+      "epoch": 1.5735987427972762,
+      "grad_norm": 2.2389168739318848,
+      "learning_rate": 1.5203140217373862e-05,
+      "loss": 0.11,
+      "step": 3004
+    },
+    {
+      "epoch": 1.5735987427972762,
+      "loss_partial": 0.047607421875,
+      "step": 3004
+    },
+    {
+      "epoch": 1.5735987427972762,
+      "loss_partial": 0.044189453125,
+      "step": 3004
+    },
+    {
+      "epoch": 1.574122577265584,
+      "grad_norm": 2.428636074066162,
+      "learning_rate": 1.5199235655142735e-05,
+      "loss": 0.1067,
+      "step": 3005
+    },
+    {
+      "epoch": 1.574122577265584,
+      "loss_partial": 0.051513671875,
+      "step": 3005
+    },
+    {
+      "epoch": 1.574122577265584,
+      "loss_partial": 0.056640625,
+      "step": 3005
+    },
+    {
+      "epoch": 1.574646411733892,
+      "grad_norm": 2.492175579071045,
+      "learning_rate": 1.5195330006303504e-05,
+      "loss": 0.1082,
+      "step": 3006
+    },
+    {
+      "epoch": 1.574646411733892,
+      "loss_partial": 0.06103515625,
+      "step": 3006
+    },
+    {
+      "epoch": 1.574646411733892,
+      "loss_partial": 0.0693359375,
+      "step": 3006
+    },
+    {
+      "epoch": 1.5751702462022001,
+      "grad_norm": 2.317754030227661,
+      "learning_rate": 1.5191423271672433e-05,
+      "loss": 0.1071,
+      "step": 3007
+    },
+    {
+      "epoch": 1.5751702462022001,
+      "loss_partial": 0.04931640625,
+      "step": 3007
+    },
+    {
+      "epoch": 1.5751702462022001,
+      "loss_partial": 0.0458984375,
+      "step": 3007
+    },
+    {
+      "epoch": 1.575694080670508,
+      "grad_norm": 2.4899628162384033,
+      "learning_rate": 1.5187515452065998e-05,
+      "loss": 0.1035,
+      "step": 3008
+    },
+    {
+      "epoch": 1.575694080670508,
+      "loss_partial": 0.05322265625,
+      "step": 3008
+    },
+    {
+      "epoch": 1.575694080670508,
+      "loss_partial": 0.050048828125,
+      "step": 3008
+    },
+    {
+      "epoch": 1.576217915138816,
+      "grad_norm": 1.5168063640594482,
+      "learning_rate": 1.5183606548300911e-05,
+      "loss": 0.1041,
+      "step": 3009
+    },
+    {
+      "epoch": 1.576217915138816,
+      "loss_partial": 0.043212890625,
+      "step": 3009
+    },
+    {
+      "epoch": 1.576217915138816,
+      "loss_partial": 0.057373046875,
+      "step": 3009
+    },
+    {
+      "epoch": 1.576741749607124,
+      "grad_norm": 2.598140001296997,
+      "learning_rate": 1.517969656119411e-05,
+      "loss": 0.1018,
+      "step": 3010
+    },
+    {
+      "epoch": 1.576741749607124,
+      "loss_partial": 0.053466796875,
+      "step": 3010
+    },
+    {
+      "epoch": 1.576741749607124,
+      "loss_partial": 0.050537109375,
+      "step": 3010
+    },
+    {
+      "epoch": 1.5772655840754322,
+      "grad_norm": 1.6365808248519897,
+      "learning_rate": 1.517578549156276e-05,
+      "loss": 0.1042,
+      "step": 3011
+    },
+    {
+      "epoch": 1.5772655840754322,
+      "loss_partial": 0.057861328125,
+      "step": 3011
+    },
+    {
+      "epoch": 1.5772655840754322,
+      "loss_partial": 0.0458984375,
+      "step": 3011
+    },
+    {
+      "epoch": 1.5777894185437402,
+      "grad_norm": 2.2033467292785645,
+      "learning_rate": 1.5171873340224248e-05,
+      "loss": 0.1046,
+      "step": 3012
+    },
+    {
+      "epoch": 1.5777894185437402,
+      "loss_partial": 0.054931640625,
+      "step": 3012
+    },
+    {
+      "epoch": 1.5777894185437402,
+      "loss_partial": 0.048828125,
+      "step": 3012
+    },
+    {
+      "epoch": 1.5783132530120483,
+      "grad_norm": 1.6499271392822266,
+      "learning_rate": 1.516796010799619e-05,
+      "loss": 0.1009,
+      "step": 3013
+    },
+    {
+      "epoch": 1.5783132530120483,
+      "loss_partial": 0.049072265625,
+      "step": 3013
+    },
+    {
+      "epoch": 1.5783132530120483,
+      "loss_partial": 0.0556640625,
+      "step": 3013
+    },
+    {
+      "epoch": 1.5788370874803563,
+      "grad_norm": 1.9913208484649658,
+      "learning_rate": 1.5164045795696427e-05,
+      "loss": 0.1073,
+      "step": 3014
+    },
+    {
+      "epoch": 1.5788370874803563,
+      "loss_partial": 0.04443359375,
+      "step": 3014
+    },
+    {
+      "epoch": 1.5788370874803563,
+      "loss_partial": 0.053955078125,
+      "step": 3014
+    },
+    {
+      "epoch": 1.5793609219486642,
+      "grad_norm": 1.5935945510864258,
+      "learning_rate": 1.5160130404143027e-05,
+      "loss": 0.1028,
+      "step": 3015
+    },
+    {
+      "epoch": 1.5793609219486642,
+      "loss_partial": 0.056884765625,
+      "step": 3015
+    },
+    {
+      "epoch": 1.5793609219486642,
+      "loss_partial": 0.0458984375,
+      "step": 3015
+    },
+    {
+      "epoch": 1.5798847564169722,
+      "grad_norm": 1.6876134872436523,
+      "learning_rate": 1.5156213934154278e-05,
+      "loss": 0.1037,
+      "step": 3016
+    },
+    {
+      "epoch": 1.5798847564169722,
+      "loss_partial": 0.046875,
+      "step": 3016
+    },
+    {
+      "epoch": 1.5798847564169722,
+      "loss_partial": 0.044189453125,
+      "step": 3016
+    },
+    {
+      "epoch": 1.5804085908852803,
+      "grad_norm": 1.5896507501602173,
+      "learning_rate": 1.5152296386548706e-05,
+      "loss": 0.1027,
+      "step": 3017
+    },
+    {
+      "epoch": 1.5804085908852803,
+      "loss_partial": 0.05224609375,
+      "step": 3017
+    },
+    {
+      "epoch": 1.5804085908852803,
+      "loss_partial": 0.058837890625,
+      "step": 3017
+    },
+    {
+      "epoch": 1.5809324253535881,
+      "grad_norm": 1.880576252937317,
+      "learning_rate": 1.5148377762145048e-05,
+      "loss": 0.107,
+      "step": 3018
+    },
+    {
+      "epoch": 1.5809324253535881,
+      "loss_partial": 0.043701171875,
+      "step": 3018
+    },
+    {
+      "epoch": 1.5809324253535881,
+      "loss_partial": 0.05029296875,
+      "step": 3018
+    },
+    {
+      "epoch": 1.5814562598218962,
+      "grad_norm": 1.714317798614502,
+      "learning_rate": 1.5144458061762275e-05,
+      "loss": 0.1046,
+      "step": 3019
+    },
+    {
+      "epoch": 1.5814562598218962,
+      "loss_partial": 0.052001953125,
+      "step": 3019
+    },
+    {
+      "epoch": 1.5814562598218962,
+      "loss_partial": 0.0517578125,
+      "step": 3019
+    },
+    {
+      "epoch": 1.5819800942902043,
+      "grad_norm": 2.2115814685821533,
+      "learning_rate": 1.5140537286219582e-05,
+      "loss": 0.1056,
+      "step": 3020
+    },
+    {
+      "epoch": 1.5819800942902043,
+      "loss_partial": 0.043212890625,
+      "step": 3020
+    },
+    {
+      "epoch": 1.5819800942902043,
+      "loss_partial": 0.048583984375,
+      "step": 3020
+    },
+    {
+      "epoch": 1.5825039287585123,
+      "grad_norm": 1.844746470451355,
+      "learning_rate": 1.513661543633638e-05,
+      "loss": 0.1057,
+      "step": 3021
+    },
+    {
+      "epoch": 1.5825039287585123,
+      "loss_partial": 0.044189453125,
+      "step": 3021
+    },
+    {
+      "epoch": 1.5825039287585123,
+      "loss_partial": 0.04736328125,
+      "step": 3021
+    },
+    {
+      "epoch": 1.5830277632268204,
+      "grad_norm": 1.7749125957489014,
+      "learning_rate": 1.5132692512932314e-05,
+      "loss": 0.106,
+      "step": 3022
+    },
+    {
+      "epoch": 1.5830277632268204,
+      "loss_partial": 0.052490234375,
+      "step": 3022
+    },
+    {
+      "epoch": 1.5830277632268204,
+      "loss_partial": 0.05908203125,
+      "step": 3022
+    },
+    {
+      "epoch": 1.5835515976951284,
+      "grad_norm": 1.850675106048584,
+      "learning_rate": 1.5128768516827257e-05,
+      "loss": 0.1068,
+      "step": 3023
+    },
+    {
+      "epoch": 1.5835515976951284,
+      "loss_partial": 0.056884765625,
+      "step": 3023
+    },
+    {
+      "epoch": 1.5835515976951284,
+      "loss_partial": 0.0517578125,
+      "step": 3023
+    },
+    {
+      "epoch": 1.5840754321634365,
+      "grad_norm": 1.6982181072235107,
+      "learning_rate": 1.5124843448841296e-05,
+      "loss": 0.1037,
+      "step": 3024
+    },
+    {
+      "epoch": 1.5840754321634365,
+      "loss_partial": 0.05078125,
+      "step": 3024
+    },
+    {
+      "epoch": 1.5840754321634365,
+      "loss_partial": 0.05322265625,
+      "step": 3024
+    },
+    {
+      "epoch": 1.5845992666317443,
+      "grad_norm": 1.3446094989776611,
+      "learning_rate": 1.5120917309794743e-05,
+      "loss": 0.1051,
+      "step": 3025
+    },
+    {
+      "epoch": 1.5845992666317443,
+      "loss_partial": 0.053955078125,
+      "step": 3025
+    },
+    {
+      "epoch": 1.5845992666317443,
+      "loss_partial": 0.04541015625,
+      "step": 3025
+    },
+    {
+      "epoch": 1.5851231011000524,
+      "grad_norm": 1.5299328565597534,
+      "learning_rate": 1.5116990100508137e-05,
+      "loss": 0.1029,
+      "step": 3026
+    },
+    {
+      "epoch": 1.5851231011000524,
+      "loss_partial": 0.05419921875,
+      "step": 3026
+    },
+    {
+      "epoch": 1.5851231011000524,
+      "loss_partial": 0.060791015625,
+      "step": 3026
+    },
+    {
+      "epoch": 1.5856469355683604,
+      "grad_norm": 1.4539567232131958,
+      "learning_rate": 1.5113061821802246e-05,
+      "loss": 0.105,
+      "step": 3027
+    },
+    {
+      "epoch": 1.5856469355683604,
+      "loss_partial": 0.046630859375,
+      "step": 3027
+    },
+    {
+      "epoch": 1.5856469355683604,
+      "loss_partial": 0.05419921875,
+      "step": 3027
+    },
+    {
+      "epoch": 1.5861707700366683,
+      "grad_norm": 1.5370943546295166,
+      "learning_rate": 1.5109132474498052e-05,
+      "loss": 0.1083,
+      "step": 3028
+    },
+    {
+      "epoch": 1.5861707700366683,
+      "loss_partial": 0.04931640625,
+      "step": 3028
+    },
+    {
+      "epoch": 1.5861707700366683,
+      "loss_partial": 0.05078125,
+      "step": 3028
+    },
+    {
+      "epoch": 1.5866946045049763,
+      "grad_norm": 1.7007743120193481,
+      "learning_rate": 1.5105202059416764e-05,
+      "loss": 0.1059,
+      "step": 3029
+    },
+    {
+      "epoch": 1.5866946045049763,
+      "loss_partial": 0.05078125,
+      "step": 3029
+    },
+    {
+      "epoch": 1.5866946045049763,
+      "loss_partial": 0.050048828125,
+      "step": 3029
+    },
+    {
+      "epoch": 1.5872184389732844,
+      "grad_norm": 1.5404162406921387,
+      "learning_rate": 1.5101270577379814e-05,
+      "loss": 0.106,
+      "step": 3030
+    },
+    {
+      "epoch": 1.5872184389732844,
+      "loss_partial": 0.052001953125,
+      "step": 3030
+    },
+    {
+      "epoch": 1.5872184389732844,
+      "loss_partial": 0.05224609375,
+      "step": 3030
+    },
+    {
+      "epoch": 1.5877422734415925,
+      "grad_norm": 1.5323355197906494,
+      "learning_rate": 1.5097338029208861e-05,
+      "loss": 0.1058,
+      "step": 3031
+    },
+    {
+      "epoch": 1.5877422734415925,
+      "loss_partial": 0.05126953125,
+      "step": 3031
+    },
+    {
+      "epoch": 1.5877422734415925,
+      "loss_partial": 0.04345703125,
+      "step": 3031
+    },
+    {
+      "epoch": 1.5882661079099005,
+      "grad_norm": 1.7917555570602417,
+      "learning_rate": 1.509340441572578e-05,
+      "loss": 0.1063,
+      "step": 3032
+    },
+    {
+      "epoch": 1.5882661079099005,
+      "loss_partial": 0.0498046875,
+      "step": 3032
+    },
+    {
+      "epoch": 1.5882661079099005,
+      "loss_partial": 0.048583984375,
+      "step": 3032
+    },
+    {
+      "epoch": 1.5887899423782086,
+      "grad_norm": 2.5598161220550537,
+      "learning_rate": 1.5089469737752673e-05,
+      "loss": 0.111,
+      "step": 3033
+    },
+    {
+      "epoch": 1.5887899423782086,
+      "loss_partial": 0.044189453125,
+      "step": 3033
+    },
+    {
+      "epoch": 1.5887899423782086,
+      "loss_partial": 0.06884765625,
+      "step": 3033
+    },
+    {
+      "epoch": 1.5893137768465166,
+      "grad_norm": 1.5529793500900269,
+      "learning_rate": 1.508553399611186e-05,
+      "loss": 0.1053,
+      "step": 3034
+    },
+    {
+      "epoch": 1.5893137768465166,
+      "loss_partial": 0.058349609375,
+      "step": 3034
+    },
+    {
+      "epoch": 1.5893137768465166,
+      "loss_partial": 0.0546875,
+      "step": 3034
+    },
+    {
+      "epoch": 1.5898376113148245,
+      "grad_norm": 2.028146982192993,
+      "learning_rate": 1.5081597191625897e-05,
+      "loss": 0.1034,
+      "step": 3035
+    },
+    {
+      "epoch": 1.5898376113148245,
+      "loss_partial": 0.041015625,
+      "step": 3035
+    },
+    {
+      "epoch": 1.5898376113148245,
+      "loss_partial": 0.055908203125,
+      "step": 3035
+    },
+    {
+      "epoch": 1.5903614457831325,
+      "grad_norm": 2.1726646423339844,
+      "learning_rate": 1.5077659325117538e-05,
+      "loss": 0.1031,
+      "step": 3036
+    },
+    {
+      "epoch": 1.5903614457831325,
+      "loss_partial": 0.047119140625,
+      "step": 3036
+    },
+    {
+      "epoch": 1.5903614457831325,
+      "loss_partial": 0.060791015625,
+      "step": 3036
+    },
+    {
+      "epoch": 1.5908852802514404,
+      "grad_norm": 1.6407470703125,
+      "learning_rate": 1.507372039740978e-05,
+      "loss": 0.1049,
+      "step": 3037
+    },
+    {
+      "epoch": 1.5908852802514404,
+      "loss_partial": 0.048583984375,
+      "step": 3037
+    },
+    {
+      "epoch": 1.5908852802514404,
+      "loss_partial": 0.054931640625,
+      "step": 3037
+    },
+    {
+      "epoch": 1.5914091147197484,
+      "grad_norm": 2.042222499847412,
+      "learning_rate": 1.5069780409325839e-05,
+      "loss": 0.1024,
+      "step": 3038
+    },
+    {
+      "epoch": 1.5914091147197484,
+      "loss_partial": 0.047119140625,
+      "step": 3038
+    },
+    {
+      "epoch": 1.5914091147197484,
+      "loss_partial": 0.0517578125,
+      "step": 3038
+    },
+    {
+      "epoch": 1.5919329491880565,
+      "grad_norm": 1.670885682106018,
+      "learning_rate": 1.5065839361689139e-05,
+      "loss": 0.1048,
+      "step": 3039
+    },
+    {
+      "epoch": 1.5919329491880565,
+      "loss_partial": 0.049072265625,
+      "step": 3039
+    },
+    {
+      "epoch": 1.5919329491880565,
+      "loss_partial": 0.05517578125,
+      "step": 3039
+    },
+    {
+      "epoch": 1.5924567836563646,
+      "grad_norm": 2.093353509902954,
+      "learning_rate": 1.5061897255323338e-05,
+      "loss": 0.1042,
+      "step": 3040
+    },
+    {
+      "epoch": 1.5924567836563646,
+      "loss_partial": 0.054443359375,
+      "step": 3040
+    },
+    {
+      "epoch": 1.5924567836563646,
+      "loss_partial": 0.05419921875,
+      "step": 3040
+    },
+    {
+      "epoch": 1.5929806181246726,
+      "grad_norm": 1.7724199295043945,
+      "learning_rate": 1.5057954091052311e-05,
+      "loss": 0.1048,
+      "step": 3041
+    },
+    {
+      "epoch": 1.5929806181246726,
+      "loss_partial": 0.04833984375,
+      "step": 3041
+    },
+    {
+      "epoch": 1.5929806181246726,
+      "loss_partial": 0.06103515625,
+      "step": 3041
+    },
+    {
+      "epoch": 1.5935044525929807,
+      "grad_norm": 2.0290367603302,
+      "learning_rate": 1.5054009869700159e-05,
+      "loss": 0.1041,
+      "step": 3042
+    },
+    {
+      "epoch": 1.5935044525929807,
+      "loss_partial": 0.044677734375,
+      "step": 3042
+    },
+    {
+      "epoch": 1.5935044525929807,
+      "loss_partial": 0.04541015625,
+      "step": 3042
+    },
+    {
+      "epoch": 1.5940282870612887,
+      "grad_norm": 1.3996270895004272,
+      "learning_rate": 1.5050064592091199e-05,
+      "loss": 0.1028,
+      "step": 3043
+    },
+    {
+      "epoch": 1.5940282870612887,
+      "loss_partial": 0.06884765625,
+      "step": 3043
+    },
+    {
+      "epoch": 1.5940282870612887,
+      "loss_partial": 0.05615234375,
+      "step": 3043
+    },
+    {
+      "epoch": 1.5945521215295968,
+      "grad_norm": 1.8780568838119507,
+      "learning_rate": 1.5046118259049967e-05,
+      "loss": 0.1029,
+      "step": 3044
+    },
+    {
+      "epoch": 1.5945521215295968,
+      "loss_partial": 0.057373046875,
+      "step": 3044
+    },
+    {
+      "epoch": 1.5945521215295968,
+      "loss_partial": 0.050537109375,
+      "step": 3044
+    },
+    {
+      "epoch": 1.5950759559979046,
+      "grad_norm": 1.6743319034576416,
+      "learning_rate": 1.504217087140122e-05,
+      "loss": 0.1041,
+      "step": 3045
+    },
+    {
+      "epoch": 1.5950759559979046,
+      "loss_partial": 0.05322265625,
+      "step": 3045
+    },
+    {
+      "epoch": 1.5950759559979046,
+      "loss_partial": 0.0517578125,
+      "step": 3045
+    },
+    {
+      "epoch": 1.5955997904662127,
+      "grad_norm": 1.9992804527282715,
+      "learning_rate": 1.5038222429969944e-05,
+      "loss": 0.1038,
+      "step": 3046
+    },
+    {
+      "epoch": 1.5955997904662127,
+      "loss_partial": 0.05126953125,
+      "step": 3046
+    },
+    {
+      "epoch": 1.5955997904662127,
+      "loss_partial": 0.055908203125,
+      "step": 3046
+    },
+    {
+      "epoch": 1.5961236249345205,
+      "grad_norm": 1.6696581840515137,
+      "learning_rate": 1.5034272935581334e-05,
+      "loss": 0.1028,
+      "step": 3047
+    },
+    {
+      "epoch": 1.5961236249345205,
+      "loss_partial": 0.049560546875,
+      "step": 3047
+    },
+    {
+      "epoch": 1.5961236249345205,
+      "loss_partial": 0.04296875,
+      "step": 3047
+    },
+    {
+      "epoch": 1.5966474594028286,
+      "grad_norm": 2.1744539737701416,
+      "learning_rate": 1.5030322389060812e-05,
+      "loss": 0.1052,
+      "step": 3048
+    },
+    {
+      "epoch": 1.5966474594028286,
+      "loss_partial": 0.054931640625,
+      "step": 3048
+    },
+    {
+      "epoch": 1.5966474594028286,
+      "loss_partial": 0.052490234375,
+      "step": 3048
+    },
+    {
+      "epoch": 1.5971712938711367,
+      "grad_norm": 2.5373482704162598,
+      "learning_rate": 1.5026370791234017e-05,
+      "loss": 0.106,
+      "step": 3049
+    },
+    {
+      "epoch": 1.5971712938711367,
+      "loss_partial": 0.064453125,
+      "step": 3049
+    },
+    {
+      "epoch": 1.5971712938711367,
+      "loss_partial": 0.04443359375,
+      "step": 3049
+    },
+    {
+      "epoch": 1.5976951283394447,
+      "grad_norm": 2.3389892578125,
+      "learning_rate": 1.5022418142926811e-05,
+      "loss": 0.1075,
+      "step": 3050
+    },
+    {
+      "epoch": 1.5976951283394447,
+      "loss_partial": 0.04931640625,
+      "step": 3050
+    },
+    {
+      "epoch": 1.5976951283394447,
+      "loss_partial": 0.048828125,
+      "step": 3050
+    },
+    {
+      "epoch": 1.5982189628077528,
+      "grad_norm": 2.40480375289917,
+      "learning_rate": 1.5018464444965269e-05,
+      "loss": 0.1063,
+      "step": 3051
+    },
+    {
+      "epoch": 1.5982189628077528,
+      "loss_partial": 0.048095703125,
+      "step": 3051
+    },
+    {
+      "epoch": 1.5982189628077528,
+      "loss_partial": 0.055419921875,
+      "step": 3051
+    },
+    {
+      "epoch": 1.5987427972760608,
+      "grad_norm": 2.1913321018218994,
+      "learning_rate": 1.501450969817569e-05,
+      "loss": 0.1052,
+      "step": 3052
+    },
+    {
+      "epoch": 1.5987427972760608,
+      "loss_partial": 0.048828125,
+      "step": 3052
+    },
+    {
+      "epoch": 1.5987427972760608,
+      "loss_partial": 0.044921875,
+      "step": 3052
+    },
+    {
+      "epoch": 1.599266631744369,
+      "grad_norm": 1.9409773349761963,
+      "learning_rate": 1.5010553903384595e-05,
+      "loss": 0.1096,
+      "step": 3053
+    },
+    {
+      "epoch": 1.599266631744369,
+      "loss_partial": 0.052490234375,
+      "step": 3053
+    },
+    {
+      "epoch": 1.599266631744369,
+      "loss_partial": 0.047607421875,
+      "step": 3053
+    },
+    {
+      "epoch": 1.599790466212677,
+      "grad_norm": 1.4812493324279785,
+      "learning_rate": 1.5006597061418714e-05,
+      "loss": 0.1031,
+      "step": 3054
+    },
+    {
+      "epoch": 1.599790466212677,
+      "loss_partial": 0.056396484375,
+      "step": 3054
+    },
+    {
+      "epoch": 1.599790466212677,
+      "loss_partial": 0.048828125,
+      "step": 3054
+    },
+    {
+      "epoch": 1.6003143006809848,
+      "grad_norm": 1.7859792709350586,
+      "learning_rate": 1.5002639173105014e-05,
+      "loss": 0.1061,
+      "step": 3055
+    },
+    {
+      "epoch": 1.6003143006809848,
+      "loss_partial": 0.06591796875,
+      "step": 3055
+    },
+    {
+      "epoch": 1.6003143006809848,
+      "loss_partial": 0.05517578125,
+      "step": 3055
+    },
+    {
+      "epoch": 1.6008381351492929,
+      "grad_norm": 1.5376718044281006,
+      "learning_rate": 1.4998680239270658e-05,
+      "loss": 0.1081,
+      "step": 3056
+    },
+    {
+      "epoch": 1.6008381351492929,
+      "loss_partial": 0.048583984375,
+      "step": 3056
+    },
+    {
+      "epoch": 1.6008381351492929,
+      "loss_partial": 0.056396484375,
+      "step": 3056
+    },
+    {
+      "epoch": 1.6013619696176007,
+      "grad_norm": 1.7157949209213257,
+      "learning_rate": 1.4994720260743045e-05,
+      "loss": 0.1051,
+      "step": 3057
+    },
+    {
+      "epoch": 1.6013619696176007,
+      "loss_partial": 0.052978515625,
+      "step": 3057
+    },
+    {
+      "epoch": 1.6013619696176007,
+      "loss_partial": 0.045166015625,
+      "step": 3057
+    },
+    {
+      "epoch": 1.6018858040859087,
+      "grad_norm": 2.097249746322632,
+      "learning_rate": 1.4990759238349784e-05,
+      "loss": 0.1076,
+      "step": 3058
+    },
+    {
+      "epoch": 1.6018858040859087,
+      "loss_partial": 0.051025390625,
+      "step": 3058
+    },
+    {
+      "epoch": 1.6018858040859087,
+      "loss_partial": 0.05859375,
+      "step": 3058
+    },
+    {
+      "epoch": 1.6024096385542168,
+      "grad_norm": 1.4737701416015625,
+      "learning_rate": 1.4986797172918698e-05,
+      "loss": 0.1071,
+      "step": 3059
+    },
+    {
+      "epoch": 1.6024096385542168,
+      "loss_partial": 0.0458984375,
+      "step": 3059
+    },
+    {
+      "epoch": 1.6024096385542168,
+      "loss_partial": 0.046630859375,
+      "step": 3059
+    },
+    {
+      "epoch": 1.6029334730225249,
+      "grad_norm": 1.9828764200210571,
+      "learning_rate": 1.4982834065277847e-05,
+      "loss": 0.1054,
+      "step": 3060
+    },
+    {
+      "epoch": 1.6029334730225249,
+      "loss_partial": 0.062255859375,
+      "step": 3060
+    },
+    {
+      "epoch": 1.6029334730225249,
+      "loss_partial": 0.04345703125,
+      "step": 3060
+    },
+    {
+      "epoch": 1.603457307490833,
+      "grad_norm": 1.8693126440048218,
+      "learning_rate": 1.4978869916255487e-05,
+      "loss": 0.1075,
+      "step": 3061
+    },
+    {
+      "epoch": 1.603457307490833,
+      "loss_partial": 0.05810546875,
+      "step": 3061
+    },
+    {
+      "epoch": 1.603457307490833,
+      "loss_partial": 0.0615234375,
+      "step": 3061
+    },
+    {
+      "epoch": 1.603981141959141,
+      "grad_norm": 1.813262939453125,
+      "learning_rate": 1.4974904726680101e-05,
+      "loss": 0.1048,
+      "step": 3062
+    },
+    {
+      "epoch": 1.603981141959141,
+      "loss_partial": 0.0517578125,
+      "step": 3062
+    },
+    {
+      "epoch": 1.603981141959141,
+      "loss_partial": 0.04443359375,
+      "step": 3062
+    },
+    {
+      "epoch": 1.604504976427449,
+      "grad_norm": 1.8483166694641113,
+      "learning_rate": 1.4970938497380396e-05,
+      "loss": 0.1061,
+      "step": 3063
+    },
+    {
+      "epoch": 1.604504976427449,
+      "loss_partial": 0.04638671875,
+      "step": 3063
+    },
+    {
+      "epoch": 1.604504976427449,
+      "loss_partial": 0.051025390625,
+      "step": 3063
+    },
+    {
+      "epoch": 1.6050288108957569,
+      "grad_norm": 2.2348666191101074,
+      "learning_rate": 1.496697122918528e-05,
+      "loss": 0.1022,
+      "step": 3064
+    },
+    {
+      "epoch": 1.6050288108957569,
+      "loss_partial": 0.053466796875,
+      "step": 3064
+    },
+    {
+      "epoch": 1.6050288108957569,
+      "loss_partial": 0.05810546875,
+      "step": 3064
+    },
+    {
+      "epoch": 1.605552645364065,
+      "grad_norm": 1.7202099561691284,
+      "learning_rate": 1.4963002922923895e-05,
+      "loss": 0.1084,
+      "step": 3065
+    },
+    {
+      "epoch": 1.605552645364065,
+      "loss_partial": 0.0634765625,
+      "step": 3065
+    },
+    {
+      "epoch": 1.605552645364065,
+      "loss_partial": 0.06298828125,
+      "step": 3065
+    },
+    {
+      "epoch": 1.606076479832373,
+      "grad_norm": 1.740931510925293,
+      "learning_rate": 1.4959033579425589e-05,
+      "loss": 0.1068,
+      "step": 3066
+    },
+    {
+      "epoch": 1.606076479832373,
+      "loss_partial": 0.052978515625,
+      "step": 3066
+    },
+    {
+      "epoch": 1.606076479832373,
+      "loss_partial": 0.060546875,
+      "step": 3066
+    },
+    {
+      "epoch": 1.6066003143006808,
+      "grad_norm": 2.342031478881836,
+      "learning_rate": 1.4955063199519928e-05,
+      "loss": 0.1042,
+      "step": 3067
+    },
+    {
+      "epoch": 1.6066003143006808,
+      "loss_partial": 0.056884765625,
+      "step": 3067
+    },
+    {
+      "epoch": 1.6066003143006808,
+      "loss_partial": 0.0517578125,
+      "step": 3067
+    },
+    {
+      "epoch": 1.607124148768989,
+      "grad_norm": 2.178830862045288,
+      "learning_rate": 1.49510917840367e-05,
+      "loss": 0.1007,
+      "step": 3068
+    },
+    {
+      "epoch": 1.607124148768989,
+      "loss_partial": 0.052734375,
+      "step": 3068
+    },
+    {
+      "epoch": 1.607124148768989,
+      "loss_partial": 0.05078125,
+      "step": 3068
+    },
+    {
+      "epoch": 1.607647983237297,
+      "grad_norm": 1.8553320169448853,
+      "learning_rate": 1.4947119333805908e-05,
+      "loss": 0.1034,
+      "step": 3069
+    },
+    {
+      "epoch": 1.607647983237297,
+      "loss_partial": 0.0498046875,
+      "step": 3069
+    },
+    {
+      "epoch": 1.607647983237297,
+      "loss_partial": 0.055908203125,
+      "step": 3069
+    },
+    {
+      "epoch": 1.608171817705605,
+      "grad_norm": 2.3696563243865967,
+      "learning_rate": 1.4943145849657768e-05,
+      "loss": 0.1049,
+      "step": 3070
+    },
+    {
+      "epoch": 1.608171817705605,
+      "loss_partial": 0.04541015625,
+      "step": 3070
+    },
+    {
+      "epoch": 1.608171817705605,
+      "loss_partial": 0.046630859375,
+      "step": 3070
+    },
+    {
+      "epoch": 1.608695652173913,
+      "grad_norm": 1.6174510717391968,
+      "learning_rate": 1.4939171332422708e-05,
+      "loss": 0.1011,
+      "step": 3071
+    },
+    {
+      "epoch": 1.608695652173913,
+      "loss_partial": 0.05078125,
+      "step": 3071
+    },
+    {
+      "epoch": 1.608695652173913,
+      "loss_partial": 0.04296875,
+      "step": 3071
+    },
+    {
+      "epoch": 1.6092194866422211,
+      "grad_norm": 1.4384078979492188,
+      "learning_rate": 1.4935195782931386e-05,
+      "loss": 0.1028,
+      "step": 3072
+    },
+    {
+      "epoch": 1.6092194866422211,
+      "loss_partial": 0.04248046875,
+      "step": 3072
+    },
+    {
+      "epoch": 1.6092194866422211,
+      "loss_partial": 0.053466796875,
+      "step": 3072
+    },
+    {
+      "epoch": 1.6097433211105292,
+      "grad_norm": 1.7279943227767944,
+      "learning_rate": 1.4931219202014662e-05,
+      "loss": 0.1044,
+      "step": 3073
+    },
+    {
+      "epoch": 1.6097433211105292,
+      "loss_partial": 0.0517578125,
+      "step": 3073
+    },
+    {
+      "epoch": 1.6097433211105292,
+      "loss_partial": 0.051025390625,
+      "step": 3073
+    },
+    {
+      "epoch": 1.610267155578837,
+      "grad_norm": 1.7295862436294556,
+      "learning_rate": 1.492724159050362e-05,
+      "loss": 0.1071,
+      "step": 3074
+    },
+    {
+      "epoch": 1.610267155578837,
+      "loss_partial": 0.049072265625,
+      "step": 3074
+    },
+    {
+      "epoch": 1.610267155578837,
+      "loss_partial": 0.06103515625,
+      "step": 3074
+    },
+    {
+      "epoch": 1.610790990047145,
+      "grad_norm": 1.9064863920211792,
+      "learning_rate": 1.4923262949229555e-05,
+      "loss": 0.1064,
+      "step": 3075
+    },
+    {
+      "epoch": 1.610790990047145,
+      "loss_partial": 0.049072265625,
+      "step": 3075
+    },
+    {
+      "epoch": 1.610790990047145,
+      "loss_partial": 0.0546875,
+      "step": 3075
+    },
+    {
+      "epoch": 1.6113148245154532,
+      "grad_norm": 1.6361057758331299,
+      "learning_rate": 1.4919283279023979e-05,
+      "loss": 0.1063,
+      "step": 3076
+    },
+    {
+      "epoch": 1.6113148245154532,
+      "loss_partial": 0.05224609375,
+      "step": 3076
+    },
+    {
+      "epoch": 1.6113148245154532,
+      "loss_partial": 0.052001953125,
+      "step": 3076
+    },
+    {
+      "epoch": 1.611838658983761,
+      "grad_norm": 1.5747592449188232,
+      "learning_rate": 1.4915302580718614e-05,
+      "loss": 0.104,
+      "step": 3077
+    },
+    {
+      "epoch": 1.611838658983761,
+      "loss_partial": 0.046142578125,
+      "step": 3077
+    },
+    {
+      "epoch": 1.611838658983761,
+      "loss_partial": 0.054443359375,
+      "step": 3077
+    },
+    {
+      "epoch": 1.612362493452069,
+      "grad_norm": 1.7253270149230957,
+      "learning_rate": 1.4911320855145407e-05,
+      "loss": 0.1068,
+      "step": 3078
+    },
+    {
+      "epoch": 1.612362493452069,
+      "loss_partial": 0.050537109375,
+      "step": 3078
+    },
+    {
+      "epoch": 1.612362493452069,
+      "loss_partial": 0.0537109375,
+      "step": 3078
+    },
+    {
+      "epoch": 1.6128863279203771,
+      "grad_norm": 1.315211296081543,
+      "learning_rate": 1.4907338103136513e-05,
+      "loss": 0.1067,
+      "step": 3079
+    },
+    {
+      "epoch": 1.6128863279203771,
+      "loss_partial": 0.057373046875,
+      "step": 3079
+    },
+    {
+      "epoch": 1.6128863279203771,
+      "loss_partial": 0.0478515625,
+      "step": 3079
+    },
+    {
+      "epoch": 1.6134101623886852,
+      "grad_norm": 2.1868724822998047,
+      "learning_rate": 1.49033543255243e-05,
+      "loss": 0.1054,
+      "step": 3080
+    },
+    {
+      "epoch": 1.6134101623886852,
+      "loss_partial": 0.051025390625,
+      "step": 3080
+    },
+    {
+      "epoch": 1.6134101623886852,
+      "loss_partial": 0.05810546875,
+      "step": 3080
+    },
+    {
+      "epoch": 1.6139339968569932,
+      "grad_norm": 1.62059485912323,
+      "learning_rate": 1.4899369523141357e-05,
+      "loss": 0.1034,
+      "step": 3081
+    },
+    {
+      "epoch": 1.6139339968569932,
+      "loss_partial": 0.045654296875,
+      "step": 3081
+    },
+    {
+      "epoch": 1.6139339968569932,
+      "loss_partial": 0.047607421875,
+      "step": 3081
+    },
+    {
+      "epoch": 1.6144578313253013,
+      "grad_norm": 1.12147855758667,
+      "learning_rate": 1.4895383696820478e-05,
+      "loss": 0.1015,
+      "step": 3082
+    },
+    {
+      "epoch": 1.6144578313253013,
+      "loss_partial": 0.05615234375,
+      "step": 3082
+    },
+    {
+      "epoch": 1.6144578313253013,
+      "loss_partial": 0.0498046875,
+      "step": 3082
+    },
+    {
+      "epoch": 1.6149816657936094,
+      "grad_norm": 1.9661871194839478,
+      "learning_rate": 1.4891396847394676e-05,
+      "loss": 0.1042,
+      "step": 3083
+    },
+    {
+      "epoch": 1.6149816657936094,
+      "loss_partial": 0.05859375,
+      "step": 3083
+    },
+    {
+      "epoch": 1.6149816657936094,
+      "loss_partial": 0.06298828125,
+      "step": 3083
+    },
+    {
+      "epoch": 1.6155055002619172,
+      "grad_norm": 2.516359329223633,
+      "learning_rate": 1.4887408975697187e-05,
+      "loss": 0.1056,
+      "step": 3084
+    },
+    {
+      "epoch": 1.6155055002619172,
+      "loss_partial": 0.058837890625,
+      "step": 3084
+    },
+    {
+      "epoch": 1.6155055002619172,
+      "loss_partial": 0.04833984375,
+      "step": 3084
+    },
+    {
+      "epoch": 1.6160293347302253,
+      "grad_norm": 2.029656410217285,
+      "learning_rate": 1.488342008256144e-05,
+      "loss": 0.1058,
+      "step": 3085
+    },
+    {
+      "epoch": 1.6160293347302253,
+      "loss_partial": 0.050048828125,
+      "step": 3085
+    },
+    {
+      "epoch": 1.6160293347302253,
+      "loss_partial": 0.054931640625,
+      "step": 3085
+    },
+    {
+      "epoch": 1.6165531691985333,
+      "grad_norm": 1.6344599723815918,
+      "learning_rate": 1.4879430168821094e-05,
+      "loss": 0.1055,
+      "step": 3086
+    },
+    {
+      "epoch": 1.6165531691985333,
+      "loss_partial": 0.040283203125,
+      "step": 3086
+    },
+    {
+      "epoch": 1.6165531691985333,
+      "loss_partial": 0.04931640625,
+      "step": 3086
+    },
+    {
+      "epoch": 1.6170770036668412,
+      "grad_norm": 2.194282293319702,
+      "learning_rate": 1.4875439235310017e-05,
+      "loss": 0.1047,
+      "step": 3087
+    },
+    {
+      "epoch": 1.6170770036668412,
+      "loss_partial": 0.04150390625,
+      "step": 3087
+    },
+    {
+      "epoch": 1.6170770036668412,
+      "loss_partial": 0.05322265625,
+      "step": 3087
+    },
+    {
+      "epoch": 1.6176008381351492,
+      "grad_norm": 1.532181739807129,
+      "learning_rate": 1.487144728286229e-05,
+      "loss": 0.1022,
+      "step": 3088
+    },
+    {
+      "epoch": 1.6176008381351492,
+      "loss_partial": 0.06005859375,
+      "step": 3088
+    },
+    {
+      "epoch": 1.6176008381351492,
+      "loss_partial": 0.050048828125,
+      "step": 3088
+    },
+    {
+      "epoch": 1.6181246726034573,
+      "grad_norm": 1.6402721405029297,
+      "learning_rate": 1.4867454312312203e-05,
+      "loss": 0.1072,
+      "step": 3089
+    },
+    {
+      "epoch": 1.6181246726034573,
+      "loss_partial": 0.0439453125,
+      "step": 3089
+    },
+    {
+      "epoch": 1.6181246726034573,
+      "loss_partial": 0.05029296875,
+      "step": 3089
+    },
+    {
+      "epoch": 1.6186485070717653,
+      "grad_norm": 1.9804131984710693,
+      "learning_rate": 1.4863460324494265e-05,
+      "loss": 0.1048,
+      "step": 3090
+    },
+    {
+      "epoch": 1.6186485070717653,
+      "loss_partial": 0.056640625,
+      "step": 3090
+    },
+    {
+      "epoch": 1.6186485070717653,
+      "loss_partial": 0.053955078125,
+      "step": 3090
+    },
+    {
+      "epoch": 1.6191723415400734,
+      "grad_norm": 2.146890640258789,
+      "learning_rate": 1.4859465320243194e-05,
+      "loss": 0.1031,
+      "step": 3091
+    },
+    {
+      "epoch": 1.6191723415400734,
+      "loss_partial": 0.05517578125,
+      "step": 3091
+    },
+    {
+      "epoch": 1.6191723415400734,
+      "loss_partial": 0.051025390625,
+      "step": 3091
+    },
+    {
+      "epoch": 1.6196961760083814,
+      "grad_norm": 2.203068494796753,
+      "learning_rate": 1.4855469300393921e-05,
+      "loss": 0.1059,
+      "step": 3092
+    },
+    {
+      "epoch": 1.6196961760083814,
+      "loss_partial": 0.05078125,
+      "step": 3092
+    },
+    {
+      "epoch": 1.6196961760083814,
+      "loss_partial": 0.05126953125,
+      "step": 3092
+    },
+    {
+      "epoch": 1.6202200104766895,
+      "grad_norm": 1.9852708578109741,
+      "learning_rate": 1.485147226578159e-05,
+      "loss": 0.1042,
+      "step": 3093
+    },
+    {
+      "epoch": 1.6202200104766895,
+      "loss_partial": 0.052734375,
+      "step": 3093
+    },
+    {
+      "epoch": 1.6202200104766895,
+      "loss_partial": 0.05517578125,
+      "step": 3093
+    },
+    {
+      "epoch": 1.6207438449449973,
+      "grad_norm": 2.306446075439453,
+      "learning_rate": 1.4847474217241557e-05,
+      "loss": 0.1031,
+      "step": 3094
+    },
+    {
+      "epoch": 1.6207438449449973,
+      "loss_partial": 0.050048828125,
+      "step": 3094
+    },
+    {
+      "epoch": 1.6207438449449973,
+      "loss_partial": 0.049072265625,
+      "step": 3094
+    },
+    {
+      "epoch": 1.6212676794133054,
+      "grad_norm": 1.603951334953308,
+      "learning_rate": 1.4843475155609386e-05,
+      "loss": 0.1046,
+      "step": 3095
+    },
+    {
+      "epoch": 1.6212676794133054,
+      "loss_partial": 0.052978515625,
+      "step": 3095
+    },
+    {
+      "epoch": 1.6212676794133054,
+      "loss_partial": 0.055419921875,
+      "step": 3095
+    },
+    {
+      "epoch": 1.6217915138816132,
+      "grad_norm": 2.422511577606201,
+      "learning_rate": 1.483947508172086e-05,
+      "loss": 0.1036,
+      "step": 3096
+    },
+    {
+      "epoch": 1.6217915138816132,
+      "loss_partial": 0.05126953125,
+      "step": 3096
+    },
+    {
+      "epoch": 1.6217915138816132,
+      "loss_partial": 0.05078125,
+      "step": 3096
+    },
+    {
+      "epoch": 1.6223153483499213,
+      "grad_norm": 2.1307082176208496,
+      "learning_rate": 1.483547399641197e-05,
+      "loss": 0.1047,
+      "step": 3097
+    },
+    {
+      "epoch": 1.6223153483499213,
+      "loss_partial": 0.0576171875,
+      "step": 3097
+    },
+    {
+      "epoch": 1.6223153483499213,
+      "loss_partial": 0.056884765625,
+      "step": 3097
+    },
+    {
+      "epoch": 1.6228391828182294,
+      "grad_norm": 2.1957337856292725,
+      "learning_rate": 1.4831471900518915e-05,
+      "loss": 0.1022,
+      "step": 3098
+    },
+    {
+      "epoch": 1.6228391828182294,
+      "loss_partial": 0.05078125,
+      "step": 3098
+    },
+    {
+      "epoch": 1.6228391828182294,
+      "loss_partial": 0.054931640625,
+      "step": 3098
+    },
+    {
+      "epoch": 1.6233630172865374,
+      "grad_norm": 1.8619388341903687,
+      "learning_rate": 1.4827468794878106e-05,
+      "loss": 0.1032,
+      "step": 3099
+    },
+    {
+      "epoch": 1.6233630172865374,
+      "loss_partial": 0.047607421875,
+      "step": 3099
+    },
+    {
+      "epoch": 1.6233630172865374,
+      "loss_partial": 0.05126953125,
+      "step": 3099
+    },
+    {
+      "epoch": 1.6238868517548455,
+      "grad_norm": 1.6846622228622437,
+      "learning_rate": 1.4823464680326178e-05,
+      "loss": 0.1043,
+      "step": 3100
+    },
+    {
+      "epoch": 1.6238868517548455,
+      "loss_partial": 0.05322265625,
+      "step": 3100
+    },
+    {
+      "epoch": 1.6238868517548455,
+      "loss_partial": 0.055908203125,
+      "step": 3100
+    },
+    {
+      "epoch": 1.6244106862231535,
+      "grad_norm": 2.106419801712036,
+      "learning_rate": 1.4819459557699954e-05,
+      "loss": 0.1051,
+      "step": 3101
+    },
+    {
+      "epoch": 1.6244106862231535,
+      "loss_partial": 0.058837890625,
+      "step": 3101
+    },
+    {
+      "epoch": 1.6244106862231535,
+      "loss_partial": 0.05322265625,
+      "step": 3101
+    },
+    {
+      "epoch": 1.6249345206914616,
+      "grad_norm": 1.690216302871704,
+      "learning_rate": 1.4815453427836488e-05,
+      "loss": 0.1073,
+      "step": 3102
+    },
+    {
+      "epoch": 1.6249345206914616,
+      "loss_partial": 0.0556640625,
+      "step": 3102
+    },
+    {
+      "epoch": 1.6249345206914616,
+      "loss_partial": 0.0498046875,
+      "step": 3102
+    },
+    {
+      "epoch": 1.6254583551597697,
+      "grad_norm": 1.7978469133377075,
+      "learning_rate": 1.4811446291573028e-05,
+      "loss": 0.106,
+      "step": 3103
+    },
+    {
+      "epoch": 1.6254583551597697,
+      "loss_partial": 0.048583984375,
+      "step": 3103
+    },
+    {
+      "epoch": 1.6254583551597697,
+      "loss_partial": 0.052001953125,
+      "step": 3103
+    },
+    {
+      "epoch": 1.6259821896280775,
+      "grad_norm": 1.5960581302642822,
+      "learning_rate": 1.4807438149747054e-05,
+      "loss": 0.1059,
+      "step": 3104
+    },
+    {
+      "epoch": 1.6259821896280775,
+      "loss_partial": 0.05078125,
+      "step": 3104
+    },
+    {
+      "epoch": 1.6259821896280775,
+      "loss_partial": 0.046142578125,
+      "step": 3104
+    },
+    {
+      "epoch": 1.6265060240963856,
+      "grad_norm": 1.5880446434020996,
+      "learning_rate": 1.4803429003196229e-05,
+      "loss": 0.1063,
+      "step": 3105
+    },
+    {
+      "epoch": 1.6265060240963856,
+      "loss_partial": 0.052001953125,
+      "step": 3105
+    },
+    {
+      "epoch": 1.6265060240963856,
+      "loss_partial": 0.047607421875,
+      "step": 3105
+    },
+    {
+      "epoch": 1.6270298585646934,
+      "grad_norm": 1.698710560798645,
+      "learning_rate": 1.479941885275845e-05,
+      "loss": 0.1062,
+      "step": 3106
+    },
+    {
+      "epoch": 1.6270298585646934,
+      "loss_partial": 0.046875,
+      "step": 3106
+    },
+    {
+      "epoch": 1.6270298585646934,
+      "loss_partial": 0.0478515625,
+      "step": 3106
+    },
+    {
+      "epoch": 1.6275536930330015,
+      "grad_norm": 1.3866488933563232,
+      "learning_rate": 1.4795407699271807e-05,
+      "loss": 0.1038,
+      "step": 3107
+    },
+    {
+      "epoch": 1.6275536930330015,
+      "loss_partial": 0.053466796875,
+      "step": 3107
+    },
+    {
+      "epoch": 1.6275536930330015,
+      "loss_partial": 0.049072265625,
+      "step": 3107
+    },
+    {
+      "epoch": 1.6280775275013095,
+      "grad_norm": 2.096822500228882,
+      "learning_rate": 1.4791395543574609e-05,
+      "loss": 0.1051,
+      "step": 3108
+    },
+    {
+      "epoch": 1.6280775275013095,
+      "loss_partial": 0.052001953125,
+      "step": 3108
+    },
+    {
+      "epoch": 1.6280775275013095,
+      "loss_partial": 0.06787109375,
+      "step": 3108
+    },
+    {
+      "epoch": 1.6286013619696176,
+      "grad_norm": 1.8244819641113281,
+      "learning_rate": 1.4787382386505372e-05,
+      "loss": 0.106,
+      "step": 3109
+    },
+    {
+      "epoch": 1.6286013619696176,
+      "loss_partial": 0.044921875,
+      "step": 3109
+    },
+    {
+      "epoch": 1.6286013619696176,
+      "loss_partial": 0.0556640625,
+      "step": 3109
+    },
+    {
+      "epoch": 1.6291251964379256,
+      "grad_norm": 1.9951732158660889,
+      "learning_rate": 1.478336822890282e-05,
+      "loss": 0.1049,
+      "step": 3110
+    },
+    {
+      "epoch": 1.6291251964379256,
+      "loss_partial": 0.055419921875,
+      "step": 3110
+    },
+    {
+      "epoch": 1.6291251964379256,
+      "loss_partial": 0.04833984375,
+      "step": 3110
+    },
+    {
+      "epoch": 1.6296490309062337,
+      "grad_norm": 2.1697869300842285,
+      "learning_rate": 1.4779353071605886e-05,
+      "loss": 0.1043,
+      "step": 3111
+    },
+    {
+      "epoch": 1.6296490309062337,
+      "loss_partial": 0.045166015625,
+      "step": 3111
+    },
+    {
+      "epoch": 1.6296490309062337,
+      "loss_partial": 0.0478515625,
+      "step": 3111
+    },
+    {
+      "epoch": 1.6301728653745418,
+      "grad_norm": 1.7207926511764526,
+      "learning_rate": 1.477533691545372e-05,
+      "loss": 0.1038,
+      "step": 3112
+    },
+    {
+      "epoch": 1.6301728653745418,
+      "loss_partial": 0.058837890625,
+      "step": 3112
+    },
+    {
+      "epoch": 1.6301728653745418,
+      "loss_partial": 0.058349609375,
+      "step": 3112
+    },
+    {
+      "epoch": 1.6306966998428498,
+      "grad_norm": 1.7031654119491577,
+      "learning_rate": 1.4771319761285665e-05,
+      "loss": 0.1053,
+      "step": 3113
+    },
+    {
+      "epoch": 1.6306966998428498,
+      "loss_partial": 0.060791015625,
+      "step": 3113
+    },
+    {
+      "epoch": 1.6306966998428498,
+      "loss_partial": 0.04638671875,
+      "step": 3113
+    },
+    {
+      "epoch": 1.6312205343111577,
+      "grad_norm": 1.8059108257293701,
+      "learning_rate": 1.4767301609941284e-05,
+      "loss": 0.1031,
+      "step": 3114
+    },
+    {
+      "epoch": 1.6312205343111577,
+      "loss_partial": 0.054931640625,
+      "step": 3114
+    },
+    {
+      "epoch": 1.6312205343111577,
+      "loss_partial": 0.04736328125,
+      "step": 3114
+    },
+    {
+      "epoch": 1.6317443687794657,
+      "grad_norm": 1.755301833152771,
+      "learning_rate": 1.4763282462260349e-05,
+      "loss": 0.1059,
+      "step": 3115
+    },
+    {
+      "epoch": 1.6317443687794657,
+      "loss_partial": 0.054443359375,
+      "step": 3115
+    },
+    {
+      "epoch": 1.6317443687794657,
+      "loss_partial": 0.046142578125,
+      "step": 3115
+    },
+    {
+      "epoch": 1.6322682032477736,
+      "grad_norm": 2.3712449073791504,
+      "learning_rate": 1.4759262319082833e-05,
+      "loss": 0.1041,
+      "step": 3116
+    },
+    {
+      "epoch": 1.6322682032477736,
+      "loss_partial": 0.047607421875,
+      "step": 3116
+    },
+    {
+      "epoch": 1.6322682032477736,
+      "loss_partial": 0.059326171875,
+      "step": 3116
+    },
+    {
+      "epoch": 1.6327920377160816,
+      "grad_norm": 1.935057282447815,
+      "learning_rate": 1.4755241181248923e-05,
+      "loss": 0.1037,
+      "step": 3117
+    },
+    {
+      "epoch": 1.6327920377160816,
+      "loss_partial": 0.05517578125,
+      "step": 3117
+    },
+    {
+      "epoch": 1.6327920377160816,
+      "loss_partial": 0.052734375,
+      "step": 3117
+    },
+    {
+      "epoch": 1.6333158721843897,
+      "grad_norm": 2.272718667984009,
+      "learning_rate": 1.4751219049599011e-05,
+      "loss": 0.1072,
+      "step": 3118
+    },
+    {
+      "epoch": 1.6333158721843897,
+      "loss_partial": 0.0556640625,
+      "step": 3118
+    },
+    {
+      "epoch": 1.6333158721843897,
+      "loss_partial": 0.0517578125,
+      "step": 3118
+    },
+    {
+      "epoch": 1.6338397066526977,
+      "grad_norm": 2.0742220878601074,
+      "learning_rate": 1.4747195924973698e-05,
+      "loss": 0.1062,
+      "step": 3119
+    },
+    {
+      "epoch": 1.6338397066526977,
+      "loss_partial": 0.0595703125,
+      "step": 3119
+    },
+    {
+      "epoch": 1.6338397066526977,
+      "loss_partial": 0.058349609375,
+      "step": 3119
+    },
+    {
+      "epoch": 1.6343635411210058,
+      "grad_norm": 1.9941831827163696,
+      "learning_rate": 1.4743171808213794e-05,
+      "loss": 0.1093,
+      "step": 3120
+    },
+    {
+      "epoch": 1.6343635411210058,
+      "loss_partial": 0.051513671875,
+      "step": 3120
+    },
+    {
+      "epoch": 1.6343635411210058,
+      "loss_partial": 0.057373046875,
+      "step": 3120
+    },
+    {
+      "epoch": 1.6348873755893139,
+      "grad_norm": 1.5468393564224243,
+      "learning_rate": 1.4739146700160312e-05,
+      "loss": 0.1066,
+      "step": 3121
+    },
+    {
+      "epoch": 1.6348873755893139,
+      "loss_partial": 0.043212890625,
+      "step": 3121
+    },
+    {
+      "epoch": 1.6348873755893139,
+      "loss_partial": 0.06005859375,
+      "step": 3121
+    },
+    {
+      "epoch": 1.635411210057622,
+      "grad_norm": 1.6927024126052856,
+      "learning_rate": 1.4735120601654476e-05,
+      "loss": 0.1013,
+      "step": 3122
+    },
+    {
+      "epoch": 1.635411210057622,
+      "loss_partial": 0.047607421875,
+      "step": 3122
+    },
+    {
+      "epoch": 1.635411210057622,
+      "loss_partial": 0.0498046875,
+      "step": 3122
+    },
+    {
+      "epoch": 1.63593504452593,
+      "grad_norm": 2.1700563430786133,
+      "learning_rate": 1.4731093513537715e-05,
+      "loss": 0.1079,
+      "step": 3123
+    },
+    {
+      "epoch": 1.63593504452593,
+      "loss_partial": 0.05810546875,
+      "step": 3123
+    },
+    {
+      "epoch": 1.63593504452593,
+      "loss_partial": 0.046142578125,
+      "step": 3123
+    },
+    {
+      "epoch": 1.6364588789942378,
+      "grad_norm": 1.5910249948501587,
+      "learning_rate": 1.4727065436651664e-05,
+      "loss": 0.1081,
+      "step": 3124
+    },
+    {
+      "epoch": 1.6364588789942378,
+      "loss_partial": 0.055419921875,
+      "step": 3124
+    },
+    {
+      "epoch": 1.6364588789942378,
+      "loss_partial": 0.04541015625,
+      "step": 3124
+    },
+    {
+      "epoch": 1.6369827134625459,
+      "grad_norm": 2.052563190460205,
+      "learning_rate": 1.4723036371838173e-05,
+      "loss": 0.1033,
+      "step": 3125
+    },
+    {
+      "epoch": 1.6369827134625459,
+      "loss_partial": 0.050537109375,
+      "step": 3125
+    },
+    {
+      "epoch": 1.6369827134625459,
+      "loss_partial": 0.05224609375,
+      "step": 3125
+    },
+    {
+      "epoch": 1.6375065479308537,
+      "grad_norm": 1.5386790037155151,
+      "learning_rate": 1.4719006319939281e-05,
+      "loss": 0.1035,
+      "step": 3126
+    },
+    {
+      "epoch": 1.6375065479308537,
+      "loss_partial": 0.056396484375,
+      "step": 3126
+    },
+    {
+      "epoch": 1.6375065479308537,
+      "loss_partial": 0.0458984375,
+      "step": 3126
+    },
+    {
+      "epoch": 1.6380303823991618,
+      "grad_norm": 1.7299144268035889,
+      "learning_rate": 1.4714975281797254e-05,
+      "loss": 0.1043,
+      "step": 3127
+    },
+    {
+      "epoch": 1.6380303823991618,
+      "loss_partial": 0.0673828125,
+      "step": 3127
+    },
+    {
+      "epoch": 1.6380303823991618,
+      "loss_partial": 0.045166015625,
+      "step": 3127
+    },
+    {
+      "epoch": 1.6385542168674698,
+      "grad_norm": 2.1476542949676514,
+      "learning_rate": 1.4710943258254548e-05,
+      "loss": 0.1079,
+      "step": 3128
+    },
+    {
+      "epoch": 1.6385542168674698,
+      "loss_partial": 0.0546875,
+      "step": 3128
+    },
+    {
+      "epoch": 1.6385542168674698,
+      "loss_partial": 0.055419921875,
+      "step": 3128
+    },
+    {
+      "epoch": 1.6390780513357779,
+      "grad_norm": 1.5276087522506714,
+      "learning_rate": 1.4706910250153832e-05,
+      "loss": 0.1083,
+      "step": 3129
+    },
+    {
+      "epoch": 1.6390780513357779,
+      "loss_partial": 0.0576171875,
+      "step": 3129
+    },
+    {
+      "epoch": 1.6390780513357779,
+      "loss_partial": 0.0556640625,
+      "step": 3129
+    },
+    {
+      "epoch": 1.639601885804086,
+      "grad_norm": 1.4502754211425781,
+      "learning_rate": 1.4702876258337982e-05,
+      "loss": 0.104,
+      "step": 3130
+    },
+    {
+      "epoch": 1.639601885804086,
+      "loss_partial": 0.05810546875,
+      "step": 3130
+    },
+    {
+      "epoch": 1.639601885804086,
+      "loss_partial": 0.04638671875,
+      "step": 3130
+    },
+    {
+      "epoch": 1.640125720272394,
+      "grad_norm": 1.9776990413665771,
+      "learning_rate": 1.4698841283650076e-05,
+      "loss": 0.1093,
+      "step": 3131
+    },
+    {
+      "epoch": 1.640125720272394,
+      "loss_partial": 0.05224609375,
+      "step": 3131
+    },
+    {
+      "epoch": 1.640125720272394,
+      "loss_partial": 0.052001953125,
+      "step": 3131
+    },
+    {
+      "epoch": 1.640649554740702,
+      "grad_norm": 2.0681281089782715,
+      "learning_rate": 1.4694805326933397e-05,
+      "loss": 0.1029,
+      "step": 3132
+    },
+    {
+      "epoch": 1.640649554740702,
+      "loss_partial": 0.049560546875,
+      "step": 3132
+    },
+    {
+      "epoch": 1.640649554740702,
+      "loss_partial": 0.052734375,
+      "step": 3132
+    },
+    {
+      "epoch": 1.64117338920901,
+      "grad_norm": 1.9825751781463623,
+      "learning_rate": 1.4690768389031437e-05,
+      "loss": 0.1009,
+      "step": 3133
+    },
+    {
+      "epoch": 1.64117338920901,
+      "loss_partial": 0.04638671875,
+      "step": 3133
+    },
+    {
+      "epoch": 1.64117338920901,
+      "loss_partial": 0.054931640625,
+      "step": 3133
+    },
+    {
+      "epoch": 1.641697223677318,
+      "grad_norm": 1.6831955909729004,
+      "learning_rate": 1.4686730470787888e-05,
+      "loss": 0.1093,
+      "step": 3134
+    },
+    {
+      "epoch": 1.641697223677318,
+      "loss_partial": 0.05224609375,
+      "step": 3134
+    },
+    {
+      "epoch": 1.641697223677318,
+      "loss_partial": 0.045166015625,
+      "step": 3134
+    },
+    {
+      "epoch": 1.642221058145626,
+      "grad_norm": 1.5450941324234009,
+      "learning_rate": 1.4682691573046657e-05,
+      "loss": 0.1,
+      "step": 3135
+    },
+    {
+      "epoch": 1.642221058145626,
+      "loss_partial": 0.0537109375,
+      "step": 3135
+    },
+    {
+      "epoch": 1.642221058145626,
+      "loss_partial": 0.056396484375,
+      "step": 3135
+    },
+    {
+      "epoch": 1.6427448926139339,
+      "grad_norm": 1.3594846725463867,
+      "learning_rate": 1.4678651696651839e-05,
+      "loss": 0.1044,
+      "step": 3136
+    },
+    {
+      "epoch": 1.6427448926139339,
+      "loss_partial": 0.052734375,
+      "step": 3136
+    },
+    {
+      "epoch": 1.6427448926139339,
+      "loss_partial": 0.059814453125,
+      "step": 3136
+    },
+    {
+      "epoch": 1.643268727082242,
+      "grad_norm": 2.0384349822998047,
+      "learning_rate": 1.4674610842447754e-05,
+      "loss": 0.105,
+      "step": 3137
+    },
+    {
+      "epoch": 1.643268727082242,
+      "loss_partial": 0.05029296875,
+      "step": 3137
+    },
+    {
+      "epoch": 1.643268727082242,
+      "loss_partial": 0.0439453125,
+      "step": 3137
+    },
+    {
+      "epoch": 1.64379256155055,
+      "grad_norm": 1.6359848976135254,
+      "learning_rate": 1.4670569011278906e-05,
+      "loss": 0.1048,
+      "step": 3138
+    },
+    {
+      "epoch": 1.64379256155055,
+      "loss_partial": 0.056884765625,
+      "step": 3138
+    },
+    {
+      "epoch": 1.64379256155055,
+      "loss_partial": 0.049072265625,
+      "step": 3138
+    },
+    {
+      "epoch": 1.644316396018858,
+      "grad_norm": 1.895721673965454,
+      "learning_rate": 1.4666526203990018e-05,
+      "loss": 0.1053,
+      "step": 3139
+    },
+    {
+      "epoch": 1.644316396018858,
+      "loss_partial": 0.05712890625,
+      "step": 3139
+    },
+    {
+      "epoch": 1.644316396018858,
+      "loss_partial": 0.04931640625,
+      "step": 3139
+    },
+    {
+      "epoch": 1.644840230487166,
+      "grad_norm": 1.4557347297668457,
+      "learning_rate": 1.466248242142601e-05,
+      "loss": 0.1046,
+      "step": 3140
+    },
+    {
+      "epoch": 1.644840230487166,
+      "loss_partial": 0.05322265625,
+      "step": 3140
+    },
+    {
+      "epoch": 1.644840230487166,
+      "loss_partial": 0.046875,
+      "step": 3140
+    },
+    {
+      "epoch": 1.6453640649554742,
+      "grad_norm": 1.5139497518539429,
+      "learning_rate": 1.4658437664432008e-05,
+      "loss": 0.1051,
+      "step": 3141
+    },
+    {
+      "epoch": 1.6453640649554742,
+      "loss_partial": 0.04736328125,
+      "step": 3141
+    },
+    {
+      "epoch": 1.6453640649554742,
+      "loss_partial": 0.0517578125,
+      "step": 3141
+    },
+    {
+      "epoch": 1.6458878994237822,
+      "grad_norm": 1.5248359441757202,
+      "learning_rate": 1.4654391933853338e-05,
+      "loss": 0.106,
+      "step": 3142
+    },
+    {
+      "epoch": 1.6458878994237822,
+      "loss_partial": 0.055419921875,
+      "step": 3142
+    },
+    {
+      "epoch": 1.6458878994237822,
+      "loss_partial": 0.05029296875,
+      "step": 3142
+    },
+    {
+      "epoch": 1.64641173389209,
+      "grad_norm": 1.8808314800262451,
+      "learning_rate": 1.465034523053554e-05,
+      "loss": 0.1063,
+      "step": 3143
+    },
+    {
+      "epoch": 1.64641173389209,
+      "loss_partial": 0.040771484375,
+      "step": 3143
+    },
+    {
+      "epoch": 1.64641173389209,
+      "loss_partial": 0.052490234375,
+      "step": 3143
+    },
+    {
+      "epoch": 1.6469355683603981,
+      "grad_norm": 1.9820364713668823,
+      "learning_rate": 1.4646297555324345e-05,
+      "loss": 0.1053,
+      "step": 3144
+    },
+    {
+      "epoch": 1.6469355683603981,
+      "loss_partial": 0.04736328125,
+      "step": 3144
+    },
+    {
+      "epoch": 1.6469355683603981,
+      "loss_partial": 0.046142578125,
+      "step": 3144
+    },
+    {
+      "epoch": 1.6474594028287062,
+      "grad_norm": 1.3443384170532227,
+      "learning_rate": 1.4642248909065693e-05,
+      "loss": 0.1041,
+      "step": 3145
+    },
+    {
+      "epoch": 1.6474594028287062,
+      "loss_partial": 0.050048828125,
+      "step": 3145
+    },
+    {
+      "epoch": 1.6474594028287062,
+      "loss_partial": 0.0634765625,
+      "step": 3145
+    },
+    {
+      "epoch": 1.647983237297014,
+      "grad_norm": 1.9554699659347534,
+      "learning_rate": 1.4638199292605723e-05,
+      "loss": 0.1035,
+      "step": 3146
+    },
+    {
+      "epoch": 1.647983237297014,
+      "loss_partial": 0.045166015625,
+      "step": 3146
+    },
+    {
+      "epoch": 1.647983237297014,
+      "loss_partial": 0.05859375,
+      "step": 3146
+    },
+    {
+      "epoch": 1.648507071765322,
+      "grad_norm": 1.3809906244277954,
+      "learning_rate": 1.4634148706790785e-05,
+      "loss": 0.1021,
+      "step": 3147
+    },
+    {
+      "epoch": 1.648507071765322,
+      "loss_partial": 0.04541015625,
+      "step": 3147
+    },
+    {
+      "epoch": 1.648507071765322,
+      "loss_partial": 0.05615234375,
+      "step": 3147
+    },
+    {
+      "epoch": 1.6490309062336301,
+      "grad_norm": 1.8794208765029907,
+      "learning_rate": 1.4630097152467422e-05,
+      "loss": 0.1041,
+      "step": 3148
+    },
+    {
+      "epoch": 1.6490309062336301,
+      "loss_partial": 0.0595703125,
+      "step": 3148
+    },
+    {
+      "epoch": 1.6490309062336301,
+      "loss_partial": 0.05078125,
+      "step": 3148
+    },
+    {
+      "epoch": 1.6495547407019382,
+      "grad_norm": 1.7769187688827515,
+      "learning_rate": 1.462604463048239e-05,
+      "loss": 0.1021,
+      "step": 3149
+    },
+    {
+      "epoch": 1.6495547407019382,
+      "loss_partial": 0.053466796875,
+      "step": 3149
+    },
+    {
+      "epoch": 1.6495547407019382,
+      "loss_partial": 0.068359375,
+      "step": 3149
+    },
+    {
+      "epoch": 1.6500785751702463,
+      "grad_norm": 1.4893949031829834,
+      "learning_rate": 1.4621991141682632e-05,
+      "loss": 0.1055,
+      "step": 3150
+    },
+    {
+      "epoch": 1.6500785751702463,
+      "loss_partial": 0.045654296875,
+      "step": 3150
+    },
+    {
+      "epoch": 1.6500785751702463,
+      "loss_partial": 0.049072265625,
+      "step": 3150
+    },
+    {
+      "epoch": 1.6506024096385543,
+      "grad_norm": 2.670872449874878,
+      "learning_rate": 1.461793668691531e-05,
+      "loss": 0.1031,
+      "step": 3151
+    },
+    {
+      "epoch": 1.6506024096385543,
+      "loss_partial": 0.06396484375,
+      "step": 3151
+    },
+    {
+      "epoch": 1.6506024096385543,
+      "loss_partial": 0.05126953125,
+      "step": 3151
+    },
+    {
+      "epoch": 1.6511262441068624,
+      "grad_norm": 1.5204659700393677,
+      "learning_rate": 1.4613881267027774e-05,
+      "loss": 0.1078,
+      "step": 3152
+    },
+    {
+      "epoch": 1.6511262441068624,
+      "loss_partial": 0.048828125,
+      "step": 3152
+    },
+    {
+      "epoch": 1.6511262441068624,
+      "loss_partial": 0.052734375,
+      "step": 3152
+    },
+    {
+      "epoch": 1.6516500785751702,
+      "grad_norm": 1.8585385084152222,
+      "learning_rate": 1.4609824882867587e-05,
+      "loss": 0.1078,
+      "step": 3153
+    },
+    {
+      "epoch": 1.6516500785751702,
+      "loss_partial": 0.0595703125,
+      "step": 3153
+    },
+    {
+      "epoch": 1.6516500785751702,
+      "loss_partial": 0.053955078125,
+      "step": 3153
+    },
+    {
+      "epoch": 1.6521739130434783,
+      "grad_norm": 1.7628045082092285,
+      "learning_rate": 1.4605767535282505e-05,
+      "loss": 0.1065,
+      "step": 3154
+    },
+    {
+      "epoch": 1.6521739130434783,
+      "loss_partial": 0.05029296875,
+      "step": 3154
+    },
+    {
+      "epoch": 1.6521739130434783,
+      "loss_partial": 0.048583984375,
+      "step": 3154
+    },
+    {
+      "epoch": 1.6526977475117863,
+      "grad_norm": 1.671938419342041,
+      "learning_rate": 1.4601709225120489e-05,
+      "loss": 0.1077,
+      "step": 3155
+    },
+    {
+      "epoch": 1.6526977475117863,
+      "loss_partial": 0.04541015625,
+      "step": 3155
+    },
+    {
+      "epoch": 1.6526977475117863,
+      "loss_partial": 0.053466796875,
+      "step": 3155
+    },
+    {
+      "epoch": 1.6532215819800942,
+      "grad_norm": 2.0406391620635986,
+      "learning_rate": 1.4597649953229703e-05,
+      "loss": 0.1068,
+      "step": 3156
+    },
+    {
+      "epoch": 1.6532215819800942,
+      "loss_partial": 0.04150390625,
+      "step": 3156
+    },
+    {
+      "epoch": 1.6532215819800942,
+      "loss_partial": 0.054443359375,
+      "step": 3156
+    },
+    {
+      "epoch": 1.6537454164484022,
+      "grad_norm": 1.6428422927856445,
+      "learning_rate": 1.4593589720458507e-05,
+      "loss": 0.1036,
+      "step": 3157
+    },
+    {
+      "epoch": 1.6537454164484022,
+      "loss_partial": 0.0537109375,
+      "step": 3157
+    },
+    {
+      "epoch": 1.6537454164484022,
+      "loss_partial": 0.0517578125,
+      "step": 3157
+    },
+    {
+      "epoch": 1.6542692509167103,
+      "grad_norm": 1.729944109916687,
+      "learning_rate": 1.4589528527655463e-05,
+      "loss": 0.1062,
+      "step": 3158
+    },
+    {
+      "epoch": 1.6542692509167103,
+      "loss_partial": 0.0673828125,
+      "step": 3158
+    },
+    {
+      "epoch": 1.6542692509167103,
+      "loss_partial": 0.04931640625,
+      "step": 3158
+    },
+    {
+      "epoch": 1.6547930853850183,
+      "grad_norm": 1.6213082075119019,
+      "learning_rate": 1.458546637566934e-05,
+      "loss": 0.1084,
+      "step": 3159
+    },
+    {
+      "epoch": 1.6547930853850183,
+      "loss_partial": 0.050048828125,
+      "step": 3159
+    },
+    {
+      "epoch": 1.6547930853850183,
+      "loss_partial": 0.060546875,
+      "step": 3159
+    },
+    {
+      "epoch": 1.6553169198533264,
+      "grad_norm": 2.304030656814575,
+      "learning_rate": 1.4581403265349099e-05,
+      "loss": 0.1062,
+      "step": 3160
+    },
+    {
+      "epoch": 1.6553169198533264,
+      "loss_partial": 0.049560546875,
+      "step": 3160
+    },
+    {
+      "epoch": 1.6553169198533264,
+      "loss_partial": 0.060546875,
+      "step": 3160
+    },
+    {
+      "epoch": 1.6558407543216345,
+      "grad_norm": 1.4327527284622192,
+      "learning_rate": 1.4577339197543908e-05,
+      "loss": 0.1012,
+      "step": 3161
+    },
+    {
+      "epoch": 1.6558407543216345,
+      "loss_partial": 0.0556640625,
+      "step": 3161
+    },
+    {
+      "epoch": 1.6558407543216345,
+      "loss_partial": 0.0478515625,
+      "step": 3161
+    },
+    {
+      "epoch": 1.6563645887899425,
+      "grad_norm": 2.1497631072998047,
+      "learning_rate": 1.4573274173103126e-05,
+      "loss": 0.105,
+      "step": 3162
+    },
+    {
+      "epoch": 1.6563645887899425,
+      "loss_partial": 0.05126953125,
+      "step": 3162
+    },
+    {
+      "epoch": 1.6563645887899425,
+      "loss_partial": 0.0498046875,
+      "step": 3162
+    },
+    {
+      "epoch": 1.6568884232582504,
+      "grad_norm": 2.26224684715271,
+      "learning_rate": 1.4569208192876327e-05,
+      "loss": 0.1041,
+      "step": 3163
+    },
+    {
+      "epoch": 1.6568884232582504,
+      "loss_partial": 0.053955078125,
+      "step": 3163
+    },
+    {
+      "epoch": 1.6568884232582504,
+      "loss_partial": 0.054931640625,
+      "step": 3163
+    },
+    {
+      "epoch": 1.6574122577265584,
+      "grad_norm": 1.783689022064209,
+      "learning_rate": 1.4565141257713267e-05,
+      "loss": 0.1028,
+      "step": 3164
+    },
+    {
+      "epoch": 1.6574122577265584,
+      "loss_partial": 0.04638671875,
+      "step": 3164
+    },
+    {
+      "epoch": 1.6574122577265584,
+      "loss_partial": 0.050537109375,
+      "step": 3164
+    },
+    {
+      "epoch": 1.6579360921948663,
+      "grad_norm": 1.809430480003357,
+      "learning_rate": 1.4561073368463916e-05,
+      "loss": 0.1022,
+      "step": 3165
+    },
+    {
+      "epoch": 1.6579360921948663,
+      "loss_partial": 0.04736328125,
+      "step": 3165
+    },
+    {
+      "epoch": 1.6579360921948663,
+      "loss_partial": 0.050048828125,
+      "step": 3165
+    },
+    {
+      "epoch": 1.6584599266631743,
+      "grad_norm": 1.7609670162200928,
+      "learning_rate": 1.4557004525978435e-05,
+      "loss": 0.1006,
+      "step": 3166
+    },
+    {
+      "epoch": 1.6584599266631743,
+      "loss_partial": 0.05224609375,
+      "step": 3166
+    },
+    {
+      "epoch": 1.6584599266631743,
+      "loss_partial": 0.04736328125,
+      "step": 3166
+    },
+    {
+      "epoch": 1.6589837611314824,
+      "grad_norm": 1.653128743171692,
+      "learning_rate": 1.4552934731107189e-05,
+      "loss": 0.1077,
+      "step": 3167
+    },
+    {
+      "epoch": 1.6589837611314824,
+      "loss_partial": 0.05126953125,
+      "step": 3167
+    },
+    {
+      "epoch": 1.6589837611314824,
+      "loss_partial": 0.06005859375,
+      "step": 3167
+    },
+    {
+      "epoch": 1.6595075955997904,
+      "grad_norm": 1.6432245969772339,
+      "learning_rate": 1.4548863984700739e-05,
+      "loss": 0.0987,
+      "step": 3168
+    },
+    {
+      "epoch": 1.6595075955997904,
+      "loss_partial": 0.051513671875,
+      "step": 3168
+    },
+    {
+      "epoch": 1.6595075955997904,
+      "loss_partial": 0.045654296875,
+      "step": 3168
+    },
+    {
+      "epoch": 1.6600314300680985,
+      "grad_norm": 1.7717682123184204,
+      "learning_rate": 1.4544792287609844e-05,
+      "loss": 0.1049,
+      "step": 3169
+    },
+    {
+      "epoch": 1.6600314300680985,
+      "loss_partial": 0.049072265625,
+      "step": 3169
+    },
+    {
+      "epoch": 1.6600314300680985,
+      "loss_partial": 0.0673828125,
+      "step": 3169
+    },
+    {
+      "epoch": 1.6605552645364066,
+      "grad_norm": 1.453141689300537,
+      "learning_rate": 1.4540719640685464e-05,
+      "loss": 0.1047,
+      "step": 3170
+    },
+    {
+      "epoch": 1.6605552645364066,
+      "loss_partial": 0.04150390625,
+      "step": 3170
+    },
+    {
+      "epoch": 1.6605552645364066,
+      "loss_partial": 0.04296875,
+      "step": 3170
+    },
+    {
+      "epoch": 1.6610790990047146,
+      "grad_norm": 2.5520195960998535,
+      "learning_rate": 1.453664604477876e-05,
+      "loss": 0.1049,
+      "step": 3171
+    },
+    {
+      "epoch": 1.6610790990047146,
+      "loss_partial": 0.04833984375,
+      "step": 3171
+    },
+    {
+      "epoch": 1.6610790990047146,
+      "loss_partial": 0.046630859375,
+      "step": 3171
+    },
+    {
+      "epoch": 1.6616029334730227,
+      "grad_norm": 1.6931109428405762,
+      "learning_rate": 1.4532571500741087e-05,
+      "loss": 0.1078,
+      "step": 3172
+    },
+    {
+      "epoch": 1.6616029334730227,
+      "loss_partial": 0.045654296875,
+      "step": 3172
+    },
+    {
+      "epoch": 1.6616029334730227,
+      "loss_partial": 0.06005859375,
+      "step": 3172
+    },
+    {
+      "epoch": 1.6621267679413305,
+      "grad_norm": 2.0037682056427,
+      "learning_rate": 1.4528496009423996e-05,
+      "loss": 0.1063,
+      "step": 3173
+    },
+    {
+      "epoch": 1.6621267679413305,
+      "loss_partial": 0.056396484375,
+      "step": 3173
+    },
+    {
+      "epoch": 1.6621267679413305,
+      "loss_partial": 0.06005859375,
+      "step": 3173
+    },
+    {
+      "epoch": 1.6626506024096386,
+      "grad_norm": 1.637986183166504,
+      "learning_rate": 1.4524419571679246e-05,
+      "loss": 0.1093,
+      "step": 3174
+    },
+    {
+      "epoch": 1.6626506024096386,
+      "loss_partial": 0.055419921875,
+      "step": 3174
+    },
+    {
+      "epoch": 1.6626506024096386,
+      "loss_partial": 0.04345703125,
+      "step": 3174
+    },
+    {
+      "epoch": 1.6631744368779464,
+      "grad_norm": 1.4178208112716675,
+      "learning_rate": 1.4520342188358783e-05,
+      "loss": 0.1035,
+      "step": 3175
+    },
+    {
+      "epoch": 1.6631744368779464,
+      "loss_partial": 0.055419921875,
+      "step": 3175
+    },
+    {
+      "epoch": 1.6631744368779464,
+      "loss_partial": 0.051513671875,
+      "step": 3175
+    },
+    {
+      "epoch": 1.6636982713462545,
+      "grad_norm": 2.2428956031799316,
+      "learning_rate": 1.4516263860314755e-05,
+      "loss": 0.1055,
+      "step": 3176
+    },
+    {
+      "epoch": 1.6636982713462545,
+      "loss_partial": 0.056396484375,
+      "step": 3176
+    },
+    {
+      "epoch": 1.6636982713462545,
+      "loss_partial": 0.048095703125,
+      "step": 3176
+    },
+    {
+      "epoch": 1.6642221058145625,
+      "grad_norm": 1.3266733884811401,
+      "learning_rate": 1.4512184588399505e-05,
+      "loss": 0.1006,
+      "step": 3177
+    },
+    {
+      "epoch": 1.6642221058145625,
+      "loss_partial": 0.055908203125,
+      "step": 3177
+    },
+    {
+      "epoch": 1.6642221058145625,
+      "loss_partial": 0.049560546875,
+      "step": 3177
+    },
+    {
+      "epoch": 1.6647459402828706,
+      "grad_norm": 2.0142295360565186,
+      "learning_rate": 1.4508104373465582e-05,
+      "loss": 0.1051,
+      "step": 3178
+    },
+    {
+      "epoch": 1.6647459402828706,
+      "loss_partial": 0.054931640625,
+      "step": 3178
+    },
+    {
+      "epoch": 1.6647459402828706,
+      "loss_partial": 0.042236328125,
+      "step": 3178
+    },
+    {
+      "epoch": 1.6652697747511787,
+      "grad_norm": 1.5387035608291626,
+      "learning_rate": 1.450402321636572e-05,
+      "loss": 0.1064,
+      "step": 3179
+    },
+    {
+      "epoch": 1.6652697747511787,
+      "loss_partial": 0.050537109375,
+      "step": 3179
+    },
+    {
+      "epoch": 1.6652697747511787,
+      "loss_partial": 0.052490234375,
+      "step": 3179
+    },
+    {
+      "epoch": 1.6657936092194867,
+      "grad_norm": 1.7241934537887573,
+      "learning_rate": 1.449994111795286e-05,
+      "loss": 0.1054,
+      "step": 3180
+    },
+    {
+      "epoch": 1.6657936092194867,
+      "loss_partial": 0.0419921875,
+      "step": 3180
+    },
+    {
+      "epoch": 1.6657936092194867,
+      "loss_partial": 0.049072265625,
+      "step": 3180
+    },
+    {
+      "epoch": 1.6663174436877948,
+      "grad_norm": 1.8231449127197266,
+      "learning_rate": 1.4495858079080132e-05,
+      "loss": 0.1044,
+      "step": 3181
+    },
+    {
+      "epoch": 1.6663174436877948,
+      "loss_partial": 0.04443359375,
+      "step": 3181
+    },
+    {
+      "epoch": 1.6663174436877948,
+      "loss_partial": 0.052490234375,
+      "step": 3181
+    },
+    {
+      "epoch": 1.6668412781561028,
+      "grad_norm": 1.657021164894104,
+      "learning_rate": 1.4491774100600868e-05,
+      "loss": 0.1081,
+      "step": 3182
+    },
+    {
+      "epoch": 1.6668412781561028,
+      "loss_partial": 0.046142578125,
+      "step": 3182
+    },
+    {
+      "epoch": 1.6668412781561028,
+      "loss_partial": 0.0498046875,
+      "step": 3182
+    },
+    {
+      "epoch": 1.6673651126244107,
+      "grad_norm": 1.5076442956924438,
+      "learning_rate": 1.4487689183368591e-05,
+      "loss": 0.1036,
+      "step": 3183
+    },
+    {
+      "epoch": 1.6673651126244107,
+      "loss_partial": 0.054443359375,
+      "step": 3183
+    },
+    {
+      "epoch": 1.6673651126244107,
+      "loss_partial": 0.05810546875,
+      "step": 3183
+    },
+    {
+      "epoch": 1.6678889470927187,
+      "grad_norm": 1.6305441856384277,
+      "learning_rate": 1.4483603328237029e-05,
+      "loss": 0.1094,
+      "step": 3184
+    },
+    {
+      "epoch": 1.6678889470927187,
+      "loss_partial": 0.0556640625,
+      "step": 3184
+    },
+    {
+      "epoch": 1.6678889470927187,
+      "loss_partial": 0.04833984375,
+      "step": 3184
+    },
+    {
+      "epoch": 1.6684127815610266,
+      "grad_norm": 1.4130265712738037,
+      "learning_rate": 1.4479516536060093e-05,
+      "loss": 0.1021,
+      "step": 3185
+    },
+    {
+      "epoch": 1.6684127815610266,
+      "loss_partial": 0.05224609375,
+      "step": 3185
+    },
+    {
+      "epoch": 1.6684127815610266,
+      "loss_partial": 0.052978515625,
+      "step": 3185
+    },
+    {
+      "epoch": 1.6689366160293346,
+      "grad_norm": 1.5223193168640137,
+      "learning_rate": 1.4475428807691904e-05,
+      "loss": 0.1051,
+      "step": 3186
+    },
+    {
+      "epoch": 1.6689366160293346,
+      "loss_partial": 0.046875,
+      "step": 3186
+    },
+    {
+      "epoch": 1.6689366160293346,
+      "loss_partial": 0.05224609375,
+      "step": 3186
+    },
+    {
+      "epoch": 1.6694604504976427,
+      "grad_norm": 1.8731982707977295,
+      "learning_rate": 1.4471340143986769e-05,
+      "loss": 0.1028,
+      "step": 3187
+    },
+    {
+      "epoch": 1.6694604504976427,
+      "loss_partial": 0.050048828125,
+      "step": 3187
+    },
+    {
+      "epoch": 1.6694604504976427,
+      "loss_partial": 0.051513671875,
+      "step": 3187
+    },
+    {
+      "epoch": 1.6699842849659507,
+      "grad_norm": 1.7796638011932373,
+      "learning_rate": 1.4467250545799196e-05,
+      "loss": 0.1034,
+      "step": 3188
+    },
+    {
+      "epoch": 1.6699842849659507,
+      "loss_partial": 0.052978515625,
+      "step": 3188
+    },
+    {
+      "epoch": 1.6699842849659507,
+      "loss_partial": 0.0537109375,
+      "step": 3188
+    },
+    {
+      "epoch": 1.6705081194342588,
+      "grad_norm": 2.2181272506713867,
+      "learning_rate": 1.4463160013983882e-05,
+      "loss": 0.1066,
+      "step": 3189
+    },
+    {
+      "epoch": 1.6705081194342588,
+      "loss_partial": 0.05078125,
+      "step": 3189
+    },
+    {
+      "epoch": 1.6705081194342588,
+      "loss_partial": 0.048095703125,
+      "step": 3189
+    },
+    {
+      "epoch": 1.6710319539025669,
+      "grad_norm": 1.6576074361801147,
+      "learning_rate": 1.4459068549395723e-05,
+      "loss": 0.107,
+      "step": 3190
+    },
+    {
+      "epoch": 1.6710319539025669,
+      "loss_partial": 0.041259765625,
+      "step": 3190
+    },
+    {
+      "epoch": 1.6710319539025669,
+      "loss_partial": 0.054931640625,
+      "step": 3190
+    },
+    {
+      "epoch": 1.671555788370875,
+      "grad_norm": 2.32450270652771,
+      "learning_rate": 1.445497615288981e-05,
+      "loss": 0.1036,
+      "step": 3191
+    },
+    {
+      "epoch": 1.671555788370875,
+      "loss_partial": 0.056884765625,
+      "step": 3191
+    },
+    {
+      "epoch": 1.671555788370875,
+      "loss_partial": 0.060791015625,
+      "step": 3191
+    },
+    {
+      "epoch": 1.6720796228391828,
+      "grad_norm": 2.4508798122406006,
+      "learning_rate": 1.4450882825321433e-05,
+      "loss": 0.1053,
+      "step": 3192
+    },
+    {
+      "epoch": 1.6720796228391828,
+      "loss_partial": 0.048828125,
+      "step": 3192
+    },
+    {
+      "epoch": 1.6720796228391828,
+      "loss_partial": 0.052001953125,
+      "step": 3192
+    },
+    {
+      "epoch": 1.6726034573074908,
+      "grad_norm": 2.7103538513183594,
+      "learning_rate": 1.444678856754607e-05,
+      "loss": 0.1055,
+      "step": 3193
+    },
+    {
+      "epoch": 1.6726034573074908,
+      "loss_partial": 0.04443359375,
+      "step": 3193
+    },
+    {
+      "epoch": 1.6726034573074908,
+      "loss_partial": 0.0517578125,
+      "step": 3193
+    },
+    {
+      "epoch": 1.6731272917757989,
+      "grad_norm": 1.308054804801941,
+      "learning_rate": 1.4442693380419392e-05,
+      "loss": 0.1034,
+      "step": 3194
+    },
+    {
+      "epoch": 1.6731272917757989,
+      "loss_partial": 0.045654296875,
+      "step": 3194
+    },
+    {
+      "epoch": 1.6731272917757989,
+      "loss_partial": 0.06103515625,
+      "step": 3194
+    },
+    {
+      "epoch": 1.6736511262441067,
+      "grad_norm": 1.831186294555664,
+      "learning_rate": 1.443859726479727e-05,
+      "loss": 0.1038,
+      "step": 3195
+    },
+    {
+      "epoch": 1.6736511262441067,
+      "loss_partial": 0.0625,
+      "step": 3195
+    },
+    {
+      "epoch": 1.6736511262441067,
+      "loss_partial": 0.05859375,
+      "step": 3195
+    },
+    {
+      "epoch": 1.6741749607124148,
+      "grad_norm": 1.5642191171646118,
+      "learning_rate": 1.4434500221535767e-05,
+      "loss": 0.1064,
+      "step": 3196
+    },
+    {
+      "epoch": 1.6741749607124148,
+      "loss_partial": 0.0576171875,
+      "step": 3196
+    },
+    {
+      "epoch": 1.6741749607124148,
+      "loss_partial": 0.04931640625,
+      "step": 3196
+    },
+    {
+      "epoch": 1.6746987951807228,
+      "grad_norm": 1.800004005432129,
+      "learning_rate": 1.443040225149114e-05,
+      "loss": 0.1011,
+      "step": 3197
+    },
+    {
+      "epoch": 1.6746987951807228,
+      "loss_partial": 0.0673828125,
+      "step": 3197
+    },
+    {
+      "epoch": 1.6746987951807228,
+      "loss_partial": 0.045166015625,
+      "step": 3197
+    },
+    {
+      "epoch": 1.675222629649031,
+      "grad_norm": 1.2609957456588745,
+      "learning_rate": 1.4426303355519841e-05,
+      "loss": 0.1045,
+      "step": 3198
+    },
+    {
+      "epoch": 1.675222629649031,
+      "loss_partial": 0.057373046875,
+      "step": 3198
+    },
+    {
+      "epoch": 1.675222629649031,
+      "loss_partial": 0.05712890625,
+      "step": 3198
+    },
+    {
+      "epoch": 1.675746464117339,
+      "grad_norm": 1.666051983833313,
+      "learning_rate": 1.4422203534478512e-05,
+      "loss": 0.1081,
+      "step": 3199
+    },
+    {
+      "epoch": 1.675746464117339,
+      "loss_partial": 0.0556640625,
+      "step": 3199
+    },
+    {
+      "epoch": 1.675746464117339,
+      "loss_partial": 0.05078125,
+      "step": 3199
+    },
+    {
+      "epoch": 1.676270298585647,
+      "grad_norm": 2.000880718231201,
+      "learning_rate": 1.4418102789223991e-05,
+      "loss": 0.1048,
+      "step": 3200
+    },
+    {
+      "epoch": 1.676270298585647,
+      "loss_partial": 0.0537109375,
+      "step": 3200
+    },
+    {
+      "epoch": 1.676270298585647,
+      "loss_partial": 0.05126953125,
+      "step": 3200
+    },
+    {
+      "epoch": 1.676794133053955,
+      "grad_norm": 1.099478840827942,
+      "learning_rate": 1.4414001120613303e-05,
+      "loss": 0.102,
+      "step": 3201
+    },
+    {
+      "epoch": 1.676794133053955,
+      "loss_partial": 0.055419921875,
+      "step": 3201
+    },
+    {
+      "epoch": 1.676794133053955,
+      "loss_partial": 0.0556640625,
+      "step": 3201
+    },
+    {
+      "epoch": 1.677317967522263,
+      "grad_norm": 1.8267444372177124,
+      "learning_rate": 1.440989852950368e-05,
+      "loss": 0.1028,
+      "step": 3202
+    },
+    {
+      "epoch": 1.677317967522263,
+      "loss_partial": 0.0595703125,
+      "step": 3202
+    },
+    {
+      "epoch": 1.677317967522263,
+      "loss_partial": 0.049560546875,
+      "step": 3202
+    },
+    {
+      "epoch": 1.677841801990571,
+      "grad_norm": 1.8956390619277954,
+      "learning_rate": 1.4405795016752533e-05,
+      "loss": 0.1057,
+      "step": 3203
+    },
+    {
+      "epoch": 1.677841801990571,
+      "loss_partial": 0.055908203125,
+      "step": 3203
+    },
+    {
+      "epoch": 1.677841801990571,
+      "loss_partial": 0.04638671875,
+      "step": 3203
+    },
+    {
+      "epoch": 1.678365636458879,
+      "grad_norm": 1.9899399280548096,
+      "learning_rate": 1.4401690583217474e-05,
+      "loss": 0.1053,
+      "step": 3204
+    },
+    {
+      "epoch": 1.678365636458879,
+      "loss_partial": 0.0556640625,
+      "step": 3204
+    },
+    {
+      "epoch": 1.678365636458879,
+      "loss_partial": 0.05712890625,
+      "step": 3204
+    },
+    {
+      "epoch": 1.6788894709271869,
+      "grad_norm": 1.766801118850708,
+      "learning_rate": 1.4397585229756301e-05,
+      "loss": 0.1052,
+      "step": 3205
+    },
+    {
+      "epoch": 1.6788894709271869,
+      "loss_partial": 0.052978515625,
+      "step": 3205
+    },
+    {
+      "epoch": 1.6788894709271869,
+      "loss_partial": 0.054931640625,
+      "step": 3205
+    },
+    {
+      "epoch": 1.679413305395495,
+      "grad_norm": 1.9482274055480957,
+      "learning_rate": 1.4393478957227011e-05,
+      "loss": 0.1027,
+      "step": 3206
+    },
+    {
+      "epoch": 1.679413305395495,
+      "loss_partial": 0.05712890625,
+      "step": 3206
+    },
+    {
+      "epoch": 1.679413305395495,
+      "loss_partial": 0.04833984375,
+      "step": 3206
+    },
+    {
+      "epoch": 1.679937139863803,
+      "grad_norm": 2.074709177017212,
+      "learning_rate": 1.4389371766487784e-05,
+      "loss": 0.104,
+      "step": 3207
+    },
+    {
+      "epoch": 1.679937139863803,
+      "loss_partial": 0.052490234375,
+      "step": 3207
+    },
+    {
+      "epoch": 1.679937139863803,
+      "loss_partial": 0.04248046875,
+      "step": 3207
+    },
+    {
+      "epoch": 1.680460974332111,
+      "grad_norm": 1.453563928604126,
+      "learning_rate": 1.4385263658397004e-05,
+      "loss": 0.1024,
+      "step": 3208
+    },
+    {
+      "epoch": 1.680460974332111,
+      "loss_partial": 0.05908203125,
+      "step": 3208
+    },
+    {
+      "epoch": 1.680460974332111,
+      "loss_partial": 0.05615234375,
+      "step": 3208
+    },
+    {
+      "epoch": 1.6809848088004191,
+      "grad_norm": 1.8699653148651123,
+      "learning_rate": 1.4381154633813233e-05,
+      "loss": 0.1075,
+      "step": 3209
+    },
+    {
+      "epoch": 1.6809848088004191,
+      "loss_partial": 0.043701171875,
+      "step": 3209
+    },
+    {
+      "epoch": 1.6809848088004191,
+      "loss_partial": 0.050537109375,
+      "step": 3209
+    },
+    {
+      "epoch": 1.6815086432687272,
+      "grad_norm": 1.599482774734497,
+      "learning_rate": 1.437704469359524e-05,
+      "loss": 0.1015,
+      "step": 3210
+    },
+    {
+      "epoch": 1.6815086432687272,
+      "loss_partial": 0.05419921875,
+      "step": 3210
+    },
+    {
+      "epoch": 1.6815086432687272,
+      "loss_partial": 0.05908203125,
+      "step": 3210
+    },
+    {
+      "epoch": 1.6820324777370352,
+      "grad_norm": 1.6048641204833984,
+      "learning_rate": 1.437293383860197e-05,
+      "loss": 0.1063,
+      "step": 3211
+    },
+    {
+      "epoch": 1.6820324777370352,
+      "loss_partial": 0.051025390625,
+      "step": 3211
+    },
+    {
+      "epoch": 1.6820324777370352,
+      "loss_partial": 0.048095703125,
+      "step": 3211
+    },
+    {
+      "epoch": 1.682556312205343,
+      "grad_norm": 1.8925129175186157,
+      "learning_rate": 1.4368822069692572e-05,
+      "loss": 0.1053,
+      "step": 3212
+    },
+    {
+      "epoch": 1.682556312205343,
+      "loss_partial": 0.05224609375,
+      "step": 3212
+    },
+    {
+      "epoch": 1.682556312205343,
+      "loss_partial": 0.046875,
+      "step": 3212
+    },
+    {
+      "epoch": 1.6830801466736511,
+      "grad_norm": 1.4229692220687866,
+      "learning_rate": 1.4364709387726372e-05,
+      "loss": 0.1034,
+      "step": 3213
+    },
+    {
+      "epoch": 1.6830801466736511,
+      "loss_partial": 0.053955078125,
+      "step": 3213
+    },
+    {
+      "epoch": 1.6830801466736511,
+      "loss_partial": 0.047119140625,
+      "step": 3213
+    },
+    {
+      "epoch": 1.6836039811419592,
+      "grad_norm": 1.906759262084961,
+      "learning_rate": 1.4360595793562906e-05,
+      "loss": 0.1048,
+      "step": 3214
+    },
+    {
+      "epoch": 1.6836039811419592,
+      "loss_partial": 0.037841796875,
+      "step": 3214
+    },
+    {
+      "epoch": 1.6836039811419592,
+      "loss_partial": 0.05859375,
+      "step": 3214
+    },
+    {
+      "epoch": 1.684127815610267,
+      "grad_norm": 1.2447763681411743,
+      "learning_rate": 1.435648128806188e-05,
+      "loss": 0.1051,
+      "step": 3215
+    },
+    {
+      "epoch": 1.684127815610267,
+      "loss_partial": 0.061767578125,
+      "step": 3215
+    },
+    {
+      "epoch": 1.684127815610267,
+      "loss_partial": 0.052490234375,
+      "step": 3215
+    },
+    {
+      "epoch": 1.684651650078575,
+      "grad_norm": 1.3994485139846802,
+      "learning_rate": 1.4352365872083209e-05,
+      "loss": 0.1052,
+      "step": 3216
+    },
+    {
+      "epoch": 1.684651650078575,
+      "loss_partial": 0.047607421875,
+      "step": 3216
+    },
+    {
+      "epoch": 1.684651650078575,
+      "loss_partial": 0.047119140625,
+      "step": 3216
+    },
+    {
+      "epoch": 1.6851754845468832,
+      "grad_norm": 1.4930006265640259,
+      "learning_rate": 1.4348249546486984e-05,
+      "loss": 0.1039,
+      "step": 3217
+    },
+    {
+      "epoch": 1.6851754845468832,
+      "loss_partial": 0.0517578125,
+      "step": 3217
+    },
+    {
+      "epoch": 1.6851754845468832,
+      "loss_partial": 0.053466796875,
+      "step": 3217
+    },
+    {
+      "epoch": 1.6856993190151912,
+      "grad_norm": 1.490067481994629,
+      "learning_rate": 1.4344132312133489e-05,
+      "loss": 0.1053,
+      "step": 3218
+    },
+    {
+      "epoch": 1.6856993190151912,
+      "loss_partial": 0.044189453125,
+      "step": 3218
+    },
+    {
+      "epoch": 1.6856993190151912,
+      "loss_partial": 0.04736328125,
+      "step": 3218
+    },
+    {
+      "epoch": 1.6862231534834993,
+      "grad_norm": 1.6215474605560303,
+      "learning_rate": 1.4340014169883208e-05,
+      "loss": 0.1011,
+      "step": 3219
+    },
+    {
+      "epoch": 1.6862231534834993,
+      "loss_partial": 0.052001953125,
+      "step": 3219
+    },
+    {
+      "epoch": 1.6862231534834993,
+      "loss_partial": 0.055419921875,
+      "step": 3219
+    },
+    {
+      "epoch": 1.6867469879518073,
+      "grad_norm": 1.9268869161605835,
+      "learning_rate": 1.4335895120596801e-05,
+      "loss": 0.1079,
+      "step": 3220
+    },
+    {
+      "epoch": 1.6867469879518073,
+      "loss_partial": 0.04833984375,
+      "step": 3220
+    },
+    {
+      "epoch": 1.6867469879518073,
+      "loss_partial": 0.0498046875,
+      "step": 3220
+    },
+    {
+      "epoch": 1.6872708224201154,
+      "grad_norm": 1.7341039180755615,
+      "learning_rate": 1.4331775165135125e-05,
+      "loss": 0.1049,
+      "step": 3221
+    },
+    {
+      "epoch": 1.6872708224201154,
+      "loss_partial": 0.05908203125,
+      "step": 3221
+    },
+    {
+      "epoch": 1.6872708224201154,
+      "loss_partial": 0.053955078125,
+      "step": 3221
+    },
+    {
+      "epoch": 1.6877946568884232,
+      "grad_norm": 1.5418428182601929,
+      "learning_rate": 1.4327654304359229e-05,
+      "loss": 0.1063,
+      "step": 3222
+    },
+    {
+      "epoch": 1.6877946568884232,
+      "loss_partial": 0.064453125,
+      "step": 3222
+    },
+    {
+      "epoch": 1.6877946568884232,
+      "loss_partial": 0.06005859375,
+      "step": 3222
+    },
+    {
+      "epoch": 1.6883184913567313,
+      "grad_norm": 1.633489727973938,
+      "learning_rate": 1.4323532539130345e-05,
+      "loss": 0.1063,
+      "step": 3223
+    },
+    {
+      "epoch": 1.6883184913567313,
+      "loss_partial": 0.047119140625,
+      "step": 3223
+    },
+    {
+      "epoch": 1.6883184913567313,
+      "loss_partial": 0.052001953125,
+      "step": 3223
+    },
+    {
+      "epoch": 1.6888423258250391,
+      "grad_norm": 1.4026304483413696,
+      "learning_rate": 1.4319409870309896e-05,
+      "loss": 0.1039,
+      "step": 3224
+    },
+    {
+      "epoch": 1.6888423258250391,
+      "loss_partial": 0.05029296875,
+      "step": 3224
+    },
+    {
+      "epoch": 1.6888423258250391,
+      "loss_partial": 0.049072265625,
+      "step": 3224
+    },
+    {
+      "epoch": 1.6893661602933472,
+      "grad_norm": 1.8178961277008057,
+      "learning_rate": 1.4315286298759491e-05,
+      "loss": 0.1036,
+      "step": 3225
+    },
+    {
+      "epoch": 1.6893661602933472,
+      "loss_partial": 0.05224609375,
+      "step": 3225
+    },
+    {
+      "epoch": 1.6893661602933472,
+      "loss_partial": 0.045654296875,
+      "step": 3225
+    },
+    {
+      "epoch": 1.6898899947616552,
+      "grad_norm": 1.4212411642074585,
+      "learning_rate": 1.431116182534094e-05,
+      "loss": 0.1046,
+      "step": 3226
+    },
+    {
+      "epoch": 1.6898899947616552,
+      "loss_partial": 0.055908203125,
+      "step": 3226
+    },
+    {
+      "epoch": 1.6898899947616552,
+      "loss_partial": 0.050048828125,
+      "step": 3226
+    },
+    {
+      "epoch": 1.6904138292299633,
+      "grad_norm": 1.9806852340698242,
+      "learning_rate": 1.4307036450916223e-05,
+      "loss": 0.1087,
+      "step": 3227
+    },
+    {
+      "epoch": 1.6904138292299633,
+      "loss_partial": 0.044921875,
+      "step": 3227
+    },
+    {
+      "epoch": 1.6904138292299633,
+      "loss_partial": 0.044677734375,
+      "step": 3227
+    },
+    {
+      "epoch": 1.6909376636982714,
+      "grad_norm": 1.5569826364517212,
+      "learning_rate": 1.4302910176347522e-05,
+      "loss": 0.1039,
+      "step": 3228
+    },
+    {
+      "epoch": 1.6909376636982714,
+      "loss_partial": 0.046875,
+      "step": 3228
+    },
+    {
+      "epoch": 1.6909376636982714,
+      "loss_partial": 0.06103515625,
+      "step": 3228
+    },
+    {
+      "epoch": 1.6914614981665794,
+      "grad_norm": 1.8782984018325806,
+      "learning_rate": 1.4298783002497203e-05,
+      "loss": 0.107,
+      "step": 3229
+    },
+    {
+      "epoch": 1.6914614981665794,
+      "loss_partial": 0.052978515625,
+      "step": 3229
+    },
+    {
+      "epoch": 1.6914614981665794,
+      "loss_partial": 0.052001953125,
+      "step": 3229
+    },
+    {
+      "epoch": 1.6919853326348875,
+      "grad_norm": 1.8413364887237549,
+      "learning_rate": 1.4294654930227823e-05,
+      "loss": 0.1037,
+      "step": 3230
+    },
+    {
+      "epoch": 1.6919853326348875,
+      "loss_partial": 0.05419921875,
+      "step": 3230
+    },
+    {
+      "epoch": 1.6919853326348875,
+      "loss_partial": 0.056640625,
+      "step": 3230
+    },
+    {
+      "epoch": 1.6925091671031955,
+      "grad_norm": 1.9259735345840454,
+      "learning_rate": 1.4290525960402115e-05,
+      "loss": 0.1054,
+      "step": 3231
+    },
+    {
+      "epoch": 1.6925091671031955,
+      "loss_partial": 0.049072265625,
+      "step": 3231
+    },
+    {
+      "epoch": 1.6925091671031955,
+      "loss_partial": 0.047607421875,
+      "step": 3231
+    },
+    {
+      "epoch": 1.6930330015715034,
+      "grad_norm": 2.0866622924804688,
+      "learning_rate": 1.4286396093883014e-05,
+      "loss": 0.1011,
+      "step": 3232
+    },
+    {
+      "epoch": 1.6930330015715034,
+      "loss_partial": 0.05419921875,
+      "step": 3232
+    },
+    {
+      "epoch": 1.6930330015715034,
+      "loss_partial": 0.0498046875,
+      "step": 3232
+    },
+    {
+      "epoch": 1.6935568360398114,
+      "grad_norm": 1.4175066947937012,
+      "learning_rate": 1.4282265331533639e-05,
+      "loss": 0.1034,
+      "step": 3233
+    },
+    {
+      "epoch": 1.6935568360398114,
+      "loss_partial": 0.041748046875,
+      "step": 3233
+    },
+    {
+      "epoch": 1.6935568360398114,
+      "loss_partial": 0.055419921875,
+      "step": 3233
+    },
+    {
+      "epoch": 1.6940806705081193,
+      "grad_norm": 1.6577701568603516,
+      "learning_rate": 1.4278133674217286e-05,
+      "loss": 0.1071,
+      "step": 3234
+    },
+    {
+      "epoch": 1.6940806705081193,
+      "loss_partial": 0.052978515625,
+      "step": 3234
+    },
+    {
+      "epoch": 1.6940806705081193,
+      "loss_partial": 0.048583984375,
+      "step": 3234
+    },
+    {
+      "epoch": 1.6946045049764273,
+      "grad_norm": 1.7167961597442627,
+      "learning_rate": 1.4274001122797452e-05,
+      "loss": 0.1056,
+      "step": 3235
+    },
+    {
+      "epoch": 1.6946045049764273,
+      "loss_partial": 0.060546875,
+      "step": 3235
+    },
+    {
+      "epoch": 1.6946045049764273,
+      "loss_partial": 0.061279296875,
+      "step": 3235
+    },
+    {
+      "epoch": 1.6951283394447354,
+      "grad_norm": 1.5972479581832886,
+      "learning_rate": 1.4269867678137815e-05,
+      "loss": 0.1065,
+      "step": 3236
+    },
+    {
+      "epoch": 1.6951283394447354,
+      "loss_partial": 0.0537109375,
+      "step": 3236
+    },
+    {
+      "epoch": 1.6951283394447354,
+      "loss_partial": 0.04638671875,
+      "step": 3236
+    },
+    {
+      "epoch": 1.6956521739130435,
+      "grad_norm": 1.2812377214431763,
+      "learning_rate": 1.4265733341102235e-05,
+      "loss": 0.1018,
+      "step": 3237
+    },
+    {
+      "epoch": 1.6956521739130435,
+      "loss_partial": 0.05517578125,
+      "step": 3237
+    },
+    {
+      "epoch": 1.6956521739130435,
+      "loss_partial": 0.060302734375,
+      "step": 3237
+    },
+    {
+      "epoch": 1.6961760083813515,
+      "grad_norm": 1.5311816930770874,
+      "learning_rate": 1.4261598112554765e-05,
+      "loss": 0.1035,
+      "step": 3238
+    },
+    {
+      "epoch": 1.6961760083813515,
+      "loss_partial": 0.05029296875,
+      "step": 3238
+    },
+    {
+      "epoch": 1.6961760083813515,
+      "loss_partial": 0.0537109375,
+      "step": 3238
+    },
+    {
+      "epoch": 1.6966998428496596,
+      "grad_norm": 1.6605706214904785,
+      "learning_rate": 1.4257461993359643e-05,
+      "loss": 0.1021,
+      "step": 3239
+    },
+    {
+      "epoch": 1.6966998428496596,
+      "loss_partial": 0.0517578125,
+      "step": 3239
+    },
+    {
+      "epoch": 1.6966998428496596,
+      "loss_partial": 0.048095703125,
+      "step": 3239
+    },
+    {
+      "epoch": 1.6972236773179676,
+      "grad_norm": 1.8864448070526123,
+      "learning_rate": 1.4253324984381291e-05,
+      "loss": 0.1037,
+      "step": 3240
+    },
+    {
+      "epoch": 1.6972236773179676,
+      "loss_partial": 0.06103515625,
+      "step": 3240
+    },
+    {
+      "epoch": 1.6972236773179676,
+      "loss_partial": 0.04443359375,
+      "step": 3240
+    },
+    {
+      "epoch": 1.6977475117862757,
+      "grad_norm": 1.3077402114868164,
+      "learning_rate": 1.4249187086484319e-05,
+      "loss": 0.1046,
+      "step": 3241
+    },
+    {
+      "epoch": 1.6977475117862757,
+      "loss_partial": 0.05712890625,
+      "step": 3241
+    },
+    {
+      "epoch": 1.6977475117862757,
+      "loss_partial": 0.042724609375,
+      "step": 3241
+    },
+    {
+      "epoch": 1.6982713462545835,
+      "grad_norm": 1.8889269828796387,
+      "learning_rate": 1.4245048300533523e-05,
+      "loss": 0.1021,
+      "step": 3242
+    },
+    {
+      "epoch": 1.6982713462545835,
+      "loss_partial": 0.04833984375,
+      "step": 3242
+    },
+    {
+      "epoch": 1.6982713462545835,
+      "loss_partial": 0.054443359375,
+      "step": 3242
+    },
+    {
+      "epoch": 1.6987951807228916,
+      "grad_norm": 1.5355477333068848,
+      "learning_rate": 1.4240908627393885e-05,
+      "loss": 0.1067,
+      "step": 3243
+    },
+    {
+      "epoch": 1.6987951807228916,
+      "loss_partial": 0.048095703125,
+      "step": 3243
+    },
+    {
+      "epoch": 1.6987951807228916,
+      "loss_partial": 0.050048828125,
+      "step": 3243
+    },
+    {
+      "epoch": 1.6993190151911994,
+      "grad_norm": 1.9361810684204102,
+      "learning_rate": 1.4236768067930565e-05,
+      "loss": 0.1029,
+      "step": 3244
+    },
+    {
+      "epoch": 1.6993190151911994,
+      "loss_partial": 0.050048828125,
+      "step": 3244
+    },
+    {
+      "epoch": 1.6993190151911994,
+      "loss_partial": 0.046875,
+      "step": 3244
+    },
+    {
+      "epoch": 1.6998428496595075,
+      "grad_norm": 2.2787976264953613,
+      "learning_rate": 1.423262662300892e-05,
+      "loss": 0.1035,
+      "step": 3245
+    },
+    {
+      "epoch": 1.6998428496595075,
+      "loss_partial": 0.05419921875,
+      "step": 3245
+    },
+    {
+      "epoch": 1.6998428496595075,
+      "loss_partial": 0.0537109375,
+      "step": 3245
+    },
+    {
+      "epoch": 1.7003666841278156,
+      "grad_norm": 1.3690067529678345,
+      "learning_rate": 1.4228484293494482e-05,
+      "loss": 0.1039,
+      "step": 3246
+    },
+    {
+      "epoch": 1.7003666841278156,
+      "loss_partial": 0.05615234375,
+      "step": 3246
+    },
+    {
+      "epoch": 1.7003666841278156,
+      "loss_partial": 0.041748046875,
+      "step": 3246
+    },
+    {
+      "epoch": 1.7008905185961236,
+      "grad_norm": 1.9607667922973633,
+      "learning_rate": 1.422434108025298e-05,
+      "loss": 0.1073,
+      "step": 3247
+    },
+    {
+      "epoch": 1.7008905185961236,
+      "loss_partial": 0.044921875,
+      "step": 3247
+    },
+    {
+      "epoch": 1.7008905185961236,
+      "loss_partial": 0.050537109375,
+      "step": 3247
+    },
+    {
+      "epoch": 1.7014143530644317,
+      "grad_norm": 1.5728684663772583,
+      "learning_rate": 1.4220196984150311e-05,
+      "loss": 0.1049,
+      "step": 3248
+    },
+    {
+      "epoch": 1.7014143530644317,
+      "loss_partial": 0.045654296875,
+      "step": 3248
+    },
+    {
+      "epoch": 1.7014143530644317,
+      "loss_partial": 0.04736328125,
+      "step": 3248
+    },
+    {
+      "epoch": 1.7019381875327397,
+      "grad_norm": 2.2740113735198975,
+      "learning_rate": 1.4216052006052572e-05,
+      "loss": 0.1033,
+      "step": 3249
+    },
+    {
+      "epoch": 1.7019381875327397,
+      "loss_partial": 0.050048828125,
+      "step": 3249
+    },
+    {
+      "epoch": 1.7019381875327397,
+      "loss_partial": 0.051025390625,
+      "step": 3249
+    },
+    {
+      "epoch": 1.7024620220010478,
+      "grad_norm": 1.507567286491394,
+      "learning_rate": 1.4211906146826031e-05,
+      "loss": 0.107,
+      "step": 3250
+    },
+    {
+      "epoch": 1.7024620220010478,
+      "loss_partial": 0.041748046875,
+      "step": 3250
+    },
+    {
+      "epoch": 1.7024620220010478,
+      "loss_partial": 0.06689453125,
+      "step": 3250
+    },
+    {
+      "epoch": 1.7029858564693556,
+      "grad_norm": 2.0596659183502197,
+      "learning_rate": 1.4207759407337154e-05,
+      "loss": 0.1064,
+      "step": 3251
+    },
+    {
+      "epoch": 1.7029858564693556,
+      "loss_partial": 0.0478515625,
+      "step": 3251
+    },
+    {
+      "epoch": 1.7029858564693556,
+      "loss_partial": 0.0576171875,
+      "step": 3251
+    },
+    {
+      "epoch": 1.7035096909376637,
+      "grad_norm": 1.823217511177063,
+      "learning_rate": 1.4203611788452578e-05,
+      "loss": 0.1054,
+      "step": 3252
+    },
+    {
+      "epoch": 1.7035096909376637,
+      "loss_partial": 0.052978515625,
+      "step": 3252
+    },
+    {
+      "epoch": 1.7035096909376637,
+      "loss_partial": 0.0576171875,
+      "step": 3252
+    },
+    {
+      "epoch": 1.7040335254059717,
+      "grad_norm": 1.687685251235962,
+      "learning_rate": 1.4199463291039136e-05,
+      "loss": 0.1069,
+      "step": 3253
+    },
+    {
+      "epoch": 1.7040335254059717,
+      "loss_partial": 0.051513671875,
+      "step": 3253
+    },
+    {
+      "epoch": 1.7040335254059717,
+      "loss_partial": 0.0546875,
+      "step": 3253
+    },
+    {
+      "epoch": 1.7045573598742796,
+      "grad_norm": 1.546135663986206,
+      "learning_rate": 1.4195313915963835e-05,
+      "loss": 0.1052,
+      "step": 3254
+    },
+    {
+      "epoch": 1.7045573598742796,
+      "loss_partial": 0.0537109375,
+      "step": 3254
+    },
+    {
+      "epoch": 1.7045573598742796,
+      "loss_partial": 0.050048828125,
+      "step": 3254
+    },
+    {
+      "epoch": 1.7050811943425876,
+      "grad_norm": 2.1646199226379395,
+      "learning_rate": 1.4191163664093867e-05,
+      "loss": 0.1058,
+      "step": 3255
+    },
+    {
+      "epoch": 1.7050811943425876,
+      "loss_partial": 0.046142578125,
+      "step": 3255
+    },
+    {
+      "epoch": 1.7050811943425876,
+      "loss_partial": 0.053466796875,
+      "step": 3255
+    },
+    {
+      "epoch": 1.7056050288108957,
+      "grad_norm": 2.1421475410461426,
+      "learning_rate": 1.418701253629661e-05,
+      "loss": 0.1033,
+      "step": 3256
+    },
+    {
+      "epoch": 1.7056050288108957,
+      "loss_partial": 0.045654296875,
+      "step": 3256
+    },
+    {
+      "epoch": 1.7056050288108957,
+      "loss_partial": 0.056640625,
+      "step": 3256
+    },
+    {
+      "epoch": 1.7061288632792038,
+      "grad_norm": 1.6567481756210327,
+      "learning_rate": 1.4182860533439624e-05,
+      "loss": 0.1029,
+      "step": 3257
+    },
+    {
+      "epoch": 1.7061288632792038,
+      "loss_partial": 0.041748046875,
+      "step": 3257
+    },
+    {
+      "epoch": 1.7061288632792038,
+      "loss_partial": 0.051513671875,
+      "step": 3257
+    },
+    {
+      "epoch": 1.7066526977475118,
+      "grad_norm": 1.5575780868530273,
+      "learning_rate": 1.4178707656390653e-05,
+      "loss": 0.1064,
+      "step": 3258
+    },
+    {
+      "epoch": 1.7066526977475118,
+      "loss_partial": 0.057861328125,
+      "step": 3258
+    },
+    {
+      "epoch": 1.7066526977475118,
+      "loss_partial": 0.044921875,
+      "step": 3258
+    },
+    {
+      "epoch": 1.7071765322158199,
+      "grad_norm": 1.6875563859939575,
+      "learning_rate": 1.4174553906017626e-05,
+      "loss": 0.1038,
+      "step": 3259
+    },
+    {
+      "epoch": 1.7071765322158199,
+      "loss_partial": 0.057861328125,
+      "step": 3259
+    },
+    {
+      "epoch": 1.7071765322158199,
+      "loss_partial": 0.05517578125,
+      "step": 3259
+    },
+    {
+      "epoch": 1.707700366684128,
+      "grad_norm": 1.543225884437561,
+      "learning_rate": 1.4170399283188643e-05,
+      "loss": 0.1053,
+      "step": 3260
+    },
+    {
+      "epoch": 1.707700366684128,
+      "loss_partial": 0.04833984375,
+      "step": 3260
+    },
+    {
+      "epoch": 1.707700366684128,
+      "loss_partial": 0.048583984375,
+      "step": 3260
+    },
+    {
+      "epoch": 1.7082242011524358,
+      "grad_norm": 1.5136460065841675,
+      "learning_rate": 1.4166243788772001e-05,
+      "loss": 0.1019,
+      "step": 3261
+    },
+    {
+      "epoch": 1.7082242011524358,
+      "loss_partial": 0.050048828125,
+      "step": 3261
+    },
+    {
+      "epoch": 1.7082242011524358,
+      "loss_partial": 0.05322265625,
+      "step": 3261
+    },
+    {
+      "epoch": 1.7087480356207438,
+      "grad_norm": 1.861923098564148,
+      "learning_rate": 1.4162087423636168e-05,
+      "loss": 0.1026,
+      "step": 3262
+    },
+    {
+      "epoch": 1.7087480356207438,
+      "loss_partial": 0.052734375,
+      "step": 3262
+    },
+    {
+      "epoch": 1.7087480356207438,
+      "loss_partial": 0.05615234375,
+      "step": 3262
+    },
+    {
+      "epoch": 1.709271870089052,
+      "grad_norm": 1.54845130443573,
+      "learning_rate": 1.4157930188649802e-05,
+      "loss": 0.1052,
+      "step": 3263
+    },
+    {
+      "epoch": 1.709271870089052,
+      "loss_partial": 0.053955078125,
+      "step": 3263
+    },
+    {
+      "epoch": 1.709271870089052,
+      "loss_partial": 0.050048828125,
+      "step": 3263
+    },
+    {
+      "epoch": 1.7097957045573597,
+      "grad_norm": 2.007089853286743,
+      "learning_rate": 1.4153772084681738e-05,
+      "loss": 0.1024,
+      "step": 3264
+    },
+    {
+      "epoch": 1.7097957045573597,
+      "loss_partial": 0.04736328125,
+      "step": 3264
+    },
+    {
+      "epoch": 1.7097957045573597,
+      "loss_partial": 0.050048828125,
+      "step": 3264
+    },
+    {
+      "epoch": 1.7103195390256678,
+      "grad_norm": 1.812349796295166,
+      "learning_rate": 1.4149613112600995e-05,
+      "loss": 0.1031,
+      "step": 3265
+    },
+    {
+      "epoch": 1.7103195390256678,
+      "loss_partial": 0.05517578125,
+      "step": 3265
+    },
+    {
+      "epoch": 1.7103195390256678,
+      "loss_partial": 0.051513671875,
+      "step": 3265
+    },
+    {
+      "epoch": 1.7108433734939759,
+      "grad_norm": 2.263056755065918,
+      "learning_rate": 1.4145453273276775e-05,
+      "loss": 0.1046,
+      "step": 3266
+    },
+    {
+      "epoch": 1.7108433734939759,
+      "loss_partial": 0.052978515625,
+      "step": 3266
+    },
+    {
+      "epoch": 1.7108433734939759,
+      "loss_partial": 0.04638671875,
+      "step": 3266
+    },
+    {
+      "epoch": 1.711367207962284,
+      "grad_norm": 1.6113723516464233,
+      "learning_rate": 1.4141292567578453e-05,
+      "loss": 0.1051,
+      "step": 3267
+    },
+    {
+      "epoch": 1.711367207962284,
+      "loss_partial": 0.059326171875,
+      "step": 3267
+    },
+    {
+      "epoch": 1.711367207962284,
+      "loss_partial": 0.050537109375,
+      "step": 3267
+    },
+    {
+      "epoch": 1.711891042430592,
+      "grad_norm": 1.653840184211731,
+      "learning_rate": 1.4137130996375592e-05,
+      "loss": 0.1045,
+      "step": 3268
+    },
+    {
+      "epoch": 1.711891042430592,
+      "loss_partial": 0.06005859375,
+      "step": 3268
+    },
+    {
+      "epoch": 1.711891042430592,
+      "loss_partial": 0.05029296875,
+      "step": 3268
+    },
+    {
+      "epoch": 1.7124148768989,
+      "grad_norm": 1.0648510456085205,
+      "learning_rate": 1.4132968560537937e-05,
+      "loss": 0.1038,
+      "step": 3269
+    },
+    {
+      "epoch": 1.7124148768989,
+      "loss_partial": 0.058349609375,
+      "step": 3269
+    },
+    {
+      "epoch": 1.7124148768989,
+      "loss_partial": 0.05029296875,
+      "step": 3269
+    },
+    {
+      "epoch": 1.712938711367208,
+      "grad_norm": 1.9634754657745361,
+      "learning_rate": 1.4128805260935409e-05,
+      "loss": 0.1058,
+      "step": 3270
+    },
+    {
+      "epoch": 1.712938711367208,
+      "loss_partial": 0.05029296875,
+      "step": 3270
+    },
+    {
+      "epoch": 1.712938711367208,
+      "loss_partial": 0.054443359375,
+      "step": 3270
+    },
+    {
+      "epoch": 1.713462545835516,
+      "grad_norm": 1.5226141214370728,
+      "learning_rate": 1.4124641098438116e-05,
+      "loss": 0.1097,
+      "step": 3271
+    },
+    {
+      "epoch": 1.713462545835516,
+      "loss_partial": 0.057373046875,
+      "step": 3271
+    },
+    {
+      "epoch": 1.713462545835516,
+      "loss_partial": 0.05322265625,
+      "step": 3271
+    },
+    {
+      "epoch": 1.713986380303824,
+      "grad_norm": 1.3709218502044678,
+      "learning_rate": 1.4120476073916338e-05,
+      "loss": 0.1043,
+      "step": 3272
+    },
+    {
+      "epoch": 1.713986380303824,
+      "loss_partial": 0.05029296875,
+      "step": 3272
+    },
+    {
+      "epoch": 1.713986380303824,
+      "loss_partial": 0.052490234375,
+      "step": 3272
+    },
+    {
+      "epoch": 1.714510214772132,
+      "grad_norm": 1.498529076576233,
+      "learning_rate": 1.4116310188240543e-05,
+      "loss": 0.1038,
+      "step": 3273
+    },
+    {
+      "epoch": 1.714510214772132,
+      "loss_partial": 0.045166015625,
+      "step": 3273
+    },
+    {
+      "epoch": 1.714510214772132,
+      "loss_partial": 0.05078125,
+      "step": 3273
+    },
+    {
+      "epoch": 1.71503404924044,
+      "grad_norm": 1.8797979354858398,
+      "learning_rate": 1.411214344228137e-05,
+      "loss": 0.1027,
+      "step": 3274
+    },
+    {
+      "epoch": 1.71503404924044,
+      "loss_partial": 0.055419921875,
+      "step": 3274
+    },
+    {
+      "epoch": 1.71503404924044,
+      "loss_partial": 0.05810546875,
+      "step": 3274
+    },
+    {
+      "epoch": 1.715557883708748,
+      "grad_norm": 1.5051801204681396,
+      "learning_rate": 1.410797583690965e-05,
+      "loss": 0.1049,
+      "step": 3275
+    },
+    {
+      "epoch": 1.715557883708748,
+      "loss_partial": 0.047607421875,
+      "step": 3275
+    },
+    {
+      "epoch": 1.715557883708748,
+      "loss_partial": 0.057861328125,
+      "step": 3275
+    },
+    {
+      "epoch": 1.716081718177056,
+      "grad_norm": 1.9681791067123413,
+      "learning_rate": 1.4103807372996377e-05,
+      "loss": 0.1042,
+      "step": 3276
+    },
+    {
+      "epoch": 1.716081718177056,
+      "loss_partial": 0.04931640625,
+      "step": 3276
+    },
+    {
+      "epoch": 1.716081718177056,
+      "loss_partial": 0.053466796875,
+      "step": 3276
+    },
+    {
+      "epoch": 1.716605552645364,
+      "grad_norm": 2.135143756866455,
+      "learning_rate": 1.4099638051412745e-05,
+      "loss": 0.108,
+      "step": 3277
+    },
+    {
+      "epoch": 1.716605552645364,
+      "loss_partial": 0.05712890625,
+      "step": 3277
+    },
+    {
+      "epoch": 1.716605552645364,
+      "loss_partial": 0.064453125,
+      "step": 3277
+    },
+    {
+      "epoch": 1.7171293871136721,
+      "grad_norm": 2.144109010696411,
+      "learning_rate": 1.409546787303011e-05,
+      "loss": 0.1034,
+      "step": 3278
+    },
+    {
+      "epoch": 1.7171293871136721,
+      "loss_partial": 0.048583984375,
+      "step": 3278
+    },
+    {
+      "epoch": 1.7171293871136721,
+      "loss_partial": 0.05078125,
+      "step": 3278
+    },
+    {
+      "epoch": 1.7176532215819802,
+      "grad_norm": 1.304150104522705,
+      "learning_rate": 1.4091296838720016e-05,
+      "loss": 0.103,
+      "step": 3279
+    },
+    {
+      "epoch": 1.7176532215819802,
+      "loss_partial": 0.050048828125,
+      "step": 3279
+    },
+    {
+      "epoch": 1.7176532215819802,
+      "loss_partial": 0.0615234375,
+      "step": 3279
+    },
+    {
+      "epoch": 1.7181770560502883,
+      "grad_norm": 1.5226199626922607,
+      "learning_rate": 1.408712494935418e-05,
+      "loss": 0.1034,
+      "step": 3280
+    },
+    {
+      "epoch": 1.7181770560502883,
+      "loss_partial": 0.046875,
+      "step": 3280
+    },
+    {
+      "epoch": 1.7181770560502883,
+      "loss_partial": 0.05224609375,
+      "step": 3280
+    },
+    {
+      "epoch": 1.718700890518596,
+      "grad_norm": 1.4711381196975708,
+      "learning_rate": 1.4082952205804507e-05,
+      "loss": 0.1051,
+      "step": 3281
+    },
+    {
+      "epoch": 1.718700890518596,
+      "loss_partial": 0.054443359375,
+      "step": 3281
+    },
+    {
+      "epoch": 1.718700890518596,
+      "loss_partial": 0.058837890625,
+      "step": 3281
+    },
+    {
+      "epoch": 1.7192247249869042,
+      "grad_norm": 1.7292104959487915,
+      "learning_rate": 1.4078778608943071e-05,
+      "loss": 0.1027,
+      "step": 3282
+    },
+    {
+      "epoch": 1.7192247249869042,
+      "loss_partial": 0.05126953125,
+      "step": 3282
+    },
+    {
+      "epoch": 1.7192247249869042,
+      "loss_partial": 0.04443359375,
+      "step": 3282
+    },
+    {
+      "epoch": 1.719748559455212,
+      "grad_norm": 1.1870197057724,
+      "learning_rate": 1.4074604159642127e-05,
+      "loss": 0.1056,
+      "step": 3283
+    },
+    {
+      "epoch": 1.719748559455212,
+      "loss_partial": 0.07666015625,
+      "step": 3283
+    },
+    {
+      "epoch": 1.719748559455212,
+      "loss_partial": 0.0478515625,
+      "step": 3283
+    },
+    {
+      "epoch": 1.72027239392352,
+      "grad_norm": 1.9334920644760132,
+      "learning_rate": 1.407042885877411e-05,
+      "loss": 0.1098,
+      "step": 3284
+    },
+    {
+      "epoch": 1.72027239392352,
+      "loss_partial": 0.0546875,
+      "step": 3284
+    },
+    {
+      "epoch": 1.72027239392352,
+      "loss_partial": 0.06396484375,
+      "step": 3284
+    },
+    {
+      "epoch": 1.720796228391828,
+      "grad_norm": 1.4889740943908691,
+      "learning_rate": 1.4066252707211635e-05,
+      "loss": 0.1088,
+      "step": 3285
+    },
+    {
+      "epoch": 1.720796228391828,
+      "loss_partial": 0.052978515625,
+      "step": 3285
+    },
+    {
+      "epoch": 1.720796228391828,
+      "loss_partial": 0.0498046875,
+      "step": 3285
+    },
+    {
+      "epoch": 1.7213200628601362,
+      "grad_norm": 1.4193400144577026,
+      "learning_rate": 1.4062075705827485e-05,
+      "loss": 0.1089,
+      "step": 3286
+    },
+    {
+      "epoch": 1.7213200628601362,
+      "loss_partial": 0.0517578125,
+      "step": 3286
+    },
+    {
+      "epoch": 1.7213200628601362,
+      "loss_partial": 0.055908203125,
+      "step": 3286
+    },
+    {
+      "epoch": 1.7218438973284442,
+      "grad_norm": 1.4402977228164673,
+      "learning_rate": 1.4057897855494633e-05,
+      "loss": 0.1055,
+      "step": 3287
+    },
+    {
+      "epoch": 1.7218438973284442,
+      "loss_partial": 0.039794921875,
+      "step": 3287
+    },
+    {
+      "epoch": 1.7218438973284442,
+      "loss_partial": 0.0439453125,
+      "step": 3287
+    },
+    {
+      "epoch": 1.7223677317967523,
+      "grad_norm": 1.6536791324615479,
+      "learning_rate": 1.4053719157086223e-05,
+      "loss": 0.1027,
+      "step": 3288
+    },
+    {
+      "epoch": 1.7223677317967523,
+      "loss_partial": 0.056640625,
+      "step": 3288
+    },
+    {
+      "epoch": 1.7223677317967523,
+      "loss_partial": 0.048095703125,
+      "step": 3288
+    },
+    {
+      "epoch": 1.7228915662650603,
+      "grad_norm": 1.320971131324768,
+      "learning_rate": 1.4049539611475575e-05,
+      "loss": 0.1099,
+      "step": 3289
+    },
+    {
+      "epoch": 1.7228915662650603,
+      "loss_partial": 0.053466796875,
+      "step": 3289
+    },
+    {
+      "epoch": 1.7228915662650603,
+      "loss_partial": 0.052978515625,
+      "step": 3289
+    },
+    {
+      "epoch": 1.7234154007333684,
+      "grad_norm": 1.7705793380737305,
+      "learning_rate": 1.4045359219536192e-05,
+      "loss": 0.1071,
+      "step": 3290
+    },
+    {
+      "epoch": 1.7234154007333684,
+      "loss_partial": 0.056884765625,
+      "step": 3290
+    },
+    {
+      "epoch": 1.7234154007333684,
+      "loss_partial": 0.043212890625,
+      "step": 3290
+    },
+    {
+      "epoch": 1.7239392352016762,
+      "grad_norm": 1.3032792806625366,
+      "learning_rate": 1.4041177982141747e-05,
+      "loss": 0.1043,
+      "step": 3291
+    },
+    {
+      "epoch": 1.7239392352016762,
+      "loss_partial": 0.048583984375,
+      "step": 3291
+    },
+    {
+      "epoch": 1.7239392352016762,
+      "loss_partial": 0.052978515625,
+      "step": 3291
+    },
+    {
+      "epoch": 1.7244630696699843,
+      "grad_norm": 1.3347338438034058,
+      "learning_rate": 1.4036995900166095e-05,
+      "loss": 0.1067,
+      "step": 3292
+    },
+    {
+      "epoch": 1.7244630696699843,
+      "loss_partial": 0.05224609375,
+      "step": 3292
+    },
+    {
+      "epoch": 1.7244630696699843,
+      "loss_partial": 0.054931640625,
+      "step": 3292
+    },
+    {
+      "epoch": 1.7249869041382921,
+      "grad_norm": 1.480056881904602,
+      "learning_rate": 1.4032812974483262e-05,
+      "loss": 0.1065,
+      "step": 3293
+    },
+    {
+      "epoch": 1.7249869041382921,
+      "loss_partial": 0.0546875,
+      "step": 3293
+    },
+    {
+      "epoch": 1.7249869041382921,
+      "loss_partial": 0.05029296875,
+      "step": 3293
+    },
+    {
+      "epoch": 1.7255107386066002,
+      "grad_norm": 1.3861596584320068,
+      "learning_rate": 1.4028629205967456e-05,
+      "loss": 0.1039,
+      "step": 3294
+    },
+    {
+      "epoch": 1.7255107386066002,
+      "loss_partial": 0.053466796875,
+      "step": 3294
+    },
+    {
+      "epoch": 1.7255107386066002,
+      "loss_partial": 0.04443359375,
+      "step": 3294
+    },
+    {
+      "epoch": 1.7260345730749083,
+      "grad_norm": 1.7306983470916748,
+      "learning_rate": 1.4024444595493059e-05,
+      "loss": 0.1026,
+      "step": 3295
+    },
+    {
+      "epoch": 1.7260345730749083,
+      "loss_partial": 0.0673828125,
+      "step": 3295
+    },
+    {
+      "epoch": 1.7260345730749083,
+      "loss_partial": 0.04736328125,
+      "step": 3295
+    },
+    {
+      "epoch": 1.7265584075432163,
+      "grad_norm": 2.1242408752441406,
+      "learning_rate": 1.4020259143934626e-05,
+      "loss": 0.1003,
+      "step": 3296
+    },
+    {
+      "epoch": 1.7265584075432163,
+      "loss_partial": 0.0517578125,
+      "step": 3296
+    },
+    {
+      "epoch": 1.7265584075432163,
+      "loss_partial": 0.044189453125,
+      "step": 3296
+    },
+    {
+      "epoch": 1.7270822420115244,
+      "grad_norm": 1.8691836595535278,
+      "learning_rate": 1.4016072852166895e-05,
+      "loss": 0.1042,
+      "step": 3297
+    },
+    {
+      "epoch": 1.7270822420115244,
+      "loss_partial": 0.054931640625,
+      "step": 3297
+    },
+    {
+      "epoch": 1.7270822420115244,
+      "loss_partial": 0.051025390625,
+      "step": 3297
+    },
+    {
+      "epoch": 1.7276060764798324,
+      "grad_norm": 3.2595436573028564,
+      "learning_rate": 1.4011885721064773e-05,
+      "loss": 0.106,
+      "step": 3298
+    },
+    {
+      "epoch": 1.7276060764798324,
+      "loss_partial": 0.050048828125,
+      "step": 3298
+    },
+    {
+      "epoch": 1.7276060764798324,
+      "loss_partial": 0.04931640625,
+      "step": 3298
+    },
+    {
+      "epoch": 1.7281299109481405,
+      "grad_norm": 2.024649143218994,
+      "learning_rate": 1.400769775150334e-05,
+      "loss": 0.1033,
+      "step": 3299
+    },
+    {
+      "epoch": 1.7281299109481405,
+      "loss_partial": 0.05712890625,
+      "step": 3299
+    },
+    {
+      "epoch": 1.7281299109481405,
+      "loss_partial": 0.05224609375,
+      "step": 3299
+    },
+    {
+      "epoch": 1.7286537454164486,
+      "grad_norm": 1.9246351718902588,
+      "learning_rate": 1.4003508944357864e-05,
+      "loss": 0.1031,
+      "step": 3300
+    },
+    {
+      "epoch": 1.7286537454164486,
+      "loss_partial": 0.057861328125,
+      "step": 3300
+    },
+    {
+      "epoch": 1.7286537454164486,
+      "loss_partial": 0.053466796875,
+      "step": 3300
+    },
+    {
+      "epoch": 1.7291775798847564,
+      "grad_norm": 1.6110477447509766,
+      "learning_rate": 1.3999319300503771e-05,
+      "loss": 0.1039,
+      "step": 3301
+    },
+    {
+      "epoch": 1.7291775798847564,
+      "loss_partial": 0.0556640625,
+      "step": 3301
+    },
+    {
+      "epoch": 1.7291775798847564,
+      "loss_partial": 0.053466796875,
+      "step": 3301
+    },
+    {
+      "epoch": 1.7297014143530645,
+      "grad_norm": 1.6944396495819092,
+      "learning_rate": 1.3995128820816679e-05,
+      "loss": 0.1069,
+      "step": 3302
+    },
+    {
+      "epoch": 1.7297014143530645,
+      "loss_partial": 0.041015625,
+      "step": 3302
+    },
+    {
+      "epoch": 1.7297014143530645,
+      "loss_partial": 0.042724609375,
+      "step": 3302
+    },
+    {
+      "epoch": 1.7302252488213723,
+      "grad_norm": 1.6804496049880981,
+      "learning_rate": 1.3990937506172369e-05,
+      "loss": 0.1059,
+      "step": 3303
+    },
+    {
+      "epoch": 1.7302252488213723,
+      "loss_partial": 0.052001953125,
+      "step": 3303
+    },
+    {
+      "epoch": 1.7302252488213723,
+      "loss_partial": 0.050537109375,
+      "step": 3303
+    },
+    {
+      "epoch": 1.7307490832896804,
+      "grad_norm": 1.9076529741287231,
+      "learning_rate": 1.3986745357446798e-05,
+      "loss": 0.1091,
+      "step": 3304
+    },
+    {
+      "epoch": 1.7307490832896804,
+      "loss_partial": 0.052490234375,
+      "step": 3304
+    },
+    {
+      "epoch": 1.7307490832896804,
+      "loss_partial": 0.0576171875,
+      "step": 3304
+    },
+    {
+      "epoch": 1.7312729177579884,
+      "grad_norm": 2.07802677154541,
+      "learning_rate": 1.3982552375516098e-05,
+      "loss": 0.1043,
+      "step": 3305
+    },
+    {
+      "epoch": 1.7312729177579884,
+      "loss_partial": 0.046875,
+      "step": 3305
+    },
+    {
+      "epoch": 1.7312729177579884,
+      "loss_partial": 0.05322265625,
+      "step": 3305
+    },
+    {
+      "epoch": 1.7317967522262965,
+      "grad_norm": 1.5468494892120361,
+      "learning_rate": 1.3978358561256582e-05,
+      "loss": 0.1052,
+      "step": 3306
+    },
+    {
+      "epoch": 1.7317967522262965,
+      "loss_partial": 0.051513671875,
+      "step": 3306
+    },
+    {
+      "epoch": 1.7317967522262965,
+      "loss_partial": 0.055419921875,
+      "step": 3306
+    },
+    {
+      "epoch": 1.7323205866946045,
+      "grad_norm": 2.082845687866211,
+      "learning_rate": 1.3974163915544727e-05,
+      "loss": 0.1035,
+      "step": 3307
+    },
+    {
+      "epoch": 1.7323205866946045,
+      "loss_partial": 0.04443359375,
+      "step": 3307
+    },
+    {
+      "epoch": 1.7323205866946045,
+      "loss_partial": 0.053466796875,
+      "step": 3307
+    },
+    {
+      "epoch": 1.7328444211629126,
+      "grad_norm": 1.7265700101852417,
+      "learning_rate": 1.3969968439257191e-05,
+      "loss": 0.105,
+      "step": 3308
+    },
+    {
+      "epoch": 1.7328444211629126,
+      "loss_partial": 0.0634765625,
+      "step": 3308
+    },
+    {
+      "epoch": 1.7328444211629126,
+      "loss_partial": 0.05078125,
+      "step": 3308
+    },
+    {
+      "epoch": 1.7333682556312207,
+      "grad_norm": 1.9086679220199585,
+      "learning_rate": 1.3965772133270798e-05,
+      "loss": 0.1088,
+      "step": 3309
+    },
+    {
+      "epoch": 1.7333682556312207,
+      "loss_partial": 0.0537109375,
+      "step": 3309
+    },
+    {
+      "epoch": 1.7333682556312207,
+      "loss_partial": 0.052490234375,
+      "step": 3309
+    },
+    {
+      "epoch": 1.7338920900995285,
+      "grad_norm": 1.8519335985183716,
+      "learning_rate": 1.3961574998462558e-05,
+      "loss": 0.1044,
+      "step": 3310
+    },
+    {
+      "epoch": 1.7338920900995285,
+      "loss_partial": 0.05029296875,
+      "step": 3310
+    },
+    {
+      "epoch": 1.7338920900995285,
+      "loss_partial": 0.06494140625,
+      "step": 3310
+    },
+    {
+      "epoch": 1.7344159245678366,
+      "grad_norm": 1.6477057933807373,
+      "learning_rate": 1.3957377035709633e-05,
+      "loss": 0.103,
+      "step": 3311
+    },
+    {
+      "epoch": 1.7344159245678366,
+      "loss_partial": 0.048095703125,
+      "step": 3311
+    },
+    {
+      "epoch": 1.7344159245678366,
+      "loss_partial": 0.052001953125,
+      "step": 3311
+    },
+    {
+      "epoch": 1.7349397590361446,
+      "grad_norm": 1.6582473516464233,
+      "learning_rate": 1.3953178245889386e-05,
+      "loss": 0.1044,
+      "step": 3312
+    },
+    {
+      "epoch": 1.7349397590361446,
+      "loss_partial": 0.052490234375,
+      "step": 3312
+    },
+    {
+      "epoch": 1.7349397590361446,
+      "loss_partial": 0.044921875,
+      "step": 3312
+    },
+    {
+      "epoch": 1.7354635935044525,
+      "grad_norm": 1.8321303129196167,
+      "learning_rate": 1.3948978629879327e-05,
+      "loss": 0.1029,
+      "step": 3313
+    },
+    {
+      "epoch": 1.7354635935044525,
+      "loss_partial": 0.052978515625,
+      "step": 3313
+    },
+    {
+      "epoch": 1.7354635935044525,
+      "loss_partial": 0.048828125,
+      "step": 3313
+    },
+    {
+      "epoch": 1.7359874279727605,
+      "grad_norm": 1.9915637969970703,
+      "learning_rate": 1.3944778188557156e-05,
+      "loss": 0.1035,
+      "step": 3314
+    },
+    {
+      "epoch": 1.7359874279727605,
+      "loss_partial": 0.0458984375,
+      "step": 3314
+    },
+    {
+      "epoch": 1.7359874279727605,
+      "loss_partial": 0.05126953125,
+      "step": 3314
+    },
+    {
+      "epoch": 1.7365112624410686,
+      "grad_norm": 2.443587303161621,
+      "learning_rate": 1.394057692280074e-05,
+      "loss": 0.1031,
+      "step": 3315
+    },
+    {
+      "epoch": 1.7365112624410686,
+      "loss_partial": 0.057861328125,
+      "step": 3315
+    },
+    {
+      "epoch": 1.7365112624410686,
+      "loss_partial": 0.0419921875,
+      "step": 3315
+    },
+    {
+      "epoch": 1.7370350969093766,
+      "grad_norm": 1.565863013267517,
+      "learning_rate": 1.3936374833488111e-05,
+      "loss": 0.1003,
+      "step": 3316
+    },
+    {
+      "epoch": 1.7370350969093766,
+      "loss_partial": 0.05517578125,
+      "step": 3316
+    },
+    {
+      "epoch": 1.7370350969093766,
+      "loss_partial": 0.06884765625,
+      "step": 3316
+    },
+    {
+      "epoch": 1.7375589313776847,
+      "grad_norm": 2.845791816711426,
+      "learning_rate": 1.3932171921497483e-05,
+      "loss": 0.1078,
+      "step": 3317
+    },
+    {
+      "epoch": 1.7375589313776847,
+      "loss_partial": 0.0615234375,
+      "step": 3317
+    },
+    {
+      "epoch": 1.7375589313776847,
+      "loss_partial": 0.053466796875,
+      "step": 3317
+    },
+    {
+      "epoch": 1.7380827658459927,
+      "grad_norm": 1.957029938697815,
+      "learning_rate": 1.3927968187707241e-05,
+      "loss": 0.1057,
+      "step": 3318
+    },
+    {
+      "epoch": 1.7380827658459927,
+      "loss_partial": 0.051513671875,
+      "step": 3318
+    },
+    {
+      "epoch": 1.7380827658459927,
+      "loss_partial": 0.042724609375,
+      "step": 3318
+    },
+    {
+      "epoch": 1.7386066003143008,
+      "grad_norm": 2.3660988807678223,
+      "learning_rate": 1.3923763632995937e-05,
+      "loss": 0.1039,
+      "step": 3319
+    },
+    {
+      "epoch": 1.7386066003143008,
+      "loss_partial": 0.044189453125,
+      "step": 3319
+    },
+    {
+      "epoch": 1.7386066003143008,
+      "loss_partial": 0.049072265625,
+      "step": 3319
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "grad_norm": 1.4311447143554688,
+      "learning_rate": 1.3919558258242299e-05,
+      "loss": 0.101,
+      "step": 3320
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "loss_partial": 0.048828125,
+      "step": 3320
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "loss_partial": 0.059326171875,
+      "step": 3320
+    },
+    {
+      "epoch": 1.7396542692509167,
+      "grad_norm": 2.3004071712493896,
+      "learning_rate": 1.3915352064325223e-05,
+      "loss": 0.1076,
+      "step": 3321
+    },
+    {
+      "epoch": 1.7396542692509167,
+      "loss_partial": 0.049560546875,
+      "step": 3321
+    },
+    {
+      "epoch": 1.7396542692509167,
+      "loss_partial": 0.04248046875,
+      "step": 3321
+    },
+    {
+      "epoch": 1.7401781037192248,
+      "grad_norm": 2.1898844242095947,
+      "learning_rate": 1.3911145052123778e-05,
+      "loss": 0.1021,
+      "step": 3322
+    },
+    {
+      "epoch": 1.7401781037192248,
+      "loss_partial": 0.048095703125,
+      "step": 3322
+    },
+    {
+      "epoch": 1.7401781037192248,
+      "loss_partial": 0.057373046875,
+      "step": 3322
+    },
+    {
+      "epoch": 1.7407019381875326,
+      "grad_norm": 1.8550668954849243,
+      "learning_rate": 1.3906937222517199e-05,
+      "loss": 0.1026,
+      "step": 3323
+    },
+    {
+      "epoch": 1.7407019381875326,
+      "loss_partial": 0.05712890625,
+      "step": 3323
+    },
+    {
+      "epoch": 1.7407019381875326,
+      "loss_partial": 0.0576171875,
+      "step": 3323
+    },
+    {
+      "epoch": 1.7412257726558407,
+      "grad_norm": 1.8363721370697021,
+      "learning_rate": 1.3902728576384902e-05,
+      "loss": 0.1051,
+      "step": 3324
+    },
+    {
+      "epoch": 1.7412257726558407,
+      "loss_partial": 0.041748046875,
+      "step": 3324
+    },
+    {
+      "epoch": 1.7412257726558407,
+      "loss_partial": 0.04541015625,
+      "step": 3324
+    },
+    {
+      "epoch": 1.7417496071241487,
+      "grad_norm": 1.4758825302124023,
+      "learning_rate": 1.3898519114606466e-05,
+      "loss": 0.1026,
+      "step": 3325
+    },
+    {
+      "epoch": 1.7417496071241487,
+      "loss_partial": 0.045166015625,
+      "step": 3325
+    },
+    {
+      "epoch": 1.7417496071241487,
+      "loss_partial": 0.057861328125,
+      "step": 3325
+    },
+    {
+      "epoch": 1.7422734415924568,
+      "grad_norm": 1.486741542816162,
+      "learning_rate": 1.3894308838061644e-05,
+      "loss": 0.1044,
+      "step": 3326
+    },
+    {
+      "epoch": 1.7422734415924568,
+      "loss_partial": 0.0537109375,
+      "step": 3326
+    },
+    {
+      "epoch": 1.7422734415924568,
+      "loss_partial": 0.0625,
+      "step": 3326
+    },
+    {
+      "epoch": 1.7427972760607648,
+      "grad_norm": 1.8372341394424438,
+      "learning_rate": 1.3890097747630357e-05,
+      "loss": 0.1062,
+      "step": 3327
+    },
+    {
+      "epoch": 1.7427972760607648,
+      "loss_partial": 0.0537109375,
+      "step": 3327
+    },
+    {
+      "epoch": 1.7427972760607648,
+      "loss_partial": 0.0458984375,
+      "step": 3327
+    },
+    {
+      "epoch": 1.743321110529073,
+      "grad_norm": 1.8216320276260376,
+      "learning_rate": 1.3885885844192694e-05,
+      "loss": 0.1028,
+      "step": 3328
+    },
+    {
+      "epoch": 1.743321110529073,
+      "loss_partial": 0.058837890625,
+      "step": 3328
+    },
+    {
+      "epoch": 1.743321110529073,
+      "loss_partial": 0.050048828125,
+      "step": 3328
+    },
+    {
+      "epoch": 1.743844944997381,
+      "grad_norm": 1.5824323892593384,
+      "learning_rate": 1.3881673128628921e-05,
+      "loss": 0.1051,
+      "step": 3329
+    },
+    {
+      "epoch": 1.743844944997381,
+      "loss_partial": 0.054443359375,
+      "step": 3329
+    },
+    {
+      "epoch": 1.743844944997381,
+      "loss_partial": 0.064453125,
+      "step": 3329
+    },
+    {
+      "epoch": 1.7443687794656888,
+      "grad_norm": 1.720883846282959,
+      "learning_rate": 1.3877459601819467e-05,
+      "loss": 0.1021,
+      "step": 3330
+    },
+    {
+      "epoch": 1.7443687794656888,
+      "loss_partial": 0.0576171875,
+      "step": 3330
+    },
+    {
+      "epoch": 1.7443687794656888,
+      "loss_partial": 0.051025390625,
+      "step": 3330
+    },
+    {
+      "epoch": 1.7448926139339969,
+      "grad_norm": 1.5061415433883667,
+      "learning_rate": 1.3873245264644934e-05,
+      "loss": 0.103,
+      "step": 3331
+    },
+    {
+      "epoch": 1.7448926139339969,
+      "loss_partial": 0.05322265625,
+      "step": 3331
+    },
+    {
+      "epoch": 1.7448926139339969,
+      "loss_partial": 0.045654296875,
+      "step": 3331
+    },
+    {
+      "epoch": 1.745416448402305,
+      "grad_norm": 2.096836566925049,
+      "learning_rate": 1.3869030117986092e-05,
+      "loss": 0.1038,
+      "step": 3332
+    },
+    {
+      "epoch": 1.745416448402305,
+      "loss_partial": 0.048828125,
+      "step": 3332
+    },
+    {
+      "epoch": 1.745416448402305,
+      "loss_partial": 0.064453125,
+      "step": 3332
+    },
+    {
+      "epoch": 1.7459402828706128,
+      "grad_norm": 1.4160401821136475,
+      "learning_rate": 1.386481416272388e-05,
+      "loss": 0.1061,
+      "step": 3333
+    },
+    {
+      "epoch": 1.7459402828706128,
+      "loss_partial": 0.052490234375,
+      "step": 3333
+    },
+    {
+      "epoch": 1.7459402828706128,
+      "loss_partial": 0.04345703125,
+      "step": 3333
+    },
+    {
+      "epoch": 1.7464641173389208,
+      "grad_norm": 2.461848735809326,
+      "learning_rate": 1.3860597399739411e-05,
+      "loss": 0.1051,
+      "step": 3334
+    },
+    {
+      "epoch": 1.7464641173389208,
+      "loss_partial": 0.04052734375,
+      "step": 3334
+    },
+    {
+      "epoch": 1.7464641173389208,
+      "loss_partial": 0.04833984375,
+      "step": 3334
+    },
+    {
+      "epoch": 1.7469879518072289,
+      "grad_norm": 2.2715466022491455,
+      "learning_rate": 1.3856379829913953e-05,
+      "loss": 0.1049,
+      "step": 3335
+    },
+    {
+      "epoch": 1.7469879518072289,
+      "loss_partial": 0.045654296875,
+      "step": 3335
+    },
+    {
+      "epoch": 1.7469879518072289,
+      "loss_partial": 0.048828125,
+      "step": 3335
+    },
+    {
+      "epoch": 1.747511786275537,
+      "grad_norm": 2.1553335189819336,
+      "learning_rate": 1.3852161454128964e-05,
+      "loss": 0.1008,
+      "step": 3336
+    },
+    {
+      "epoch": 1.747511786275537,
+      "loss_partial": 0.061767578125,
+      "step": 3336
+    },
+    {
+      "epoch": 1.747511786275537,
+      "loss_partial": 0.0556640625,
+      "step": 3336
+    },
+    {
+      "epoch": 1.748035620743845,
+      "grad_norm": 2.208735227584839,
+      "learning_rate": 1.3847942273266044e-05,
+      "loss": 0.1017,
+      "step": 3337
+    },
+    {
+      "epoch": 1.748035620743845,
+      "loss_partial": 0.0537109375,
+      "step": 3337
+    },
+    {
+      "epoch": 1.748035620743845,
+      "loss_partial": 0.054443359375,
+      "step": 3337
+    },
+    {
+      "epoch": 1.748559455212153,
+      "grad_norm": 2.28847074508667,
+      "learning_rate": 1.3843722288206988e-05,
+      "loss": 0.1009,
+      "step": 3338
+    },
+    {
+      "epoch": 1.748559455212153,
+      "loss_partial": 0.058837890625,
+      "step": 3338
+    },
+    {
+      "epoch": 1.748559455212153,
+      "loss_partial": 0.062255859375,
+      "step": 3338
+    },
+    {
+      "epoch": 1.7490832896804611,
+      "grad_norm": 1.6844403743743896,
+      "learning_rate": 1.3839501499833745e-05,
+      "loss": 0.1082,
+      "step": 3339
+    },
+    {
+      "epoch": 1.7490832896804611,
+      "loss_partial": 0.047119140625,
+      "step": 3339
+    },
+    {
+      "epoch": 1.7490832896804611,
+      "loss_partial": 0.048828125,
+      "step": 3339
+    },
+    {
+      "epoch": 1.749607124148769,
+      "grad_norm": 1.9456998109817505,
+      "learning_rate": 1.383527990902843e-05,
+      "loss": 0.104,
+      "step": 3340
+    },
+    {
+      "epoch": 1.749607124148769,
+      "loss_partial": 0.060546875,
+      "step": 3340
+    },
+    {
+      "epoch": 1.749607124148769,
+      "loss_partial": 0.052978515625,
+      "step": 3340
+    },
+    {
+      "epoch": 1.750130958617077,
+      "grad_norm": 1.9462772607803345,
+      "learning_rate": 1.383105751667333e-05,
+      "loss": 0.1067,
+      "step": 3341
+    },
+    {
+      "epoch": 1.750130958617077,
+      "loss_partial": 0.047607421875,
+      "step": 3341
+    },
+    {
+      "epoch": 1.750130958617077,
+      "loss_partial": 0.0546875,
+      "step": 3341
+    },
+    {
+      "epoch": 1.7506547930853849,
+      "grad_norm": 2.0114235877990723,
+      "learning_rate": 1.3826834323650899e-05,
+      "loss": 0.1038,
+      "step": 3342
+    },
+    {
+      "epoch": 1.7506547930853849,
+      "loss_partial": 0.048095703125,
+      "step": 3342
+    },
+    {
+      "epoch": 1.7506547930853849,
+      "loss_partial": 0.048828125,
+      "step": 3342
+    },
+    {
+      "epoch": 1.751178627553693,
+      "grad_norm": 1.45498526096344,
+      "learning_rate": 1.382261033084376e-05,
+      "loss": 0.1052,
+      "step": 3343
+    },
+    {
+      "epoch": 1.751178627553693,
+      "loss_partial": 0.05908203125,
+      "step": 3343
+    },
+    {
+      "epoch": 1.751178627553693,
+      "loss_partial": 0.05126953125,
+      "step": 3343
+    },
+    {
+      "epoch": 1.751702462022001,
+      "grad_norm": 1.7970212697982788,
+      "learning_rate": 1.3818385539134698e-05,
+      "loss": 0.1028,
+      "step": 3344
+    },
+    {
+      "epoch": 1.751702462022001,
+      "loss_partial": 0.039794921875,
+      "step": 3344
+    },
+    {
+      "epoch": 1.751702462022001,
+      "loss_partial": 0.048828125,
+      "step": 3344
+    },
+    {
+      "epoch": 1.752226296490309,
+      "grad_norm": 1.795263409614563,
+      "learning_rate": 1.3814159949406674e-05,
+      "loss": 0.1025,
+      "step": 3345
+    },
+    {
+      "epoch": 1.752226296490309,
+      "loss_partial": 0.0595703125,
+      "step": 3345
+    },
+    {
+      "epoch": 1.752226296490309,
+      "loss_partial": 0.052001953125,
+      "step": 3345
+    },
+    {
+      "epoch": 1.752750130958617,
+      "grad_norm": 1.8453953266143799,
+      "learning_rate": 1.3809933562542805e-05,
+      "loss": 0.1059,
+      "step": 3346
+    },
+    {
+      "epoch": 1.752750130958617,
+      "loss_partial": 0.059814453125,
+      "step": 3346
+    },
+    {
+      "epoch": 1.752750130958617,
+      "loss_partial": 0.052490234375,
+      "step": 3346
+    },
+    {
+      "epoch": 1.7532739654269252,
+      "grad_norm": 2.2320311069488525,
+      "learning_rate": 1.380570637942638e-05,
+      "loss": 0.1059,
+      "step": 3347
+    },
+    {
+      "epoch": 1.7532739654269252,
+      "loss_partial": 0.051025390625,
+      "step": 3347
+    },
+    {
+      "epoch": 1.7532739654269252,
+      "loss_partial": 0.044189453125,
+      "step": 3347
+    },
+    {
+      "epoch": 1.7537977998952332,
+      "grad_norm": 1.8962596654891968,
+      "learning_rate": 1.3801478400940855e-05,
+      "loss": 0.1024,
+      "step": 3348
+    },
+    {
+      "epoch": 1.7537977998952332,
+      "loss_partial": 0.049072265625,
+      "step": 3348
+    },
+    {
+      "epoch": 1.7537977998952332,
+      "loss_partial": 0.07080078125,
+      "step": 3348
+    },
+    {
+      "epoch": 1.7543216343635413,
+      "grad_norm": 1.3109620809555054,
+      "learning_rate": 1.3797249627969856e-05,
+      "loss": 0.1072,
+      "step": 3349
+    },
+    {
+      "epoch": 1.7543216343635413,
+      "loss_partial": 0.052734375,
+      "step": 3349
+    },
+    {
+      "epoch": 1.7543216343635413,
+      "loss_partial": 0.05078125,
+      "step": 3349
+    },
+    {
+      "epoch": 1.754845468831849,
+      "grad_norm": 1.7344610691070557,
+      "learning_rate": 1.3793020061397159e-05,
+      "loss": 0.1054,
+      "step": 3350
+    },
+    {
+      "epoch": 1.754845468831849,
+      "loss_partial": 0.05322265625,
+      "step": 3350
+    },
+    {
+      "epoch": 1.754845468831849,
+      "loss_partial": 0.05224609375,
+      "step": 3350
+    },
+    {
+      "epoch": 1.7553693033001572,
+      "grad_norm": 1.797244906425476,
+      "learning_rate": 1.378878970210673e-05,
+      "loss": 0.1042,
+      "step": 3351
+    },
+    {
+      "epoch": 1.7553693033001572,
+      "loss_partial": 0.05810546875,
+      "step": 3351
+    },
+    {
+      "epoch": 1.7553693033001572,
+      "loss_partial": 0.057861328125,
+      "step": 3351
+    },
+    {
+      "epoch": 1.755893137768465,
+      "grad_norm": 2.0582520961761475,
+      "learning_rate": 1.3784558550982684e-05,
+      "loss": 0.1062,
+      "step": 3352
+    },
+    {
+      "epoch": 1.755893137768465,
+      "loss_partial": 0.06103515625,
+      "step": 3352
+    },
+    {
+      "epoch": 1.755893137768465,
+      "loss_partial": 0.05419921875,
+      "step": 3352
+    },
+    {
+      "epoch": 1.756416972236773,
+      "grad_norm": 1.8405956029891968,
+      "learning_rate": 1.3780326608909303e-05,
+      "loss": 0.1041,
+      "step": 3353
+    },
+    {
+      "epoch": 1.756416972236773,
+      "loss_partial": 0.049072265625,
+      "step": 3353
+    },
+    {
+      "epoch": 1.756416972236773,
+      "loss_partial": 0.0546875,
+      "step": 3353
+    },
+    {
+      "epoch": 1.7569408067050811,
+      "grad_norm": 1.7202844619750977,
+      "learning_rate": 1.3776093876771033e-05,
+      "loss": 0.1031,
+      "step": 3354
+    },
+    {
+      "epoch": 1.7569408067050811,
+      "loss_partial": 0.049072265625,
+      "step": 3354
+    },
+    {
+      "epoch": 1.7569408067050811,
+      "loss_partial": 0.048583984375,
+      "step": 3354
+    },
+    {
+      "epoch": 1.7574646411733892,
+      "grad_norm": 1.960274577140808,
+      "learning_rate": 1.3771860355452497e-05,
+      "loss": 0.104,
+      "step": 3355
+    },
+    {
+      "epoch": 1.7574646411733892,
+      "loss_partial": 0.0478515625,
+      "step": 3355
+    },
+    {
+      "epoch": 1.7574646411733892,
+      "loss_partial": 0.05615234375,
+      "step": 3355
+    },
+    {
+      "epoch": 1.7579884756416972,
+      "grad_norm": 1.8882496356964111,
+      "learning_rate": 1.376762604583847e-05,
+      "loss": 0.1082,
+      "step": 3356
+    },
+    {
+      "epoch": 1.7579884756416972,
+      "loss_partial": 0.05224609375,
+      "step": 3356
+    },
+    {
+      "epoch": 1.7579884756416972,
+      "loss_partial": 0.05615234375,
+      "step": 3356
+    },
+    {
+      "epoch": 1.7585123101100053,
+      "grad_norm": 1.552277684211731,
+      "learning_rate": 1.3763390948813897e-05,
+      "loss": 0.1104,
+      "step": 3357
+    },
+    {
+      "epoch": 1.7585123101100053,
+      "loss_partial": 0.05615234375,
+      "step": 3357
+    },
+    {
+      "epoch": 1.7585123101100053,
+      "loss_partial": 0.049560546875,
+      "step": 3357
+    },
+    {
+      "epoch": 1.7590361445783134,
+      "grad_norm": 1.2888076305389404,
+      "learning_rate": 1.3759155065263891e-05,
+      "loss": 0.1037,
+      "step": 3358
+    },
+    {
+      "epoch": 1.7590361445783134,
+      "loss_partial": 0.044189453125,
+      "step": 3358
+    },
+    {
+      "epoch": 1.7590361445783134,
+      "loss_partial": 0.058349609375,
+      "step": 3358
+    },
+    {
+      "epoch": 1.7595599790466214,
+      "grad_norm": 1.6356544494628906,
+      "learning_rate": 1.375491839607372e-05,
+      "loss": 0.1052,
+      "step": 3359
+    },
+    {
+      "epoch": 1.7595599790466214,
+      "loss_partial": 0.05078125,
+      "step": 3359
+    },
+    {
+      "epoch": 1.7595599790466214,
+      "loss_partial": 0.056396484375,
+      "step": 3359
+    },
+    {
+      "epoch": 1.7600838135149293,
+      "grad_norm": 2.2535181045532227,
+      "learning_rate": 1.3750680942128825e-05,
+      "loss": 0.1018,
+      "step": 3360
+    },
+    {
+      "epoch": 1.7600838135149293,
+      "loss_partial": 0.050048828125,
+      "step": 3360
+    },
+    {
+      "epoch": 1.7600838135149293,
+      "loss_partial": 0.048828125,
+      "step": 3360
+    },
+    {
+      "epoch": 1.7606076479832373,
+      "grad_norm": 1.4807660579681396,
+      "learning_rate": 1.3746442704314803e-05,
+      "loss": 0.1034,
+      "step": 3361
+    },
+    {
+      "epoch": 1.7606076479832373,
+      "loss_partial": 0.052734375,
+      "step": 3361
+    },
+    {
+      "epoch": 1.7606076479832373,
+      "loss_partial": 0.054931640625,
+      "step": 3361
+    },
+    {
+      "epoch": 1.7611314824515452,
+      "grad_norm": 1.8275761604309082,
+      "learning_rate": 1.3742203683517428e-05,
+      "loss": 0.1046,
+      "step": 3362
+    },
+    {
+      "epoch": 1.7611314824515452,
+      "loss_partial": 0.051513671875,
+      "step": 3362
+    },
+    {
+      "epoch": 1.7611314824515452,
+      "loss_partial": 0.04833984375,
+      "step": 3362
+    },
+    {
+      "epoch": 1.7616553169198532,
+      "grad_norm": 1.6432766914367676,
+      "learning_rate": 1.3737963880622622e-05,
+      "loss": 0.1039,
+      "step": 3363
+    },
+    {
+      "epoch": 1.7616553169198532,
+      "loss_partial": 0.04736328125,
+      "step": 3363
+    },
+    {
+      "epoch": 1.7616553169198532,
+      "loss_partial": 0.05224609375,
+      "step": 3363
+    },
+    {
+      "epoch": 1.7621791513881613,
+      "grad_norm": 2.500072479248047,
+      "learning_rate": 1.3733723296516482e-05,
+      "loss": 0.0993,
+      "step": 3364
+    },
+    {
+      "epoch": 1.7621791513881613,
+      "loss_partial": 0.0517578125,
+      "step": 3364
+    },
+    {
+      "epoch": 1.7621791513881613,
+      "loss_partial": 0.050048828125,
+      "step": 3364
+    },
+    {
+      "epoch": 1.7627029858564693,
+      "grad_norm": 1.2355388402938843,
+      "learning_rate": 1.3729481932085258e-05,
+      "loss": 0.1048,
+      "step": 3365
+    },
+    {
+      "epoch": 1.7627029858564693,
+      "loss_partial": 0.046875,
+      "step": 3365
+    },
+    {
+      "epoch": 1.7627029858564693,
+      "loss_partial": 0.047607421875,
+      "step": 3365
+    },
+    {
+      "epoch": 1.7632268203247774,
+      "grad_norm": 1.7962719202041626,
+      "learning_rate": 1.3725239788215374e-05,
+      "loss": 0.0967,
+      "step": 3366
+    },
+    {
+      "epoch": 1.7632268203247774,
+      "loss_partial": 0.045654296875,
+      "step": 3366
+    },
+    {
+      "epoch": 1.7632268203247774,
+      "loss_partial": 0.055419921875,
+      "step": 3366
+    },
+    {
+      "epoch": 1.7637506547930855,
+      "grad_norm": 1.6000107526779175,
+      "learning_rate": 1.372099686579341e-05,
+      "loss": 0.0997,
+      "step": 3367
+    },
+    {
+      "epoch": 1.7637506547930855,
+      "loss_partial": 0.04638671875,
+      "step": 3367
+    },
+    {
+      "epoch": 1.7637506547930855,
+      "loss_partial": 0.057373046875,
+      "step": 3367
+    },
+    {
+      "epoch": 1.7642744892613935,
+      "grad_norm": 2.4981017112731934,
+      "learning_rate": 1.3716753165706111e-05,
+      "loss": 0.1064,
+      "step": 3368
+    },
+    {
+      "epoch": 1.7642744892613935,
+      "loss_partial": 0.059814453125,
+      "step": 3368
+    },
+    {
+      "epoch": 1.7642744892613935,
+      "loss_partial": 0.06396484375,
+      "step": 3368
+    },
+    {
+      "epoch": 1.7647983237297016,
+      "grad_norm": 1.691039800643921,
+      "learning_rate": 1.3712508688840384e-05,
+      "loss": 0.1085,
+      "step": 3369
+    },
+    {
+      "epoch": 1.7647983237297016,
+      "loss_partial": 0.06494140625,
+      "step": 3369
+    },
+    {
+      "epoch": 1.7647983237297016,
+      "loss_partial": 0.049072265625,
+      "step": 3369
+    },
+    {
+      "epoch": 1.7653221581980094,
+      "grad_norm": 2.2018468379974365,
+      "learning_rate": 1.3708263436083303e-05,
+      "loss": 0.1071,
+      "step": 3370
+    },
+    {
+      "epoch": 1.7653221581980094,
+      "loss_partial": 0.0537109375,
+      "step": 3370
+    },
+    {
+      "epoch": 1.7653221581980094,
+      "loss_partial": 0.04248046875,
+      "step": 3370
+    },
+    {
+      "epoch": 1.7658459926663175,
+      "grad_norm": 1.2016593217849731,
+      "learning_rate": 1.3704017408322093e-05,
+      "loss": 0.1022,
+      "step": 3371
+    },
+    {
+      "epoch": 1.7658459926663175,
+      "loss_partial": 0.052734375,
+      "step": 3371
+    },
+    {
+      "epoch": 1.7658459926663175,
+      "loss_partial": 0.060302734375,
+      "step": 3371
+    },
+    {
+      "epoch": 1.7663698271346253,
+      "grad_norm": 2.087737560272217,
+      "learning_rate": 1.3699770606444145e-05,
+      "loss": 0.1035,
+      "step": 3372
+    },
+    {
+      "epoch": 1.7663698271346253,
+      "loss_partial": 0.047119140625,
+      "step": 3372
+    },
+    {
+      "epoch": 1.7663698271346253,
+      "loss_partial": 0.052490234375,
+      "step": 3372
+    },
+    {
+      "epoch": 1.7668936616029334,
+      "grad_norm": 1.3526767492294312,
+      "learning_rate": 1.3695523031337023e-05,
+      "loss": 0.1057,
+      "step": 3373
+    },
+    {
+      "epoch": 1.7668936616029334,
+      "loss_partial": 0.05615234375,
+      "step": 3373
+    },
+    {
+      "epoch": 1.7668936616029334,
+      "loss_partial": 0.050537109375,
+      "step": 3373
+    },
+    {
+      "epoch": 1.7674174960712414,
+      "grad_norm": 1.877537488937378,
+      "learning_rate": 1.3691274683888436e-05,
+      "loss": 0.106,
+      "step": 3374
+    },
+    {
+      "epoch": 1.7674174960712414,
+      "loss_partial": 0.046875,
+      "step": 3374
+    },
+    {
+      "epoch": 1.7674174960712414,
+      "loss_partial": 0.05322265625,
+      "step": 3374
+    },
+    {
+      "epoch": 1.7679413305395495,
+      "grad_norm": 1.6992740631103516,
+      "learning_rate": 1.3687025564986271e-05,
+      "loss": 0.1046,
+      "step": 3375
+    },
+    {
+      "epoch": 1.7679413305395495,
+      "loss_partial": 0.0693359375,
+      "step": 3375
+    },
+    {
+      "epoch": 1.7679413305395495,
+      "loss_partial": 0.040283203125,
+      "step": 3375
+    },
+    {
+      "epoch": 1.7684651650078576,
+      "grad_norm": 2.1479592323303223,
+      "learning_rate": 1.368277567551856e-05,
+      "loss": 0.1084,
+      "step": 3376
+    },
+    {
+      "epoch": 1.7684651650078576,
+      "loss_partial": 0.04541015625,
+      "step": 3376
+    },
+    {
+      "epoch": 1.7684651650078576,
+      "loss_partial": 0.048095703125,
+      "step": 3376
+    },
+    {
+      "epoch": 1.7689889994761656,
+      "grad_norm": 1.503637433052063,
+      "learning_rate": 1.3678525016373509e-05,
+      "loss": 0.0993,
+      "step": 3377
+    },
+    {
+      "epoch": 1.7689889994761656,
+      "loss_partial": 0.046630859375,
+      "step": 3377
+    },
+    {
+      "epoch": 1.7689889994761656,
+      "loss_partial": 0.060791015625,
+      "step": 3377
+    },
+    {
+      "epoch": 1.7695128339444737,
+      "grad_norm": 1.4210515022277832,
+      "learning_rate": 1.3674273588439472e-05,
+      "loss": 0.1044,
+      "step": 3378
+    },
+    {
+      "epoch": 1.7695128339444737,
+      "loss_partial": 0.060302734375,
+      "step": 3378
+    },
+    {
+      "epoch": 1.7695128339444737,
+      "loss_partial": 0.0595703125,
+      "step": 3378
+    },
+    {
+      "epoch": 1.7700366684127815,
+      "grad_norm": 1.74802827835083,
+      "learning_rate": 1.3670021392604982e-05,
+      "loss": 0.1042,
+      "step": 3379
+    },
+    {
+      "epoch": 1.7700366684127815,
+      "loss_partial": 0.051513671875,
+      "step": 3379
+    },
+    {
+      "epoch": 1.7700366684127815,
+      "loss_partial": 0.0498046875,
+      "step": 3379
+    },
+    {
+      "epoch": 1.7705605028810896,
+      "grad_norm": 1.6431986093521118,
+      "learning_rate": 1.3665768429758713e-05,
+      "loss": 0.102,
+      "step": 3380
+    },
+    {
+      "epoch": 1.7705605028810896,
+      "loss_partial": 0.0546875,
+      "step": 3380
+    },
+    {
+      "epoch": 1.7705605028810896,
+      "loss_partial": 0.0439453125,
+      "step": 3380
+    },
+    {
+      "epoch": 1.7710843373493976,
+      "grad_norm": 2.247055768966675,
+      "learning_rate": 1.3661514700789508e-05,
+      "loss": 0.1049,
+      "step": 3381
+    },
+    {
+      "epoch": 1.7710843373493976,
+      "loss_partial": 0.04638671875,
+      "step": 3381
+    },
+    {
+      "epoch": 1.7710843373493976,
+      "loss_partial": 0.054443359375,
+      "step": 3381
+    },
+    {
+      "epoch": 1.7716081718177055,
+      "grad_norm": 1.6446259021759033,
+      "learning_rate": 1.3657260206586375e-05,
+      "loss": 0.0995,
+      "step": 3382
+    },
+    {
+      "epoch": 1.7716081718177055,
+      "loss_partial": 0.04931640625,
+      "step": 3382
+    },
+    {
+      "epoch": 1.7716081718177055,
+      "loss_partial": 0.06787109375,
+      "step": 3382
+    },
+    {
+      "epoch": 1.7721320062860135,
+      "grad_norm": 1.6841627359390259,
+      "learning_rate": 1.3653004948038477e-05,
+      "loss": 0.1073,
+      "step": 3383
+    },
+    {
+      "epoch": 1.7721320062860135,
+      "loss_partial": 0.052978515625,
+      "step": 3383
+    },
+    {
+      "epoch": 1.7721320062860135,
+      "loss_partial": 0.052490234375,
+      "step": 3383
+    },
+    {
+      "epoch": 1.7726558407543216,
+      "grad_norm": 1.1672677993774414,
+      "learning_rate": 1.364874892603513e-05,
+      "loss": 0.1025,
+      "step": 3384
+    },
+    {
+      "epoch": 1.7726558407543216,
+      "loss_partial": 0.046630859375,
+      "step": 3384
+    },
+    {
+      "epoch": 1.7726558407543216,
+      "loss_partial": 0.05322265625,
+      "step": 3384
+    },
+    {
+      "epoch": 1.7731796752226296,
+      "grad_norm": 1.9948687553405762,
+      "learning_rate": 1.3644492141465823e-05,
+      "loss": 0.1057,
+      "step": 3385
+    },
+    {
+      "epoch": 1.7731796752226296,
+      "loss_partial": 0.052490234375,
+      "step": 3385
+    },
+    {
+      "epoch": 1.7731796752226296,
+      "loss_partial": 0.052734375,
+      "step": 3385
+    },
+    {
+      "epoch": 1.7737035096909377,
+      "grad_norm": 2.2065885066986084,
+      "learning_rate": 1.3640234595220196e-05,
+      "loss": 0.1087,
+      "step": 3386
+    },
+    {
+      "epoch": 1.7737035096909377,
+      "loss_partial": 0.05029296875,
+      "step": 3386
+    },
+    {
+      "epoch": 1.7737035096909377,
+      "loss_partial": 0.048095703125,
+      "step": 3386
+    },
+    {
+      "epoch": 1.7742273441592458,
+      "grad_norm": 1.827948808670044,
+      "learning_rate": 1.3635976288188048e-05,
+      "loss": 0.1003,
+      "step": 3387
+    },
+    {
+      "epoch": 1.7742273441592458,
+      "loss_partial": 0.05810546875,
+      "step": 3387
+    },
+    {
+      "epoch": 1.7742273441592458,
+      "loss_partial": 0.057373046875,
+      "step": 3387
+    },
+    {
+      "epoch": 1.7747511786275538,
+      "grad_norm": 2.0018317699432373,
+      "learning_rate": 1.3631717221259341e-05,
+      "loss": 0.1059,
+      "step": 3388
+    },
+    {
+      "epoch": 1.7747511786275538,
+      "loss_partial": 0.053955078125,
+      "step": 3388
+    },
+    {
+      "epoch": 1.7747511786275538,
+      "loss_partial": 0.046875,
+      "step": 3388
+    },
+    {
+      "epoch": 1.7752750130958617,
+      "grad_norm": 1.658531665802002,
+      "learning_rate": 1.3627457395324193e-05,
+      "loss": 0.1028,
+      "step": 3389
+    },
+    {
+      "epoch": 1.7752750130958617,
+      "loss_partial": 0.060546875,
+      "step": 3389
+    },
+    {
+      "epoch": 1.7752750130958617,
+      "loss_partial": 0.048095703125,
+      "step": 3389
+    },
+    {
+      "epoch": 1.7757988475641697,
+      "grad_norm": 2.1023919582366943,
+      "learning_rate": 1.3623196811272879e-05,
+      "loss": 0.1035,
+      "step": 3390
+    },
+    {
+      "epoch": 1.7757988475641697,
+      "loss_partial": 0.057373046875,
+      "step": 3390
+    },
+    {
+      "epoch": 1.7757988475641697,
+      "loss_partial": 0.047607421875,
+      "step": 3390
+    },
+    {
+      "epoch": 1.7763226820324778,
+      "grad_norm": 1.5600957870483398,
+      "learning_rate": 1.3618935469995838e-05,
+      "loss": 0.1082,
+      "step": 3391
+    },
+    {
+      "epoch": 1.7763226820324778,
+      "loss_partial": 0.05712890625,
+      "step": 3391
+    },
+    {
+      "epoch": 1.7763226820324778,
+      "loss_partial": 0.048828125,
+      "step": 3391
+    },
+    {
+      "epoch": 1.7768465165007856,
+      "grad_norm": 2.5458922386169434,
+      "learning_rate": 1.361467337238366e-05,
+      "loss": 0.1024,
+      "step": 3392
+    },
+    {
+      "epoch": 1.7768465165007856,
+      "loss_partial": 0.057373046875,
+      "step": 3392
+    },
+    {
+      "epoch": 1.7768465165007856,
+      "loss_partial": 0.052490234375,
+      "step": 3392
+    },
+    {
+      "epoch": 1.7773703509690937,
+      "grad_norm": 1.5650073289871216,
+      "learning_rate": 1.3610410519327101e-05,
+      "loss": 0.1052,
+      "step": 3393
+    },
+    {
+      "epoch": 1.7773703509690937,
+      "loss_partial": 0.048095703125,
+      "step": 3393
+    },
+    {
+      "epoch": 1.7773703509690937,
+      "loss_partial": 0.047119140625,
+      "step": 3393
+    },
+    {
+      "epoch": 1.7778941854374017,
+      "grad_norm": 2.0354785919189453,
+      "learning_rate": 1.3606146911717071e-05,
+      "loss": 0.1008,
+      "step": 3394
+    },
+    {
+      "epoch": 1.7778941854374017,
+      "loss_partial": 0.042724609375,
+      "step": 3394
+    },
+    {
+      "epoch": 1.7778941854374017,
+      "loss_partial": 0.04931640625,
+      "step": 3394
+    },
+    {
+      "epoch": 1.7784180199057098,
+      "grad_norm": 2.06020450592041,
+      "learning_rate": 1.3601882550444635e-05,
+      "loss": 0.1018,
+      "step": 3395
+    },
+    {
+      "epoch": 1.7784180199057098,
+      "loss_partial": 0.046630859375,
+      "step": 3395
+    },
+    {
+      "epoch": 1.7784180199057098,
+      "loss_partial": 0.0537109375,
+      "step": 3395
+    },
+    {
+      "epoch": 1.7789418543740179,
+      "grad_norm": 2.2288687229156494,
+      "learning_rate": 1.3597617436401019e-05,
+      "loss": 0.1032,
+      "step": 3396
+    },
+    {
+      "epoch": 1.7789418543740179,
+      "loss_partial": 0.055908203125,
+      "step": 3396
+    },
+    {
+      "epoch": 1.7789418543740179,
+      "loss_partial": 0.05078125,
+      "step": 3396
+    },
+    {
+      "epoch": 1.779465688842326,
+      "grad_norm": 1.5742734670639038,
+      "learning_rate": 1.3593351570477608e-05,
+      "loss": 0.1035,
+      "step": 3397
+    },
+    {
+      "epoch": 1.779465688842326,
+      "loss_partial": 0.044921875,
+      "step": 3397
+    },
+    {
+      "epoch": 1.779465688842326,
+      "loss_partial": 0.059326171875,
+      "step": 3397
+    },
+    {
+      "epoch": 1.779989523310634,
+      "grad_norm": 2.001847267150879,
+      "learning_rate": 1.3589084953565936e-05,
+      "loss": 0.1024,
+      "step": 3398
+    },
+    {
+      "epoch": 1.779989523310634,
+      "loss_partial": 0.0537109375,
+      "step": 3398
+    },
+    {
+      "epoch": 1.779989523310634,
+      "loss_partial": 0.05078125,
+      "step": 3398
+    },
+    {
+      "epoch": 1.7805133577789418,
+      "grad_norm": 2.0826661586761475,
+      "learning_rate": 1.3584817586557706e-05,
+      "loss": 0.1041,
+      "step": 3399
+    },
+    {
+      "epoch": 1.7805133577789418,
+      "loss_partial": 0.05615234375,
+      "step": 3399
+    },
+    {
+      "epoch": 1.7805133577789418,
+      "loss_partial": 0.0498046875,
+      "step": 3399
+    },
+    {
+      "epoch": 1.7810371922472499,
+      "grad_norm": 2.263427734375,
+      "learning_rate": 1.3580549470344769e-05,
+      "loss": 0.1048,
+      "step": 3400
+    },
+    {
+      "epoch": 1.7810371922472499,
+      "loss_partial": 0.048095703125,
+      "step": 3400
+    },
+    {
+      "epoch": 1.7810371922472499,
+      "loss_partial": 0.05322265625,
+      "step": 3400
+    },
+    {
+      "epoch": 1.781561026715558,
+      "grad_norm": 2.070455551147461,
+      "learning_rate": 1.357628060581913e-05,
+      "loss": 0.104,
+      "step": 3401
+    },
+    {
+      "epoch": 1.781561026715558,
+      "loss_partial": 0.053466796875,
+      "step": 3401
+    },
+    {
+      "epoch": 1.781561026715558,
+      "loss_partial": 0.04833984375,
+      "step": 3401
+    },
+    {
+      "epoch": 1.7820848611838658,
+      "grad_norm": 1.6581332683563232,
+      "learning_rate": 1.3572010993872963e-05,
+      "loss": 0.1024,
+      "step": 3402
+    },
+    {
+      "epoch": 1.7820848611838658,
+      "loss_partial": 0.059326171875,
+      "step": 3402
+    },
+    {
+      "epoch": 1.7820848611838658,
+      "loss_partial": 0.0478515625,
+      "step": 3402
+    },
+    {
+      "epoch": 1.7826086956521738,
+      "grad_norm": 1.451131820678711,
+      "learning_rate": 1.3567740635398586e-05,
+      "loss": 0.1054,
+      "step": 3403
+    },
+    {
+      "epoch": 1.7826086956521738,
+      "loss_partial": 0.052978515625,
+      "step": 3403
+    },
+    {
+      "epoch": 1.7826086956521738,
+      "loss_partial": 0.052978515625,
+      "step": 3403
+    },
+    {
+      "epoch": 1.783132530120482,
+      "grad_norm": 1.6049253940582275,
+      "learning_rate": 1.3563469531288479e-05,
+      "loss": 0.1047,
+      "step": 3404
+    },
+    {
+      "epoch": 1.783132530120482,
+      "loss_partial": 0.045166015625,
+      "step": 3404
+    },
+    {
+      "epoch": 1.783132530120482,
+      "loss_partial": 0.046142578125,
+      "step": 3404
+    },
+    {
+      "epoch": 1.78365636458879,
+      "grad_norm": 1.3816055059432983,
+      "learning_rate": 1.3559197682435273e-05,
+      "loss": 0.1023,
+      "step": 3405
+    },
+    {
+      "epoch": 1.78365636458879,
+      "loss_partial": 0.06396484375,
+      "step": 3405
+    },
+    {
+      "epoch": 1.78365636458879,
+      "loss_partial": 0.0537109375,
+      "step": 3405
+    },
+    {
+      "epoch": 1.784180199057098,
+      "grad_norm": 1.8022923469543457,
+      "learning_rate": 1.3554925089731761e-05,
+      "loss": 0.1025,
+      "step": 3406
+    },
+    {
+      "epoch": 1.784180199057098,
+      "loss_partial": 0.05322265625,
+      "step": 3406
+    },
+    {
+      "epoch": 1.784180199057098,
+      "loss_partial": 0.052978515625,
+      "step": 3406
+    },
+    {
+      "epoch": 1.784704033525406,
+      "grad_norm": 1.397724986076355,
+      "learning_rate": 1.3550651754070891e-05,
+      "loss": 0.1053,
+      "step": 3407
+    },
+    {
+      "epoch": 1.784704033525406,
+      "loss_partial": 0.05517578125,
+      "step": 3407
+    },
+    {
+      "epoch": 1.784704033525406,
+      "loss_partial": 0.0546875,
+      "step": 3407
+    },
+    {
+      "epoch": 1.7852278679937141,
+      "grad_norm": 1.8641406297683716,
+      "learning_rate": 1.3546377676345761e-05,
+      "loss": 0.1057,
+      "step": 3408
+    },
+    {
+      "epoch": 1.7852278679937141,
+      "loss_partial": 0.046142578125,
+      "step": 3408
+    },
+    {
+      "epoch": 1.7852278679937141,
+      "loss_partial": 0.053955078125,
+      "step": 3408
+    },
+    {
+      "epoch": 1.785751702462022,
+      "grad_norm": 1.4891246557235718,
+      "learning_rate": 1.3542102857449624e-05,
+      "loss": 0.1056,
+      "step": 3409
+    },
+    {
+      "epoch": 1.785751702462022,
+      "loss_partial": 0.0517578125,
+      "step": 3409
+    },
+    {
+      "epoch": 1.785751702462022,
+      "loss_partial": 0.048095703125,
+      "step": 3409
+    },
+    {
+      "epoch": 1.78627553693033,
+      "grad_norm": 1.7005970478057861,
+      "learning_rate": 1.3537827298275896e-05,
+      "loss": 0.1021,
+      "step": 3410
+    },
+    {
+      "epoch": 1.78627553693033,
+      "loss_partial": 0.051025390625,
+      "step": 3410
+    },
+    {
+      "epoch": 1.78627553693033,
+      "loss_partial": 0.057373046875,
+      "step": 3410
+    },
+    {
+      "epoch": 1.7867993713986379,
+      "grad_norm": 1.3232823610305786,
+      "learning_rate": 1.3533550999718138e-05,
+      "loss": 0.1035,
+      "step": 3411
+    },
+    {
+      "epoch": 1.7867993713986379,
+      "loss_partial": 0.048583984375,
+      "step": 3411
+    },
+    {
+      "epoch": 1.7867993713986379,
+      "loss_partial": 0.056640625,
+      "step": 3411
+    },
+    {
+      "epoch": 1.787323205866946,
+      "grad_norm": 1.682060718536377,
+      "learning_rate": 1.3529273962670074e-05,
+      "loss": 0.105,
+      "step": 3412
+    },
+    {
+      "epoch": 1.787323205866946,
+      "loss_partial": 0.0537109375,
+      "step": 3412
+    },
+    {
+      "epoch": 1.787323205866946,
+      "loss_partial": 0.044189453125,
+      "step": 3412
+    },
+    {
+      "epoch": 1.787847040335254,
+      "grad_norm": 1.5685988664627075,
+      "learning_rate": 1.3524996188025574e-05,
+      "loss": 0.1027,
+      "step": 3413
+    },
+    {
+      "epoch": 1.787847040335254,
+      "loss_partial": 0.05078125,
+      "step": 3413
+    },
+    {
+      "epoch": 1.787847040335254,
+      "loss_partial": 0.049560546875,
+      "step": 3413
+    },
+    {
+      "epoch": 1.788370874803562,
+      "grad_norm": 2.0190961360931396,
+      "learning_rate": 1.3520717676678673e-05,
+      "loss": 0.0999,
+      "step": 3414
+    },
+    {
+      "epoch": 1.788370874803562,
+      "loss_partial": 0.046875,
+      "step": 3414
+    },
+    {
+      "epoch": 1.788370874803562,
+      "loss_partial": 0.044677734375,
+      "step": 3414
+    },
+    {
+      "epoch": 1.78889470927187,
+      "grad_norm": 2.1455607414245605,
+      "learning_rate": 1.3516438429523542e-05,
+      "loss": 0.1039,
+      "step": 3415
+    },
+    {
+      "epoch": 1.78889470927187,
+      "loss_partial": 0.068359375,
+      "step": 3415
+    },
+    {
+      "epoch": 1.78889470927187,
+      "loss_partial": 0.0439453125,
+      "step": 3415
+    },
+    {
+      "epoch": 1.7894185437401782,
+      "grad_norm": 1.984678030014038,
+      "learning_rate": 1.3512158447454529e-05,
+      "loss": 0.1069,
+      "step": 3416
+    },
+    {
+      "epoch": 1.7894185437401782,
+      "loss_partial": 0.052001953125,
+      "step": 3416
+    },
+    {
+      "epoch": 1.7894185437401782,
+      "loss_partial": 0.05126953125,
+      "step": 3416
+    },
+    {
+      "epoch": 1.7899423782084862,
+      "grad_norm": 1.7251983880996704,
+      "learning_rate": 1.3507877731366116e-05,
+      "loss": 0.1051,
+      "step": 3417
+    },
+    {
+      "epoch": 1.7899423782084862,
+      "loss_partial": 0.0478515625,
+      "step": 3417
+    },
+    {
+      "epoch": 1.7899423782084862,
+      "loss_partial": 0.050048828125,
+      "step": 3417
+    },
+    {
+      "epoch": 1.7904662126767943,
+      "grad_norm": 1.657008409500122,
+      "learning_rate": 1.3503596282152951e-05,
+      "loss": 0.1056,
+      "step": 3418
+    },
+    {
+      "epoch": 1.7904662126767943,
+      "loss_partial": 0.044189453125,
+      "step": 3418
+    },
+    {
+      "epoch": 1.7904662126767943,
+      "loss_partial": 0.064453125,
+      "step": 3418
+    },
+    {
+      "epoch": 1.7909900471451021,
+      "grad_norm": 1.7917205095291138,
+      "learning_rate": 1.349931410070983e-05,
+      "loss": 0.1075,
+      "step": 3419
+    },
+    {
+      "epoch": 1.7909900471451021,
+      "loss_partial": 0.04296875,
+      "step": 3419
+    },
+    {
+      "epoch": 1.7909900471451021,
+      "loss_partial": 0.050537109375,
+      "step": 3419
+    },
+    {
+      "epoch": 1.7915138816134102,
+      "grad_norm": 1.0845152139663696,
+      "learning_rate": 1.34950311879317e-05,
+      "loss": 0.1005,
+      "step": 3420
+    },
+    {
+      "epoch": 1.7915138816134102,
+      "loss_partial": 0.048828125,
+      "step": 3420
+    },
+    {
+      "epoch": 1.7915138816134102,
+      "loss_partial": 0.043212890625,
+      "step": 3420
+    },
+    {
+      "epoch": 1.792037716081718,
+      "grad_norm": 1.9161115884780884,
+      "learning_rate": 1.3490747544713662e-05,
+      "loss": 0.1017,
+      "step": 3421
+    },
+    {
+      "epoch": 1.792037716081718,
+      "loss_partial": 0.042724609375,
+      "step": 3421
+    },
+    {
+      "epoch": 1.792037716081718,
+      "loss_partial": 0.061767578125,
+      "step": 3421
+    },
+    {
+      "epoch": 1.792561550550026,
+      "grad_norm": 2.0853896141052246,
+      "learning_rate": 1.3486463171950979e-05,
+      "loss": 0.1047,
+      "step": 3422
+    },
+    {
+      "epoch": 1.792561550550026,
+      "loss_partial": 0.049072265625,
+      "step": 3422
+    },
+    {
+      "epoch": 1.792561550550026,
+      "loss_partial": 0.04052734375,
+      "step": 3422
+    },
+    {
+      "epoch": 1.7930853850183341,
+      "grad_norm": 2.202889919281006,
+      "learning_rate": 1.3482178070539046e-05,
+      "loss": 0.1033,
+      "step": 3423
+    },
+    {
+      "epoch": 1.7930853850183341,
+      "loss_partial": 0.048583984375,
+      "step": 3423
+    },
+    {
+      "epoch": 1.7930853850183341,
+      "loss_partial": 0.040771484375,
+      "step": 3423
+    },
+    {
+      "epoch": 1.7936092194866422,
+      "grad_norm": 2.0118165016174316,
+      "learning_rate": 1.3477892241373434e-05,
+      "loss": 0.1033,
+      "step": 3424
+    },
+    {
+      "epoch": 1.7936092194866422,
+      "loss_partial": 0.050048828125,
+      "step": 3424
+    },
+    {
+      "epoch": 1.7936092194866422,
+      "loss_partial": 0.0546875,
+      "step": 3424
+    },
+    {
+      "epoch": 1.7941330539549503,
+      "grad_norm": 2.267324924468994,
+      "learning_rate": 1.347360568534985e-05,
+      "loss": 0.1056,
+      "step": 3425
+    },
+    {
+      "epoch": 1.7941330539549503,
+      "loss_partial": 0.0478515625,
+      "step": 3425
+    },
+    {
+      "epoch": 1.7941330539549503,
+      "loss_partial": 0.05322265625,
+      "step": 3425
+    },
+    {
+      "epoch": 1.7946568884232583,
+      "grad_norm": 1.7486255168914795,
+      "learning_rate": 1.3469318403364158e-05,
+      "loss": 0.0999,
+      "step": 3426
+    },
+    {
+      "epoch": 1.7946568884232583,
+      "loss_partial": 0.0634765625,
+      "step": 3426
+    },
+    {
+      "epoch": 1.7946568884232583,
+      "loss_partial": 0.0419921875,
+      "step": 3426
+    },
+    {
+      "epoch": 1.7951807228915664,
+      "grad_norm": 1.8113235235214233,
+      "learning_rate": 1.3465030396312368e-05,
+      "loss": 0.1003,
+      "step": 3427
+    },
+    {
+      "epoch": 1.7951807228915664,
+      "loss_partial": 0.048828125,
+      "step": 3427
+    },
+    {
+      "epoch": 1.7951807228915664,
+      "loss_partial": 0.054443359375,
+      "step": 3427
+    },
+    {
+      "epoch": 1.7957045573598744,
+      "grad_norm": 2.116044521331787,
+      "learning_rate": 1.3460741665090658e-05,
+      "loss": 0.1041,
+      "step": 3428
+    },
+    {
+      "epoch": 1.7957045573598744,
+      "loss_partial": 0.052001953125,
+      "step": 3428
+    },
+    {
+      "epoch": 1.7957045573598744,
+      "loss_partial": 0.040283203125,
+      "step": 3428
+    },
+    {
+      "epoch": 1.7962283918281823,
+      "grad_norm": 2.0119247436523438,
+      "learning_rate": 1.3456452210595336e-05,
+      "loss": 0.1013,
+      "step": 3429
+    },
+    {
+      "epoch": 1.7962283918281823,
+      "loss_partial": 0.048095703125,
+      "step": 3429
+    },
+    {
+      "epoch": 1.7962283918281823,
+      "loss_partial": 0.05517578125,
+      "step": 3429
+    },
+    {
+      "epoch": 1.7967522262964903,
+      "grad_norm": 2.2236080169677734,
+      "learning_rate": 1.3452162033722881e-05,
+      "loss": 0.1059,
+      "step": 3430
+    },
+    {
+      "epoch": 1.7967522262964903,
+      "loss_partial": 0.048828125,
+      "step": 3430
+    },
+    {
+      "epoch": 1.7967522262964903,
+      "loss_partial": 0.05322265625,
+      "step": 3430
+    },
+    {
+      "epoch": 1.7972760607647982,
+      "grad_norm": 1.8512706756591797,
+      "learning_rate": 1.3447871135369908e-05,
+      "loss": 0.1003,
+      "step": 3431
+    },
+    {
+      "epoch": 1.7972760607647982,
+      "loss_partial": 0.053955078125,
+      "step": 3431
+    },
+    {
+      "epoch": 1.7972760607647982,
+      "loss_partial": 0.049072265625,
+      "step": 3431
+    },
+    {
+      "epoch": 1.7977998952331062,
+      "grad_norm": 1.940999150276184,
+      "learning_rate": 1.3443579516433188e-05,
+      "loss": 0.1048,
+      "step": 3432
+    },
+    {
+      "epoch": 1.7977998952331062,
+      "loss_partial": 0.056396484375,
+      "step": 3432
+    },
+    {
+      "epoch": 1.7977998952331062,
+      "loss_partial": 0.047607421875,
+      "step": 3432
+    },
+    {
+      "epoch": 1.7983237297014143,
+      "grad_norm": 1.599428653717041,
+      "learning_rate": 1.3439287177809639e-05,
+      "loss": 0.1039,
+      "step": 3433
+    },
+    {
+      "epoch": 1.7983237297014143,
+      "loss_partial": 0.05859375,
+      "step": 3433
+    },
+    {
+      "epoch": 1.7983237297014143,
+      "loss_partial": 0.048095703125,
+      "step": 3433
+    },
+    {
+      "epoch": 1.7988475641697224,
+      "grad_norm": 1.8009244203567505,
+      "learning_rate": 1.3434994120396343e-05,
+      "loss": 0.107,
+      "step": 3434
+    },
+    {
+      "epoch": 1.7988475641697224,
+      "loss_partial": 0.054931640625,
+      "step": 3434
+    },
+    {
+      "epoch": 1.7988475641697224,
+      "loss_partial": 0.04833984375,
+      "step": 3434
+    },
+    {
+      "epoch": 1.7993713986380304,
+      "grad_norm": 1.625229001045227,
+      "learning_rate": 1.3430700345090513e-05,
+      "loss": 0.1037,
+      "step": 3435
+    },
+    {
+      "epoch": 1.7993713986380304,
+      "loss_partial": 0.04833984375,
+      "step": 3435
+    },
+    {
+      "epoch": 1.7993713986380304,
+      "loss_partial": 0.05224609375,
+      "step": 3435
+    },
+    {
+      "epoch": 1.7998952331063385,
+      "grad_norm": 1.3241984844207764,
+      "learning_rate": 1.342640585278953e-05,
+      "loss": 0.105,
+      "step": 3436
+    },
+    {
+      "epoch": 1.7998952331063385,
+      "loss_partial": 0.057861328125,
+      "step": 3436
+    },
+    {
+      "epoch": 1.7998952331063385,
+      "loss_partial": 0.056396484375,
+      "step": 3436
+    },
+    {
+      "epoch": 1.8004190675746465,
+      "grad_norm": 1.6548863649368286,
+      "learning_rate": 1.3422110644390911e-05,
+      "loss": 0.1057,
+      "step": 3437
+    },
+    {
+      "epoch": 1.8004190675746465,
+      "loss_partial": 0.04736328125,
+      "step": 3437
+    },
+    {
+      "epoch": 1.8004190675746465,
+      "loss_partial": 0.048095703125,
+      "step": 3437
+    },
+    {
+      "epoch": 1.8009429020429544,
+      "grad_norm": 1.5867960453033447,
+      "learning_rate": 1.3417814720792328e-05,
+      "loss": 0.101,
+      "step": 3438
+    },
+    {
+      "epoch": 1.8009429020429544,
+      "loss_partial": 0.051513671875,
+      "step": 3438
+    },
+    {
+      "epoch": 1.8009429020429544,
+      "loss_partial": 0.0458984375,
+      "step": 3438
+    },
+    {
+      "epoch": 1.8014667365112624,
+      "grad_norm": 1.3977134227752686,
+      "learning_rate": 1.3413518082891604e-05,
+      "loss": 0.1026,
+      "step": 3439
+    },
+    {
+      "epoch": 1.8014667365112624,
+      "loss_partial": 0.048095703125,
+      "step": 3439
+    },
+    {
+      "epoch": 1.8014667365112624,
+      "loss_partial": 0.056884765625,
+      "step": 3439
+    },
+    {
+      "epoch": 1.8019905709795705,
+      "grad_norm": 2.264728546142578,
+      "learning_rate": 1.3409220731586707e-05,
+      "loss": 0.1013,
+      "step": 3440
+    },
+    {
+      "epoch": 1.8019905709795705,
+      "loss_partial": 0.05029296875,
+      "step": 3440
+    },
+    {
+      "epoch": 1.8019905709795705,
+      "loss_partial": 0.060791015625,
+      "step": 3440
+    },
+    {
+      "epoch": 1.8025144054478783,
+      "grad_norm": 1.8782877922058105,
+      "learning_rate": 1.3404922667775762e-05,
+      "loss": 0.1017,
+      "step": 3441
+    },
+    {
+      "epoch": 1.8025144054478783,
+      "loss_partial": 0.049072265625,
+      "step": 3441
+    },
+    {
+      "epoch": 1.8025144054478783,
+      "loss_partial": 0.06298828125,
+      "step": 3441
+    },
+    {
+      "epoch": 1.8030382399161864,
+      "grad_norm": 2.2540018558502197,
+      "learning_rate": 1.3400623892357034e-05,
+      "loss": 0.1101,
+      "step": 3442
+    },
+    {
+      "epoch": 1.8030382399161864,
+      "loss_partial": 0.054931640625,
+      "step": 3442
+    },
+    {
+      "epoch": 1.8030382399161864,
+      "loss_partial": 0.06103515625,
+      "step": 3442
+    },
+    {
+      "epoch": 1.8035620743844945,
+      "grad_norm": 2.8122408390045166,
+      "learning_rate": 1.3396324406228943e-05,
+      "loss": 0.1048,
+      "step": 3443
+    },
+    {
+      "epoch": 1.8035620743844945,
+      "loss_partial": 0.053466796875,
+      "step": 3443
+    },
+    {
+      "epoch": 1.8035620743844945,
+      "loss_partial": 0.046142578125,
+      "step": 3443
+    },
+    {
+      "epoch": 1.8040859088528025,
+      "grad_norm": 1.4555286169052124,
+      "learning_rate": 1.3392024210290051e-05,
+      "loss": 0.1056,
+      "step": 3444
+    },
+    {
+      "epoch": 1.8040859088528025,
+      "loss_partial": 0.05810546875,
+      "step": 3444
+    },
+    {
+      "epoch": 1.8040859088528025,
+      "loss_partial": 0.054443359375,
+      "step": 3444
+    },
+    {
+      "epoch": 1.8046097433211106,
+      "grad_norm": 2.0844149589538574,
+      "learning_rate": 1.3387723305439073e-05,
+      "loss": 0.104,
+      "step": 3445
+    },
+    {
+      "epoch": 1.8046097433211106,
+      "loss_partial": 0.04296875,
+      "step": 3445
+    },
+    {
+      "epoch": 1.8046097433211106,
+      "loss_partial": 0.04833984375,
+      "step": 3445
+    },
+    {
+      "epoch": 1.8051335777894186,
+      "grad_norm": 1.4508798122406006,
+      "learning_rate": 1.3383421692574875e-05,
+      "loss": 0.1025,
+      "step": 3446
+    },
+    {
+      "epoch": 1.8051335777894186,
+      "loss_partial": 0.049560546875,
+      "step": 3446
+    },
+    {
+      "epoch": 1.8051335777894186,
+      "loss_partial": 0.057373046875,
+      "step": 3446
+    },
+    {
+      "epoch": 1.8056574122577267,
+      "grad_norm": 2.000927448272705,
+      "learning_rate": 1.3379119372596465e-05,
+      "loss": 0.1077,
+      "step": 3447
+    },
+    {
+      "epoch": 1.8056574122577267,
+      "loss_partial": 0.0458984375,
+      "step": 3447
+    },
+    {
+      "epoch": 1.8056574122577267,
+      "loss_partial": 0.0576171875,
+      "step": 3447
+    },
+    {
+      "epoch": 1.8061812467260345,
+      "grad_norm": 1.957266092300415,
+      "learning_rate": 1.3374816346403006e-05,
+      "loss": 0.1016,
+      "step": 3448
+    },
+    {
+      "epoch": 1.8061812467260345,
+      "loss_partial": 0.060546875,
+      "step": 3448
+    },
+    {
+      "epoch": 1.8061812467260345,
+      "loss_partial": 0.048095703125,
+      "step": 3448
+    },
+    {
+      "epoch": 1.8067050811943426,
+      "grad_norm": 1.9551365375518799,
+      "learning_rate": 1.3370512614893798e-05,
+      "loss": 0.1037,
+      "step": 3449
+    },
+    {
+      "epoch": 1.8067050811943426,
+      "loss_partial": 0.05078125,
+      "step": 3449
+    },
+    {
+      "epoch": 1.8067050811943426,
+      "loss_partial": 0.0498046875,
+      "step": 3449
+    },
+    {
+      "epoch": 1.8072289156626506,
+      "grad_norm": 1.5959274768829346,
+      "learning_rate": 1.3366208178968295e-05,
+      "loss": 0.1025,
+      "step": 3450
+    },
+    {
+      "epoch": 1.8072289156626506,
+      "loss_partial": 0.0546875,
+      "step": 3450
+    },
+    {
+      "epoch": 1.8072289156626506,
+      "loss_partial": 0.0400390625,
+      "step": 3450
+    },
+    {
+      "epoch": 1.8077527501309585,
+      "grad_norm": 1.7423290014266968,
+      "learning_rate": 1.3361903039526096e-05,
+      "loss": 0.104,
+      "step": 3451
+    },
+    {
+      "epoch": 1.8077527501309585,
+      "loss_partial": 0.05615234375,
+      "step": 3451
+    },
+    {
+      "epoch": 1.8077527501309585,
+      "loss_partial": 0.039794921875,
+      "step": 3451
+    },
+    {
+      "epoch": 1.8082765845992665,
+      "grad_norm": 1.4955675601959229,
+      "learning_rate": 1.3357597197466955e-05,
+      "loss": 0.1075,
+      "step": 3452
+    },
+    {
+      "epoch": 1.8082765845992665,
+      "loss_partial": 0.042724609375,
+      "step": 3452
+    },
+    {
+      "epoch": 1.8082765845992665,
+      "loss_partial": 0.049560546875,
+      "step": 3452
+    },
+    {
+      "epoch": 1.8088004190675746,
+      "grad_norm": 1.7758243083953857,
+      "learning_rate": 1.3353290653690761e-05,
+      "loss": 0.103,
+      "step": 3453
+    },
+    {
+      "epoch": 1.8088004190675746,
+      "loss_partial": 0.04931640625,
+      "step": 3453
+    },
+    {
+      "epoch": 1.8088004190675746,
+      "loss_partial": 0.04833984375,
+      "step": 3453
+    },
+    {
+      "epoch": 1.8093242535358827,
+      "grad_norm": 1.560744047164917,
+      "learning_rate": 1.3348983409097556e-05,
+      "loss": 0.1029,
+      "step": 3454
+    },
+    {
+      "epoch": 1.8093242535358827,
+      "loss_partial": 0.059326171875,
+      "step": 3454
+    },
+    {
+      "epoch": 1.8093242535358827,
+      "loss_partial": 0.042236328125,
+      "step": 3454
+    },
+    {
+      "epoch": 1.8098480880041907,
+      "grad_norm": 1.5729283094406128,
+      "learning_rate": 1.334467546458753e-05,
+      "loss": 0.1048,
+      "step": 3455
+    },
+    {
+      "epoch": 1.8098480880041907,
+      "loss_partial": 0.04833984375,
+      "step": 3455
+    },
+    {
+      "epoch": 1.8098480880041907,
+      "loss_partial": 0.056884765625,
+      "step": 3455
+    },
+    {
+      "epoch": 1.8103719224724988,
+      "grad_norm": 1.4267603158950806,
+      "learning_rate": 1.3340366821061015e-05,
+      "loss": 0.1049,
+      "step": 3456
+    },
+    {
+      "epoch": 1.8103719224724988,
+      "loss_partial": 0.052490234375,
+      "step": 3456
+    },
+    {
+      "epoch": 1.8103719224724988,
+      "loss_partial": 0.05078125,
+      "step": 3456
+    },
+    {
+      "epoch": 1.8108957569408068,
+      "grad_norm": 1.1915416717529297,
+      "learning_rate": 1.333605747941849e-05,
+      "loss": 0.1032,
+      "step": 3457
+    },
+    {
+      "epoch": 1.8108957569408068,
+      "loss_partial": 0.04833984375,
+      "step": 3457
+    },
+    {
+      "epoch": 1.8108957569408068,
+      "loss_partial": 0.04052734375,
+      "step": 3457
+    },
+    {
+      "epoch": 1.8114195914091147,
+      "grad_norm": 1.6994565725326538,
+      "learning_rate": 1.3331747440560579e-05,
+      "loss": 0.1056,
+      "step": 3458
+    },
+    {
+      "epoch": 1.8114195914091147,
+      "loss_partial": 0.0458984375,
+      "step": 3458
+    },
+    {
+      "epoch": 1.8114195914091147,
+      "loss_partial": 0.05419921875,
+      "step": 3458
+    },
+    {
+      "epoch": 1.8119434258774227,
+      "grad_norm": 1.4912583827972412,
+      "learning_rate": 1.3327436705388056e-05,
+      "loss": 0.1026,
+      "step": 3459
+    },
+    {
+      "epoch": 1.8119434258774227,
+      "loss_partial": 0.0390625,
+      "step": 3459
+    },
+    {
+      "epoch": 1.8119434258774227,
+      "loss_partial": 0.057373046875,
+      "step": 3459
+    },
+    {
+      "epoch": 1.8124672603457308,
+      "grad_norm": 1.5680958032608032,
+      "learning_rate": 1.3323125274801838e-05,
+      "loss": 0.1064,
+      "step": 3460
+    },
+    {
+      "epoch": 1.8124672603457308,
+      "loss_partial": 0.0517578125,
+      "step": 3460
+    },
+    {
+      "epoch": 1.8124672603457308,
+      "loss_partial": 0.04443359375,
+      "step": 3460
+    },
+    {
+      "epoch": 1.8129910948140386,
+      "grad_norm": 1.74692702293396,
+      "learning_rate": 1.331881314970299e-05,
+      "loss": 0.1053,
+      "step": 3461
+    },
+    {
+      "epoch": 1.8129910948140386,
+      "loss_partial": 0.05126953125,
+      "step": 3461
+    },
+    {
+      "epoch": 1.8129910948140386,
+      "loss_partial": 0.04541015625,
+      "step": 3461
+    },
+    {
+      "epoch": 1.8135149292823467,
+      "grad_norm": 1.480643630027771,
+      "learning_rate": 1.3314500330992714e-05,
+      "loss": 0.103,
+      "step": 3462
+    },
+    {
+      "epoch": 1.8135149292823467,
+      "loss_partial": 0.04248046875,
+      "step": 3462
+    },
+    {
+      "epoch": 1.8135149292823467,
+      "loss_partial": 0.048095703125,
+      "step": 3462
+    },
+    {
+      "epoch": 1.8140387637506548,
+      "grad_norm": 1.741967797279358,
+      "learning_rate": 1.3310186819572368e-05,
+      "loss": 0.1046,
+      "step": 3463
+    },
+    {
+      "epoch": 1.8140387637506548,
+      "loss_partial": 0.04443359375,
+      "step": 3463
+    },
+    {
+      "epoch": 1.8140387637506548,
+      "loss_partial": 0.04931640625,
+      "step": 3463
+    },
+    {
+      "epoch": 1.8145625982189628,
+      "grad_norm": 1.9007090330123901,
+      "learning_rate": 1.3305872616343442e-05,
+      "loss": 0.1012,
+      "step": 3464
+    },
+    {
+      "epoch": 1.8145625982189628,
+      "loss_partial": 0.055419921875,
+      "step": 3464
+    },
+    {
+      "epoch": 1.8145625982189628,
+      "loss_partial": 0.048095703125,
+      "step": 3464
+    },
+    {
+      "epoch": 1.8150864326872709,
+      "grad_norm": 1.8132257461547852,
+      "learning_rate": 1.3301557722207586e-05,
+      "loss": 0.1009,
+      "step": 3465
+    },
+    {
+      "epoch": 1.8150864326872709,
+      "loss_partial": 0.041748046875,
+      "step": 3465
+    },
+    {
+      "epoch": 1.8150864326872709,
+      "loss_partial": 0.06884765625,
+      "step": 3465
+    },
+    {
+      "epoch": 1.815610267155579,
+      "grad_norm": 1.7073415517807007,
+      "learning_rate": 1.3297242138066579e-05,
+      "loss": 0.1,
+      "step": 3466
+    },
+    {
+      "epoch": 1.815610267155579,
+      "loss_partial": 0.04638671875,
+      "step": 3466
+    },
+    {
+      "epoch": 1.815610267155579,
+      "loss_partial": 0.0615234375,
+      "step": 3466
+    },
+    {
+      "epoch": 1.816134101623887,
+      "grad_norm": 1.6719945669174194,
+      "learning_rate": 1.3292925864822356e-05,
+      "loss": 0.1032,
+      "step": 3467
+    },
+    {
+      "epoch": 1.816134101623887,
+      "loss_partial": 0.05126953125,
+      "step": 3467
+    },
+    {
+      "epoch": 1.816134101623887,
+      "loss_partial": 0.052001953125,
+      "step": 3467
+    },
+    {
+      "epoch": 1.8166579360921948,
+      "grad_norm": 1.2356096506118774,
+      "learning_rate": 1.3288608903376991e-05,
+      "loss": 0.1045,
+      "step": 3468
+    },
+    {
+      "epoch": 1.8166579360921948,
+      "loss_partial": 0.050048828125,
+      "step": 3468
+    },
+    {
+      "epoch": 1.8166579360921948,
+      "loss_partial": 0.04931640625,
+      "step": 3468
+    },
+    {
+      "epoch": 1.817181770560503,
+      "grad_norm": 1.4245591163635254,
+      "learning_rate": 1.3284291254632701e-05,
+      "loss": 0.1055,
+      "step": 3469
+    },
+    {
+      "epoch": 1.817181770560503,
+      "loss_partial": 0.044921875,
+      "step": 3469
+    },
+    {
+      "epoch": 1.817181770560503,
+      "loss_partial": 0.054443359375,
+      "step": 3469
+    },
+    {
+      "epoch": 1.8177056050288107,
+      "grad_norm": 1.8519150018692017,
+      "learning_rate": 1.3279972919491849e-05,
+      "loss": 0.102,
+      "step": 3470
+    },
+    {
+      "epoch": 1.8177056050288107,
+      "loss_partial": 0.059814453125,
+      "step": 3470
+    },
+    {
+      "epoch": 1.8177056050288107,
+      "loss_partial": 0.052978515625,
+      "step": 3470
+    },
+    {
+      "epoch": 1.8182294394971188,
+      "grad_norm": 1.6448618173599243,
+      "learning_rate": 1.3275653898856939e-05,
+      "loss": 0.1084,
+      "step": 3471
+    },
+    {
+      "epoch": 1.8182294394971188,
+      "loss_partial": 0.054931640625,
+      "step": 3471
+    },
+    {
+      "epoch": 1.8182294394971188,
+      "loss_partial": 0.051513671875,
+      "step": 3471
+    },
+    {
+      "epoch": 1.8187532739654269,
+      "grad_norm": 1.7541135549545288,
+      "learning_rate": 1.3271334193630618e-05,
+      "loss": 0.1026,
+      "step": 3472
+    },
+    {
+      "epoch": 1.8187532739654269,
+      "loss_partial": 0.052734375,
+      "step": 3472
+    },
+    {
+      "epoch": 1.8187532739654269,
+      "loss_partial": 0.05078125,
+      "step": 3472
+    },
+    {
+      "epoch": 1.819277108433735,
+      "grad_norm": 2.1506078243255615,
+      "learning_rate": 1.3267013804715683e-05,
+      "loss": 0.1026,
+      "step": 3473
+    },
+    {
+      "epoch": 1.819277108433735,
+      "loss_partial": 0.051513671875,
+      "step": 3473
+    },
+    {
+      "epoch": 1.819277108433735,
+      "loss_partial": 0.04736328125,
+      "step": 3473
+    },
+    {
+      "epoch": 1.819800942902043,
+      "grad_norm": 1.6740885972976685,
+      "learning_rate": 1.3262692733015063e-05,
+      "loss": 0.1044,
+      "step": 3474
+    },
+    {
+      "epoch": 1.819800942902043,
+      "loss_partial": 0.055419921875,
+      "step": 3474
+    },
+    {
+      "epoch": 1.819800942902043,
+      "loss_partial": 0.04736328125,
+      "step": 3474
+    },
+    {
+      "epoch": 1.820324777370351,
+      "grad_norm": 2.4858226776123047,
+      "learning_rate": 1.325837097943184e-05,
+      "loss": 0.1076,
+      "step": 3475
+    },
+    {
+      "epoch": 1.820324777370351,
+      "loss_partial": 0.054931640625,
+      "step": 3475
+    },
+    {
+      "epoch": 1.820324777370351,
+      "loss_partial": 0.050537109375,
+      "step": 3475
+    },
+    {
+      "epoch": 1.820848611838659,
+      "grad_norm": 2.940753936767578,
+      "learning_rate": 1.3254048544869229e-05,
+      "loss": 0.1033,
+      "step": 3476
+    },
+    {
+      "epoch": 1.820848611838659,
+      "loss_partial": 0.064453125,
+      "step": 3476
+    },
+    {
+      "epoch": 1.820848611838659,
+      "loss_partial": 0.06689453125,
+      "step": 3476
+    },
+    {
+      "epoch": 1.8213724463069672,
+      "grad_norm": 1.891904354095459,
+      "learning_rate": 1.3249725430230595e-05,
+      "loss": 0.1025,
+      "step": 3477
+    },
+    {
+      "epoch": 1.8213724463069672,
+      "loss_partial": 0.05419921875,
+      "step": 3477
+    },
+    {
+      "epoch": 1.8213724463069672,
+      "loss_partial": 0.050537109375,
+      "step": 3477
+    },
+    {
+      "epoch": 1.821896280775275,
+      "grad_norm": 1.9260891675949097,
+      "learning_rate": 1.3245401636419441e-05,
+      "loss": 0.1075,
+      "step": 3478
+    },
+    {
+      "epoch": 1.821896280775275,
+      "loss_partial": 0.041748046875,
+      "step": 3478
+    },
+    {
+      "epoch": 1.821896280775275,
+      "loss_partial": 0.037353515625,
+      "step": 3478
+    },
+    {
+      "epoch": 1.822420115243583,
+      "grad_norm": 2.2736313343048096,
+      "learning_rate": 1.3241077164339418e-05,
+      "loss": 0.0974,
+      "step": 3479
+    },
+    {
+      "epoch": 1.822420115243583,
+      "loss_partial": 0.051513671875,
+      "step": 3479
+    },
+    {
+      "epoch": 1.822420115243583,
+      "loss_partial": 0.048583984375,
+      "step": 3479
+    },
+    {
+      "epoch": 1.8229439497118909,
+      "grad_norm": 1.996047854423523,
+      "learning_rate": 1.3236752014894307e-05,
+      "loss": 0.1002,
+      "step": 3480
+    },
+    {
+      "epoch": 1.8229439497118909,
+      "loss_partial": 0.053955078125,
+      "step": 3480
+    },
+    {
+      "epoch": 1.8229439497118909,
+      "loss_partial": 0.049560546875,
+      "step": 3480
+    },
+    {
+      "epoch": 1.823467784180199,
+      "grad_norm": 1.8842825889587402,
+      "learning_rate": 1.3232426188988041e-05,
+      "loss": 0.1068,
+      "step": 3481
+    },
+    {
+      "epoch": 1.823467784180199,
+      "loss_partial": 0.06103515625,
+      "step": 3481
+    },
+    {
+      "epoch": 1.823467784180199,
+      "loss_partial": 0.04833984375,
+      "step": 3481
+    },
+    {
+      "epoch": 1.823991618648507,
+      "grad_norm": 1.702142357826233,
+      "learning_rate": 1.3228099687524688e-05,
+      "loss": 0.1033,
+      "step": 3482
+    },
+    {
+      "epoch": 1.823991618648507,
+      "loss_partial": 0.056884765625,
+      "step": 3482
+    },
+    {
+      "epoch": 1.823991618648507,
+      "loss_partial": 0.039794921875,
+      "step": 3482
+    },
+    {
+      "epoch": 1.824515453116815,
+      "grad_norm": 2.1850745677948,
+      "learning_rate": 1.3223772511408467e-05,
+      "loss": 0.1032,
+      "step": 3483
+    },
+    {
+      "epoch": 1.824515453116815,
+      "loss_partial": 0.04248046875,
+      "step": 3483
+    },
+    {
+      "epoch": 1.824515453116815,
+      "loss_partial": 0.050537109375,
+      "step": 3483
+    },
+    {
+      "epoch": 1.8250392875851231,
+      "grad_norm": 1.6512846946716309,
+      "learning_rate": 1.321944466154372e-05,
+      "loss": 0.1046,
+      "step": 3484
+    },
+    {
+      "epoch": 1.8250392875851231,
+      "loss_partial": 0.05029296875,
+      "step": 3484
+    },
+    {
+      "epoch": 1.8250392875851231,
+      "loss_partial": 0.06396484375,
+      "step": 3484
+    },
+    {
+      "epoch": 1.8255631220534312,
+      "grad_norm": 1.9539052248001099,
+      "learning_rate": 1.3215116138834953e-05,
+      "loss": 0.1005,
+      "step": 3485
+    },
+    {
+      "epoch": 1.8255631220534312,
+      "loss_partial": 0.046142578125,
+      "step": 3485
+    },
+    {
+      "epoch": 1.8255631220534312,
+      "loss_partial": 0.044921875,
+      "step": 3485
+    },
+    {
+      "epoch": 1.8260869565217392,
+      "grad_norm": 1.726051926612854,
+      "learning_rate": 1.3210786944186793e-05,
+      "loss": 0.1009,
+      "step": 3486
+    },
+    {
+      "epoch": 1.8260869565217392,
+      "loss_partial": 0.045166015625,
+      "step": 3486
+    },
+    {
+      "epoch": 1.8260869565217392,
+      "loss_partial": 0.051513671875,
+      "step": 3486
+    },
+    {
+      "epoch": 1.8266107909900473,
+      "grad_norm": 1.7445255517959595,
+      "learning_rate": 1.3206457078504015e-05,
+      "loss": 0.1033,
+      "step": 3487
+    },
+    {
+      "epoch": 1.8266107909900473,
+      "loss_partial": 0.061279296875,
+      "step": 3487
+    },
+    {
+      "epoch": 1.8266107909900473,
+      "loss_partial": 0.049560546875,
+      "step": 3487
+    },
+    {
+      "epoch": 1.8271346254583551,
+      "grad_norm": 1.697070837020874,
+      "learning_rate": 1.3202126542691536e-05,
+      "loss": 0.1,
+      "step": 3488
+    },
+    {
+      "epoch": 1.8271346254583551,
+      "loss_partial": 0.050537109375,
+      "step": 3488
+    },
+    {
+      "epoch": 1.8271346254583551,
+      "loss_partial": 0.052001953125,
+      "step": 3488
+    },
+    {
+      "epoch": 1.8276584599266632,
+      "grad_norm": 1.6930406093597412,
+      "learning_rate": 1.3197795337654413e-05,
+      "loss": 0.1095,
+      "step": 3489
+    },
+    {
+      "epoch": 1.8276584599266632,
+      "loss_partial": 0.0546875,
+      "step": 3489
+    },
+    {
+      "epoch": 1.8276584599266632,
+      "loss_partial": 0.047607421875,
+      "step": 3489
+    },
+    {
+      "epoch": 1.828182294394971,
+      "grad_norm": 2.205095052719116,
+      "learning_rate": 1.3193463464297836e-05,
+      "loss": 0.1057,
+      "step": 3490
+    },
+    {
+      "epoch": 1.828182294394971,
+      "loss_partial": 0.048095703125,
+      "step": 3490
+    },
+    {
+      "epoch": 1.828182294394971,
+      "loss_partial": 0.05078125,
+      "step": 3490
+    },
+    {
+      "epoch": 1.828706128863279,
+      "grad_norm": 1.594042181968689,
+      "learning_rate": 1.3189130923527145e-05,
+      "loss": 0.1011,
+      "step": 3491
+    },
+    {
+      "epoch": 1.828706128863279,
+      "loss_partial": 0.057861328125,
+      "step": 3491
+    },
+    {
+      "epoch": 1.828706128863279,
+      "loss_partial": 0.04736328125,
+      "step": 3491
+    },
+    {
+      "epoch": 1.8292299633315872,
+      "grad_norm": 1.5664504766464233,
+      "learning_rate": 1.318479771624781e-05,
+      "loss": 0.1015,
+      "step": 3492
+    },
+    {
+      "epoch": 1.8292299633315872,
+      "loss_partial": 0.053955078125,
+      "step": 3492
+    },
+    {
+      "epoch": 1.8292299633315872,
+      "loss_partial": 0.045654296875,
+      "step": 3492
+    },
+    {
+      "epoch": 1.8297537977998952,
+      "grad_norm": 1.7699040174484253,
+      "learning_rate": 1.318046384336545e-05,
+      "loss": 0.1007,
+      "step": 3493
+    },
+    {
+      "epoch": 1.8297537977998952,
+      "loss_partial": 0.0654296875,
+      "step": 3493
+    },
+    {
+      "epoch": 1.8297537977998952,
+      "loss_partial": 0.04833984375,
+      "step": 3493
+    },
+    {
+      "epoch": 1.8302776322682033,
+      "grad_norm": 1.5905261039733887,
+      "learning_rate": 1.317612930578581e-05,
+      "loss": 0.104,
+      "step": 3494
+    },
+    {
+      "epoch": 1.8302776322682033,
+      "loss_partial": 0.052490234375,
+      "step": 3494
+    },
+    {
+      "epoch": 1.8302776322682033,
+      "loss_partial": 0.04833984375,
+      "step": 3494
+    },
+    {
+      "epoch": 1.8308014667365113,
+      "grad_norm": 1.4565174579620361,
+      "learning_rate": 1.317179410441479e-05,
+      "loss": 0.0995,
+      "step": 3495
+    },
+    {
+      "epoch": 1.8308014667365113,
+      "loss_partial": 0.0517578125,
+      "step": 3495
+    },
+    {
+      "epoch": 1.8308014667365113,
+      "loss_partial": 0.05322265625,
+      "step": 3495
+    },
+    {
+      "epoch": 1.8313253012048194,
+      "grad_norm": 1.4309685230255127,
+      "learning_rate": 1.316745824015841e-05,
+      "loss": 0.1031,
+      "step": 3496
+    },
+    {
+      "epoch": 1.8313253012048194,
+      "loss_partial": 0.054443359375,
+      "step": 3496
+    },
+    {
+      "epoch": 1.8313253012048194,
+      "loss_partial": 0.0576171875,
+      "step": 3496
+    },
+    {
+      "epoch": 1.8318491356731272,
+      "grad_norm": 1.930712342262268,
+      "learning_rate": 1.3163121713922851e-05,
+      "loss": 0.1013,
+      "step": 3497
+    },
+    {
+      "epoch": 1.8318491356731272,
+      "loss_partial": 0.057861328125,
+      "step": 3497
+    },
+    {
+      "epoch": 1.8318491356731272,
+      "loss_partial": 0.0576171875,
+      "step": 3497
+    },
+    {
+      "epoch": 1.8323729701414353,
+      "grad_norm": 2.2525339126586914,
+      "learning_rate": 1.315878452661441e-05,
+      "loss": 0.104,
+      "step": 3498
+    },
+    {
+      "epoch": 1.8323729701414353,
+      "loss_partial": 0.04345703125,
+      "step": 3498
+    },
+    {
+      "epoch": 1.8323729701414353,
+      "loss_partial": 0.044677734375,
+      "step": 3498
+    },
+    {
+      "epoch": 1.8328968046097434,
+      "grad_norm": 1.6350946426391602,
+      "learning_rate": 1.3154446679139538e-05,
+      "loss": 0.1058,
+      "step": 3499
+    },
+    {
+      "epoch": 1.8328968046097434,
+      "loss_partial": 0.04931640625,
+      "step": 3499
+    },
+    {
+      "epoch": 1.8328968046097434,
+      "loss_partial": 0.049072265625,
+      "step": 3499
+    },
+    {
+      "epoch": 1.8334206390780512,
+      "grad_norm": 2.3286850452423096,
+      "learning_rate": 1.3150108172404817e-05,
+      "loss": 0.1086,
+      "step": 3500
+    },
+    {
+      "epoch": 1.8334206390780512,
+      "loss_partial": 0.050537109375,
+      "step": 3500
+    },
+    {
+      "epoch": 1.8334206390780512,
+      "loss_partial": 0.04736328125,
+      "step": 3500
+    },
+    {
+      "epoch": 1.8339444735463593,
+      "grad_norm": 1.9732191562652588,
+      "learning_rate": 1.3145769007316966e-05,
+      "loss": 0.1081,
+      "step": 3501
+    },
+    {
+      "epoch": 1.8339444735463593,
+      "loss_partial": 0.04833984375,
+      "step": 3501
+    },
+    {
+      "epoch": 1.8339444735463593,
+      "loss_partial": 0.040771484375,
+      "step": 3501
+    },
+    {
+      "epoch": 1.8344683080146673,
+      "grad_norm": 2.0734925270080566,
+      "learning_rate": 1.3141429184782845e-05,
+      "loss": 0.1019,
+      "step": 3502
+    },
+    {
+      "epoch": 1.8344683080146673,
+      "loss_partial": 0.05615234375,
+      "step": 3502
+    },
+    {
+      "epoch": 1.8344683080146673,
+      "loss_partial": 0.04443359375,
+      "step": 3502
+    },
+    {
+      "epoch": 1.8349921424829754,
+      "grad_norm": 2.1774611473083496,
+      "learning_rate": 1.3137088705709453e-05,
+      "loss": 0.1036,
+      "step": 3503
+    },
+    {
+      "epoch": 1.8349921424829754,
+      "loss_partial": 0.05029296875,
+      "step": 3503
+    },
+    {
+      "epoch": 1.8349921424829754,
+      "loss_partial": 0.05419921875,
+      "step": 3503
+    },
+    {
+      "epoch": 1.8355159769512834,
+      "grad_norm": 1.8405975103378296,
+      "learning_rate": 1.3132747571003916e-05,
+      "loss": 0.1048,
+      "step": 3504
+    },
+    {
+      "epoch": 1.8355159769512834,
+      "loss_partial": 0.06103515625,
+      "step": 3504
+    },
+    {
+      "epoch": 1.8355159769512834,
+      "loss_partial": 0.04443359375,
+      "step": 3504
+    },
+    {
+      "epoch": 1.8360398114195915,
+      "grad_norm": 2.0103137493133545,
+      "learning_rate": 1.3128405781573516e-05,
+      "loss": 0.106,
+      "step": 3505
+    },
+    {
+      "epoch": 1.8360398114195915,
+      "loss_partial": 0.053466796875,
+      "step": 3505
+    },
+    {
+      "epoch": 1.8360398114195915,
+      "loss_partial": 0.042724609375,
+      "step": 3505
+    },
+    {
+      "epoch": 1.8365636458878996,
+      "grad_norm": 2.0809097290039062,
+      "learning_rate": 1.3124063338325648e-05,
+      "loss": 0.1004,
+      "step": 3506
+    },
+    {
+      "epoch": 1.8365636458878996,
+      "loss_partial": 0.053466796875,
+      "step": 3506
+    },
+    {
+      "epoch": 1.8365636458878996,
+      "loss_partial": 0.044189453125,
+      "step": 3506
+    },
+    {
+      "epoch": 1.8370874803562074,
+      "grad_norm": 1.819396734237671,
+      "learning_rate": 1.3119720242167864e-05,
+      "loss": 0.1057,
+      "step": 3507
+    },
+    {
+      "epoch": 1.8370874803562074,
+      "loss_partial": 0.049560546875,
+      "step": 3507
+    },
+    {
+      "epoch": 1.8370874803562074,
+      "loss_partial": 0.05419921875,
+      "step": 3507
+    },
+    {
+      "epoch": 1.8376113148245155,
+      "grad_norm": 2.2465884685516357,
+      "learning_rate": 1.3115376494007837e-05,
+      "loss": 0.1049,
+      "step": 3508
+    },
+    {
+      "epoch": 1.8376113148245155,
+      "loss_partial": 0.045166015625,
+      "step": 3508
+    },
+    {
+      "epoch": 1.8376113148245155,
+      "loss_partial": 0.049072265625,
+      "step": 3508
+    },
+    {
+      "epoch": 1.8381351492928235,
+      "grad_norm": 1.71309232711792,
+      "learning_rate": 1.3111032094753391e-05,
+      "loss": 0.1038,
+      "step": 3509
+    },
+    {
+      "epoch": 1.8381351492928235,
+      "loss_partial": 0.049560546875,
+      "step": 3509
+    },
+    {
+      "epoch": 1.8381351492928235,
+      "loss_partial": 0.0595703125,
+      "step": 3509
+    },
+    {
+      "epoch": 1.8386589837611313,
+      "grad_norm": 1.876616358757019,
+      "learning_rate": 1.3106687045312472e-05,
+      "loss": 0.1037,
+      "step": 3510
+    },
+    {
+      "epoch": 1.8386589837611313,
+      "loss_partial": 0.05615234375,
+      "step": 3510
+    },
+    {
+      "epoch": 1.8386589837611313,
+      "loss_partial": 0.048828125,
+      "step": 3510
+    },
+    {
+      "epoch": 1.8391828182294394,
+      "grad_norm": 1.944832444190979,
+      "learning_rate": 1.3102341346593177e-05,
+      "loss": 0.1055,
+      "step": 3511
+    },
+    {
+      "epoch": 1.8391828182294394,
+      "loss_partial": 0.051513671875,
+      "step": 3511
+    },
+    {
+      "epoch": 1.8391828182294394,
+      "loss_partial": 0.041259765625,
+      "step": 3511
+    },
+    {
+      "epoch": 1.8397066526977475,
+      "grad_norm": 1.75615656375885,
+      "learning_rate": 1.309799499950372e-05,
+      "loss": 0.1039,
+      "step": 3512
+    },
+    {
+      "epoch": 1.8397066526977475,
+      "loss_partial": 0.044677734375,
+      "step": 3512
+    },
+    {
+      "epoch": 1.8397066526977475,
+      "loss_partial": 0.050537109375,
+      "step": 3512
+    },
+    {
+      "epoch": 1.8402304871660555,
+      "grad_norm": 2.2075257301330566,
+      "learning_rate": 1.3093648004952469e-05,
+      "loss": 0.1019,
+      "step": 3513
+    },
+    {
+      "epoch": 1.8402304871660555,
+      "loss_partial": 0.06298828125,
+      "step": 3513
+    },
+    {
+      "epoch": 1.8402304871660555,
+      "loss_partial": 0.0517578125,
+      "step": 3513
+    },
+    {
+      "epoch": 1.8407543216343636,
+      "grad_norm": 1.7288633584976196,
+      "learning_rate": 1.3089300363847914e-05,
+      "loss": 0.1039,
+      "step": 3514
+    },
+    {
+      "epoch": 1.8407543216343636,
+      "loss_partial": 0.046875,
+      "step": 3514
+    },
+    {
+      "epoch": 1.8407543216343636,
+      "loss_partial": 0.04931640625,
+      "step": 3514
+    },
+    {
+      "epoch": 1.8412781561026716,
+      "grad_norm": 2.342735528945923,
+      "learning_rate": 1.3084952077098685e-05,
+      "loss": 0.1039,
+      "step": 3515
+    },
+    {
+      "epoch": 1.8412781561026716,
+      "loss_partial": 0.05126953125,
+      "step": 3515
+    },
+    {
+      "epoch": 1.8412781561026716,
+      "loss_partial": 0.05029296875,
+      "step": 3515
+    },
+    {
+      "epoch": 1.8418019905709797,
+      "grad_norm": 2.162097215652466,
+      "learning_rate": 1.308060314561355e-05,
+      "loss": 0.1046,
+      "step": 3516
+    },
+    {
+      "epoch": 1.8418019905709797,
+      "loss_partial": 0.05419921875,
+      "step": 3516
+    },
+    {
+      "epoch": 1.8418019905709797,
+      "loss_partial": 0.058837890625,
+      "step": 3516
+    },
+    {
+      "epoch": 1.8423258250392875,
+      "grad_norm": 1.7427990436553955,
+      "learning_rate": 1.3076253570301409e-05,
+      "loss": 0.1037,
+      "step": 3517
+    },
+    {
+      "epoch": 1.8423258250392875,
+      "loss_partial": 0.055419921875,
+      "step": 3517
+    },
+    {
+      "epoch": 1.8423258250392875,
+      "loss_partial": 0.0537109375,
+      "step": 3517
+    },
+    {
+      "epoch": 1.8428496595075956,
+      "grad_norm": 1.4171411991119385,
+      "learning_rate": 1.3071903352071294e-05,
+      "loss": 0.1003,
+      "step": 3518
+    },
+    {
+      "epoch": 1.8428496595075956,
+      "loss_partial": 0.049560546875,
+      "step": 3518
+    },
+    {
+      "epoch": 1.8428496595075956,
+      "loss_partial": 0.0517578125,
+      "step": 3518
+    },
+    {
+      "epoch": 1.8433734939759037,
+      "grad_norm": 2.2066633701324463,
+      "learning_rate": 1.3067552491832371e-05,
+      "loss": 0.1043,
+      "step": 3519
+    },
+    {
+      "epoch": 1.8433734939759037,
+      "loss_partial": 0.050537109375,
+      "step": 3519
+    },
+    {
+      "epoch": 1.8433734939759037,
+      "loss_partial": 0.050048828125,
+      "step": 3519
+    },
+    {
+      "epoch": 1.8438973284442115,
+      "grad_norm": 1.7142493724822998,
+      "learning_rate": 1.3063200990493952e-05,
+      "loss": 0.1054,
+      "step": 3520
+    },
+    {
+      "epoch": 1.8438973284442115,
+      "loss_partial": 0.05908203125,
+      "step": 3520
+    },
+    {
+      "epoch": 1.8438973284442115,
+      "loss_partial": 0.05224609375,
+      "step": 3520
+    },
+    {
+      "epoch": 1.8444211629125196,
+      "grad_norm": 2.2165849208831787,
+      "learning_rate": 1.3058848848965464e-05,
+      "loss": 0.1069,
+      "step": 3521
+    },
+    {
+      "epoch": 1.8444211629125196,
+      "loss_partial": 0.059814453125,
+      "step": 3521
+    },
+    {
+      "epoch": 1.8444211629125196,
+      "loss_partial": 0.05078125,
+      "step": 3521
+    },
+    {
+      "epoch": 1.8449449973808276,
+      "grad_norm": 1.684404730796814,
+      "learning_rate": 1.3054496068156483e-05,
+      "loss": 0.1035,
+      "step": 3522
+    },
+    {
+      "epoch": 1.8449449973808276,
+      "loss_partial": 0.044921875,
+      "step": 3522
+    },
+    {
+      "epoch": 1.8449449973808276,
+      "loss_partial": 0.045654296875,
+      "step": 3522
+    },
+    {
+      "epoch": 1.8454688318491357,
+      "grad_norm": 1.7884858846664429,
+      "learning_rate": 1.3050142648976714e-05,
+      "loss": 0.1034,
+      "step": 3523
+    },
+    {
+      "epoch": 1.8454688318491357,
+      "loss_partial": 0.042724609375,
+      "step": 3523
+    },
+    {
+      "epoch": 1.8454688318491357,
+      "loss_partial": 0.0556640625,
+      "step": 3523
+    },
+    {
+      "epoch": 1.8459926663174437,
+      "grad_norm": 1.5362823009490967,
+      "learning_rate": 1.3045788592335993e-05,
+      "loss": 0.1035,
+      "step": 3524
+    },
+    {
+      "epoch": 1.8459926663174437,
+      "loss_partial": 0.044677734375,
+      "step": 3524
+    },
+    {
+      "epoch": 1.8459926663174437,
+      "loss_partial": 0.047607421875,
+      "step": 3524
+    },
+    {
+      "epoch": 1.8465165007857518,
+      "grad_norm": 1.6689263582229614,
+      "learning_rate": 1.3041433899144287e-05,
+      "loss": 0.1029,
+      "step": 3525
+    },
+    {
+      "epoch": 1.8465165007857518,
+      "loss_partial": 0.059814453125,
+      "step": 3525
+    },
+    {
+      "epoch": 1.8465165007857518,
+      "loss_partial": 0.06201171875,
+      "step": 3525
+    },
+    {
+      "epoch": 1.8470403352540599,
+      "grad_norm": 1.2951209545135498,
+      "learning_rate": 1.3037078570311707e-05,
+      "loss": 0.105,
+      "step": 3526
+    },
+    {
+      "epoch": 1.8470403352540599,
+      "loss_partial": 0.048583984375,
+      "step": 3526
+    },
+    {
+      "epoch": 1.8470403352540599,
+      "loss_partial": 0.051025390625,
+      "step": 3526
+    },
+    {
+      "epoch": 1.8475641697223677,
+      "grad_norm": 1.9359936714172363,
+      "learning_rate": 1.3032722606748482e-05,
+      "loss": 0.1051,
+      "step": 3527
+    },
+    {
+      "epoch": 1.8475641697223677,
+      "loss_partial": 0.047119140625,
+      "step": 3527
+    },
+    {
+      "epoch": 1.8475641697223677,
+      "loss_partial": 0.05322265625,
+      "step": 3527
+    },
+    {
+      "epoch": 1.8480880041906758,
+      "grad_norm": 1.5152250528335571,
+      "learning_rate": 1.3028366009364992e-05,
+      "loss": 0.104,
+      "step": 3528
+    },
+    {
+      "epoch": 1.8480880041906758,
+      "loss_partial": 0.058349609375,
+      "step": 3528
+    },
+    {
+      "epoch": 1.8480880041906758,
+      "loss_partial": 0.0458984375,
+      "step": 3528
+    },
+    {
+      "epoch": 1.8486118386589836,
+      "grad_norm": 1.3635709285736084,
+      "learning_rate": 1.3024008779071731e-05,
+      "loss": 0.1051,
+      "step": 3529
+    },
+    {
+      "epoch": 1.8486118386589836,
+      "loss_partial": 0.044677734375,
+      "step": 3529
+    },
+    {
+      "epoch": 1.8486118386589836,
+      "loss_partial": 0.06005859375,
+      "step": 3529
+    },
+    {
+      "epoch": 1.8491356731272917,
+      "grad_norm": 1.790295124053955,
+      "learning_rate": 1.3019650916779334e-05,
+      "loss": 0.1037,
+      "step": 3530
+    },
+    {
+      "epoch": 1.8491356731272917,
+      "loss_partial": 0.046875,
+      "step": 3530
+    },
+    {
+      "epoch": 1.8491356731272917,
+      "loss_partial": 0.0537109375,
+      "step": 3530
+    },
+    {
+      "epoch": 1.8496595075955997,
+      "grad_norm": 1.6433268785476685,
+      "learning_rate": 1.3015292423398569e-05,
+      "loss": 0.103,
+      "step": 3531
+    },
+    {
+      "epoch": 1.8496595075955997,
+      "loss_partial": 0.057861328125,
+      "step": 3531
+    },
+    {
+      "epoch": 1.8496595075955997,
+      "loss_partial": 0.048583984375,
+      "step": 3531
+    },
+    {
+      "epoch": 1.8501833420639078,
+      "grad_norm": 1.8736674785614014,
+      "learning_rate": 1.3010933299840334e-05,
+      "loss": 0.1042,
+      "step": 3532
+    },
+    {
+      "epoch": 1.8501833420639078,
+      "loss_partial": 0.0498046875,
+      "step": 3532
+    },
+    {
+      "epoch": 1.8501833420639078,
+      "loss_partial": 0.039794921875,
+      "step": 3532
+    },
+    {
+      "epoch": 1.8507071765322158,
+      "grad_norm": 1.4991737604141235,
+      "learning_rate": 1.3006573547015657e-05,
+      "loss": 0.1015,
+      "step": 3533
+    },
+    {
+      "epoch": 1.8507071765322158,
+      "loss_partial": 0.0654296875,
+      "step": 3533
+    },
+    {
+      "epoch": 1.8507071765322158,
+      "loss_partial": 0.052490234375,
+      "step": 3533
+    },
+    {
+      "epoch": 1.851231011000524,
+      "grad_norm": 1.959887146949768,
+      "learning_rate": 1.3002213165835701e-05,
+      "loss": 0.107,
+      "step": 3534
+    },
+    {
+      "epoch": 1.851231011000524,
+      "loss_partial": 0.054443359375,
+      "step": 3534
+    },
+    {
+      "epoch": 1.851231011000524,
+      "loss_partial": 0.060791015625,
+      "step": 3534
+    },
+    {
+      "epoch": 1.851754845468832,
+      "grad_norm": 1.4222222566604614,
+      "learning_rate": 1.2997852157211756e-05,
+      "loss": 0.1028,
+      "step": 3535
+    },
+    {
+      "epoch": 1.851754845468832,
+      "loss_partial": 0.043701171875,
+      "step": 3535
+    },
+    {
+      "epoch": 1.851754845468832,
+      "loss_partial": 0.0595703125,
+      "step": 3535
+    },
+    {
+      "epoch": 1.85227867993714,
+      "grad_norm": 1.5000487565994263,
+      "learning_rate": 1.2993490522055257e-05,
+      "loss": 0.1021,
+      "step": 3536
+    },
+    {
+      "epoch": 1.85227867993714,
+      "loss_partial": 0.044189453125,
+      "step": 3536
+    },
+    {
+      "epoch": 1.85227867993714,
+      "loss_partial": 0.045166015625,
+      "step": 3536
+    },
+    {
+      "epoch": 1.8528025144054479,
+      "grad_norm": 1.5394266843795776,
+      "learning_rate": 1.2989128261277742e-05,
+      "loss": 0.1021,
+      "step": 3537
+    },
+    {
+      "epoch": 1.8528025144054479,
+      "loss_partial": 0.04443359375,
+      "step": 3537
+    },
+    {
+      "epoch": 1.8528025144054479,
+      "loss_partial": 0.051025390625,
+      "step": 3537
+    },
+    {
+      "epoch": 1.853326348873756,
+      "grad_norm": 1.8198848962783813,
+      "learning_rate": 1.2984765375790906e-05,
+      "loss": 0.1036,
+      "step": 3538
+    },
+    {
+      "epoch": 1.853326348873756,
+      "loss_partial": 0.044189453125,
+      "step": 3538
+    },
+    {
+      "epoch": 1.853326348873756,
+      "loss_partial": 0.048583984375,
+      "step": 3538
+    },
+    {
+      "epoch": 1.8538501833420638,
+      "grad_norm": 1.1909503936767578,
+      "learning_rate": 1.2980401866506565e-05,
+      "loss": 0.1002,
+      "step": 3539
+    },
+    {
+      "epoch": 1.8538501833420638,
+      "loss_partial": 0.06103515625,
+      "step": 3539
+    },
+    {
+      "epoch": 1.8538501833420638,
+      "loss_partial": 0.0634765625,
+      "step": 3539
+    },
+    {
+      "epoch": 1.8543740178103718,
+      "grad_norm": 1.7665375471115112,
+      "learning_rate": 1.2976037734336664e-05,
+      "loss": 0.1073,
+      "step": 3540
+    },
+    {
+      "epoch": 1.8543740178103718,
+      "loss_partial": 0.050537109375,
+      "step": 3540
+    },
+    {
+      "epoch": 1.8543740178103718,
+      "loss_partial": 0.0654296875,
+      "step": 3540
+    },
+    {
+      "epoch": 1.8548978522786799,
+      "grad_norm": 1.227291226387024,
+      "learning_rate": 1.2971672980193279e-05,
+      "loss": 0.1032,
+      "step": 3541
+    },
+    {
+      "epoch": 1.8548978522786799,
+      "loss_partial": 0.048828125,
+      "step": 3541
+    },
+    {
+      "epoch": 1.8548978522786799,
+      "loss_partial": 0.03515625,
+      "step": 3541
+    },
+    {
+      "epoch": 1.855421686746988,
+      "grad_norm": 1.612934947013855,
+      "learning_rate": 1.296730760498862e-05,
+      "loss": 0.1039,
+      "step": 3542
+    },
+    {
+      "epoch": 1.855421686746988,
+      "loss_partial": 0.042724609375,
+      "step": 3542
+    },
+    {
+      "epoch": 1.855421686746988,
+      "loss_partial": 0.041748046875,
+      "step": 3542
+    },
+    {
+      "epoch": 1.855945521215296,
+      "grad_norm": 1.6675620079040527,
+      "learning_rate": 1.2962941609635018e-05,
+      "loss": 0.1038,
+      "step": 3543
+    },
+    {
+      "epoch": 1.855945521215296,
+      "loss_partial": 0.05322265625,
+      "step": 3543
+    },
+    {
+      "epoch": 1.855945521215296,
+      "loss_partial": 0.05615234375,
+      "step": 3543
+    },
+    {
+      "epoch": 1.856469355683604,
+      "grad_norm": 1.4371885061264038,
+      "learning_rate": 1.2958574995044944e-05,
+      "loss": 0.1064,
+      "step": 3544
+    },
+    {
+      "epoch": 1.856469355683604,
+      "loss_partial": 0.05712890625,
+      "step": 3544
+    },
+    {
+      "epoch": 1.856469355683604,
+      "loss_partial": 0.053466796875,
+      "step": 3544
+    },
+    {
+      "epoch": 1.856993190151912,
+      "grad_norm": 1.8895790576934814,
+      "learning_rate": 1.2954207762130987e-05,
+      "loss": 0.1021,
+      "step": 3545
+    },
+    {
+      "epoch": 1.856993190151912,
+      "loss_partial": 0.05029296875,
+      "step": 3545
+    },
+    {
+      "epoch": 1.856993190151912,
+      "loss_partial": 0.04736328125,
+      "step": 3545
+    },
+    {
+      "epoch": 1.8575170246202202,
+      "grad_norm": 1.9095799922943115,
+      "learning_rate": 1.2949839911805881e-05,
+      "loss": 0.1018,
+      "step": 3546
+    },
+    {
+      "epoch": 1.8575170246202202,
+      "loss_partial": 0.05712890625,
+      "step": 3546
+    },
+    {
+      "epoch": 1.8575170246202202,
+      "loss_partial": 0.041259765625,
+      "step": 3546
+    },
+    {
+      "epoch": 1.858040859088528,
+      "grad_norm": 2.069920539855957,
+      "learning_rate": 1.2945471444982471e-05,
+      "loss": 0.1015,
+      "step": 3547
+    },
+    {
+      "epoch": 1.858040859088528,
+      "loss_partial": 0.049560546875,
+      "step": 3547
+    },
+    {
+      "epoch": 1.858040859088528,
+      "loss_partial": 0.046875,
+      "step": 3547
+    },
+    {
+      "epoch": 1.858564693556836,
+      "grad_norm": 2.1112639904022217,
+      "learning_rate": 1.2941102362573748e-05,
+      "loss": 0.1022,
+      "step": 3548
+    },
+    {
+      "epoch": 1.858564693556836,
+      "loss_partial": 0.050048828125,
+      "step": 3548
+    },
+    {
+      "epoch": 1.858564693556836,
+      "loss_partial": 0.060302734375,
+      "step": 3548
+    },
+    {
+      "epoch": 1.859088528025144,
+      "grad_norm": 2.8362061977386475,
+      "learning_rate": 1.2936732665492813e-05,
+      "loss": 0.1037,
+      "step": 3549
+    },
+    {
+      "epoch": 1.859088528025144,
+      "loss_partial": 0.04638671875,
+      "step": 3549
+    },
+    {
+      "epoch": 1.859088528025144,
+      "loss_partial": 0.040283203125,
+      "step": 3549
+    },
+    {
+      "epoch": 1.859612362493452,
+      "grad_norm": 2.167642593383789,
+      "learning_rate": 1.2932362354652915e-05,
+      "loss": 0.105,
+      "step": 3550
+    },
+    {
+      "epoch": 1.859612362493452,
+      "loss_partial": 0.049560546875,
+      "step": 3550
+    },
+    {
+      "epoch": 1.859612362493452,
+      "loss_partial": 0.046630859375,
+      "step": 3550
+    },
+    {
+      "epoch": 1.86013619696176,
+      "grad_norm": 2.708977699279785,
+      "learning_rate": 1.2927991430967415e-05,
+      "loss": 0.0991,
+      "step": 3551
+    },
+    {
+      "epoch": 1.86013619696176,
+      "loss_partial": 0.048828125,
+      "step": 3551
+    },
+    {
+      "epoch": 1.86013619696176,
+      "loss_partial": 0.054931640625,
+      "step": 3551
+    },
+    {
+      "epoch": 1.860660031430068,
+      "grad_norm": 1.9667545557022095,
+      "learning_rate": 1.2923619895349815e-05,
+      "loss": 0.1042,
+      "step": 3552
+    },
+    {
+      "epoch": 1.860660031430068,
+      "loss_partial": 0.041748046875,
+      "step": 3552
+    },
+    {
+      "epoch": 1.860660031430068,
+      "loss_partial": 0.05517578125,
+      "step": 3552
+    },
+    {
+      "epoch": 1.8611838658983761,
+      "grad_norm": 2.862320899963379,
+      "learning_rate": 1.2919247748713732e-05,
+      "loss": 0.0985,
+      "step": 3553
+    },
+    {
+      "epoch": 1.8611838658983761,
+      "loss_partial": 0.050048828125,
+      "step": 3553
+    },
+    {
+      "epoch": 1.8611838658983761,
+      "loss_partial": 0.062255859375,
+      "step": 3553
+    },
+    {
+      "epoch": 1.8617077003666842,
+      "grad_norm": 2.022576093673706,
+      "learning_rate": 1.2914874991972929e-05,
+      "loss": 0.1031,
+      "step": 3554
+    },
+    {
+      "epoch": 1.8617077003666842,
+      "loss_partial": 0.05810546875,
+      "step": 3554
+    },
+    {
+      "epoch": 1.8617077003666842,
+      "loss_partial": 0.06396484375,
+      "step": 3554
+    },
+    {
+      "epoch": 1.8622315348349923,
+      "grad_norm": 2.9434592723846436,
+      "learning_rate": 1.291050162604127e-05,
+      "loss": 0.1046,
+      "step": 3555
+    },
+    {
+      "epoch": 1.8622315348349923,
+      "loss_partial": 0.04248046875,
+      "step": 3555
+    },
+    {
+      "epoch": 1.8622315348349923,
+      "loss_partial": 0.054931640625,
+      "step": 3555
+    },
+    {
+      "epoch": 1.8627553693033003,
+      "grad_norm": 1.8421663045883179,
+      "learning_rate": 1.2906127651832772e-05,
+      "loss": 0.1035,
+      "step": 3556
+    },
+    {
+      "epoch": 1.8627553693033003,
+      "loss_partial": 0.05615234375,
+      "step": 3556
+    },
+    {
+      "epoch": 1.8627553693033003,
+      "loss_partial": 0.056640625,
+      "step": 3556
+    },
+    {
+      "epoch": 1.8632792037716082,
+      "grad_norm": 2.436563491821289,
+      "learning_rate": 1.2901753070261565e-05,
+      "loss": 0.1045,
+      "step": 3557
+    },
+    {
+      "epoch": 1.8632792037716082,
+      "loss_partial": 0.060791015625,
+      "step": 3557
+    },
+    {
+      "epoch": 1.8632792037716082,
+      "loss_partial": 0.055419921875,
+      "step": 3557
+    },
+    {
+      "epoch": 1.8638030382399162,
+      "grad_norm": 1.5071804523468018,
+      "learning_rate": 1.2897377882241913e-05,
+      "loss": 0.1059,
+      "step": 3558
+    },
+    {
+      "epoch": 1.8638030382399162,
+      "loss_partial": 0.04833984375,
+      "step": 3558
+    },
+    {
+      "epoch": 1.8638030382399162,
+      "loss_partial": 0.049072265625,
+      "step": 3558
+    },
+    {
+      "epoch": 1.864326872708224,
+      "grad_norm": 1.9075536727905273,
+      "learning_rate": 1.2893002088688196e-05,
+      "loss": 0.1005,
+      "step": 3559
+    },
+    {
+      "epoch": 1.864326872708224,
+      "loss_partial": 0.06298828125,
+      "step": 3559
+    },
+    {
+      "epoch": 1.864326872708224,
+      "loss_partial": 0.060546875,
+      "step": 3559
+    },
+    {
+      "epoch": 1.8648507071765321,
+      "grad_norm": 1.519271731376648,
+      "learning_rate": 1.2888625690514937e-05,
+      "loss": 0.1066,
+      "step": 3560
+    },
+    {
+      "epoch": 1.8648507071765321,
+      "loss_partial": 0.052734375,
+      "step": 3560
+    },
+    {
+      "epoch": 1.8648507071765321,
+      "loss_partial": 0.0546875,
+      "step": 3560
+    },
+    {
+      "epoch": 1.8653745416448402,
+      "grad_norm": 1.9648252725601196,
+      "learning_rate": 1.2884248688636769e-05,
+      "loss": 0.0991,
+      "step": 3561
+    },
+    {
+      "epoch": 1.8653745416448402,
+      "loss_partial": 0.047607421875,
+      "step": 3561
+    },
+    {
+      "epoch": 1.8653745416448402,
+      "loss_partial": 0.046630859375,
+      "step": 3561
+    },
+    {
+      "epoch": 1.8658983761131482,
+      "grad_norm": 1.5310251712799072,
+      "learning_rate": 1.287987108396846e-05,
+      "loss": 0.1022,
+      "step": 3562
+    },
+    {
+      "epoch": 1.8658983761131482,
+      "loss_partial": 0.052978515625,
+      "step": 3562
+    },
+    {
+      "epoch": 1.8658983761131482,
+      "loss_partial": 0.044921875,
+      "step": 3562
+    },
+    {
+      "epoch": 1.8664222105814563,
+      "grad_norm": 2.17948842048645,
+      "learning_rate": 1.2875492877424905e-05,
+      "loss": 0.1023,
+      "step": 3563
+    },
+    {
+      "epoch": 1.8664222105814563,
+      "loss_partial": 0.049560546875,
+      "step": 3563
+    },
+    {
+      "epoch": 1.8664222105814563,
+      "loss_partial": 0.044189453125,
+      "step": 3563
+    },
+    {
+      "epoch": 1.8669460450497644,
+      "grad_norm": 1.7376842498779297,
+      "learning_rate": 1.2871114069921117e-05,
+      "loss": 0.1005,
+      "step": 3564
+    },
+    {
+      "epoch": 1.8669460450497644,
+      "loss_partial": 0.05810546875,
+      "step": 3564
+    },
+    {
+      "epoch": 1.8669460450497644,
+      "loss_partial": 0.054443359375,
+      "step": 3564
+    },
+    {
+      "epoch": 1.8674698795180724,
+      "grad_norm": 2.0607380867004395,
+      "learning_rate": 1.2866734662372244e-05,
+      "loss": 0.1061,
+      "step": 3565
+    },
+    {
+      "epoch": 1.8674698795180724,
+      "loss_partial": 0.06591796875,
+      "step": 3565
+    },
+    {
+      "epoch": 1.8674698795180724,
+      "loss_partial": 0.048828125,
+      "step": 3565
+    },
+    {
+      "epoch": 1.8679937139863803,
+      "grad_norm": 2.1242475509643555,
+      "learning_rate": 1.2862354655693556e-05,
+      "loss": 0.1037,
+      "step": 3566
+    },
+    {
+      "epoch": 1.8679937139863803,
+      "loss_partial": 0.052490234375,
+      "step": 3566
+    },
+    {
+      "epoch": 1.8679937139863803,
+      "loss_partial": 0.05224609375,
+      "step": 3566
+    },
+    {
+      "epoch": 1.8685175484546883,
+      "grad_norm": 1.5825755596160889,
+      "learning_rate": 1.2857974050800445e-05,
+      "loss": 0.1026,
+      "step": 3567
+    },
+    {
+      "epoch": 1.8685175484546883,
+      "loss_partial": 0.039306640625,
+      "step": 3567
+    },
+    {
+      "epoch": 1.8685175484546883,
+      "loss_partial": 0.05322265625,
+      "step": 3567
+    },
+    {
+      "epoch": 1.8690413829229964,
+      "grad_norm": 1.6845415830612183,
+      "learning_rate": 1.2853592848608432e-05,
+      "loss": 0.1044,
+      "step": 3568
+    },
+    {
+      "epoch": 1.8690413829229964,
+      "loss_partial": 0.05224609375,
+      "step": 3568
+    },
+    {
+      "epoch": 1.8690413829229964,
+      "loss_partial": 0.05224609375,
+      "step": 3568
+    },
+    {
+      "epoch": 1.8695652173913042,
+      "grad_norm": 1.815574288368225,
+      "learning_rate": 1.2849211050033154e-05,
+      "loss": 0.103,
+      "step": 3569
+    },
+    {
+      "epoch": 1.8695652173913042,
+      "loss_partial": 0.05712890625,
+      "step": 3569
+    },
+    {
+      "epoch": 1.8695652173913042,
+      "loss_partial": 0.046875,
+      "step": 3569
+    },
+    {
+      "epoch": 1.8700890518596123,
+      "grad_norm": 1.8067140579223633,
+      "learning_rate": 1.284482865599039e-05,
+      "loss": 0.1032,
+      "step": 3570
+    },
+    {
+      "epoch": 1.8700890518596123,
+      "loss_partial": 0.045654296875,
+      "step": 3570
+    },
+    {
+      "epoch": 1.8700890518596123,
+      "loss_partial": 0.053466796875,
+      "step": 3570
+    },
+    {
+      "epoch": 1.8706128863279203,
+      "grad_norm": 1.6134756803512573,
+      "learning_rate": 1.2840445667396026e-05,
+      "loss": 0.0976,
+      "step": 3571
+    },
+    {
+      "epoch": 1.8706128863279203,
+      "loss_partial": 0.054931640625,
+      "step": 3571
+    },
+    {
+      "epoch": 1.8706128863279203,
+      "loss_partial": 0.050048828125,
+      "step": 3571
+    },
+    {
+      "epoch": 1.8711367207962284,
+      "grad_norm": 2.14243221282959,
+      "learning_rate": 1.2836062085166085e-05,
+      "loss": 0.106,
+      "step": 3572
+    },
+    {
+      "epoch": 1.8711367207962284,
+      "loss_partial": 0.05078125,
+      "step": 3572
+    },
+    {
+      "epoch": 1.8711367207962284,
+      "loss_partial": 0.052490234375,
+      "step": 3572
+    },
+    {
+      "epoch": 1.8716605552645365,
+      "grad_norm": 1.5315227508544922,
+      "learning_rate": 1.2831677910216706e-05,
+      "loss": 0.1076,
+      "step": 3573
+    },
+    {
+      "epoch": 1.8716605552645365,
+      "loss_partial": 0.055419921875,
+      "step": 3573
+    },
+    {
+      "epoch": 1.8716605552645365,
+      "loss_partial": 0.052490234375,
+      "step": 3573
+    },
+    {
+      "epoch": 1.8721843897328445,
+      "grad_norm": 1.8461523056030273,
+      "learning_rate": 1.2827293143464155e-05,
+      "loss": 0.1075,
+      "step": 3574
+    },
+    {
+      "epoch": 1.8721843897328445,
+      "loss_partial": 0.0517578125,
+      "step": 3574
+    },
+    {
+      "epoch": 1.8721843897328445,
+      "loss_partial": 0.041259765625,
+      "step": 3574
+    },
+    {
+      "epoch": 1.8727082242011526,
+      "grad_norm": 1.5719919204711914,
+      "learning_rate": 1.2822907785824818e-05,
+      "loss": 0.1029,
+      "step": 3575
+    },
+    {
+      "epoch": 1.8727082242011526,
+      "loss_partial": 0.0517578125,
+      "step": 3575
+    },
+    {
+      "epoch": 1.8727082242011526,
+      "loss_partial": 0.05810546875,
+      "step": 3575
+    },
+    {
+      "epoch": 1.8732320586694604,
+      "grad_norm": 1.8850047588348389,
+      "learning_rate": 1.281852183821521e-05,
+      "loss": 0.1049,
+      "step": 3576
+    },
+    {
+      "epoch": 1.8732320586694604,
+      "loss_partial": 0.054443359375,
+      "step": 3576
+    },
+    {
+      "epoch": 1.8732320586694604,
+      "loss_partial": 0.046142578125,
+      "step": 3576
+    },
+    {
+      "epoch": 1.8737558931377685,
+      "grad_norm": 1.8835570812225342,
+      "learning_rate": 1.2814135301551968e-05,
+      "loss": 0.104,
+      "step": 3577
+    },
+    {
+      "epoch": 1.8737558931377685,
+      "loss_partial": 0.05029296875,
+      "step": 3577
+    },
+    {
+      "epoch": 1.8737558931377685,
+      "loss_partial": 0.062255859375,
+      "step": 3577
+    },
+    {
+      "epoch": 1.8742797276060765,
+      "grad_norm": 1.8714312314987183,
+      "learning_rate": 1.2809748176751849e-05,
+      "loss": 0.1056,
+      "step": 3578
+    },
+    {
+      "epoch": 1.8742797276060765,
+      "loss_partial": 0.051513671875,
+      "step": 3578
+    },
+    {
+      "epoch": 1.8742797276060765,
+      "loss_partial": 0.05712890625,
+      "step": 3578
+    },
+    {
+      "epoch": 1.8748035620743844,
+      "grad_norm": 1.8740702867507935,
+      "learning_rate": 1.280536046473174e-05,
+      "loss": 0.1056,
+      "step": 3579
+    },
+    {
+      "epoch": 1.8748035620743844,
+      "loss_partial": 0.06201171875,
+      "step": 3579
+    },
+    {
+      "epoch": 1.8748035620743844,
+      "loss_partial": 0.047119140625,
+      "step": 3579
+    },
+    {
+      "epoch": 1.8753273965426924,
+      "grad_norm": 1.7437055110931396,
+      "learning_rate": 1.2800972166408637e-05,
+      "loss": 0.1049,
+      "step": 3580
+    },
+    {
+      "epoch": 1.8753273965426924,
+      "loss_partial": 0.052001953125,
+      "step": 3580
+    },
+    {
+      "epoch": 1.8753273965426924,
+      "loss_partial": 0.056640625,
+      "step": 3580
+    },
+    {
+      "epoch": 1.8758512310110005,
+      "grad_norm": 2.526331901550293,
+      "learning_rate": 1.2796583282699673e-05,
+      "loss": 0.1018,
+      "step": 3581
+    },
+    {
+      "epoch": 1.8758512310110005,
+      "loss_partial": 0.05224609375,
+      "step": 3581
+    },
+    {
+      "epoch": 1.8758512310110005,
+      "loss_partial": 0.05126953125,
+      "step": 3581
+    },
+    {
+      "epoch": 1.8763750654793085,
+      "grad_norm": 1.4176855087280273,
+      "learning_rate": 1.2792193814522097e-05,
+      "loss": 0.1019,
+      "step": 3582
+    },
+    {
+      "epoch": 1.8763750654793085,
+      "loss_partial": 0.04736328125,
+      "step": 3582
+    },
+    {
+      "epoch": 1.8763750654793085,
+      "loss_partial": 0.0615234375,
+      "step": 3582
+    },
+    {
+      "epoch": 1.8768988999476166,
+      "grad_norm": 1.5419855117797852,
+      "learning_rate": 1.278780376279328e-05,
+      "loss": 0.1045,
+      "step": 3583
+    },
+    {
+      "epoch": 1.8768988999476166,
+      "loss_partial": 0.053466796875,
+      "step": 3583
+    },
+    {
+      "epoch": 1.8768988999476166,
+      "loss_partial": 0.048828125,
+      "step": 3583
+    },
+    {
+      "epoch": 1.8774227344159247,
+      "grad_norm": 1.6988694667816162,
+      "learning_rate": 1.2783413128430712e-05,
+      "loss": 0.0997,
+      "step": 3584
+    },
+    {
+      "epoch": 1.8774227344159247,
+      "loss_partial": 0.051513671875,
+      "step": 3584
+    },
+    {
+      "epoch": 1.8774227344159247,
+      "loss_partial": 0.0458984375,
+      "step": 3584
+    },
+    {
+      "epoch": 1.8779465688842327,
+      "grad_norm": 1.563900113105774,
+      "learning_rate": 1.2779021912352022e-05,
+      "loss": 0.1021,
+      "step": 3585
+    },
+    {
+      "epoch": 1.8779465688842327,
+      "loss_partial": 0.06005859375,
+      "step": 3585
+    },
+    {
+      "epoch": 1.8779465688842327,
+      "loss_partial": 0.05615234375,
+      "step": 3585
+    },
+    {
+      "epoch": 1.8784704033525406,
+      "grad_norm": 1.2902953624725342,
+      "learning_rate": 1.277463011547493e-05,
+      "loss": 0.1045,
+      "step": 3586
+    },
+    {
+      "epoch": 1.8784704033525406,
+      "loss_partial": 0.05712890625,
+      "step": 3586
+    },
+    {
+      "epoch": 1.8784704033525406,
+      "loss_partial": 0.051025390625,
+      "step": 3586
+    },
+    {
+      "epoch": 1.8789942378208486,
+      "grad_norm": 1.940757155418396,
+      "learning_rate": 1.2770237738717301e-05,
+      "loss": 0.1035,
+      "step": 3587
+    },
+    {
+      "epoch": 1.8789942378208486,
+      "loss_partial": 0.053955078125,
+      "step": 3587
+    },
+    {
+      "epoch": 1.8789942378208486,
+      "loss_partial": 0.04541015625,
+      "step": 3587
+    },
+    {
+      "epoch": 1.8795180722891565,
+      "grad_norm": 1.4398363828659058,
+      "learning_rate": 1.2765844782997117e-05,
+      "loss": 0.1045,
+      "step": 3588
+    },
+    {
+      "epoch": 1.8795180722891565,
+      "loss_partial": 0.0517578125,
+      "step": 3588
+    },
+    {
+      "epoch": 1.8795180722891565,
+      "loss_partial": 0.05615234375,
+      "step": 3588
+    },
+    {
+      "epoch": 1.8800419067574645,
+      "grad_norm": 1.3868626356124878,
+      "learning_rate": 1.276145124923248e-05,
+      "loss": 0.107,
+      "step": 3589
+    },
+    {
+      "epoch": 1.8800419067574645,
+      "loss_partial": 0.0498046875,
+      "step": 3589
+    },
+    {
+      "epoch": 1.8800419067574645,
+      "loss_partial": 0.05224609375,
+      "step": 3589
+    },
+    {
+      "epoch": 1.8805657412257726,
+      "grad_norm": 2.2697746753692627,
+      "learning_rate": 1.2757057138341601e-05,
+      "loss": 0.106,
+      "step": 3590
+    },
+    {
+      "epoch": 1.8805657412257726,
+      "loss_partial": 0.0400390625,
+      "step": 3590
+    },
+    {
+      "epoch": 1.8805657412257726,
+      "loss_partial": 0.04736328125,
+      "step": 3590
+    },
+    {
+      "epoch": 1.8810895756940806,
+      "grad_norm": 1.0234349966049194,
+      "learning_rate": 1.2752662451242839e-05,
+      "loss": 0.1033,
+      "step": 3591
+    },
+    {
+      "epoch": 1.8810895756940806,
+      "loss_partial": 0.043701171875,
+      "step": 3591
+    },
+    {
+      "epoch": 1.8810895756940806,
+      "loss_partial": 0.0419921875,
+      "step": 3591
+    },
+    {
+      "epoch": 1.8816134101623887,
+      "grad_norm": 2.3167724609375,
+      "learning_rate": 1.274826718885464e-05,
+      "loss": 0.1047,
+      "step": 3592
+    },
+    {
+      "epoch": 1.8816134101623887,
+      "loss_partial": 0.051025390625,
+      "step": 3592
+    },
+    {
+      "epoch": 1.8816134101623887,
+      "loss_partial": 0.05078125,
+      "step": 3592
+    },
+    {
+      "epoch": 1.8821372446306968,
+      "grad_norm": 1.7664874792099,
+      "learning_rate": 1.2743871352095593e-05,
+      "loss": 0.1025,
+      "step": 3593
+    },
+    {
+      "epoch": 1.8821372446306968,
+      "loss_partial": 0.0517578125,
+      "step": 3593
+    },
+    {
+      "epoch": 1.8821372446306968,
+      "loss_partial": 0.0615234375,
+      "step": 3593
+    },
+    {
+      "epoch": 1.8826610790990048,
+      "grad_norm": 1.3643354177474976,
+      "learning_rate": 1.2739474941884403e-05,
+      "loss": 0.1042,
+      "step": 3594
+    },
+    {
+      "epoch": 1.8826610790990048,
+      "loss_partial": 0.05419921875,
+      "step": 3594
+    },
+    {
+      "epoch": 1.8826610790990048,
+      "loss_partial": 0.047607421875,
+      "step": 3594
+    },
+    {
+      "epoch": 1.8831849135673129,
+      "grad_norm": 2.2897865772247314,
+      "learning_rate": 1.273507795913989e-05,
+      "loss": 0.104,
+      "step": 3595
+    },
+    {
+      "epoch": 1.8831849135673129,
+      "loss_partial": 0.059326171875,
+      "step": 3595
+    },
+    {
+      "epoch": 1.8831849135673129,
+      "loss_partial": 0.05078125,
+      "step": 3595
+    },
+    {
+      "epoch": 1.8837087480356207,
+      "grad_norm": 1.8455891609191895,
+      "learning_rate": 1.2730680404780994e-05,
+      "loss": 0.1025,
+      "step": 3596
+    },
+    {
+      "epoch": 1.8837087480356207,
+      "loss_partial": 0.050048828125,
+      "step": 3596
+    },
+    {
+      "epoch": 1.8837087480356207,
+      "loss_partial": 0.048828125,
+      "step": 3596
+    },
+    {
+      "epoch": 1.8842325825039288,
+      "grad_norm": 1.5829212665557861,
+      "learning_rate": 1.2726282279726788e-05,
+      "loss": 0.1042,
+      "step": 3597
+    },
+    {
+      "epoch": 1.8842325825039288,
+      "loss_partial": 0.047607421875,
+      "step": 3597
+    },
+    {
+      "epoch": 1.8842325825039288,
+      "loss_partial": 0.0498046875,
+      "step": 3597
+    },
+    {
+      "epoch": 1.8847564169722366,
+      "grad_norm": 2.5885424613952637,
+      "learning_rate": 1.2721883584896438e-05,
+      "loss": 0.1028,
+      "step": 3598
+    },
+    {
+      "epoch": 1.8847564169722366,
+      "loss_partial": 0.0576171875,
+      "step": 3598
+    },
+    {
+      "epoch": 1.8847564169722366,
+      "loss_partial": 0.045654296875,
+      "step": 3598
+    },
+    {
+      "epoch": 1.8852802514405447,
+      "grad_norm": 1.5356537103652954,
+      "learning_rate": 1.271748432120925e-05,
+      "loss": 0.1056,
+      "step": 3599
+    },
+    {
+      "epoch": 1.8852802514405447,
+      "loss_partial": 0.0517578125,
+      "step": 3599
+    },
+    {
+      "epoch": 1.8852802514405447,
+      "loss_partial": 0.05322265625,
+      "step": 3599
+    },
+    {
+      "epoch": 1.8858040859088527,
+      "grad_norm": 2.0329580307006836,
+      "learning_rate": 1.2713084489584642e-05,
+      "loss": 0.1036,
+      "step": 3600
+    },
+    {
+      "epoch": 1.8858040859088527,
+      "loss_partial": 0.0556640625,
+      "step": 3600
+    },
+    {
+      "epoch": 1.8858040859088527,
+      "loss_partial": 0.05126953125,
+      "step": 3600
+    },
+    {
+      "epoch": 1.8863279203771608,
+      "grad_norm": 1.8290971517562866,
+      "learning_rate": 1.2708684090942155e-05,
+      "loss": 0.1061,
+      "step": 3601
+    },
+    {
+      "epoch": 1.8863279203771608,
+      "loss_partial": 0.05419921875,
+      "step": 3601
+    },
+    {
+      "epoch": 1.8863279203771608,
+      "loss_partial": 0.05517578125,
+      "step": 3601
+    },
+    {
+      "epoch": 1.8868517548454689,
+      "grad_norm": 2.126718044281006,
+      "learning_rate": 1.2704283126201444e-05,
+      "loss": 0.1065,
+      "step": 3602
+    },
+    {
+      "epoch": 1.8868517548454689,
+      "loss_partial": 0.047607421875,
+      "step": 3602
+    },
+    {
+      "epoch": 1.8868517548454689,
+      "loss_partial": 0.0458984375,
+      "step": 3602
+    },
+    {
+      "epoch": 1.887375589313777,
+      "grad_norm": 1.733028531074524,
+      "learning_rate": 1.2699881596282282e-05,
+      "loss": 0.1021,
+      "step": 3603
+    },
+    {
+      "epoch": 1.887375589313777,
+      "loss_partial": 0.046630859375,
+      "step": 3603
+    },
+    {
+      "epoch": 1.887375589313777,
+      "loss_partial": 0.059814453125,
+      "step": 3603
+    },
+    {
+      "epoch": 1.887899423782085,
+      "grad_norm": 1.4733211994171143,
+      "learning_rate": 1.2695479502104561e-05,
+      "loss": 0.1057,
+      "step": 3604
+    },
+    {
+      "epoch": 1.887899423782085,
+      "loss_partial": 0.04638671875,
+      "step": 3604
+    },
+    {
+      "epoch": 1.887899423782085,
+      "loss_partial": 0.06494140625,
+      "step": 3604
+    },
+    {
+      "epoch": 1.888423258250393,
+      "grad_norm": 2.068082094192505,
+      "learning_rate": 1.2691076844588296e-05,
+      "loss": 0.1051,
+      "step": 3605
+    },
+    {
+      "epoch": 1.888423258250393,
+      "loss_partial": 0.05224609375,
+      "step": 3605
+    },
+    {
+      "epoch": 1.888423258250393,
+      "loss_partial": 0.05419921875,
+      "step": 3605
+    },
+    {
+      "epoch": 1.8889470927187009,
+      "grad_norm": 1.945989727973938,
+      "learning_rate": 1.2686673624653606e-05,
+      "loss": 0.1046,
+      "step": 3606
+    },
+    {
+      "epoch": 1.8889470927187009,
+      "loss_partial": 0.051025390625,
+      "step": 3606
+    },
+    {
+      "epoch": 1.8889470927187009,
+      "loss_partial": 0.055908203125,
+      "step": 3606
+    },
+    {
+      "epoch": 1.889470927187009,
+      "grad_norm": 1.6425772905349731,
+      "learning_rate": 1.2682269843220744e-05,
+      "loss": 0.1059,
+      "step": 3607
+    },
+    {
+      "epoch": 1.889470927187009,
+      "loss_partial": 0.05615234375,
+      "step": 3607
+    },
+    {
+      "epoch": 1.889470927187009,
+      "loss_partial": 0.051025390625,
+      "step": 3607
+    },
+    {
+      "epoch": 1.8899947616553168,
+      "grad_norm": 2.5733723640441895,
+      "learning_rate": 1.2677865501210071e-05,
+      "loss": 0.1038,
+      "step": 3608
+    },
+    {
+      "epoch": 1.8899947616553168,
+      "loss_partial": 0.052978515625,
+      "step": 3608
+    },
+    {
+      "epoch": 1.8899947616553168,
+      "loss_partial": 0.043701171875,
+      "step": 3608
+    },
+    {
+      "epoch": 1.8905185961236248,
+      "grad_norm": 1.897334337234497,
+      "learning_rate": 1.267346059954207e-05,
+      "loss": 0.1045,
+      "step": 3609
+    },
+    {
+      "epoch": 1.8905185961236248,
+      "loss_partial": 0.0615234375,
+      "step": 3609
+    },
+    {
+      "epoch": 1.8905185961236248,
+      "loss_partial": 0.04638671875,
+      "step": 3609
+    },
+    {
+      "epoch": 1.8910424305919329,
+      "grad_norm": 2.002110481262207,
+      "learning_rate": 1.2669055139137336e-05,
+      "loss": 0.1019,
+      "step": 3610
+    },
+    {
+      "epoch": 1.8910424305919329,
+      "loss_partial": 0.041748046875,
+      "step": 3610
+    },
+    {
+      "epoch": 1.8910424305919329,
+      "loss_partial": 0.0634765625,
+      "step": 3610
+    },
+    {
+      "epoch": 1.891566265060241,
+      "grad_norm": 1.6748496294021606,
+      "learning_rate": 1.2664649120916584e-05,
+      "loss": 0.1042,
+      "step": 3611
+    },
+    {
+      "epoch": 1.891566265060241,
+      "loss_partial": 0.05029296875,
+      "step": 3611
+    },
+    {
+      "epoch": 1.891566265060241,
+      "loss_partial": 0.05078125,
+      "step": 3611
+    },
+    {
+      "epoch": 1.892090099528549,
+      "grad_norm": 1.8964760303497314,
+      "learning_rate": 1.2660242545800638e-05,
+      "loss": 0.1042,
+      "step": 3612
+    },
+    {
+      "epoch": 1.892090099528549,
+      "loss_partial": 0.046142578125,
+      "step": 3612
+    },
+    {
+      "epoch": 1.892090099528549,
+      "loss_partial": 0.053955078125,
+      "step": 3612
+    },
+    {
+      "epoch": 1.892613933996857,
+      "grad_norm": 1.789864420890808,
+      "learning_rate": 1.2655835414710456e-05,
+      "loss": 0.1034,
+      "step": 3613
+    },
+    {
+      "epoch": 1.892613933996857,
+      "loss_partial": 0.0537109375,
+      "step": 3613
+    },
+    {
+      "epoch": 1.892613933996857,
+      "loss_partial": 0.05810546875,
+      "step": 3613
+    },
+    {
+      "epoch": 1.8931377684651651,
+      "grad_norm": 2.022592306137085,
+      "learning_rate": 1.265142772856709e-05,
+      "loss": 0.1073,
+      "step": 3614
+    },
+    {
+      "epoch": 1.8931377684651651,
+      "loss_partial": 0.04150390625,
+      "step": 3614
+    },
+    {
+      "epoch": 1.8931377684651651,
+      "loss_partial": 0.050537109375,
+      "step": 3614
+    },
+    {
+      "epoch": 1.8936616029334732,
+      "grad_norm": 1.7335740327835083,
+      "learning_rate": 1.2647019488291729e-05,
+      "loss": 0.1009,
+      "step": 3615
+    },
+    {
+      "epoch": 1.8936616029334732,
+      "loss_partial": 0.0537109375,
+      "step": 3615
+    },
+    {
+      "epoch": 1.8936616029334732,
+      "loss_partial": 0.0439453125,
+      "step": 3615
+    },
+    {
+      "epoch": 1.894185437401781,
+      "grad_norm": 1.592719554901123,
+      "learning_rate": 1.2642610694805664e-05,
+      "loss": 0.1025,
+      "step": 3616
+    },
+    {
+      "epoch": 1.894185437401781,
+      "loss_partial": 0.056396484375,
+      "step": 3616
+    },
+    {
+      "epoch": 1.894185437401781,
+      "loss_partial": 0.04541015625,
+      "step": 3616
+    },
+    {
+      "epoch": 1.894709271870089,
+      "grad_norm": 1.6965513229370117,
+      "learning_rate": 1.2638201349030304e-05,
+      "loss": 0.102,
+      "step": 3617
+    },
+    {
+      "epoch": 1.894709271870089,
+      "loss_partial": 0.04833984375,
+      "step": 3617
+    },
+    {
+      "epoch": 1.894709271870089,
+      "loss_partial": 0.049560546875,
+      "step": 3617
+    },
+    {
+      "epoch": 1.895233106338397,
+      "grad_norm": 1.5341368913650513,
+      "learning_rate": 1.2633791451887174e-05,
+      "loss": 0.1045,
+      "step": 3618
+    },
+    {
+      "epoch": 1.895233106338397,
+      "loss_partial": 0.05517578125,
+      "step": 3618
+    },
+    {
+      "epoch": 1.895233106338397,
+      "loss_partial": 0.05322265625,
+      "step": 3618
+    },
+    {
+      "epoch": 1.895756940806705,
+      "grad_norm": 1.5870779752731323,
+      "learning_rate": 1.262938100429792e-05,
+      "loss": 0.1086,
+      "step": 3619
+    },
+    {
+      "epoch": 1.895756940806705,
+      "loss_partial": 0.0498046875,
+      "step": 3619
+    },
+    {
+      "epoch": 1.895756940806705,
+      "loss_partial": 0.057373046875,
+      "step": 3619
+    },
+    {
+      "epoch": 1.896280775275013,
+      "grad_norm": 1.3740516901016235,
+      "learning_rate": 1.2624970007184291e-05,
+      "loss": 0.104,
+      "step": 3620
+    },
+    {
+      "epoch": 1.896280775275013,
+      "loss_partial": 0.044921875,
+      "step": 3620
+    },
+    {
+      "epoch": 1.896280775275013,
+      "loss_partial": 0.048583984375,
+      "step": 3620
+    },
+    {
+      "epoch": 1.896804609743321,
+      "grad_norm": 1.9660522937774658,
+      "learning_rate": 1.2620558461468169e-05,
+      "loss": 0.1042,
+      "step": 3621
+    },
+    {
+      "epoch": 1.896804609743321,
+      "loss_partial": 0.052734375,
+      "step": 3621
+    },
+    {
+      "epoch": 1.896804609743321,
+      "loss_partial": 0.04833984375,
+      "step": 3621
+    },
+    {
+      "epoch": 1.8973284442116292,
+      "grad_norm": 1.7893122434616089,
+      "learning_rate": 1.2616146368071532e-05,
+      "loss": 0.104,
+      "step": 3622
+    },
+    {
+      "epoch": 1.8973284442116292,
+      "loss_partial": 0.046142578125,
+      "step": 3622
+    },
+    {
+      "epoch": 1.8973284442116292,
+      "loss_partial": 0.052001953125,
+      "step": 3622
+    },
+    {
+      "epoch": 1.8978522786799372,
+      "grad_norm": 2.129554271697998,
+      "learning_rate": 1.2611733727916483e-05,
+      "loss": 0.1013,
+      "step": 3623
+    },
+    {
+      "epoch": 1.8978522786799372,
+      "loss_partial": 0.042724609375,
+      "step": 3623
+    },
+    {
+      "epoch": 1.8978522786799372,
+      "loss_partial": 0.046875,
+      "step": 3623
+    },
+    {
+      "epoch": 1.8983761131482453,
+      "grad_norm": 1.3428629636764526,
+      "learning_rate": 1.2607320541925232e-05,
+      "loss": 0.1014,
+      "step": 3624
+    },
+    {
+      "epoch": 1.8983761131482453,
+      "loss_partial": 0.051513671875,
+      "step": 3624
+    },
+    {
+      "epoch": 1.8983761131482453,
+      "loss_partial": 0.047607421875,
+      "step": 3624
+    },
+    {
+      "epoch": 1.8988999476165531,
+      "grad_norm": 1.69319486618042,
+      "learning_rate": 1.2602906811020112e-05,
+      "loss": 0.1044,
+      "step": 3625
+    },
+    {
+      "epoch": 1.8988999476165531,
+      "loss_partial": 0.052490234375,
+      "step": 3625
+    },
+    {
+      "epoch": 1.8988999476165531,
+      "loss_partial": 0.05126953125,
+      "step": 3625
+    },
+    {
+      "epoch": 1.8994237820848612,
+      "grad_norm": 1.8162256479263306,
+      "learning_rate": 1.2598492536123564e-05,
+      "loss": 0.1,
+      "step": 3626
+    },
+    {
+      "epoch": 1.8994237820848612,
+      "loss_partial": 0.0576171875,
+      "step": 3626
+    },
+    {
+      "epoch": 1.8994237820848612,
+      "loss_partial": 0.04541015625,
+      "step": 3626
+    },
+    {
+      "epoch": 1.8999476165531692,
+      "grad_norm": 1.5575478076934814,
+      "learning_rate": 1.2594077718158147e-05,
+      "loss": 0.1042,
+      "step": 3627
+    },
+    {
+      "epoch": 1.8999476165531692,
+      "loss_partial": 0.050537109375,
+      "step": 3627
+    },
+    {
+      "epoch": 1.8999476165531692,
+      "loss_partial": 0.046630859375,
+      "step": 3627
+    },
+    {
+      "epoch": 1.900471451021477,
+      "grad_norm": 1.646298885345459,
+      "learning_rate": 1.2589662358046526e-05,
+      "loss": 0.1022,
+      "step": 3628
+    },
+    {
+      "epoch": 1.900471451021477,
+      "loss_partial": 0.0517578125,
+      "step": 3628
+    },
+    {
+      "epoch": 1.900471451021477,
+      "loss_partial": 0.04248046875,
+      "step": 3628
+    },
+    {
+      "epoch": 1.9009952854897851,
+      "grad_norm": 1.6359760761260986,
+      "learning_rate": 1.2585246456711487e-05,
+      "loss": 0.1035,
+      "step": 3629
+    },
+    {
+      "epoch": 1.9009952854897851,
+      "loss_partial": 0.047607421875,
+      "step": 3629
+    },
+    {
+      "epoch": 1.9009952854897851,
+      "loss_partial": 0.04833984375,
+      "step": 3629
+    },
+    {
+      "epoch": 1.9015191199580932,
+      "grad_norm": 1.7456717491149902,
+      "learning_rate": 1.2580830015075926e-05,
+      "loss": 0.1012,
+      "step": 3630
+    },
+    {
+      "epoch": 1.9015191199580932,
+      "loss_partial": 0.050048828125,
+      "step": 3630
+    },
+    {
+      "epoch": 1.9015191199580932,
+      "loss_partial": 0.049072265625,
+      "step": 3630
+    },
+    {
+      "epoch": 1.9020429544264013,
+      "grad_norm": 1.4911717176437378,
+      "learning_rate": 1.2576413034062848e-05,
+      "loss": 0.1024,
+      "step": 3631
+    },
+    {
+      "epoch": 1.9020429544264013,
+      "loss_partial": 0.052001953125,
+      "step": 3631
+    },
+    {
+      "epoch": 1.9020429544264013,
+      "loss_partial": 0.05029296875,
+      "step": 3631
+    },
+    {
+      "epoch": 1.9025667888947093,
+      "grad_norm": 1.5405336618423462,
+      "learning_rate": 1.257199551459538e-05,
+      "loss": 0.1041,
+      "step": 3632
+    },
+    {
+      "epoch": 1.9025667888947093,
+      "loss_partial": 0.05322265625,
+      "step": 3632
+    },
+    {
+      "epoch": 1.9025667888947093,
+      "loss_partial": 0.053466796875,
+      "step": 3632
+    },
+    {
+      "epoch": 1.9030906233630174,
+      "grad_norm": 1.4847190380096436,
+      "learning_rate": 1.2567577457596753e-05,
+      "loss": 0.1057,
+      "step": 3633
+    },
+    {
+      "epoch": 1.9030906233630174,
+      "loss_partial": 0.05126953125,
+      "step": 3633
+    },
+    {
+      "epoch": 1.9030906233630174,
+      "loss_partial": 0.06640625,
+      "step": 3633
+    },
+    {
+      "epoch": 1.9036144578313254,
+      "grad_norm": 1.8779125213623047,
+      "learning_rate": 1.2563158863990319e-05,
+      "loss": 0.1044,
+      "step": 3634
+    },
+    {
+      "epoch": 1.9036144578313254,
+      "loss_partial": 0.051513671875,
+      "step": 3634
+    },
+    {
+      "epoch": 1.9036144578313254,
+      "loss_partial": 0.061279296875,
+      "step": 3634
+    },
+    {
+      "epoch": 1.9041382922996333,
+      "grad_norm": 1.6592323780059814,
+      "learning_rate": 1.255873973469953e-05,
+      "loss": 0.1042,
+      "step": 3635
+    },
+    {
+      "epoch": 1.9041382922996333,
+      "loss_partial": 0.05419921875,
+      "step": 3635
+    },
+    {
+      "epoch": 1.9041382922996333,
+      "loss_partial": 0.047607421875,
+      "step": 3635
+    },
+    {
+      "epoch": 1.9046621267679413,
+      "grad_norm": 1.8324942588806152,
+      "learning_rate": 1.2554320070647958e-05,
+      "loss": 0.1022,
+      "step": 3636
+    },
+    {
+      "epoch": 1.9046621267679413,
+      "loss_partial": 0.05322265625,
+      "step": 3636
+    },
+    {
+      "epoch": 1.9046621267679413,
+      "loss_partial": 0.050537109375,
+      "step": 3636
+    },
+    {
+      "epoch": 1.9051859612362494,
+      "grad_norm": 2.0041656494140625,
+      "learning_rate": 1.2549899872759288e-05,
+      "loss": 0.1048,
+      "step": 3637
+    },
+    {
+      "epoch": 1.9051859612362494,
+      "loss_partial": 0.0498046875,
+      "step": 3637
+    },
+    {
+      "epoch": 1.9051859612362494,
+      "loss_partial": 0.0439453125,
+      "step": 3637
+    },
+    {
+      "epoch": 1.9057097957045572,
+      "grad_norm": 1.7367979288101196,
+      "learning_rate": 1.2545479141957312e-05,
+      "loss": 0.1039,
+      "step": 3638
+    },
+    {
+      "epoch": 1.9057097957045572,
+      "loss_partial": 0.05859375,
+      "step": 3638
+    },
+    {
+      "epoch": 1.9057097957045572,
+      "loss_partial": 0.0517578125,
+      "step": 3638
+    },
+    {
+      "epoch": 1.9062336301728653,
+      "grad_norm": 1.6862943172454834,
+      "learning_rate": 1.2541057879165935e-05,
+      "loss": 0.0982,
+      "step": 3639
+    },
+    {
+      "epoch": 1.9062336301728653,
+      "loss_partial": 0.0556640625,
+      "step": 3639
+    },
+    {
+      "epoch": 1.9062336301728653,
+      "loss_partial": 0.047119140625,
+      "step": 3639
+    },
+    {
+      "epoch": 1.9067574646411733,
+      "grad_norm": 1.4257228374481201,
+      "learning_rate": 1.2536636085309177e-05,
+      "loss": 0.105,
+      "step": 3640
+    },
+    {
+      "epoch": 1.9067574646411733,
+      "loss_partial": 0.0400390625,
+      "step": 3640
+    },
+    {
+      "epoch": 1.9067574646411733,
+      "loss_partial": 0.053955078125,
+      "step": 3640
+    },
+    {
+      "epoch": 1.9072812991094814,
+      "grad_norm": 1.5906404256820679,
+      "learning_rate": 1.2532213761311162e-05,
+      "loss": 0.1035,
+      "step": 3641
+    },
+    {
+      "epoch": 1.9072812991094814,
+      "loss_partial": 0.053466796875,
+      "step": 3641
+    },
+    {
+      "epoch": 1.9072812991094814,
+      "loss_partial": 0.046142578125,
+      "step": 3641
+    },
+    {
+      "epoch": 1.9078051335777895,
+      "grad_norm": 1.3719887733459473,
+      "learning_rate": 1.2527790908096132e-05,
+      "loss": 0.1042,
+      "step": 3642
+    },
+    {
+      "epoch": 1.9078051335777895,
+      "loss_partial": 0.05224609375,
+      "step": 3642
+    },
+    {
+      "epoch": 1.9078051335777895,
+      "loss_partial": 0.04296875,
+      "step": 3642
+    },
+    {
+      "epoch": 1.9083289680460975,
+      "grad_norm": 1.4580798149108887,
+      "learning_rate": 1.2523367526588427e-05,
+      "loss": 0.1037,
+      "step": 3643
+    },
+    {
+      "epoch": 1.9083289680460975,
+      "loss_partial": 0.049560546875,
+      "step": 3643
+    },
+    {
+      "epoch": 1.9083289680460975,
+      "loss_partial": 0.044921875,
+      "step": 3643
+    },
+    {
+      "epoch": 1.9088528025144056,
+      "grad_norm": 1.5981743335723877,
+      "learning_rate": 1.251894361771252e-05,
+      "loss": 0.1025,
+      "step": 3644
+    },
+    {
+      "epoch": 1.9088528025144056,
+      "loss_partial": 0.0498046875,
+      "step": 3644
+    },
+    {
+      "epoch": 1.9088528025144056,
+      "loss_partial": 0.04638671875,
+      "step": 3644
+    },
+    {
+      "epoch": 1.9093766369827134,
+      "grad_norm": 2.1398510932922363,
+      "learning_rate": 1.251451918239297e-05,
+      "loss": 0.1046,
+      "step": 3645
+    },
+    {
+      "epoch": 1.9093766369827134,
+      "loss_partial": 0.05078125,
+      "step": 3645
+    },
+    {
+      "epoch": 1.9093766369827134,
+      "loss_partial": 0.048095703125,
+      "step": 3645
+    },
+    {
+      "epoch": 1.9099004714510215,
+      "grad_norm": 1.862260341644287,
+      "learning_rate": 1.251009422155446e-05,
+      "loss": 0.1048,
+      "step": 3646
+    },
+    {
+      "epoch": 1.9099004714510215,
+      "loss_partial": 0.052001953125,
+      "step": 3646
+    },
+    {
+      "epoch": 1.9099004714510215,
+      "loss_partial": 0.06884765625,
+      "step": 3646
+    },
+    {
+      "epoch": 1.9104243059193295,
+      "grad_norm": 1.6089377403259277,
+      "learning_rate": 1.2505668736121783e-05,
+      "loss": 0.1039,
+      "step": 3647
+    },
+    {
+      "epoch": 1.9104243059193295,
+      "loss_partial": 0.058349609375,
+      "step": 3647
+    },
+    {
+      "epoch": 1.9104243059193295,
+      "loss_partial": 0.044677734375,
+      "step": 3647
+    },
+    {
+      "epoch": 1.9109481403876374,
+      "grad_norm": 1.6086385250091553,
+      "learning_rate": 1.2501242727019834e-05,
+      "loss": 0.1023,
+      "step": 3648
+    },
+    {
+      "epoch": 1.9109481403876374,
+      "loss_partial": 0.046142578125,
+      "step": 3648
+    },
+    {
+      "epoch": 1.9109481403876374,
+      "loss_partial": 0.048828125,
+      "step": 3648
+    },
+    {
+      "epoch": 1.9114719748559454,
+      "grad_norm": 1.6481298208236694,
+      "learning_rate": 1.2496816195173617e-05,
+      "loss": 0.1037,
+      "step": 3649
+    },
+    {
+      "epoch": 1.9114719748559454,
+      "loss_partial": 0.05078125,
+      "step": 3649
+    },
+    {
+      "epoch": 1.9114719748559454,
+      "loss_partial": 0.04248046875,
+      "step": 3649
+    },
+    {
+      "epoch": 1.9119958093242535,
+      "grad_norm": 1.6183115243911743,
+      "learning_rate": 1.249238914150826e-05,
+      "loss": 0.1022,
+      "step": 3650
+    },
+    {
+      "epoch": 1.9119958093242535,
+      "loss_partial": 0.058349609375,
+      "step": 3650
+    },
+    {
+      "epoch": 1.9119958093242535,
+      "loss_partial": 0.04931640625,
+      "step": 3650
+    },
+    {
+      "epoch": 1.9125196437925616,
+      "grad_norm": 2.0502984523773193,
+      "learning_rate": 1.2487961566948982e-05,
+      "loss": 0.1046,
+      "step": 3651
+    },
+    {
+      "epoch": 1.9125196437925616,
+      "loss_partial": 0.0615234375,
+      "step": 3651
+    },
+    {
+      "epoch": 1.9125196437925616,
+      "loss_partial": 0.048828125,
+      "step": 3651
+    },
+    {
+      "epoch": 1.9130434782608696,
+      "grad_norm": 1.3825616836547852,
+      "learning_rate": 1.2483533472421126e-05,
+      "loss": 0.1021,
+      "step": 3652
+    },
+    {
+      "epoch": 1.9130434782608696,
+      "loss_partial": 0.057373046875,
+      "step": 3652
+    },
+    {
+      "epoch": 1.9130434782608696,
+      "loss_partial": 0.04541015625,
+      "step": 3652
+    },
+    {
+      "epoch": 1.9135673127291777,
+      "grad_norm": 1.5802963972091675,
+      "learning_rate": 1.2479104858850129e-05,
+      "loss": 0.1049,
+      "step": 3653
+    },
+    {
+      "epoch": 1.9135673127291777,
+      "loss_partial": 0.054931640625,
+      "step": 3653
+    },
+    {
+      "epoch": 1.9135673127291777,
+      "loss_partial": 0.047119140625,
+      "step": 3653
+    },
+    {
+      "epoch": 1.9140911471974857,
+      "grad_norm": 1.5844980478286743,
+      "learning_rate": 1.2474675727161549e-05,
+      "loss": 0.1032,
+      "step": 3654
+    },
+    {
+      "epoch": 1.9140911471974857,
+      "loss_partial": 0.053466796875,
+      "step": 3654
+    },
+    {
+      "epoch": 1.9140911471974857,
+      "loss_partial": 0.04248046875,
+      "step": 3654
+    },
+    {
+      "epoch": 1.9146149816657936,
+      "grad_norm": 1.383912205696106,
+      "learning_rate": 1.247024607828104e-05,
+      "loss": 0.1026,
+      "step": 3655
+    },
+    {
+      "epoch": 1.9146149816657936,
+      "loss_partial": 0.051025390625,
+      "step": 3655
+    },
+    {
+      "epoch": 1.9146149816657936,
+      "loss_partial": 0.0615234375,
+      "step": 3655
+    },
+    {
+      "epoch": 1.9151388161341016,
+      "grad_norm": 2.2387351989746094,
+      "learning_rate": 1.246581591313438e-05,
+      "loss": 0.1028,
+      "step": 3656
+    },
+    {
+      "epoch": 1.9151388161341016,
+      "loss_partial": 0.044921875,
+      "step": 3656
+    },
+    {
+      "epoch": 1.9151388161341016,
+      "loss_partial": 0.04345703125,
+      "step": 3656
+    },
+    {
+      "epoch": 1.9156626506024095,
+      "grad_norm": 1.6651811599731445,
+      "learning_rate": 1.2461385232647441e-05,
+      "loss": 0.1038,
+      "step": 3657
+    },
+    {
+      "epoch": 1.9156626506024095,
+      "loss_partial": 0.0439453125,
+      "step": 3657
+    },
+    {
+      "epoch": 1.9156626506024095,
+      "loss_partial": 0.045166015625,
+      "step": 3657
+    },
+    {
+      "epoch": 1.9161864850707175,
+      "grad_norm": 1.6232390403747559,
+      "learning_rate": 1.245695403774621e-05,
+      "loss": 0.1058,
+      "step": 3658
+    },
+    {
+      "epoch": 1.9161864850707175,
+      "loss_partial": 0.052978515625,
+      "step": 3658
+    },
+    {
+      "epoch": 1.9161864850707175,
+      "loss_partial": 0.0537109375,
+      "step": 3658
+    },
+    {
+      "epoch": 1.9167103195390256,
+      "grad_norm": 1.7615864276885986,
+      "learning_rate": 1.245252232935678e-05,
+      "loss": 0.1009,
+      "step": 3659
+    },
+    {
+      "epoch": 1.9167103195390256,
+      "loss_partial": 0.045166015625,
+      "step": 3659
+    },
+    {
+      "epoch": 1.9167103195390256,
+      "loss_partial": 0.04443359375,
+      "step": 3659
+    },
+    {
+      "epoch": 1.9172341540073337,
+      "grad_norm": 1.857097864151001,
+      "learning_rate": 1.2448090108405348e-05,
+      "loss": 0.1055,
+      "step": 3660
+    },
+    {
+      "epoch": 1.9172341540073337,
+      "loss_partial": 0.054931640625,
+      "step": 3660
+    },
+    {
+      "epoch": 1.9172341540073337,
+      "loss_partial": 0.04345703125,
+      "step": 3660
+    },
+    {
+      "epoch": 1.9177579884756417,
+      "grad_norm": 1.9654662609100342,
+      "learning_rate": 1.2443657375818219e-05,
+      "loss": 0.1041,
+      "step": 3661
+    },
+    {
+      "epoch": 1.9177579884756417,
+      "loss_partial": 0.045654296875,
+      "step": 3661
+    },
+    {
+      "epoch": 1.9177579884756417,
+      "loss_partial": 0.05419921875,
+      "step": 3661
+    },
+    {
+      "epoch": 1.9182818229439498,
+      "grad_norm": 1.7292393445968628,
+      "learning_rate": 1.2439224132521811e-05,
+      "loss": 0.1028,
+      "step": 3662
+    },
+    {
+      "epoch": 1.9182818229439498,
+      "loss_partial": 0.048828125,
+      "step": 3662
+    },
+    {
+      "epoch": 1.9182818229439498,
+      "loss_partial": 0.044921875,
+      "step": 3662
+    },
+    {
+      "epoch": 1.9188056574122578,
+      "grad_norm": 1.749749779701233,
+      "learning_rate": 1.2434790379442645e-05,
+      "loss": 0.106,
+      "step": 3663
+    },
+    {
+      "epoch": 1.9188056574122578,
+      "loss_partial": 0.047607421875,
+      "step": 3663
+    },
+    {
+      "epoch": 1.9188056574122578,
+      "loss_partial": 0.035400390625,
+      "step": 3663
+    },
+    {
+      "epoch": 1.919329491880566,
+      "grad_norm": 1.6156631708145142,
+      "learning_rate": 1.2430356117507344e-05,
+      "loss": 0.1004,
+      "step": 3664
+    },
+    {
+      "epoch": 1.919329491880566,
+      "loss_partial": 0.052734375,
+      "step": 3664
+    },
+    {
+      "epoch": 1.919329491880566,
+      "loss_partial": 0.04638671875,
+      "step": 3664
+    },
+    {
+      "epoch": 1.9198533263488737,
+      "grad_norm": 1.5892438888549805,
+      "learning_rate": 1.2425921347642645e-05,
+      "loss": 0.1023,
+      "step": 3665
+    },
+    {
+      "epoch": 1.9198533263488737,
+      "loss_partial": 0.051025390625,
+      "step": 3665
+    },
+    {
+      "epoch": 1.9198533263488737,
+      "loss_partial": 0.0498046875,
+      "step": 3665
+    },
+    {
+      "epoch": 1.9203771608171818,
+      "grad_norm": 2.35719633102417,
+      "learning_rate": 1.2421486070775386e-05,
+      "loss": 0.103,
+      "step": 3666
+    },
+    {
+      "epoch": 1.9203771608171818,
+      "loss_partial": 0.046630859375,
+      "step": 3666
+    },
+    {
+      "epoch": 1.9203771608171818,
+      "loss_partial": 0.049560546875,
+      "step": 3666
+    },
+    {
+      "epoch": 1.9209009952854896,
+      "grad_norm": 1.3504724502563477,
+      "learning_rate": 1.241705028783251e-05,
+      "loss": 0.1001,
+      "step": 3667
+    },
+    {
+      "epoch": 1.9209009952854896,
+      "loss_partial": 0.056396484375,
+      "step": 3667
+    },
+    {
+      "epoch": 1.9209009952854896,
+      "loss_partial": 0.05908203125,
+      "step": 3667
+    },
+    {
+      "epoch": 1.9214248297537977,
+      "grad_norm": 2.1223888397216797,
+      "learning_rate": 1.2412613999741073e-05,
+      "loss": 0.1029,
+      "step": 3668
+    },
+    {
+      "epoch": 1.9214248297537977,
+      "loss_partial": 0.0517578125,
+      "step": 3668
+    },
+    {
+      "epoch": 1.9214248297537977,
+      "loss_partial": 0.051025390625,
+      "step": 3668
+    },
+    {
+      "epoch": 1.9219486642221058,
+      "grad_norm": 1.91873037815094,
+      "learning_rate": 1.2408177207428224e-05,
+      "loss": 0.1007,
+      "step": 3669
+    },
+    {
+      "epoch": 1.9219486642221058,
+      "loss_partial": 0.05224609375,
+      "step": 3669
+    },
+    {
+      "epoch": 1.9219486642221058,
+      "loss_partial": 0.04443359375,
+      "step": 3669
+    },
+    {
+      "epoch": 1.9224724986904138,
+      "grad_norm": 1.774730920791626,
+      "learning_rate": 1.2403739911821235e-05,
+      "loss": 0.1014,
+      "step": 3670
+    },
+    {
+      "epoch": 1.9224724986904138,
+      "loss_partial": 0.048095703125,
+      "step": 3670
+    },
+    {
+      "epoch": 1.9224724986904138,
+      "loss_partial": 0.051513671875,
+      "step": 3670
+    },
+    {
+      "epoch": 1.9229963331587219,
+      "grad_norm": 2.0459389686584473,
+      "learning_rate": 1.239930211384747e-05,
+      "loss": 0.1004,
+      "step": 3671
+    },
+    {
+      "epoch": 1.9229963331587219,
+      "loss_partial": 0.048095703125,
+      "step": 3671
+    },
+    {
+      "epoch": 1.9229963331587219,
+      "loss_partial": 0.048828125,
+      "step": 3671
+    },
+    {
+      "epoch": 1.92352016762703,
+      "grad_norm": 1.4658162593841553,
+      "learning_rate": 1.2394863814434399e-05,
+      "loss": 0.1059,
+      "step": 3672
+    },
+    {
+      "epoch": 1.92352016762703,
+      "loss_partial": 0.0458984375,
+      "step": 3672
+    },
+    {
+      "epoch": 1.92352016762703,
+      "loss_partial": 0.056396484375,
+      "step": 3672
+    },
+    {
+      "epoch": 1.924044002095338,
+      "grad_norm": 2.0273845195770264,
+      "learning_rate": 1.2390425014509597e-05,
+      "loss": 0.1044,
+      "step": 3673
+    },
+    {
+      "epoch": 1.924044002095338,
+      "loss_partial": 0.044189453125,
+      "step": 3673
+    },
+    {
+      "epoch": 1.924044002095338,
+      "loss_partial": 0.051513671875,
+      "step": 3673
+    },
+    {
+      "epoch": 1.924567836563646,
+      "grad_norm": 1.9061720371246338,
+      "learning_rate": 1.238598571500075e-05,
+      "loss": 0.1034,
+      "step": 3674
+    },
+    {
+      "epoch": 1.924567836563646,
+      "loss_partial": 0.052001953125,
+      "step": 3674
+    },
+    {
+      "epoch": 1.924567836563646,
+      "loss_partial": 0.0517578125,
+      "step": 3674
+    },
+    {
+      "epoch": 1.9250916710319539,
+      "grad_norm": 2.187915325164795,
+      "learning_rate": 1.2381545916835645e-05,
+      "loss": 0.102,
+      "step": 3675
+    },
+    {
+      "epoch": 1.9250916710319539,
+      "loss_partial": 0.0458984375,
+      "step": 3675
+    },
+    {
+      "epoch": 1.9250916710319539,
+      "loss_partial": 0.0673828125,
+      "step": 3675
+    },
+    {
+      "epoch": 1.925615505500262,
+      "grad_norm": 2.384889602661133,
+      "learning_rate": 1.2377105620942168e-05,
+      "loss": 0.1015,
+      "step": 3676
+    },
+    {
+      "epoch": 1.925615505500262,
+      "loss_partial": 0.052490234375,
+      "step": 3676
+    },
+    {
+      "epoch": 1.925615505500262,
+      "loss_partial": 0.05810546875,
+      "step": 3676
+    },
+    {
+      "epoch": 1.9261393399685698,
+      "grad_norm": 1.991015076637268,
+      "learning_rate": 1.237266482824832e-05,
+      "loss": 0.1083,
+      "step": 3677
+    },
+    {
+      "epoch": 1.9261393399685698,
+      "loss_partial": 0.054443359375,
+      "step": 3677
+    },
+    {
+      "epoch": 1.9261393399685698,
+      "loss_partial": 0.045654296875,
+      "step": 3677
+    },
+    {
+      "epoch": 1.9266631744368778,
+      "grad_norm": 2.2633056640625,
+      "learning_rate": 1.2368223539682191e-05,
+      "loss": 0.1024,
+      "step": 3678
+    },
+    {
+      "epoch": 1.9266631744368778,
+      "loss_partial": 0.04931640625,
+      "step": 3678
+    },
+    {
+      "epoch": 1.9266631744368778,
+      "loss_partial": 0.0673828125,
+      "step": 3678
+    },
+    {
+      "epoch": 1.927187008905186,
+      "grad_norm": 2.1303560733795166,
+      "learning_rate": 1.2363781756171989e-05,
+      "loss": 0.1053,
+      "step": 3679
+    },
+    {
+      "epoch": 1.927187008905186,
+      "loss_partial": 0.04736328125,
+      "step": 3679
+    },
+    {
+      "epoch": 1.927187008905186,
+      "loss_partial": 0.04345703125,
+      "step": 3679
+    },
+    {
+      "epoch": 1.927710843373494,
+      "grad_norm": 1.9377487897872925,
+      "learning_rate": 1.2359339478646015e-05,
+      "loss": 0.1063,
+      "step": 3680
+    },
+    {
+      "epoch": 1.927710843373494,
+      "loss_partial": 0.048828125,
+      "step": 3680
+    },
+    {
+      "epoch": 1.927710843373494,
+      "loss_partial": 0.048095703125,
+      "step": 3680
+    },
+    {
+      "epoch": 1.928234677841802,
+      "grad_norm": 2.3132576942443848,
+      "learning_rate": 1.235489670803268e-05,
+      "loss": 0.1011,
+      "step": 3681
+    },
+    {
+      "epoch": 1.928234677841802,
+      "loss_partial": 0.040771484375,
+      "step": 3681
+    },
+    {
+      "epoch": 1.928234677841802,
+      "loss_partial": 0.0498046875,
+      "step": 3681
+    },
+    {
+      "epoch": 1.92875851231011,
+      "grad_norm": 1.8866811990737915,
+      "learning_rate": 1.2350453445260497e-05,
+      "loss": 0.1053,
+      "step": 3682
+    },
+    {
+      "epoch": 1.92875851231011,
+      "loss_partial": 0.04736328125,
+      "step": 3682
+    },
+    {
+      "epoch": 1.92875851231011,
+      "loss_partial": 0.046630859375,
+      "step": 3682
+    },
+    {
+      "epoch": 1.9292823467784181,
+      "grad_norm": 1.75897216796875,
+      "learning_rate": 1.2346009691258079e-05,
+      "loss": 0.0999,
+      "step": 3683
+    },
+    {
+      "epoch": 1.9292823467784181,
+      "loss_partial": 0.0537109375,
+      "step": 3683
+    },
+    {
+      "epoch": 1.9292823467784181,
+      "loss_partial": 0.048583984375,
+      "step": 3683
+    },
+    {
+      "epoch": 1.929806181246726,
+      "grad_norm": 1.7411210536956787,
+      "learning_rate": 1.2341565446954143e-05,
+      "loss": 0.1002,
+      "step": 3684
+    },
+    {
+      "epoch": 1.929806181246726,
+      "loss_partial": 0.045654296875,
+      "step": 3684
+    },
+    {
+      "epoch": 1.929806181246726,
+      "loss_partial": 0.055908203125,
+      "step": 3684
+    },
+    {
+      "epoch": 1.930330015715034,
+      "grad_norm": 1.3898450136184692,
+      "learning_rate": 1.2337120713277504e-05,
+      "loss": 0.1013,
+      "step": 3685
+    },
+    {
+      "epoch": 1.930330015715034,
+      "loss_partial": 0.053955078125,
+      "step": 3685
+    },
+    {
+      "epoch": 1.930330015715034,
+      "loss_partial": 0.045166015625,
+      "step": 3685
+    },
+    {
+      "epoch": 1.930853850183342,
+      "grad_norm": 1.383149266242981,
+      "learning_rate": 1.2332675491157094e-05,
+      "loss": 0.1045,
+      "step": 3686
+    },
+    {
+      "epoch": 1.930853850183342,
+      "loss_partial": 0.0517578125,
+      "step": 3686
+    },
+    {
+      "epoch": 1.930853850183342,
+      "loss_partial": 0.04833984375,
+      "step": 3686
+    },
+    {
+      "epoch": 1.93137768465165,
+      "grad_norm": 2.3210253715515137,
+      "learning_rate": 1.2328229781521928e-05,
+      "loss": 0.1024,
+      "step": 3687
+    },
+    {
+      "epoch": 1.93137768465165,
+      "loss_partial": 0.053955078125,
+      "step": 3687
+    },
+    {
+      "epoch": 1.93137768465165,
+      "loss_partial": 0.05419921875,
+      "step": 3687
+    },
+    {
+      "epoch": 1.931901519119958,
+      "grad_norm": 1.7333241701126099,
+      "learning_rate": 1.2323783585301134e-05,
+      "loss": 0.1028,
+      "step": 3688
+    },
+    {
+      "epoch": 1.931901519119958,
+      "loss_partial": 0.05908203125,
+      "step": 3688
+    },
+    {
+      "epoch": 1.931901519119958,
+      "loss_partial": 0.0458984375,
+      "step": 3688
+    },
+    {
+      "epoch": 1.932425353588266,
+      "grad_norm": 2.024387836456299,
+      "learning_rate": 1.2319336903423941e-05,
+      "loss": 0.1063,
+      "step": 3689
+    },
+    {
+      "epoch": 1.932425353588266,
+      "loss_partial": 0.052734375,
+      "step": 3689
+    },
+    {
+      "epoch": 1.932425353588266,
+      "loss_partial": 0.0537109375,
+      "step": 3689
+    },
+    {
+      "epoch": 1.9329491880565741,
+      "grad_norm": 1.9980957508087158,
+      "learning_rate": 1.2314889736819681e-05,
+      "loss": 0.1059,
+      "step": 3690
+    },
+    {
+      "epoch": 1.9329491880565741,
+      "loss_partial": 0.0458984375,
+      "step": 3690
+    },
+    {
+      "epoch": 1.9329491880565741,
+      "loss_partial": 0.055908203125,
+      "step": 3690
+    },
+    {
+      "epoch": 1.9334730225248822,
+      "grad_norm": 1.6699573993682861,
+      "learning_rate": 1.2310442086417779e-05,
+      "loss": 0.1021,
+      "step": 3691
+    },
+    {
+      "epoch": 1.9334730225248822,
+      "loss_partial": 0.0537109375,
+      "step": 3691
+    },
+    {
+      "epoch": 1.9334730225248822,
+      "loss_partial": 0.045166015625,
+      "step": 3691
+    },
+    {
+      "epoch": 1.9339968569931902,
+      "grad_norm": 2.0117745399475098,
+      "learning_rate": 1.2305993953147766e-05,
+      "loss": 0.1064,
+      "step": 3692
+    },
+    {
+      "epoch": 1.9339968569931902,
+      "loss_partial": 0.044189453125,
+      "step": 3692
+    },
+    {
+      "epoch": 1.9339968569931902,
+      "loss_partial": 0.049560546875,
+      "step": 3692
+    },
+    {
+      "epoch": 1.9345206914614983,
+      "grad_norm": 1.4471633434295654,
+      "learning_rate": 1.230154533793928e-05,
+      "loss": 0.1068,
+      "step": 3693
+    },
+    {
+      "epoch": 1.9345206914614983,
+      "loss_partial": 0.06298828125,
+      "step": 3693
+    },
+    {
+      "epoch": 1.9345206914614983,
+      "loss_partial": 0.044677734375,
+      "step": 3693
+    },
+    {
+      "epoch": 1.9350445259298061,
+      "grad_norm": 2.2250165939331055,
+      "learning_rate": 1.2297096241722046e-05,
+      "loss": 0.0974,
+      "step": 3694
+    },
+    {
+      "epoch": 1.9350445259298061,
+      "loss_partial": 0.04736328125,
+      "step": 3694
+    },
+    {
+      "epoch": 1.9350445259298061,
+      "loss_partial": 0.0625,
+      "step": 3694
+    },
+    {
+      "epoch": 1.9355683603981142,
+      "grad_norm": 1.853442907333374,
+      "learning_rate": 1.2292646665425908e-05,
+      "loss": 0.1011,
+      "step": 3695
+    },
+    {
+      "epoch": 1.9355683603981142,
+      "loss_partial": 0.050537109375,
+      "step": 3695
+    },
+    {
+      "epoch": 1.9355683603981142,
+      "loss_partial": 0.046142578125,
+      "step": 3695
+    },
+    {
+      "epoch": 1.9360921948664223,
+      "grad_norm": 1.6658283472061157,
+      "learning_rate": 1.2288196609980792e-05,
+      "loss": 0.1029,
+      "step": 3696
+    },
+    {
+      "epoch": 1.9360921948664223,
+      "loss_partial": 0.043212890625,
+      "step": 3696
+    },
+    {
+      "epoch": 1.9360921948664223,
+      "loss_partial": 0.052001953125,
+      "step": 3696
+    },
+    {
+      "epoch": 1.93661602933473,
+      "grad_norm": 1.3556063175201416,
+      "learning_rate": 1.2283746076316733e-05,
+      "loss": 0.1054,
+      "step": 3697
+    },
+    {
+      "epoch": 1.93661602933473,
+      "loss_partial": 0.047119140625,
+      "step": 3697
+    },
+    {
+      "epoch": 1.93661602933473,
+      "loss_partial": 0.046142578125,
+      "step": 3697
+    },
+    {
+      "epoch": 1.9371398638030382,
+      "grad_norm": 1.5436655282974243,
+      "learning_rate": 1.2279295065363865e-05,
+      "loss": 0.1029,
+      "step": 3698
+    },
+    {
+      "epoch": 1.9371398638030382,
+      "loss_partial": 0.050048828125,
+      "step": 3698
+    },
+    {
+      "epoch": 1.9371398638030382,
+      "loss_partial": 0.059326171875,
+      "step": 3698
+    },
+    {
+      "epoch": 1.9376636982713462,
+      "grad_norm": 1.4554004669189453,
+      "learning_rate": 1.2274843578052426e-05,
+      "loss": 0.1038,
+      "step": 3699
+    },
+    {
+      "epoch": 1.9376636982713462,
+      "loss_partial": 0.0478515625,
+      "step": 3699
+    },
+    {
+      "epoch": 1.9376636982713462,
+      "loss_partial": 0.043212890625,
+      "step": 3699
+    },
+    {
+      "epoch": 1.9381875327396543,
+      "grad_norm": 1.3907432556152344,
+      "learning_rate": 1.2270391615312747e-05,
+      "loss": 0.1073,
+      "step": 3700
+    },
+    {
+      "epoch": 1.9381875327396543,
+      "loss_partial": 0.056640625,
+      "step": 3700
+    },
+    {
+      "epoch": 1.9381875327396543,
+      "loss_partial": 0.054931640625,
+      "step": 3700
+    },
+    {
+      "epoch": 1.9387113672079623,
+      "grad_norm": 1.6439814567565918,
+      "learning_rate": 1.2265939178075259e-05,
+      "loss": 0.1039,
+      "step": 3701
+    },
+    {
+      "epoch": 1.9387113672079623,
+      "loss_partial": 0.046142578125,
+      "step": 3701
+    },
+    {
+      "epoch": 1.9387113672079623,
+      "loss_partial": 0.050048828125,
+      "step": 3701
+    },
+    {
+      "epoch": 1.9392352016762704,
+      "grad_norm": 1.4140933752059937,
+      "learning_rate": 1.2261486267270495e-05,
+      "loss": 0.101,
+      "step": 3702
+    },
+    {
+      "epoch": 1.9392352016762704,
+      "loss_partial": 0.05078125,
+      "step": 3702
+    },
+    {
+      "epoch": 1.9392352016762704,
+      "loss_partial": 0.044921875,
+      "step": 3702
+    },
+    {
+      "epoch": 1.9397590361445785,
+      "grad_norm": 1.4980560541152954,
+      "learning_rate": 1.2257032883829086e-05,
+      "loss": 0.0996,
+      "step": 3703
+    },
+    {
+      "epoch": 1.9397590361445785,
+      "loss_partial": 0.05029296875,
+      "step": 3703
+    },
+    {
+      "epoch": 1.9397590361445785,
+      "loss_partial": 0.044921875,
+      "step": 3703
+    },
+    {
+      "epoch": 1.9402828706128863,
+      "grad_norm": 1.5109726190567017,
+      "learning_rate": 1.225257902868176e-05,
+      "loss": 0.1039,
+      "step": 3704
+    },
+    {
+      "epoch": 1.9402828706128863,
+      "loss_partial": 0.0498046875,
+      "step": 3704
+    },
+    {
+      "epoch": 1.9402828706128863,
+      "loss_partial": 0.04931640625,
+      "step": 3704
+    },
+    {
+      "epoch": 1.9408067050811943,
+      "grad_norm": 1.4719531536102295,
+      "learning_rate": 1.2248124702759348e-05,
+      "loss": 0.1069,
+      "step": 3705
+    },
+    {
+      "epoch": 1.9408067050811943,
+      "loss_partial": 0.056640625,
+      "step": 3705
+    },
+    {
+      "epoch": 1.9408067050811943,
+      "loss_partial": 0.0478515625,
+      "step": 3705
+    },
+    {
+      "epoch": 1.9413305395495024,
+      "grad_norm": 1.8450257778167725,
+      "learning_rate": 1.2243669906992773e-05,
+      "loss": 0.1027,
+      "step": 3706
+    },
+    {
+      "epoch": 1.9413305395495024,
+      "loss_partial": 0.061767578125,
+      "step": 3706
+    },
+    {
+      "epoch": 1.9413305395495024,
+      "loss_partial": 0.052978515625,
+      "step": 3706
+    },
+    {
+      "epoch": 1.9418543740178102,
+      "grad_norm": 1.6283793449401855,
+      "learning_rate": 1.2239214642313063e-05,
+      "loss": 0.106,
+      "step": 3707
+    },
+    {
+      "epoch": 1.9418543740178102,
+      "loss_partial": 0.043701171875,
+      "step": 3707
+    },
+    {
+      "epoch": 1.9418543740178102,
+      "loss_partial": 0.057373046875,
+      "step": 3707
+    },
+    {
+      "epoch": 1.9423782084861183,
+      "grad_norm": 1.5354809761047363,
+      "learning_rate": 1.223475890965134e-05,
+      "loss": 0.1035,
+      "step": 3708
+    },
+    {
+      "epoch": 1.9423782084861183,
+      "loss_partial": 0.0478515625,
+      "step": 3708
+    },
+    {
+      "epoch": 1.9423782084861183,
+      "loss_partial": 0.046142578125,
+      "step": 3708
+    },
+    {
+      "epoch": 1.9429020429544264,
+      "grad_norm": 1.6292821168899536,
+      "learning_rate": 1.2230302709938826e-05,
+      "loss": 0.1023,
+      "step": 3709
+    },
+    {
+      "epoch": 1.9429020429544264,
+      "loss_partial": 0.0615234375,
+      "step": 3709
+    },
+    {
+      "epoch": 1.9429020429544264,
+      "loss_partial": 0.06884765625,
+      "step": 3709
+    },
+    {
+      "epoch": 1.9434258774227344,
+      "grad_norm": 2.0634522438049316,
+      "learning_rate": 1.2225846044106833e-05,
+      "loss": 0.1041,
+      "step": 3710
+    },
+    {
+      "epoch": 1.9434258774227344,
+      "loss_partial": 0.0517578125,
+      "step": 3710
+    },
+    {
+      "epoch": 1.9434258774227344,
+      "loss_partial": 0.05029296875,
+      "step": 3710
+    },
+    {
+      "epoch": 1.9439497118910425,
+      "grad_norm": 1.4637259244918823,
+      "learning_rate": 1.2221388913086784e-05,
+      "loss": 0.101,
+      "step": 3711
+    },
+    {
+      "epoch": 1.9439497118910425,
+      "loss_partial": 0.0576171875,
+      "step": 3711
+    },
+    {
+      "epoch": 1.9439497118910425,
+      "loss_partial": 0.05224609375,
+      "step": 3711
+    },
+    {
+      "epoch": 1.9444735463593505,
+      "grad_norm": 1.6353387832641602,
+      "learning_rate": 1.2216931317810189e-05,
+      "loss": 0.1053,
+      "step": 3712
+    },
+    {
+      "epoch": 1.9444735463593505,
+      "loss_partial": 0.057373046875,
+      "step": 3712
+    },
+    {
+      "epoch": 1.9444735463593505,
+      "loss_partial": 0.055419921875,
+      "step": 3712
+    },
+    {
+      "epoch": 1.9449973808276586,
+      "grad_norm": 2.3473870754241943,
+      "learning_rate": 1.2212473259208655e-05,
+      "loss": 0.1047,
+      "step": 3713
+    },
+    {
+      "epoch": 1.9449973808276586,
+      "loss_partial": 0.0498046875,
+      "step": 3713
+    },
+    {
+      "epoch": 1.9449973808276586,
+      "loss_partial": 0.0517578125,
+      "step": 3713
+    },
+    {
+      "epoch": 1.9455212152959664,
+      "grad_norm": 1.8288456201553345,
+      "learning_rate": 1.2208014738213895e-05,
+      "loss": 0.1039,
+      "step": 3714
+    },
+    {
+      "epoch": 1.9455212152959664,
+      "loss_partial": 0.046630859375,
+      "step": 3714
+    },
+    {
+      "epoch": 1.9455212152959664,
+      "loss_partial": 0.043701171875,
+      "step": 3714
+    },
+    {
+      "epoch": 1.9460450497642745,
+      "grad_norm": 1.5758596658706665,
+      "learning_rate": 1.2203555755757708e-05,
+      "loss": 0.1082,
+      "step": 3715
+    },
+    {
+      "epoch": 1.9460450497642745,
+      "loss_partial": 0.054931640625,
+      "step": 3715
+    },
+    {
+      "epoch": 1.9460450497642745,
+      "loss_partial": 0.05029296875,
+      "step": 3715
+    },
+    {
+      "epoch": 1.9465688842325823,
+      "grad_norm": 1.287334680557251,
+      "learning_rate": 1.2199096312771992e-05,
+      "loss": 0.103,
+      "step": 3716
+    },
+    {
+      "epoch": 1.9465688842325823,
+      "loss_partial": 0.046875,
+      "step": 3716
+    },
+    {
+      "epoch": 1.9465688842325823,
+      "loss_partial": 0.053955078125,
+      "step": 3716
+    },
+    {
+      "epoch": 1.9470927187008904,
+      "grad_norm": 1.65269935131073,
+      "learning_rate": 1.2194636410188748e-05,
+      "loss": 0.1031,
+      "step": 3717
+    },
+    {
+      "epoch": 1.9470927187008904,
+      "loss_partial": 0.04931640625,
+      "step": 3717
+    },
+    {
+      "epoch": 1.9470927187008904,
+      "loss_partial": 0.04736328125,
+      "step": 3717
+    },
+    {
+      "epoch": 1.9476165531691985,
+      "grad_norm": 1.7743381261825562,
+      "learning_rate": 1.2190176048940061e-05,
+      "loss": 0.1032,
+      "step": 3718
+    },
+    {
+      "epoch": 1.9476165531691985,
+      "loss_partial": 0.050537109375,
+      "step": 3718
+    },
+    {
+      "epoch": 1.9476165531691985,
+      "loss_partial": 0.054443359375,
+      "step": 3718
+    },
+    {
+      "epoch": 1.9481403876375065,
+      "grad_norm": 1.3485459089279175,
+      "learning_rate": 1.2185715229958129e-05,
+      "loss": 0.0984,
+      "step": 3719
+    },
+    {
+      "epoch": 1.9481403876375065,
+      "loss_partial": 0.04736328125,
+      "step": 3719
+    },
+    {
+      "epoch": 1.9481403876375065,
+      "loss_partial": 0.043701171875,
+      "step": 3719
+    },
+    {
+      "epoch": 1.9486642221058146,
+      "grad_norm": 1.4474579095840454,
+      "learning_rate": 1.2181253954175227e-05,
+      "loss": 0.1038,
+      "step": 3720
+    },
+    {
+      "epoch": 1.9486642221058146,
+      "loss_partial": 0.057861328125,
+      "step": 3720
+    },
+    {
+      "epoch": 1.9486642221058146,
+      "loss_partial": 0.057373046875,
+      "step": 3720
+    },
+    {
+      "epoch": 1.9491880565741226,
+      "grad_norm": 1.7803213596343994,
+      "learning_rate": 1.2176792222523739e-05,
+      "loss": 0.1035,
+      "step": 3721
+    },
+    {
+      "epoch": 1.9491880565741226,
+      "loss_partial": 0.053955078125,
+      "step": 3721
+    },
+    {
+      "epoch": 1.9491880565741226,
+      "loss_partial": 0.0498046875,
+      "step": 3721
+    },
+    {
+      "epoch": 1.9497118910424307,
+      "grad_norm": 2.0068564414978027,
+      "learning_rate": 1.2172330035936136e-05,
+      "loss": 0.1033,
+      "step": 3722
+    },
+    {
+      "epoch": 1.9497118910424307,
+      "loss_partial": 0.060791015625,
+      "step": 3722
+    },
+    {
+      "epoch": 1.9497118910424307,
+      "loss_partial": 0.049072265625,
+      "step": 3722
+    },
+    {
+      "epoch": 1.9502357255107388,
+      "grad_norm": 1.2930114269256592,
+      "learning_rate": 1.216786739534499e-05,
+      "loss": 0.1032,
+      "step": 3723
+    },
+    {
+      "epoch": 1.9502357255107388,
+      "loss_partial": 0.04736328125,
+      "step": 3723
+    },
+    {
+      "epoch": 1.9502357255107388,
+      "loss_partial": 0.052490234375,
+      "step": 3723
+    },
+    {
+      "epoch": 1.9507595599790466,
+      "grad_norm": 2.6348040103912354,
+      "learning_rate": 1.216340430168296e-05,
+      "loss": 0.1021,
+      "step": 3724
+    },
+    {
+      "epoch": 1.9507595599790466,
+      "loss_partial": 0.044677734375,
+      "step": 3724
+    },
+    {
+      "epoch": 1.9507595599790466,
+      "loss_partial": 0.051025390625,
+      "step": 3724
+    },
+    {
+      "epoch": 1.9512833944473547,
+      "grad_norm": 1.557606816291809,
+      "learning_rate": 1.2158940755882813e-05,
+      "loss": 0.0982,
+      "step": 3725
+    },
+    {
+      "epoch": 1.9512833944473547,
+      "loss_partial": 0.053955078125,
+      "step": 3725
+    },
+    {
+      "epoch": 1.9512833944473547,
+      "loss_partial": 0.05810546875,
+      "step": 3725
+    },
+    {
+      "epoch": 1.9518072289156625,
+      "grad_norm": 1.9532470703125,
+      "learning_rate": 1.2154476758877398e-05,
+      "loss": 0.1066,
+      "step": 3726
+    },
+    {
+      "epoch": 1.9518072289156625,
+      "loss_partial": 0.044677734375,
+      "step": 3726
+    },
+    {
+      "epoch": 1.9518072289156625,
+      "loss_partial": 0.04931640625,
+      "step": 3726
+    },
+    {
+      "epoch": 1.9523310633839706,
+      "grad_norm": 2.0455567836761475,
+      "learning_rate": 1.2150012311599664e-05,
+      "loss": 0.1057,
+      "step": 3727
+    },
+    {
+      "epoch": 1.9523310633839706,
+      "loss_partial": 0.051513671875,
+      "step": 3727
+    },
+    {
+      "epoch": 1.9523310633839706,
+      "loss_partial": 0.046142578125,
+      "step": 3727
+    },
+    {
+      "epoch": 1.9528548978522786,
+      "grad_norm": 1.9580941200256348,
+      "learning_rate": 1.214554741498265e-05,
+      "loss": 0.1001,
+      "step": 3728
+    },
+    {
+      "epoch": 1.9528548978522786,
+      "loss_partial": 0.06640625,
+      "step": 3728
+    },
+    {
+      "epoch": 1.9528548978522786,
+      "loss_partial": 0.055908203125,
+      "step": 3728
+    },
+    {
+      "epoch": 1.9533787323205867,
+      "grad_norm": 2.064277410507202,
+      "learning_rate": 1.2141082069959496e-05,
+      "loss": 0.1098,
+      "step": 3729
+    },
+    {
+      "epoch": 1.9533787323205867,
+      "loss_partial": 0.05224609375,
+      "step": 3729
+    },
+    {
+      "epoch": 1.9533787323205867,
+      "loss_partial": 0.046630859375,
+      "step": 3729
+    },
+    {
+      "epoch": 1.9539025667888947,
+      "grad_norm": 1.5467712879180908,
+      "learning_rate": 1.2136616277463427e-05,
+      "loss": 0.1011,
+      "step": 3730
+    },
+    {
+      "epoch": 1.9539025667888947,
+      "loss_partial": 0.047607421875,
+      "step": 3730
+    },
+    {
+      "epoch": 1.9539025667888947,
+      "loss_partial": 0.056396484375,
+      "step": 3730
+    },
+    {
+      "epoch": 1.9544264012572028,
+      "grad_norm": 1.6670771837234497,
+      "learning_rate": 1.2132150038427771e-05,
+      "loss": 0.1041,
+      "step": 3731
+    },
+    {
+      "epoch": 1.9544264012572028,
+      "loss_partial": 0.048828125,
+      "step": 3731
+    },
+    {
+      "epoch": 1.9544264012572028,
+      "loss_partial": 0.05078125,
+      "step": 3731
+    },
+    {
+      "epoch": 1.9549502357255109,
+      "grad_norm": 1.5367686748504639,
+      "learning_rate": 1.2127683353785939e-05,
+      "loss": 0.106,
+      "step": 3732
+    },
+    {
+      "epoch": 1.9549502357255109,
+      "loss_partial": 0.05322265625,
+      "step": 3732
+    },
+    {
+      "epoch": 1.9549502357255109,
+      "loss_partial": 0.052734375,
+      "step": 3732
+    },
+    {
+      "epoch": 1.955474070193819,
+      "grad_norm": 1.6687893867492676,
+      "learning_rate": 1.2123216224471447e-05,
+      "loss": 0.1057,
+      "step": 3733
+    },
+    {
+      "epoch": 1.955474070193819,
+      "loss_partial": 0.0556640625,
+      "step": 3733
+    },
+    {
+      "epoch": 1.955474070193819,
+      "loss_partial": 0.045166015625,
+      "step": 3733
+    },
+    {
+      "epoch": 1.9559979046621268,
+      "grad_norm": 1.6396311521530151,
+      "learning_rate": 1.2118748651417889e-05,
+      "loss": 0.1024,
+      "step": 3734
+    },
+    {
+      "epoch": 1.9559979046621268,
+      "loss_partial": 0.054931640625,
+      "step": 3734
+    },
+    {
+      "epoch": 1.9559979046621268,
+      "loss_partial": 0.046142578125,
+      "step": 3734
+    },
+    {
+      "epoch": 1.9565217391304348,
+      "grad_norm": 1.8526471853256226,
+      "learning_rate": 1.2114280635558968e-05,
+      "loss": 0.1018,
+      "step": 3735
+    },
+    {
+      "epoch": 1.9565217391304348,
+      "loss_partial": 0.049560546875,
+      "step": 3735
+    },
+    {
+      "epoch": 1.9565217391304348,
+      "loss_partial": 0.0517578125,
+      "step": 3735
+    },
+    {
+      "epoch": 1.9570455735987426,
+      "grad_norm": 1.7301369905471802,
+      "learning_rate": 1.2109812177828469e-05,
+      "loss": 0.1007,
+      "step": 3736
+    },
+    {
+      "epoch": 1.9570455735987426,
+      "loss_partial": 0.06396484375,
+      "step": 3736
+    },
+    {
+      "epoch": 1.9570455735987426,
+      "loss_partial": 0.0498046875,
+      "step": 3736
+    },
+    {
+      "epoch": 1.9575694080670507,
+      "grad_norm": 1.8772118091583252,
+      "learning_rate": 1.2105343279160269e-05,
+      "loss": 0.1053,
+      "step": 3737
+    },
+    {
+      "epoch": 1.9575694080670507,
+      "loss_partial": 0.0498046875,
+      "step": 3737
+    },
+    {
+      "epoch": 1.9575694080670507,
+      "loss_partial": 0.055908203125,
+      "step": 3737
+    },
+    {
+      "epoch": 1.9580932425353588,
+      "grad_norm": 2.0641493797302246,
+      "learning_rate": 1.2100873940488345e-05,
+      "loss": 0.1051,
+      "step": 3738
+    },
+    {
+      "epoch": 1.9580932425353588,
+      "loss_partial": 0.04248046875,
+      "step": 3738
+    },
+    {
+      "epoch": 1.9580932425353588,
+      "loss_partial": 0.053955078125,
+      "step": 3738
+    },
+    {
+      "epoch": 1.9586170770036668,
+      "grad_norm": 1.964463233947754,
+      "learning_rate": 1.2096404162746758e-05,
+      "loss": 0.0999,
+      "step": 3739
+    },
+    {
+      "epoch": 1.9586170770036668,
+      "loss_partial": 0.055419921875,
+      "step": 3739
+    },
+    {
+      "epoch": 1.9586170770036668,
+      "loss_partial": 0.052001953125,
+      "step": 3739
+    },
+    {
+      "epoch": 1.9591409114719749,
+      "grad_norm": 2.141430139541626,
+      "learning_rate": 1.2091933946869663e-05,
+      "loss": 0.1078,
+      "step": 3740
+    },
+    {
+      "epoch": 1.9591409114719749,
+      "loss_partial": 0.0478515625,
+      "step": 3740
+    },
+    {
+      "epoch": 1.9591409114719749,
+      "loss_partial": 0.055908203125,
+      "step": 3740
+    },
+    {
+      "epoch": 1.959664745940283,
+      "grad_norm": 1.6259315013885498,
+      "learning_rate": 1.2087463293791312e-05,
+      "loss": 0.1041,
+      "step": 3741
+    },
+    {
+      "epoch": 1.959664745940283,
+      "loss_partial": 0.050537109375,
+      "step": 3741
+    },
+    {
+      "epoch": 1.959664745940283,
+      "loss_partial": 0.0478515625,
+      "step": 3741
+    },
+    {
+      "epoch": 1.960188580408591,
+      "grad_norm": 1.7625585794448853,
+      "learning_rate": 1.2082992204446038e-05,
+      "loss": 0.1014,
+      "step": 3742
+    },
+    {
+      "epoch": 1.960188580408591,
+      "loss_partial": 0.05810546875,
+      "step": 3742
+    },
+    {
+      "epoch": 1.960188580408591,
+      "loss_partial": 0.055908203125,
+      "step": 3742
+    },
+    {
+      "epoch": 1.9607124148768988,
+      "grad_norm": 1.1755284070968628,
+      "learning_rate": 1.2078520679768278e-05,
+      "loss": 0.1019,
+      "step": 3743
+    },
+    {
+      "epoch": 1.9607124148768988,
+      "loss_partial": 0.04443359375,
+      "step": 3743
+    },
+    {
+      "epoch": 1.9607124148768988,
+      "loss_partial": 0.050537109375,
+      "step": 3743
+    },
+    {
+      "epoch": 1.961236249345207,
+      "grad_norm": 1.988808512687683,
+      "learning_rate": 1.2074048720692548e-05,
+      "loss": 0.1038,
+      "step": 3744
+    },
+    {
+      "epoch": 1.961236249345207,
+      "loss_partial": 0.05517578125,
+      "step": 3744
+    },
+    {
+      "epoch": 1.961236249345207,
+      "loss_partial": 0.050537109375,
+      "step": 3744
+    },
+    {
+      "epoch": 1.961760083813515,
+      "grad_norm": 1.3767372369766235,
+      "learning_rate": 1.206957632815346e-05,
+      "loss": 0.1013,
+      "step": 3745
+    },
+    {
+      "epoch": 1.961760083813515,
+      "loss_partial": 0.0517578125,
+      "step": 3745
+    },
+    {
+      "epoch": 1.961760083813515,
+      "loss_partial": 0.051025390625,
+      "step": 3745
+    },
+    {
+      "epoch": 1.9622839182818228,
+      "grad_norm": 1.6451773643493652,
+      "learning_rate": 1.2065103503085718e-05,
+      "loss": 0.106,
+      "step": 3746
+    },
+    {
+      "epoch": 1.9622839182818228,
+      "loss_partial": 0.052734375,
+      "step": 3746
+    },
+    {
+      "epoch": 1.9622839182818228,
+      "loss_partial": 0.04345703125,
+      "step": 3746
+    },
+    {
+      "epoch": 1.9628077527501309,
+      "grad_norm": 1.2521859407424927,
+      "learning_rate": 1.2060630246424112e-05,
+      "loss": 0.1023,
+      "step": 3747
+    },
+    {
+      "epoch": 1.9628077527501309,
+      "loss_partial": 0.056396484375,
+      "step": 3747
+    },
+    {
+      "epoch": 1.9628077527501309,
+      "loss_partial": 0.043212890625,
+      "step": 3747
+    },
+    {
+      "epoch": 1.963331587218439,
+      "grad_norm": 1.767633080482483,
+      "learning_rate": 1.2056156559103529e-05,
+      "loss": 0.1003,
+      "step": 3748
+    },
+    {
+      "epoch": 1.963331587218439,
+      "loss_partial": 0.048095703125,
+      "step": 3748
+    },
+    {
+      "epoch": 1.963331587218439,
+      "loss_partial": 0.0458984375,
+      "step": 3748
+    },
+    {
+      "epoch": 1.963855421686747,
+      "grad_norm": 1.978418231010437,
+      "learning_rate": 1.2051682442058939e-05,
+      "loss": 0.1033,
+      "step": 3749
+    },
+    {
+      "epoch": 1.963855421686747,
+      "loss_partial": 0.044921875,
+      "step": 3749
+    },
+    {
+      "epoch": 1.963855421686747,
+      "loss_partial": 0.047119140625,
+      "step": 3749
+    },
+    {
+      "epoch": 1.964379256155055,
+      "grad_norm": 1.3758379220962524,
+      "learning_rate": 1.2047207896225408e-05,
+      "loss": 0.1017,
+      "step": 3750
+    },
+    {
+      "epoch": 1.964379256155055,
+      "loss_partial": 0.05029296875,
+      "step": 3750
+    },
+    {
+      "epoch": 1.964379256155055,
+      "loss_partial": 0.053466796875,
+      "step": 3750
+    },
+    {
+      "epoch": 1.964903090623363,
+      "grad_norm": 1.8632196187973022,
+      "learning_rate": 1.2042732922538086e-05,
+      "loss": 0.1022,
+      "step": 3751
+    },
+    {
+      "epoch": 1.964903090623363,
+      "loss_partial": 0.0625,
+      "step": 3751
+    },
+    {
+      "epoch": 1.964903090623363,
+      "loss_partial": 0.04541015625,
+      "step": 3751
+    },
+    {
+      "epoch": 1.9654269250916712,
+      "grad_norm": 1.5547317266464233,
+      "learning_rate": 1.2038257521932212e-05,
+      "loss": 0.1067,
+      "step": 3752
+    },
+    {
+      "epoch": 1.9654269250916712,
+      "loss_partial": 0.054931640625,
+      "step": 3752
+    },
+    {
+      "epoch": 1.9654269250916712,
+      "loss_partial": 0.05322265625,
+      "step": 3752
+    },
+    {
+      "epoch": 1.965950759559979,
+      "grad_norm": 2.0090112686157227,
+      "learning_rate": 1.203378169534312e-05,
+      "loss": 0.1067,
+      "step": 3753
+    },
+    {
+      "epoch": 1.965950759559979,
+      "loss_partial": 0.04833984375,
+      "step": 3753
+    },
+    {
+      "epoch": 1.965950759559979,
+      "loss_partial": 0.0595703125,
+      "step": 3753
+    },
+    {
+      "epoch": 1.966474594028287,
+      "grad_norm": 1.571070671081543,
+      "learning_rate": 1.2029305443706235e-05,
+      "loss": 0.1023,
+      "step": 3754
+    },
+    {
+      "epoch": 1.966474594028287,
+      "loss_partial": 0.045166015625,
+      "step": 3754
+    },
+    {
+      "epoch": 1.966474594028287,
+      "loss_partial": 0.051513671875,
+      "step": 3754
+    },
+    {
+      "epoch": 1.9669984284965951,
+      "grad_norm": 1.9593676328659058,
+      "learning_rate": 1.2024828767957055e-05,
+      "loss": 0.1002,
+      "step": 3755
+    },
+    {
+      "epoch": 1.9669984284965951,
+      "loss_partial": 0.04638671875,
+      "step": 3755
+    },
+    {
+      "epoch": 1.9669984284965951,
+      "loss_partial": 0.046875,
+      "step": 3755
+    },
+    {
+      "epoch": 1.967522262964903,
+      "grad_norm": 1.6474441289901733,
+      "learning_rate": 1.2020351669031188e-05,
+      "loss": 0.1027,
+      "step": 3756
+    },
+    {
+      "epoch": 1.967522262964903,
+      "loss_partial": 0.056884765625,
+      "step": 3756
+    },
+    {
+      "epoch": 1.967522262964903,
+      "loss_partial": 0.0634765625,
+      "step": 3756
+    },
+    {
+      "epoch": 1.968046097433211,
+      "grad_norm": 1.665024757385254,
+      "learning_rate": 1.2015874147864314e-05,
+      "loss": 0.1045,
+      "step": 3757
+    },
+    {
+      "epoch": 1.968046097433211,
+      "loss_partial": 0.052490234375,
+      "step": 3757
+    },
+    {
+      "epoch": 1.968046097433211,
+      "loss_partial": 0.0478515625,
+      "step": 3757
+    },
+    {
+      "epoch": 1.968569931901519,
+      "grad_norm": 1.7047768831253052,
+      "learning_rate": 1.2011396205392212e-05,
+      "loss": 0.0984,
+      "step": 3758
+    },
+    {
+      "epoch": 1.968569931901519,
+      "loss_partial": 0.06591796875,
+      "step": 3758
+    },
+    {
+      "epoch": 1.968569931901519,
+      "loss_partial": 0.046875,
+      "step": 3758
+    },
+    {
+      "epoch": 1.9690937663698271,
+      "grad_norm": 1.7346855401992798,
+      "learning_rate": 1.2006917842550735e-05,
+      "loss": 0.1033,
+      "step": 3759
+    },
+    {
+      "epoch": 1.9690937663698271,
+      "loss_partial": 0.049072265625,
+      "step": 3759
+    },
+    {
+      "epoch": 1.9690937663698271,
+      "loss_partial": 0.058349609375,
+      "step": 3759
+    },
+    {
+      "epoch": 1.9696176008381352,
+      "grad_norm": 1.4483327865600586,
+      "learning_rate": 1.2002439060275846e-05,
+      "loss": 0.1046,
+      "step": 3760
+    },
+    {
+      "epoch": 1.9696176008381352,
+      "loss_partial": 0.052490234375,
+      "step": 3760
+    },
+    {
+      "epoch": 1.9696176008381352,
+      "loss_partial": 0.054443359375,
+      "step": 3760
+    },
+    {
+      "epoch": 1.9701414353064433,
+      "grad_norm": 1.6137057542800903,
+      "learning_rate": 1.199795985950357e-05,
+      "loss": 0.1065,
+      "step": 3761
+    },
+    {
+      "epoch": 1.9701414353064433,
+      "loss_partial": 0.0478515625,
+      "step": 3761
+    },
+    {
+      "epoch": 1.9701414353064433,
+      "loss_partial": 0.054443359375,
+      "step": 3761
+    },
+    {
+      "epoch": 1.9706652697747513,
+      "grad_norm": 1.4274473190307617,
+      "learning_rate": 1.1993480241170041e-05,
+      "loss": 0.1019,
+      "step": 3762
+    },
+    {
+      "epoch": 1.9706652697747513,
+      "loss_partial": 0.053466796875,
+      "step": 3762
+    },
+    {
+      "epoch": 1.9706652697747513,
+      "loss_partial": 0.048583984375,
+      "step": 3762
+    },
+    {
+      "epoch": 1.9711891042430592,
+      "grad_norm": 1.5316693782806396,
+      "learning_rate": 1.1989000206211472e-05,
+      "loss": 0.1024,
+      "step": 3763
+    },
+    {
+      "epoch": 1.9711891042430592,
+      "loss_partial": 0.050048828125,
+      "step": 3763
+    },
+    {
+      "epoch": 1.9711891042430592,
+      "loss_partial": 0.06201171875,
+      "step": 3763
+    },
+    {
+      "epoch": 1.9717129387113672,
+      "grad_norm": 1.314064621925354,
+      "learning_rate": 1.1984519755564157e-05,
+      "loss": 0.1044,
+      "step": 3764
+    },
+    {
+      "epoch": 1.9717129387113672,
+      "loss_partial": 0.0546875,
+      "step": 3764
+    },
+    {
+      "epoch": 1.9717129387113672,
+      "loss_partial": 0.04931640625,
+      "step": 3764
+    },
+    {
+      "epoch": 1.9722367731796753,
+      "grad_norm": 1.2981634140014648,
+      "learning_rate": 1.1980038890164483e-05,
+      "loss": 0.1025,
+      "step": 3765
+    },
+    {
+      "epoch": 1.9722367731796753,
+      "loss_partial": 0.053955078125,
+      "step": 3765
+    },
+    {
+      "epoch": 1.9722367731796753,
+      "loss_partial": 0.049072265625,
+      "step": 3765
+    },
+    {
+      "epoch": 1.972760607647983,
+      "grad_norm": 1.2584840059280396,
+      "learning_rate": 1.1975557610948927e-05,
+      "loss": 0.1037,
+      "step": 3766
+    },
+    {
+      "epoch": 1.972760607647983,
+      "loss_partial": 0.056884765625,
+      "step": 3766
+    },
+    {
+      "epoch": 1.972760607647983,
+      "loss_partial": 0.053955078125,
+      "step": 3766
+    },
+    {
+      "epoch": 1.9732844421162912,
+      "grad_norm": 1.5998486280441284,
+      "learning_rate": 1.1971075918854044e-05,
+      "loss": 0.103,
+      "step": 3767
+    },
+    {
+      "epoch": 1.9732844421162912,
+      "loss_partial": 0.048095703125,
+      "step": 3767
+    },
+    {
+      "epoch": 1.9732844421162912,
+      "loss_partial": 0.0576171875,
+      "step": 3767
+    },
+    {
+      "epoch": 1.9738082765845992,
+      "grad_norm": 1.1518548727035522,
+      "learning_rate": 1.1966593814816487e-05,
+      "loss": 0.1024,
+      "step": 3768
+    },
+    {
+      "epoch": 1.9738082765845992,
+      "loss_partial": 0.05322265625,
+      "step": 3768
+    },
+    {
+      "epoch": 1.9738082765845992,
+      "loss_partial": 0.0400390625,
+      "step": 3768
+    },
+    {
+      "epoch": 1.9743321110529073,
+      "grad_norm": 1.9997354745864868,
+      "learning_rate": 1.1962111299772983e-05,
+      "loss": 0.1046,
+      "step": 3769
+    },
+    {
+      "epoch": 1.9743321110529073,
+      "loss_partial": 0.047607421875,
+      "step": 3769
+    },
+    {
+      "epoch": 1.9743321110529073,
+      "loss_partial": 0.0595703125,
+      "step": 3769
+    },
+    {
+      "epoch": 1.9748559455212154,
+      "grad_norm": 1.7439628839492798,
+      "learning_rate": 1.195762837466035e-05,
+      "loss": 0.1046,
+      "step": 3770
+    },
+    {
+      "epoch": 1.9748559455212154,
+      "loss_partial": 0.0517578125,
+      "step": 3770
+    },
+    {
+      "epoch": 1.9748559455212154,
+      "loss_partial": 0.05322265625,
+      "step": 3770
+    },
+    {
+      "epoch": 1.9753797799895234,
+      "grad_norm": 1.5911695957183838,
+      "learning_rate": 1.1953145040415493e-05,
+      "loss": 0.1058,
+      "step": 3771
+    },
+    {
+      "epoch": 1.9753797799895234,
+      "loss_partial": 0.048828125,
+      "step": 3771
+    },
+    {
+      "epoch": 1.9753797799895234,
+      "loss_partial": 0.0615234375,
+      "step": 3771
+    },
+    {
+      "epoch": 1.9759036144578315,
+      "grad_norm": 1.5510188341140747,
+      "learning_rate": 1.19486612979754e-05,
+      "loss": 0.11,
+      "step": 3772
+    },
+    {
+      "epoch": 1.9759036144578315,
+      "loss_partial": 0.04150390625,
+      "step": 3772
+    },
+    {
+      "epoch": 1.9759036144578315,
+      "loss_partial": 0.050048828125,
+      "step": 3772
+    },
+    {
+      "epoch": 1.9764274489261393,
+      "grad_norm": 1.3866852521896362,
+      "learning_rate": 1.1944177148277143e-05,
+      "loss": 0.1025,
+      "step": 3773
+    },
+    {
+      "epoch": 1.9764274489261393,
+      "loss_partial": 0.05029296875,
+      "step": 3773
+    },
+    {
+      "epoch": 1.9764274489261393,
+      "loss_partial": 0.055419921875,
+      "step": 3773
+    },
+    {
+      "epoch": 1.9769512833944474,
+      "grad_norm": 1.4587721824645996,
+      "learning_rate": 1.1939692592257888e-05,
+      "loss": 0.1066,
+      "step": 3774
+    },
+    {
+      "epoch": 1.9769512833944474,
+      "loss_partial": 0.05859375,
+      "step": 3774
+    },
+    {
+      "epoch": 1.9769512833944474,
+      "loss_partial": 0.044677734375,
+      "step": 3774
+    },
+    {
+      "epoch": 1.9774751178627552,
+      "grad_norm": 1.7010908126831055,
+      "learning_rate": 1.1935207630854875e-05,
+      "loss": 0.1019,
+      "step": 3775
+    },
+    {
+      "epoch": 1.9774751178627552,
+      "loss_partial": 0.047607421875,
+      "step": 3775
+    },
+    {
+      "epoch": 1.9774751178627552,
+      "loss_partial": 0.05126953125,
+      "step": 3775
+    },
+    {
+      "epoch": 1.9779989523310633,
+      "grad_norm": 2.1974141597747803,
+      "learning_rate": 1.1930722265005434e-05,
+      "loss": 0.1057,
+      "step": 3776
+    },
+    {
+      "epoch": 1.9779989523310633,
+      "loss_partial": 0.04736328125,
+      "step": 3776
+    },
+    {
+      "epoch": 1.9779989523310633,
+      "loss_partial": 0.0556640625,
+      "step": 3776
+    },
+    {
+      "epoch": 1.9785227867993713,
+      "grad_norm": 1.3928396701812744,
+      "learning_rate": 1.1926236495646977e-05,
+      "loss": 0.1034,
+      "step": 3777
+    },
+    {
+      "epoch": 1.9785227867993713,
+      "loss_partial": 0.055908203125,
+      "step": 3777
+    },
+    {
+      "epoch": 1.9785227867993713,
+      "loss_partial": 0.051025390625,
+      "step": 3777
+    },
+    {
+      "epoch": 1.9790466212676794,
+      "grad_norm": 2.0205750465393066,
+      "learning_rate": 1.1921750323717007e-05,
+      "loss": 0.1034,
+      "step": 3778
+    },
+    {
+      "epoch": 1.9790466212676794,
+      "loss_partial": 0.046630859375,
+      "step": 3778
+    },
+    {
+      "epoch": 1.9790466212676794,
+      "loss_partial": 0.053466796875,
+      "step": 3778
+    },
+    {
+      "epoch": 1.9795704557359874,
+      "grad_norm": 1.9198421239852905,
+      "learning_rate": 1.19172637501531e-05,
+      "loss": 0.1038,
+      "step": 3779
+    },
+    {
+      "epoch": 1.9795704557359874,
+      "loss_partial": 0.050537109375,
+      "step": 3779
+    },
+    {
+      "epoch": 1.9795704557359874,
+      "loss_partial": 0.044921875,
+      "step": 3779
+    },
+    {
+      "epoch": 1.9800942902042955,
+      "grad_norm": 1.3045916557312012,
+      "learning_rate": 1.1912776775892927e-05,
+      "loss": 0.104,
+      "step": 3780
+    },
+    {
+      "epoch": 1.9800942902042955,
+      "loss_partial": 0.047607421875,
+      "step": 3780
+    },
+    {
+      "epoch": 1.9800942902042955,
+      "loss_partial": 0.0595703125,
+      "step": 3780
+    },
+    {
+      "epoch": 1.9806181246726036,
+      "grad_norm": 1.8424298763275146,
+      "learning_rate": 1.1908289401874237e-05,
+      "loss": 0.1013,
+      "step": 3781
+    },
+    {
+      "epoch": 1.9806181246726036,
+      "loss_partial": 0.052978515625,
+      "step": 3781
+    },
+    {
+      "epoch": 1.9806181246726036,
+      "loss_partial": 0.052734375,
+      "step": 3781
+    },
+    {
+      "epoch": 1.9811419591409116,
+      "grad_norm": 1.4682608842849731,
+      "learning_rate": 1.1903801629034863e-05,
+      "loss": 0.1028,
+      "step": 3782
+    },
+    {
+      "epoch": 1.9811419591409116,
+      "loss_partial": 0.04541015625,
+      "step": 3782
+    },
+    {
+      "epoch": 1.9811419591409116,
+      "loss_partial": 0.046142578125,
+      "step": 3782
+    },
+    {
+      "epoch": 1.9816657936092195,
+      "grad_norm": 2.1507863998413086,
+      "learning_rate": 1.1899313458312718e-05,
+      "loss": 0.1036,
+      "step": 3783
+    },
+    {
+      "epoch": 1.9816657936092195,
+      "loss_partial": 0.049072265625,
+      "step": 3783
+    },
+    {
+      "epoch": 1.9816657936092195,
+      "loss_partial": 0.04052734375,
+      "step": 3783
+    },
+    {
+      "epoch": 1.9821896280775275,
+      "grad_norm": 1.9014586210250854,
+      "learning_rate": 1.1894824890645809e-05,
+      "loss": 0.1017,
+      "step": 3784
+    },
+    {
+      "epoch": 1.9821896280775275,
+      "loss_partial": 0.04638671875,
+      "step": 3784
+    },
+    {
+      "epoch": 1.9821896280775275,
+      "loss_partial": 0.049072265625,
+      "step": 3784
+    },
+    {
+      "epoch": 1.9827134625458354,
+      "grad_norm": 1.3450146913528442,
+      "learning_rate": 1.189033592697221e-05,
+      "loss": 0.1015,
+      "step": 3785
+    },
+    {
+      "epoch": 1.9827134625458354,
+      "loss_partial": 0.04931640625,
+      "step": 3785
+    },
+    {
+      "epoch": 1.9827134625458354,
+      "loss_partial": 0.057861328125,
+      "step": 3785
+    },
+    {
+      "epoch": 1.9832372970141434,
+      "grad_norm": 1.4241876602172852,
+      "learning_rate": 1.1885846568230096e-05,
+      "loss": 0.1038,
+      "step": 3786
+    },
+    {
+      "epoch": 1.9832372970141434,
+      "loss_partial": 0.05322265625,
+      "step": 3786
+    },
+    {
+      "epoch": 1.9832372970141434,
+      "loss_partial": 0.05322265625,
+      "step": 3786
+    },
+    {
+      "epoch": 1.9837611314824515,
+      "grad_norm": 1.3460887670516968,
+      "learning_rate": 1.188135681535771e-05,
+      "loss": 0.1035,
+      "step": 3787
+    },
+    {
+      "epoch": 1.9837611314824515,
+      "loss_partial": 0.0556640625,
+      "step": 3787
+    },
+    {
+      "epoch": 1.9837611314824515,
+      "loss_partial": 0.06103515625,
+      "step": 3787
+    },
+    {
+      "epoch": 1.9842849659507595,
+      "grad_norm": 1.4377820491790771,
+      "learning_rate": 1.1876866669293385e-05,
+      "loss": 0.1085,
+      "step": 3788
+    },
+    {
+      "epoch": 1.9842849659507595,
+      "loss_partial": 0.04443359375,
+      "step": 3788
+    },
+    {
+      "epoch": 1.9842849659507595,
+      "loss_partial": 0.057373046875,
+      "step": 3788
+    },
+    {
+      "epoch": 1.9848088004190676,
+      "grad_norm": 1.469657301902771,
+      "learning_rate": 1.187237613097553e-05,
+      "loss": 0.1033,
+      "step": 3789
+    },
+    {
+      "epoch": 1.9848088004190676,
+      "loss_partial": 0.04638671875,
+      "step": 3789
+    },
+    {
+      "epoch": 1.9848088004190676,
+      "loss_partial": 0.05224609375,
+      "step": 3789
+    },
+    {
+      "epoch": 1.9853326348873757,
+      "grad_norm": 1.658393383026123,
+      "learning_rate": 1.1867885201342641e-05,
+      "loss": 0.1021,
+      "step": 3790
+    },
+    {
+      "epoch": 1.9853326348873757,
+      "loss_partial": 0.05224609375,
+      "step": 3790
+    },
+    {
+      "epoch": 1.9853326348873757,
+      "loss_partial": 0.048583984375,
+      "step": 3790
+    },
+    {
+      "epoch": 1.9858564693556837,
+      "grad_norm": 1.6888519525527954,
+      "learning_rate": 1.1863393881333299e-05,
+      "loss": 0.1064,
+      "step": 3791
+    },
+    {
+      "epoch": 1.9858564693556837,
+      "loss_partial": 0.054931640625,
+      "step": 3791
+    },
+    {
+      "epoch": 1.9858564693556837,
+      "loss_partial": 0.050048828125,
+      "step": 3791
+    },
+    {
+      "epoch": 1.9863803038239918,
+      "grad_norm": 1.6543262004852295,
+      "learning_rate": 1.1858902171886154e-05,
+      "loss": 0.1042,
+      "step": 3792
+    },
+    {
+      "epoch": 1.9863803038239918,
+      "loss_partial": 0.0537109375,
+      "step": 3792
+    },
+    {
+      "epoch": 1.9863803038239918,
+      "loss_partial": 0.054443359375,
+      "step": 3792
+    },
+    {
+      "epoch": 1.9869041382922996,
+      "grad_norm": 1.8576622009277344,
+      "learning_rate": 1.1854410073939958e-05,
+      "loss": 0.1037,
+      "step": 3793
+    },
+    {
+      "epoch": 1.9869041382922996,
+      "loss_partial": 0.04931640625,
+      "step": 3793
+    },
+    {
+      "epoch": 1.9869041382922996,
+      "loss_partial": 0.052978515625,
+      "step": 3793
+    },
+    {
+      "epoch": 1.9874279727606077,
+      "grad_norm": 1.6179695129394531,
+      "learning_rate": 1.1849917588433517e-05,
+      "loss": 0.1044,
+      "step": 3794
+    },
+    {
+      "epoch": 1.9874279727606077,
+      "loss_partial": 0.050048828125,
+      "step": 3794
+    },
+    {
+      "epoch": 1.9874279727606077,
+      "loss_partial": 0.060302734375,
+      "step": 3794
+    },
+    {
+      "epoch": 1.9879518072289155,
+      "grad_norm": 1.423241138458252,
+      "learning_rate": 1.184542471630574e-05,
+      "loss": 0.1031,
+      "step": 3795
+    },
+    {
+      "epoch": 1.9879518072289155,
+      "loss_partial": 0.0576171875,
+      "step": 3795
+    },
+    {
+      "epoch": 1.9879518072289155,
+      "loss_partial": 0.05322265625,
+      "step": 3795
+    },
+    {
+      "epoch": 1.9884756416972236,
+      "grad_norm": 1.6267250776290894,
+      "learning_rate": 1.1840931458495613e-05,
+      "loss": 0.1048,
+      "step": 3796
+    },
+    {
+      "epoch": 1.9884756416972236,
+      "loss_partial": 0.04638671875,
+      "step": 3796
+    },
+    {
+      "epoch": 1.9884756416972236,
+      "loss_partial": 0.050048828125,
+      "step": 3796
+    },
+    {
+      "epoch": 1.9889994761655316,
+      "grad_norm": 1.509578824043274,
+      "learning_rate": 1.183643781594219e-05,
+      "loss": 0.1035,
+      "step": 3797
+    },
+    {
+      "epoch": 1.9889994761655316,
+      "loss_partial": 0.049072265625,
+      "step": 3797
+    },
+    {
+      "epoch": 1.9889994761655316,
+      "loss_partial": 0.04931640625,
+      "step": 3797
+    },
+    {
+      "epoch": 1.9895233106338397,
+      "grad_norm": 1.4602503776550293,
+      "learning_rate": 1.1831943789584623e-05,
+      "loss": 0.1025,
+      "step": 3798
+    },
+    {
+      "epoch": 1.9895233106338397,
+      "loss_partial": 0.059814453125,
+      "step": 3798
+    },
+    {
+      "epoch": 1.9895233106338397,
+      "loss_partial": 0.05224609375,
+      "step": 3798
+    },
+    {
+      "epoch": 1.9900471451021478,
+      "grad_norm": 1.4915584325790405,
+      "learning_rate": 1.1827449380362132e-05,
+      "loss": 0.1021,
+      "step": 3799
+    },
+    {
+      "epoch": 1.9900471451021478,
+      "loss_partial": 0.050537109375,
+      "step": 3799
+    },
+    {
+      "epoch": 1.9900471451021478,
+      "loss_partial": 0.061279296875,
+      "step": 3799
+    },
+    {
+      "epoch": 1.9905709795704558,
+      "grad_norm": 1.3702579736709595,
+      "learning_rate": 1.1822954589214024e-05,
+      "loss": 0.099,
+      "step": 3800
+    },
+    {
+      "epoch": 1.9905709795704558,
+      "loss_partial": 0.04736328125,
+      "step": 3800
+    },
+    {
+      "epoch": 1.9905709795704558,
+      "loss_partial": 0.0458984375,
+      "step": 3800
+    },
+    {
+      "epoch": 1.9910948140387639,
+      "grad_norm": 1.6693480014801025,
+      "learning_rate": 1.1818459417079677e-05,
+      "loss": 0.1019,
+      "step": 3801
+    },
+    {
+      "epoch": 1.9910948140387639,
+      "loss_partial": 0.05322265625,
+      "step": 3801
+    },
+    {
+      "epoch": 1.9910948140387639,
+      "loss_partial": 0.0673828125,
+      "step": 3801
+    },
+    {
+      "epoch": 1.991618648507072,
+      "grad_norm": 1.2863794565200806,
+      "learning_rate": 1.1813963864898554e-05,
+      "loss": 0.1024,
+      "step": 3802
+    },
+    {
+      "epoch": 1.991618648507072,
+      "loss_partial": 0.05859375,
+      "step": 3802
+    },
+    {
+      "epoch": 1.991618648507072,
+      "loss_partial": 0.0546875,
+      "step": 3802
+    },
+    {
+      "epoch": 1.9921424829753798,
+      "grad_norm": 1.7575840950012207,
+      "learning_rate": 1.1809467933610203e-05,
+      "loss": 0.0994,
+      "step": 3803
+    },
+    {
+      "epoch": 1.9921424829753798,
+      "loss_partial": 0.053955078125,
+      "step": 3803
+    },
+    {
+      "epoch": 1.9921424829753798,
+      "loss_partial": 0.05322265625,
+      "step": 3803
+    },
+    {
+      "epoch": 1.9926663174436878,
+      "grad_norm": 2.0164542198181152,
+      "learning_rate": 1.1804971624154242e-05,
+      "loss": 0.1023,
+      "step": 3804
+    },
+    {
+      "epoch": 1.9926663174436878,
+      "loss_partial": 0.050537109375,
+      "step": 3804
+    },
+    {
+      "epoch": 1.9926663174436878,
+      "loss_partial": 0.044189453125,
+      "step": 3804
+    },
+    {
+      "epoch": 1.9931901519119957,
+      "grad_norm": 1.6146918535232544,
+      "learning_rate": 1.1800474937470376e-05,
+      "loss": 0.1011,
+      "step": 3805
+    },
+    {
+      "epoch": 1.9931901519119957,
+      "loss_partial": 0.054931640625,
+      "step": 3805
+    },
+    {
+      "epoch": 1.9931901519119957,
+      "loss_partial": 0.04248046875,
+      "step": 3805
+    },
+    {
+      "epoch": 1.9937139863803037,
+      "grad_norm": 2.288591146469116,
+      "learning_rate": 1.1795977874498381e-05,
+      "loss": 0.1032,
+      "step": 3806
+    },
+    {
+      "epoch": 1.9937139863803037,
+      "loss_partial": 0.049560546875,
+      "step": 3806
+    },
+    {
+      "epoch": 1.9937139863803037,
+      "loss_partial": 0.04638671875,
+      "step": 3806
+    },
+    {
+      "epoch": 1.9942378208486118,
+      "grad_norm": 1.8129945993423462,
+      "learning_rate": 1.1791480436178118e-05,
+      "loss": 0.1019,
+      "step": 3807
+    },
+    {
+      "epoch": 1.9942378208486118,
+      "loss_partial": 0.052734375,
+      "step": 3807
+    },
+    {
+      "epoch": 1.9942378208486118,
+      "loss_partial": 0.054931640625,
+      "step": 3807
+    },
+    {
+      "epoch": 1.9947616553169198,
+      "grad_norm": 1.6118441820144653,
+      "learning_rate": 1.1786982623449517e-05,
+      "loss": 0.1028,
+      "step": 3808
+    },
+    {
+      "epoch": 1.9947616553169198,
+      "loss_partial": 0.0546875,
+      "step": 3808
+    },
+    {
+      "epoch": 1.9947616553169198,
+      "loss_partial": 0.051025390625,
+      "step": 3808
+    },
+    {
+      "epoch": 1.995285489785228,
+      "grad_norm": 2.0511434078216553,
+      "learning_rate": 1.1782484437252601e-05,
+      "loss": 0.1036,
+      "step": 3809
+    },
+    {
+      "epoch": 1.995285489785228,
+      "loss_partial": 0.056396484375,
+      "step": 3809
+    },
+    {
+      "epoch": 1.995285489785228,
+      "loss_partial": 0.053955078125,
+      "step": 3809
+    },
+    {
+      "epoch": 1.995809324253536,
+      "grad_norm": 1.4506785869598389,
+      "learning_rate": 1.1777985878527459e-05,
+      "loss": 0.1039,
+      "step": 3810
+    },
+    {
+      "epoch": 1.995809324253536,
+      "loss_partial": 0.04931640625,
+      "step": 3810
+    },
+    {
+      "epoch": 1.995809324253536,
+      "loss_partial": 0.048828125,
+      "step": 3810
+    },
+    {
+      "epoch": 1.996333158721844,
+      "grad_norm": 1.4165338277816772,
+      "learning_rate": 1.1773486948214263e-05,
+      "loss": 0.1005,
+      "step": 3811
+    },
+    {
+      "epoch": 1.996333158721844,
+      "loss_partial": 0.042236328125,
+      "step": 3811
+    },
+    {
+      "epoch": 1.996333158721844,
+      "loss_partial": 0.060546875,
+      "step": 3811
+    },
+    {
+      "epoch": 1.9968569931901519,
+      "grad_norm": 1.5458524227142334,
+      "learning_rate": 1.1768987647253265e-05,
+      "loss": 0.1009,
+      "step": 3812
+    },
+    {
+      "epoch": 1.9968569931901519,
+      "loss_partial": 0.039794921875,
+      "step": 3812
+    },
+    {
+      "epoch": 1.9968569931901519,
+      "loss_partial": 0.048583984375,
+      "step": 3812
+    },
+    {
+      "epoch": 1.99738082765846,
+      "grad_norm": 2.1625754833221436,
+      "learning_rate": 1.1764487976584787e-05,
+      "loss": 0.1068,
+      "step": 3813
+    },
+    {
+      "epoch": 1.99738082765846,
+      "loss_partial": 0.047119140625,
+      "step": 3813
+    },
+    {
+      "epoch": 1.99738082765846,
+      "loss_partial": 0.05126953125,
+      "step": 3813
+    },
+    {
+      "epoch": 1.997904662126768,
+      "grad_norm": 1.6346458196640015,
+      "learning_rate": 1.1759987937149228e-05,
+      "loss": 0.1017,
+      "step": 3814
+    },
+    {
+      "epoch": 1.997904662126768,
+      "loss_partial": 0.048095703125,
+      "step": 3814
+    },
+    {
+      "epoch": 1.997904662126768,
+      "loss_partial": 0.05419921875,
+      "step": 3814
+    },
+    {
+      "epoch": 1.9984284965950758,
+      "grad_norm": 1.9467390775680542,
+      "learning_rate": 1.1755487529887077e-05,
+      "loss": 0.0997,
+      "step": 3815
+    },
+    {
+      "epoch": 1.9984284965950758,
+      "loss_partial": 0.050048828125,
+      "step": 3815
+    },
+    {
+      "epoch": 1.9984284965950758,
+      "loss_partial": 0.05517578125,
+      "step": 3815
+    },
+    {
+      "epoch": 1.9989523310633839,
+      "grad_norm": 1.9445374011993408,
+      "learning_rate": 1.1750986755738887e-05,
+      "loss": 0.102,
+      "step": 3816
+    },
+    {
+      "epoch": 1.9989523310633839,
+      "loss_partial": 0.056396484375,
+      "step": 3816
+    },
+    {
+      "epoch": 1.9989523310633839,
+      "loss_partial": 0.043701171875,
+      "step": 3816
+    },
+    {
+      "epoch": 1.999476165531692,
+      "grad_norm": 1.7701895236968994,
+      "learning_rate": 1.1746485615645295e-05,
+      "loss": 0.107,
+      "step": 3817
+    },
+    {
+      "epoch": 1.999476165531692,
+      "loss_partial": 0.06298828125,
+      "step": 3817
+    },
+    {
+      "epoch": 1.999476165531692,
+      "loss_partial": 0.046630859375,
+      "step": 3817
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.9606728553771973,
+      "learning_rate": 1.1741984110547009e-05,
+      "loss": 0.1023,
+      "step": 3818
+    },
+    {
+      "epoch": 2.0,
+      "loss_partial": 0.05322265625,
+      "step": 3818
+    },
+    {
+      "epoch": 2.0,
+      "loss_partial": 0.0634765625,
+      "step": 3818
+    },
+    {
+      "epoch": 2.000523834468308,
+      "grad_norm": 1.6141618490219116,
+      "learning_rate": 1.1737482241384817e-05,
+      "loss": 0.1063,
+      "step": 3819
+    },
+    {
+      "epoch": 2.000523834468308,
+      "loss_partial": 0.056884765625,
+      "step": 3819
+    },
+    {
+      "epoch": 2.000523834468308,
+      "loss_partial": 0.046142578125,
+      "step": 3819
+    },
+    {
+      "epoch": 2.001047668936616,
+      "grad_norm": 1.8151497840881348,
+      "learning_rate": 1.1732980009099576e-05,
+      "loss": 0.102,
+      "step": 3820
+    },
+    {
+      "epoch": 2.001047668936616,
+      "loss_partial": 0.0439453125,
+      "step": 3820
+    },
+    {
+      "epoch": 2.001047668936616,
+      "loss_partial": 0.0498046875,
+      "step": 3820
+    },
+    {
+      "epoch": 2.001571503404924,
+      "grad_norm": 1.457637906074524,
+      "learning_rate": 1.1728477414632233e-05,
+      "loss": 0.1039,
+      "step": 3821
+    },
+    {
+      "epoch": 2.001571503404924,
+      "loss_partial": 0.042724609375,
+      "step": 3821
+    },
+    {
+      "epoch": 2.001571503404924,
+      "loss_partial": 0.047607421875,
+      "step": 3821
+    },
+    {
+      "epoch": 2.0020953378732322,
+      "grad_norm": 2.1773781776428223,
+      "learning_rate": 1.1723974458923796e-05,
+      "loss": 0.103,
+      "step": 3822
+    },
+    {
+      "epoch": 2.0020953378732322,
+      "loss_partial": 0.048828125,
+      "step": 3822
+    },
+    {
+      "epoch": 2.0020953378732322,
+      "loss_partial": 0.047607421875,
+      "step": 3822
+    },
+    {
+      "epoch": 2.00261917234154,
+      "grad_norm": 1.2896745204925537,
+      "learning_rate": 1.1719471142915362e-05,
+      "loss": 0.1041,
+      "step": 3823
+    },
+    {
+      "epoch": 2.00261917234154,
+      "loss_partial": 0.044677734375,
+      "step": 3823
+    },
+    {
+      "epoch": 2.00261917234154,
+      "loss_partial": 0.060546875,
+      "step": 3823
+    },
+    {
+      "epoch": 2.003143006809848,
+      "grad_norm": 1.9294331073760986,
+      "learning_rate": 1.171496746754809e-05,
+      "loss": 0.1052,
+      "step": 3824
+    },
+    {
+      "epoch": 2.003143006809848,
+      "loss_partial": 0.0546875,
+      "step": 3824
+    },
+    {
+      "epoch": 2.003143006809848,
+      "loss_partial": 0.046875,
+      "step": 3824
+    },
+    {
+      "epoch": 2.003666841278156,
+      "grad_norm": 1.5053311586380005,
+      "learning_rate": 1.1710463433763224e-05,
+      "loss": 0.1061,
+      "step": 3825
+    },
+    {
+      "epoch": 2.003666841278156,
+      "loss_partial": 0.047119140625,
+      "step": 3825
+    },
+    {
+      "epoch": 2.003666841278156,
+      "loss_partial": 0.052490234375,
+      "step": 3825
+    },
+    {
+      "epoch": 2.004190675746464,
+      "grad_norm": 1.8100781440734863,
+      "learning_rate": 1.1705959042502076e-05,
+      "loss": 0.1024,
+      "step": 3826
+    },
+    {
+      "epoch": 2.004190675746464,
+      "loss_partial": 0.052978515625,
+      "step": 3826
+    },
+    {
+      "epoch": 2.004190675746464,
+      "loss_partial": 0.044677734375,
+      "step": 3826
+    },
+    {
+      "epoch": 2.004714510214772,
+      "grad_norm": 1.4244567155838013,
+      "learning_rate": 1.1701454294706038e-05,
+      "loss": 0.1039,
+      "step": 3827
+    },
+    {
+      "epoch": 2.004714510214772,
+      "loss_partial": 0.05029296875,
+      "step": 3827
+    },
+    {
+      "epoch": 2.004714510214772,
+      "loss_partial": 0.048828125,
+      "step": 3827
+    },
+    {
+      "epoch": 2.00523834468308,
+      "grad_norm": 1.356160283088684,
+      "learning_rate": 1.1696949191316571e-05,
+      "loss": 0.1024,
+      "step": 3828
+    },
+    {
+      "epoch": 2.00523834468308,
+      "loss_partial": 0.048095703125,
+      "step": 3828
+    },
+    {
+      "epoch": 2.00523834468308,
+      "loss_partial": 0.057861328125,
+      "step": 3828
+    },
+    {
+      "epoch": 2.005762179151388,
+      "grad_norm": 1.5181236267089844,
+      "learning_rate": 1.169244373327522e-05,
+      "loss": 0.1067,
+      "step": 3829
+    },
+    {
+      "epoch": 2.005762179151388,
+      "loss_partial": 0.05078125,
+      "step": 3829
+    },
+    {
+      "epoch": 2.005762179151388,
+      "loss_partial": 0.0517578125,
+      "step": 3829
+    },
+    {
+      "epoch": 2.0062860136196963,
+      "grad_norm": 1.4340286254882812,
+      "learning_rate": 1.1687937921523596e-05,
+      "loss": 0.1079,
+      "step": 3830
+    },
+    {
+      "epoch": 2.0062860136196963,
+      "loss_partial": 0.056640625,
+      "step": 3830
+    },
+    {
+      "epoch": 2.0062860136196963,
+      "loss_partial": 0.050048828125,
+      "step": 3830
+    },
+    {
+      "epoch": 2.0068098480880043,
+      "grad_norm": 1.4859195947647095,
+      "learning_rate": 1.1683431757003381e-05,
+      "loss": 0.1003,
+      "step": 3831
+    },
+    {
+      "epoch": 2.0068098480880043,
+      "loss_partial": 0.0498046875,
+      "step": 3831
+    },
+    {
+      "epoch": 2.0068098480880043,
+      "loss_partial": 0.054443359375,
+      "step": 3831
+    },
+    {
+      "epoch": 2.0073336825563124,
+      "grad_norm": 1.257773995399475,
+      "learning_rate": 1.1678925240656337e-05,
+      "loss": 0.1063,
+      "step": 3832
+    },
+    {
+      "epoch": 2.0073336825563124,
+      "loss_partial": 0.0537109375,
+      "step": 3832
+    },
+    {
+      "epoch": 2.0073336825563124,
+      "loss_partial": 0.057373046875,
+      "step": 3832
+    },
+    {
+      "epoch": 2.00785751702462,
+      "grad_norm": 1.9005005359649658,
+      "learning_rate": 1.1674418373424306e-05,
+      "loss": 0.1034,
+      "step": 3833
+    },
+    {
+      "epoch": 2.00785751702462,
+      "loss_partial": 0.05078125,
+      "step": 3833
+    },
+    {
+      "epoch": 2.00785751702462,
+      "loss_partial": 0.047119140625,
+      "step": 3833
+    },
+    {
+      "epoch": 2.008381351492928,
+      "grad_norm": 1.8268593549728394,
+      "learning_rate": 1.1669911156249183e-05,
+      "loss": 0.1035,
+      "step": 3834
+    },
+    {
+      "epoch": 2.008381351492928,
+      "loss_partial": 0.054931640625,
+      "step": 3834
+    },
+    {
+      "epoch": 2.008381351492928,
+      "loss_partial": 0.04443359375,
+      "step": 3834
+    },
+    {
+      "epoch": 2.008905185961236,
+      "grad_norm": 1.491845726966858,
+      "learning_rate": 1.1665403590072958e-05,
+      "loss": 0.1012,
+      "step": 3835
+    },
+    {
+      "epoch": 2.008905185961236,
+      "loss_partial": 0.046875,
+      "step": 3835
+    },
+    {
+      "epoch": 2.008905185961236,
+      "loss_partial": 0.05224609375,
+      "step": 3835
+    },
+    {
+      "epoch": 2.009429020429544,
+      "grad_norm": 1.7842400074005127,
+      "learning_rate": 1.1660895675837682e-05,
+      "loss": 0.1031,
+      "step": 3836
+    },
+    {
+      "epoch": 2.009429020429544,
+      "loss_partial": 0.052734375,
+      "step": 3836
+    },
+    {
+      "epoch": 2.009429020429544,
+      "loss_partial": 0.044677734375,
+      "step": 3836
+    },
+    {
+      "epoch": 2.0099528548978522,
+      "grad_norm": 2.098881721496582,
+      "learning_rate": 1.165638741448548e-05,
+      "loss": 0.1031,
+      "step": 3837
+    },
+    {
+      "epoch": 2.0099528548978522,
+      "loss_partial": 0.046630859375,
+      "step": 3837
+    },
+    {
+      "epoch": 2.0099528548978522,
+      "loss_partial": 0.05029296875,
+      "step": 3837
+    },
+    {
+      "epoch": 2.0104766893661603,
+      "grad_norm": 2.047577381134033,
+      "learning_rate": 1.1651878806958547e-05,
+      "loss": 0.0993,
+      "step": 3838
+    },
+    {
+      "epoch": 2.0104766893661603,
+      "loss_partial": 0.050537109375,
+      "step": 3838
+    },
+    {
+      "epoch": 2.0104766893661603,
+      "loss_partial": 0.046142578125,
+      "step": 3838
+    },
+    {
+      "epoch": 2.0110005238344684,
+      "grad_norm": 1.8531569242477417,
+      "learning_rate": 1.1647369854199164e-05,
+      "loss": 0.1031,
+      "step": 3839
+    },
+    {
+      "epoch": 2.0110005238344684,
+      "loss_partial": 0.04833984375,
+      "step": 3839
+    },
+    {
+      "epoch": 2.0110005238344684,
+      "loss_partial": 0.043701171875,
+      "step": 3839
+    },
+    {
+      "epoch": 2.0115243583027764,
+      "grad_norm": 1.6642943620681763,
+      "learning_rate": 1.1642860557149664e-05,
+      "loss": 0.1012,
+      "step": 3840
+    },
+    {
+      "epoch": 2.0115243583027764,
+      "loss_partial": 0.04638671875,
+      "step": 3840
+    },
+    {
+      "epoch": 2.0115243583027764,
+      "loss_partial": 0.043212890625,
+      "step": 3840
+    },
+    {
+      "epoch": 2.0120481927710845,
+      "grad_norm": 1.977705478668213,
+      "learning_rate": 1.163835091675247e-05,
+      "loss": 0.0997,
+      "step": 3841
+    },
+    {
+      "epoch": 2.0120481927710845,
+      "loss_partial": 0.050048828125,
+      "step": 3841
+    },
+    {
+      "epoch": 2.0120481927710845,
+      "loss_partial": 0.045654296875,
+      "step": 3841
+    },
+    {
+      "epoch": 2.0125720272393925,
+      "grad_norm": 1.6529297828674316,
+      "learning_rate": 1.1633840933950064e-05,
+      "loss": 0.1035,
+      "step": 3842
+    },
+    {
+      "epoch": 2.0125720272393925,
+      "loss_partial": 0.046875,
+      "step": 3842
+    },
+    {
+      "epoch": 2.0125720272393925,
+      "loss_partial": 0.04736328125,
+      "step": 3842
+    },
+    {
+      "epoch": 2.0130958617077,
+      "grad_norm": 1.8696359395980835,
+      "learning_rate": 1.1629330609685008e-05,
+      "loss": 0.1027,
+      "step": 3843
+    },
+    {
+      "epoch": 2.0130958617077,
+      "loss_partial": 0.05126953125,
+      "step": 3843
+    },
+    {
+      "epoch": 2.0130958617077,
+      "loss_partial": 0.054931640625,
+      "step": 3843
+    },
+    {
+      "epoch": 2.0136196961760082,
+      "grad_norm": 1.8566350936889648,
+      "learning_rate": 1.1624819944899928e-05,
+      "loss": 0.1004,
+      "step": 3844
+    },
+    {
+      "epoch": 2.0136196961760082,
+      "loss_partial": 0.052001953125,
+      "step": 3844
+    },
+    {
+      "epoch": 2.0136196961760082,
+      "loss_partial": 0.06689453125,
+      "step": 3844
+    },
+    {
+      "epoch": 2.0141435306443163,
+      "grad_norm": 1.6099997758865356,
+      "learning_rate": 1.1620308940537527e-05,
+      "loss": 0.1037,
+      "step": 3845
+    },
+    {
+      "epoch": 2.0141435306443163,
+      "loss_partial": 0.052001953125,
+      "step": 3845
+    },
+    {
+      "epoch": 2.0141435306443163,
+      "loss_partial": 0.060302734375,
+      "step": 3845
+    },
+    {
+      "epoch": 2.0146673651126243,
+      "grad_norm": 1.7751051187515259,
+      "learning_rate": 1.1615797597540577e-05,
+      "loss": 0.1018,
+      "step": 3846
+    },
+    {
+      "epoch": 2.0146673651126243,
+      "loss_partial": 0.054931640625,
+      "step": 3846
+    },
+    {
+      "epoch": 2.0146673651126243,
+      "loss_partial": 0.054443359375,
+      "step": 3846
+    },
+    {
+      "epoch": 2.0151911995809324,
+      "grad_norm": 1.4348556995391846,
+      "learning_rate": 1.1611285916851922e-05,
+      "loss": 0.1064,
+      "step": 3847
+    },
+    {
+      "epoch": 2.0151911995809324,
+      "loss_partial": 0.0458984375,
+      "step": 3847
+    },
+    {
+      "epoch": 2.0151911995809324,
+      "loss_partial": 0.0615234375,
+      "step": 3847
+    },
+    {
+      "epoch": 2.0157150340492405,
+      "grad_norm": 2.091855764389038,
+      "learning_rate": 1.1606773899414476e-05,
+      "loss": 0.1042,
+      "step": 3848
+    },
+    {
+      "epoch": 2.0157150340492405,
+      "loss_partial": 0.0556640625,
+      "step": 3848
+    },
+    {
+      "epoch": 2.0157150340492405,
+      "loss_partial": 0.04833984375,
+      "step": 3848
+    },
+    {
+      "epoch": 2.0162388685175485,
+      "grad_norm": 1.2741467952728271,
+      "learning_rate": 1.1602261546171221e-05,
+      "loss": 0.1045,
+      "step": 3849
+    },
+    {
+      "epoch": 2.0162388685175485,
+      "loss_partial": 0.043701171875,
+      "step": 3849
+    },
+    {
+      "epoch": 2.0162388685175485,
+      "loss_partial": 0.04638671875,
+      "step": 3849
+    },
+    {
+      "epoch": 2.0167627029858566,
+      "grad_norm": 1.9412782192230225,
+      "learning_rate": 1.1597748858065205e-05,
+      "loss": 0.1031,
+      "step": 3850
+    },
+    {
+      "epoch": 2.0167627029858566,
+      "loss_partial": 0.045654296875,
+      "step": 3850
+    },
+    {
+      "epoch": 2.0167627029858566,
+      "loss_partial": 0.053466796875,
+      "step": 3850
+    },
+    {
+      "epoch": 2.0172865374541646,
+      "grad_norm": 1.5244251489639282,
+      "learning_rate": 1.1593235836039564e-05,
+      "loss": 0.1031,
+      "step": 3851
+    },
+    {
+      "epoch": 2.0172865374541646,
+      "loss_partial": 0.04931640625,
+      "step": 3851
+    },
+    {
+      "epoch": 2.0172865374541646,
+      "loss_partial": 0.059326171875,
+      "step": 3851
+    },
+    {
+      "epoch": 2.0178103719224727,
+      "grad_norm": 1.9703848361968994,
+      "learning_rate": 1.1588722481037481e-05,
+      "loss": 0.1057,
+      "step": 3852
+    },
+    {
+      "epoch": 2.0178103719224727,
+      "loss_partial": 0.0576171875,
+      "step": 3852
+    },
+    {
+      "epoch": 2.0178103719224727,
+      "loss_partial": 0.049072265625,
+      "step": 3852
+    },
+    {
+      "epoch": 2.0183342063907803,
+      "grad_norm": 1.6158605813980103,
+      "learning_rate": 1.1584208794002227e-05,
+      "loss": 0.1052,
+      "step": 3853
+    },
+    {
+      "epoch": 2.0183342063907803,
+      "loss_partial": 0.056640625,
+      "step": 3853
+    },
+    {
+      "epoch": 2.0183342063907803,
+      "loss_partial": 0.05322265625,
+      "step": 3853
+    },
+    {
+      "epoch": 2.0188580408590884,
+      "grad_norm": 2.373288869857788,
+      "learning_rate": 1.1579694775877134e-05,
+      "loss": 0.1043,
+      "step": 3854
+    },
+    {
+      "epoch": 2.0188580408590884,
+      "loss_partial": 0.0439453125,
+      "step": 3854
+    },
+    {
+      "epoch": 2.0188580408590884,
+      "loss_partial": 0.056396484375,
+      "step": 3854
+    },
+    {
+      "epoch": 2.0193818753273964,
+      "grad_norm": 2.1631407737731934,
+      "learning_rate": 1.1575180427605597e-05,
+      "loss": 0.1024,
+      "step": 3855
+    },
+    {
+      "epoch": 2.0193818753273964,
+      "loss_partial": 0.0498046875,
+      "step": 3855
+    },
+    {
+      "epoch": 2.0193818753273964,
+      "loss_partial": 0.048828125,
+      "step": 3855
+    },
+    {
+      "epoch": 2.0199057097957045,
+      "grad_norm": 1.8477598428726196,
+      "learning_rate": 1.157066575013109e-05,
+      "loss": 0.1062,
+      "step": 3856
+    },
+    {
+      "epoch": 2.0199057097957045,
+      "loss_partial": 0.056396484375,
+      "step": 3856
+    },
+    {
+      "epoch": 2.0199057097957045,
+      "loss_partial": 0.045654296875,
+      "step": 3856
+    },
+    {
+      "epoch": 2.0204295442640126,
+      "grad_norm": 1.5384107828140259,
+      "learning_rate": 1.1566150744397158e-05,
+      "loss": 0.1007,
+      "step": 3857
+    },
+    {
+      "epoch": 2.0204295442640126,
+      "loss_partial": 0.047607421875,
+      "step": 3857
+    },
+    {
+      "epoch": 2.0204295442640126,
+      "loss_partial": 0.04150390625,
+      "step": 3857
+    },
+    {
+      "epoch": 2.0209533787323206,
+      "grad_norm": 2.06825852394104,
+      "learning_rate": 1.1561635411347406e-05,
+      "loss": 0.101,
+      "step": 3858
+    },
+    {
+      "epoch": 2.0209533787323206,
+      "loss_partial": 0.051513671875,
+      "step": 3858
+    },
+    {
+      "epoch": 2.0209533787323206,
+      "loss_partial": 0.0546875,
+      "step": 3858
+    },
+    {
+      "epoch": 2.0214772132006287,
+      "grad_norm": 1.3603748083114624,
+      "learning_rate": 1.1557119751925505e-05,
+      "loss": 0.1069,
+      "step": 3859
+    },
+    {
+      "epoch": 2.0214772132006287,
+      "loss_partial": 0.04736328125,
+      "step": 3859
+    },
+    {
+      "epoch": 2.0214772132006287,
+      "loss_partial": 0.057861328125,
+      "step": 3859
+    },
+    {
+      "epoch": 2.0220010476689367,
+      "grad_norm": 1.962710976600647,
+      "learning_rate": 1.1552603767075207e-05,
+      "loss": 0.1055,
+      "step": 3860
+    },
+    {
+      "epoch": 2.0220010476689367,
+      "loss_partial": 0.07177734375,
+      "step": 3860
+    },
+    {
+      "epoch": 2.0220010476689367,
+      "loss_partial": 0.0498046875,
+      "step": 3860
+    },
+    {
+      "epoch": 2.022524882137245,
+      "grad_norm": 1.6922783851623535,
+      "learning_rate": 1.1548087457740325e-05,
+      "loss": 0.1055,
+      "step": 3861
+    },
+    {
+      "epoch": 2.022524882137245,
+      "loss_partial": 0.051513671875,
+      "step": 3861
+    },
+    {
+      "epoch": 2.022524882137245,
+      "loss_partial": 0.054443359375,
+      "step": 3861
+    },
+    {
+      "epoch": 2.023048716605553,
+      "grad_norm": 1.705483078956604,
+      "learning_rate": 1.1543570824864733e-05,
+      "loss": 0.1032,
+      "step": 3862
+    },
+    {
+      "epoch": 2.023048716605553,
+      "loss_partial": 0.05029296875,
+      "step": 3862
+    },
+    {
+      "epoch": 2.023048716605553,
+      "loss_partial": 0.041259765625,
+      "step": 3862
+    },
+    {
+      "epoch": 2.0235725510738605,
+      "grad_norm": 1.5265690088272095,
+      "learning_rate": 1.1539053869392387e-05,
+      "loss": 0.0997,
+      "step": 3863
+    },
+    {
+      "epoch": 2.0235725510738605,
+      "loss_partial": 0.058349609375,
+      "step": 3863
+    },
+    {
+      "epoch": 2.0235725510738605,
+      "loss_partial": 0.0419921875,
+      "step": 3863
+    },
+    {
+      "epoch": 2.0240963855421685,
+      "grad_norm": 2.0449841022491455,
+      "learning_rate": 1.1534536592267298e-05,
+      "loss": 0.1013,
+      "step": 3864
+    },
+    {
+      "epoch": 2.0240963855421685,
+      "loss_partial": 0.05322265625,
+      "step": 3864
+    },
+    {
+      "epoch": 2.0240963855421685,
+      "loss_partial": 0.048828125,
+      "step": 3864
+    },
+    {
+      "epoch": 2.0246202200104766,
+      "grad_norm": 1.4224085807800293,
+      "learning_rate": 1.153001899443355e-05,
+      "loss": 0.1045,
+      "step": 3865
+    },
+    {
+      "epoch": 2.0246202200104766,
+      "loss_partial": 0.06201171875,
+      "step": 3865
+    },
+    {
+      "epoch": 2.0246202200104766,
+      "loss_partial": 0.041015625,
+      "step": 3865
+    },
+    {
+      "epoch": 2.0251440544787846,
+      "grad_norm": 1.5231618881225586,
+      "learning_rate": 1.1525501076835295e-05,
+      "loss": 0.1058,
+      "step": 3866
+    },
+    {
+      "epoch": 2.0251440544787846,
+      "loss_partial": 0.05029296875,
+      "step": 3866
+    },
+    {
+      "epoch": 2.0251440544787846,
+      "loss_partial": 0.05029296875,
+      "step": 3866
+    },
+    {
+      "epoch": 2.0256678889470927,
+      "grad_norm": 1.3764619827270508,
+      "learning_rate": 1.152098284041675e-05,
+      "loss": 0.0994,
+      "step": 3867
+    },
+    {
+      "epoch": 2.0256678889470927,
+      "loss_partial": 0.0478515625,
+      "step": 3867
+    },
+    {
+      "epoch": 2.0256678889470927,
+      "loss_partial": 0.053466796875,
+      "step": 3867
+    },
+    {
+      "epoch": 2.0261917234154008,
+      "grad_norm": 1.9590307474136353,
+      "learning_rate": 1.1516464286122196e-05,
+      "loss": 0.1053,
+      "step": 3868
+    },
+    {
+      "epoch": 2.0261917234154008,
+      "loss_partial": 0.0517578125,
+      "step": 3868
+    },
+    {
+      "epoch": 2.0261917234154008,
+      "loss_partial": 0.04345703125,
+      "step": 3868
+    },
+    {
+      "epoch": 2.026715557883709,
+      "grad_norm": 1.537001132965088,
+      "learning_rate": 1.151194541489598e-05,
+      "loss": 0.0991,
+      "step": 3869
+    },
+    {
+      "epoch": 2.026715557883709,
+      "loss_partial": 0.05712890625,
+      "step": 3869
+    },
+    {
+      "epoch": 2.026715557883709,
+      "loss_partial": 0.056396484375,
+      "step": 3869
+    },
+    {
+      "epoch": 2.027239392352017,
+      "grad_norm": 1.9722352027893066,
+      "learning_rate": 1.1507426227682526e-05,
+      "loss": 0.1014,
+      "step": 3870
+    },
+    {
+      "epoch": 2.027239392352017,
+      "loss_partial": 0.05224609375,
+      "step": 3870
+    },
+    {
+      "epoch": 2.027239392352017,
+      "loss_partial": 0.048828125,
+      "step": 3870
+    },
+    {
+      "epoch": 2.027763226820325,
+      "grad_norm": 1.6995984315872192,
+      "learning_rate": 1.1502906725426308e-05,
+      "loss": 0.1015,
+      "step": 3871
+    },
+    {
+      "epoch": 2.027763226820325,
+      "loss_partial": 0.0400390625,
+      "step": 3871
+    },
+    {
+      "epoch": 2.027763226820325,
+      "loss_partial": 0.053466796875,
+      "step": 3871
+    },
+    {
+      "epoch": 2.028287061288633,
+      "grad_norm": 2.0951666831970215,
+      "learning_rate": 1.1498386909071881e-05,
+      "loss": 0.1053,
+      "step": 3872
+    },
+    {
+      "epoch": 2.028287061288633,
+      "loss_partial": 0.042724609375,
+      "step": 3872
+    },
+    {
+      "epoch": 2.028287061288633,
+      "loss_partial": 0.048583984375,
+      "step": 3872
+    },
+    {
+      "epoch": 2.0288108957569406,
+      "grad_norm": 1.991392970085144,
+      "learning_rate": 1.1493866779563856e-05,
+      "loss": 0.1018,
+      "step": 3873
+    },
+    {
+      "epoch": 2.0288108957569406,
+      "loss_partial": 0.044677734375,
+      "step": 3873
+    },
+    {
+      "epoch": 2.0288108957569406,
+      "loss_partial": 0.052001953125,
+      "step": 3873
+    },
+    {
+      "epoch": 2.0293347302252487,
+      "grad_norm": 1.8120362758636475,
+      "learning_rate": 1.1489346337846912e-05,
+      "loss": 0.1033,
+      "step": 3874
+    },
+    {
+      "epoch": 2.0293347302252487,
+      "loss_partial": 0.05224609375,
+      "step": 3874
+    },
+    {
+      "epoch": 2.0293347302252487,
+      "loss_partial": 0.052978515625,
+      "step": 3874
+    },
+    {
+      "epoch": 2.0298585646935567,
+      "grad_norm": 2.079119920730591,
+      "learning_rate": 1.1484825584865792e-05,
+      "loss": 0.1032,
+      "step": 3875
+    },
+    {
+      "epoch": 2.0298585646935567,
+      "loss_partial": 0.05322265625,
+      "step": 3875
+    },
+    {
+      "epoch": 2.0298585646935567,
+      "loss_partial": 0.049072265625,
+      "step": 3875
+    },
+    {
+      "epoch": 2.030382399161865,
+      "grad_norm": 1.8018975257873535,
+      "learning_rate": 1.1480304521565306e-05,
+      "loss": 0.1029,
+      "step": 3876
+    },
+    {
+      "epoch": 2.030382399161865,
+      "loss_partial": 0.056640625,
+      "step": 3876
+    },
+    {
+      "epoch": 2.030382399161865,
+      "loss_partial": 0.05078125,
+      "step": 3876
+    },
+    {
+      "epoch": 2.030906233630173,
+      "grad_norm": 2.2429280281066895,
+      "learning_rate": 1.147578314889033e-05,
+      "loss": 0.1066,
+      "step": 3877
+    },
+    {
+      "epoch": 2.030906233630173,
+      "loss_partial": 0.04931640625,
+      "step": 3877
+    },
+    {
+      "epoch": 2.030906233630173,
+      "loss_partial": 0.045654296875,
+      "step": 3877
+    },
+    {
+      "epoch": 2.031430068098481,
+      "grad_norm": 1.871033787727356,
+      "learning_rate": 1.1471261467785802e-05,
+      "loss": 0.1015,
+      "step": 3878
+    },
+    {
+      "epoch": 2.031430068098481,
+      "loss_partial": 0.0546875,
+      "step": 3878
+    },
+    {
+      "epoch": 2.031430068098481,
+      "loss_partial": 0.048828125,
+      "step": 3878
+    },
+    {
+      "epoch": 2.031953902566789,
+      "grad_norm": 1.9860341548919678,
+      "learning_rate": 1.1466739479196725e-05,
+      "loss": 0.1064,
+      "step": 3879
+    },
+    {
+      "epoch": 2.031953902566789,
+      "loss_partial": 0.05126953125,
+      "step": 3879
+    },
+    {
+      "epoch": 2.031953902566789,
+      "loss_partial": 0.04833984375,
+      "step": 3879
+    },
+    {
+      "epoch": 2.032477737035097,
+      "grad_norm": 1.685318946838379,
+      "learning_rate": 1.1462217184068167e-05,
+      "loss": 0.104,
+      "step": 3880
+    },
+    {
+      "epoch": 2.032477737035097,
+      "loss_partial": 0.049560546875,
+      "step": 3880
+    },
+    {
+      "epoch": 2.032477737035097,
+      "loss_partial": 0.046142578125,
+      "step": 3880
+    },
+    {
+      "epoch": 2.033001571503405,
+      "grad_norm": 1.8241225481033325,
+      "learning_rate": 1.145769458334526e-05,
+      "loss": 0.0985,
+      "step": 3881
+    },
+    {
+      "epoch": 2.033001571503405,
+      "loss_partial": 0.049072265625,
+      "step": 3881
+    },
+    {
+      "epoch": 2.033001571503405,
+      "loss_partial": 0.048095703125,
+      "step": 3881
+    },
+    {
+      "epoch": 2.033525405971713,
+      "grad_norm": 1.7483168840408325,
+      "learning_rate": 1.1453171677973199e-05,
+      "loss": 0.1004,
+      "step": 3882
+    },
+    {
+      "epoch": 2.033525405971713,
+      "loss_partial": 0.048095703125,
+      "step": 3882
+    },
+    {
+      "epoch": 2.033525405971713,
+      "loss_partial": 0.04638671875,
+      "step": 3882
+    },
+    {
+      "epoch": 2.034049240440021,
+      "grad_norm": 2.118739366531372,
+      "learning_rate": 1.1448648468897242e-05,
+      "loss": 0.1027,
+      "step": 3883
+    },
+    {
+      "epoch": 2.034049240440021,
+      "loss_partial": 0.05419921875,
+      "step": 3883
+    },
+    {
+      "epoch": 2.034049240440021,
+      "loss_partial": 0.043701171875,
+      "step": 3883
+    },
+    {
+      "epoch": 2.034573074908329,
+      "grad_norm": 1.6318352222442627,
+      "learning_rate": 1.1444124957062715e-05,
+      "loss": 0.1012,
+      "step": 3884
+    },
+    {
+      "epoch": 2.034573074908329,
+      "loss_partial": 0.057373046875,
+      "step": 3884
+    },
+    {
+      "epoch": 2.034573074908329,
+      "loss_partial": 0.049560546875,
+      "step": 3884
+    },
+    {
+      "epoch": 2.035096909376637,
+      "grad_norm": 2.1829919815063477,
+      "learning_rate": 1.1439601143415003e-05,
+      "loss": 0.105,
+      "step": 3885
+    },
+    {
+      "epoch": 2.035096909376637,
+      "loss_partial": 0.04443359375,
+      "step": 3885
+    },
+    {
+      "epoch": 2.035096909376637,
+      "loss_partial": 0.05859375,
+      "step": 3885
+    },
+    {
+      "epoch": 2.035620743844945,
+      "grad_norm": 1.7502796649932861,
+      "learning_rate": 1.1435077028899553e-05,
+      "loss": 0.1066,
+      "step": 3886
+    },
+    {
+      "epoch": 2.035620743844945,
+      "loss_partial": 0.052490234375,
+      "step": 3886
+    },
+    {
+      "epoch": 2.035620743844945,
+      "loss_partial": 0.05859375,
+      "step": 3886
+    },
+    {
+      "epoch": 2.036144578313253,
+      "grad_norm": 1.515309453010559,
+      "learning_rate": 1.1430552614461877e-05,
+      "loss": 0.1022,
+      "step": 3887
+    },
+    {
+      "epoch": 2.036144578313253,
+      "loss_partial": 0.0556640625,
+      "step": 3887
+    },
+    {
+      "epoch": 2.036144578313253,
+      "loss_partial": 0.04541015625,
+      "step": 3887
+    },
+    {
+      "epoch": 2.036668412781561,
+      "grad_norm": 1.4591270685195923,
+      "learning_rate": 1.1426027901047551e-05,
+      "loss": 0.1044,
+      "step": 3888
+    },
+    {
+      "epoch": 2.036668412781561,
+      "loss_partial": 0.06396484375,
+      "step": 3888
+    },
+    {
+      "epoch": 2.036668412781561,
+      "loss_partial": 0.054931640625,
+      "step": 3888
+    },
+    {
+      "epoch": 2.037192247249869,
+      "grad_norm": 2.0943760871887207,
+      "learning_rate": 1.1421502889602212e-05,
+      "loss": 0.1057,
+      "step": 3889
+    },
+    {
+      "epoch": 2.037192247249869,
+      "loss_partial": 0.05712890625,
+      "step": 3889
+    },
+    {
+      "epoch": 2.037192247249869,
+      "loss_partial": 0.045166015625,
+      "step": 3889
+    },
+    {
+      "epoch": 2.037716081718177,
+      "grad_norm": 1.5506174564361572,
+      "learning_rate": 1.141697758107156e-05,
+      "loss": 0.1027,
+      "step": 3890
+    },
+    {
+      "epoch": 2.037716081718177,
+      "loss_partial": 0.049560546875,
+      "step": 3890
+    },
+    {
+      "epoch": 2.037716081718177,
+      "loss_partial": 0.049560546875,
+      "step": 3890
+    },
+    {
+      "epoch": 2.0382399161864853,
+      "grad_norm": 2.2700085639953613,
+      "learning_rate": 1.1412451976401357e-05,
+      "loss": 0.1049,
+      "step": 3891
+    },
+    {
+      "epoch": 2.0382399161864853,
+      "loss_partial": 0.0576171875,
+      "step": 3891
+    },
+    {
+      "epoch": 2.0382399161864853,
+      "loss_partial": 0.047119140625,
+      "step": 3891
+    },
+    {
+      "epoch": 2.038763750654793,
+      "grad_norm": 1.683738350868225,
+      "learning_rate": 1.1407926076537425e-05,
+      "loss": 0.106,
+      "step": 3892
+    },
+    {
+      "epoch": 2.038763750654793,
+      "loss_partial": 0.046630859375,
+      "step": 3892
+    },
+    {
+      "epoch": 2.038763750654793,
+      "loss_partial": 0.04736328125,
+      "step": 3892
+    },
+    {
+      "epoch": 2.039287585123101,
+      "grad_norm": 1.96676504611969,
+      "learning_rate": 1.1403399882425647e-05,
+      "loss": 0.1075,
+      "step": 3893
+    },
+    {
+      "epoch": 2.039287585123101,
+      "loss_partial": 0.0478515625,
+      "step": 3893
+    },
+    {
+      "epoch": 2.039287585123101,
+      "loss_partial": 0.052490234375,
+      "step": 3893
+    },
+    {
+      "epoch": 2.039811419591409,
+      "grad_norm": 2.300421953201294,
+      "learning_rate": 1.1398873395011973e-05,
+      "loss": 0.1022,
+      "step": 3894
+    },
+    {
+      "epoch": 2.039811419591409,
+      "loss_partial": 0.055908203125,
+      "step": 3894
+    },
+    {
+      "epoch": 2.039811419591409,
+      "loss_partial": 0.05224609375,
+      "step": 3894
+    },
+    {
+      "epoch": 2.040335254059717,
+      "grad_norm": 1.9097695350646973,
+      "learning_rate": 1.1394346615242411e-05,
+      "loss": 0.1011,
+      "step": 3895
+    },
+    {
+      "epoch": 2.040335254059717,
+      "loss_partial": 0.05615234375,
+      "step": 3895
+    },
+    {
+      "epoch": 2.040335254059717,
+      "loss_partial": 0.0654296875,
+      "step": 3895
+    },
+    {
+      "epoch": 2.040859088528025,
+      "grad_norm": 1.8692134618759155,
+      "learning_rate": 1.1389819544063027e-05,
+      "loss": 0.1035,
+      "step": 3896
+    },
+    {
+      "epoch": 2.040859088528025,
+      "loss_partial": 0.04833984375,
+      "step": 3896
+    },
+    {
+      "epoch": 2.040859088528025,
+      "loss_partial": 0.050048828125,
+      "step": 3896
+    },
+    {
+      "epoch": 2.041382922996333,
+      "grad_norm": 2.2393507957458496,
+      "learning_rate": 1.1385292182419955e-05,
+      "loss": 0.1019,
+      "step": 3897
+    },
+    {
+      "epoch": 2.041382922996333,
+      "loss_partial": 0.04345703125,
+      "step": 3897
+    },
+    {
+      "epoch": 2.041382922996333,
+      "loss_partial": 0.05517578125,
+      "step": 3897
+    },
+    {
+      "epoch": 2.0419067574646412,
+      "grad_norm": 2.254582405090332,
+      "learning_rate": 1.1380764531259386e-05,
+      "loss": 0.1054,
+      "step": 3898
+    },
+    {
+      "epoch": 2.0419067574646412,
+      "loss_partial": 0.046630859375,
+      "step": 3898
+    },
+    {
+      "epoch": 2.0419067574646412,
+      "loss_partial": 0.055419921875,
+      "step": 3898
+    },
+    {
+      "epoch": 2.0424305919329493,
+      "grad_norm": 2.1799519062042236,
+      "learning_rate": 1.137623659152756e-05,
+      "loss": 0.1043,
+      "step": 3899
+    },
+    {
+      "epoch": 2.0424305919329493,
+      "loss_partial": 0.0517578125,
+      "step": 3899
+    },
+    {
+      "epoch": 2.0424305919329493,
+      "loss_partial": 0.055908203125,
+      "step": 3899
+    },
+    {
+      "epoch": 2.0429544264012574,
+      "grad_norm": 1.793286681175232,
+      "learning_rate": 1.1371708364170804e-05,
+      "loss": 0.105,
+      "step": 3900
+    },
+    {
+      "epoch": 2.0429544264012574,
+      "loss_partial": 0.047607421875,
+      "step": 3900
+    },
+    {
+      "epoch": 2.0429544264012574,
+      "loss_partial": 0.041015625,
+      "step": 3900
+    },
+    {
+      "epoch": 2.0434782608695654,
+      "grad_norm": 1.7471271753311157,
+      "learning_rate": 1.1367179850135478e-05,
+      "loss": 0.1086,
+      "step": 3901
+    },
+    {
+      "epoch": 2.0434782608695654,
+      "loss_partial": 0.05615234375,
+      "step": 3901
+    },
+    {
+      "epoch": 2.0434782608695654,
+      "loss_partial": 0.05078125,
+      "step": 3901
+    },
+    {
+      "epoch": 2.044002095337873,
+      "grad_norm": 1.556159496307373,
+      "learning_rate": 1.136265105036802e-05,
+      "loss": 0.1,
+      "step": 3902
+    },
+    {
+      "epoch": 2.044002095337873,
+      "loss_partial": 0.054931640625,
+      "step": 3902
+    },
+    {
+      "epoch": 2.044002095337873,
+      "loss_partial": 0.044677734375,
+      "step": 3902
+    },
+    {
+      "epoch": 2.044525929806181,
+      "grad_norm": 2.2746546268463135,
+      "learning_rate": 1.1358121965814918e-05,
+      "loss": 0.1011,
+      "step": 3903
+    },
+    {
+      "epoch": 2.044525929806181,
+      "loss_partial": 0.048095703125,
+      "step": 3903
+    },
+    {
+      "epoch": 2.044525929806181,
+      "loss_partial": 0.04296875,
+      "step": 3903
+    },
+    {
+      "epoch": 2.045049764274489,
+      "grad_norm": 1.2786346673965454,
+      "learning_rate": 1.1353592597422725e-05,
+      "loss": 0.1032,
+      "step": 3904
+    },
+    {
+      "epoch": 2.045049764274489,
+      "loss_partial": 0.049560546875,
+      "step": 3904
+    },
+    {
+      "epoch": 2.045049764274489,
+      "loss_partial": 0.050048828125,
+      "step": 3904
+    },
+    {
+      "epoch": 2.045573598742797,
+      "grad_norm": 2.034836530685425,
+      "learning_rate": 1.1349062946138046e-05,
+      "loss": 0.1078,
+      "step": 3905
+    },
+    {
+      "epoch": 2.045573598742797,
+      "loss_partial": 0.048095703125,
+      "step": 3905
+    },
+    {
+      "epoch": 2.045573598742797,
+      "loss_partial": 0.046630859375,
+      "step": 3905
+    },
+    {
+      "epoch": 2.0460974332111053,
+      "grad_norm": 2.0093467235565186,
+      "learning_rate": 1.1344533012907557e-05,
+      "loss": 0.1008,
+      "step": 3906
+    },
+    {
+      "epoch": 2.0460974332111053,
+      "loss_partial": 0.05712890625,
+      "step": 3906
+    },
+    {
+      "epoch": 2.0460974332111053,
+      "loss_partial": 0.04833984375,
+      "step": 3906
+    },
+    {
+      "epoch": 2.0466212676794133,
+      "grad_norm": 1.4986815452575684,
+      "learning_rate": 1.134000279867798e-05,
+      "loss": 0.1038,
+      "step": 3907
+    },
+    {
+      "epoch": 2.0466212676794133,
+      "loss_partial": 0.04541015625,
+      "step": 3907
+    },
+    {
+      "epoch": 2.0466212676794133,
+      "loss_partial": 0.057861328125,
+      "step": 3907
+    },
+    {
+      "epoch": 2.0471451021477214,
+      "grad_norm": 2.039639711380005,
+      "learning_rate": 1.1335472304396106e-05,
+      "loss": 0.1024,
+      "step": 3908
+    },
+    {
+      "epoch": 2.0471451021477214,
+      "loss_partial": 0.052001953125,
+      "step": 3908
+    },
+    {
+      "epoch": 2.0471451021477214,
+      "loss_partial": 0.058837890625,
+      "step": 3908
+    },
+    {
+      "epoch": 2.0476689366160294,
+      "grad_norm": 1.666052222251892,
+      "learning_rate": 1.1330941531008779e-05,
+      "loss": 0.105,
+      "step": 3909
+    },
+    {
+      "epoch": 2.0476689366160294,
+      "loss_partial": 0.054931640625,
+      "step": 3909
+    },
+    {
+      "epoch": 2.0476689366160294,
+      "loss_partial": 0.045654296875,
+      "step": 3909
+    },
+    {
+      "epoch": 2.0481927710843375,
+      "grad_norm": 1.681930661201477,
+      "learning_rate": 1.1326410479462904e-05,
+      "loss": 0.1039,
+      "step": 3910
+    },
+    {
+      "epoch": 2.0481927710843375,
+      "loss_partial": 0.04541015625,
+      "step": 3910
+    },
+    {
+      "epoch": 2.0481927710843375,
+      "loss_partial": 0.05517578125,
+      "step": 3910
+    },
+    {
+      "epoch": 2.0487166055526456,
+      "grad_norm": 1.6540364027023315,
+      "learning_rate": 1.1321879150705437e-05,
+      "loss": 0.1055,
+      "step": 3911
+    },
+    {
+      "epoch": 2.0487166055526456,
+      "loss_partial": 0.053955078125,
+      "step": 3911
+    },
+    {
+      "epoch": 2.0487166055526456,
+      "loss_partial": 0.05419921875,
+      "step": 3911
+    },
+    {
+      "epoch": 2.049240440020953,
+      "grad_norm": 2.2906405925750732,
+      "learning_rate": 1.1317347545683403e-05,
+      "loss": 0.1073,
+      "step": 3912
+    },
+    {
+      "epoch": 2.049240440020953,
+      "loss_partial": 0.05908203125,
+      "step": 3912
+    },
+    {
+      "epoch": 2.049240440020953,
+      "loss_partial": 0.045654296875,
+      "step": 3912
+    },
+    {
+      "epoch": 2.0497642744892612,
+      "grad_norm": 1.8064889907836914,
+      "learning_rate": 1.131281566534388e-05,
+      "loss": 0.1001,
+      "step": 3913
+    },
+    {
+      "epoch": 2.0497642744892612,
+      "loss_partial": 0.0546875,
+      "step": 3913
+    },
+    {
+      "epoch": 2.0497642744892612,
+      "loss_partial": 0.055908203125,
+      "step": 3913
+    },
+    {
+      "epoch": 2.0502881089575693,
+      "grad_norm": 1.6198713779449463,
+      "learning_rate": 1.1308283510634e-05,
+      "loss": 0.1028,
+      "step": 3914
+    },
+    {
+      "epoch": 2.0502881089575693,
+      "loss_partial": 0.0439453125,
+      "step": 3914
+    },
+    {
+      "epoch": 2.0502881089575693,
+      "loss_partial": 0.03955078125,
+      "step": 3914
+    },
+    {
+      "epoch": 2.0508119434258774,
+      "grad_norm": 1.2663826942443848,
+      "learning_rate": 1.1303751082500956e-05,
+      "loss": 0.1004,
+      "step": 3915
+    },
+    {
+      "epoch": 2.0508119434258774,
+      "loss_partial": 0.04931640625,
+      "step": 3915
+    },
+    {
+      "epoch": 2.0508119434258774,
+      "loss_partial": 0.0498046875,
+      "step": 3915
+    },
+    {
+      "epoch": 2.0513357778941854,
+      "grad_norm": 1.5204627513885498,
+      "learning_rate": 1.1299218381891998e-05,
+      "loss": 0.1043,
+      "step": 3916
+    },
+    {
+      "epoch": 2.0513357778941854,
+      "loss_partial": 0.049072265625,
+      "step": 3916
+    },
+    {
+      "epoch": 2.0513357778941854,
+      "loss_partial": 0.05419921875,
+      "step": 3916
+    },
+    {
+      "epoch": 2.0518596123624935,
+      "grad_norm": 1.2633016109466553,
+      "learning_rate": 1.1294685409754434e-05,
+      "loss": 0.1032,
+      "step": 3917
+    },
+    {
+      "epoch": 2.0518596123624935,
+      "loss_partial": 0.046142578125,
+      "step": 3917
+    },
+    {
+      "epoch": 2.0518596123624935,
+      "loss_partial": 0.055908203125,
+      "step": 3917
+    },
+    {
+      "epoch": 2.0523834468308015,
+      "grad_norm": 1.689273715019226,
+      "learning_rate": 1.1290152167035623e-05,
+      "loss": 0.1014,
+      "step": 3918
+    },
+    {
+      "epoch": 2.0523834468308015,
+      "loss_partial": 0.046142578125,
+      "step": 3918
+    },
+    {
+      "epoch": 2.0523834468308015,
+      "loss_partial": 0.053955078125,
+      "step": 3918
+    },
+    {
+      "epoch": 2.0529072812991096,
+      "grad_norm": 1.2310404777526855,
+      "learning_rate": 1.1285618654682988e-05,
+      "loss": 0.1018,
+      "step": 3919
+    },
+    {
+      "epoch": 2.0529072812991096,
+      "loss_partial": 0.050048828125,
+      "step": 3919
+    },
+    {
+      "epoch": 2.0529072812991096,
+      "loss_partial": 0.05126953125,
+      "step": 3919
+    },
+    {
+      "epoch": 2.0534311157674177,
+      "grad_norm": 1.36723792552948,
+      "learning_rate": 1.1281084873644004e-05,
+      "loss": 0.1013,
+      "step": 3920
+    },
+    {
+      "epoch": 2.0534311157674177,
+      "loss_partial": 0.054931640625,
+      "step": 3920
+    },
+    {
+      "epoch": 2.0534311157674177,
+      "loss_partial": 0.048095703125,
+      "step": 3920
+    },
+    {
+      "epoch": 2.0539549502357257,
+      "grad_norm": 1.5507285594940186,
+      "learning_rate": 1.1276550824866206e-05,
+      "loss": 0.0994,
+      "step": 3921
+    },
+    {
+      "epoch": 2.0539549502357257,
+      "loss_partial": 0.043701171875,
+      "step": 3921
+    },
+    {
+      "epoch": 2.0539549502357257,
+      "loss_partial": 0.048095703125,
+      "step": 3921
+    },
+    {
+      "epoch": 2.0544787847040333,
+      "grad_norm": 1.6534068584442139,
+      "learning_rate": 1.1272016509297176e-05,
+      "loss": 0.1028,
+      "step": 3922
+    },
+    {
+      "epoch": 2.0544787847040333,
+      "loss_partial": 0.056884765625,
+      "step": 3922
+    },
+    {
+      "epoch": 2.0544787847040333,
+      "loss_partial": 0.054931640625,
+      "step": 3922
+    },
+    {
+      "epoch": 2.0550026191723414,
+      "grad_norm": 1.6322187185287476,
+      "learning_rate": 1.1267481927884564e-05,
+      "loss": 0.1014,
+      "step": 3923
+    },
+    {
+      "epoch": 2.0550026191723414,
+      "loss_partial": 0.05615234375,
+      "step": 3923
+    },
+    {
+      "epoch": 2.0550026191723414,
+      "loss_partial": 0.040771484375,
+      "step": 3923
+    },
+    {
+      "epoch": 2.0555264536406495,
+      "grad_norm": 1.4522418975830078,
+      "learning_rate": 1.1262947081576065e-05,
+      "loss": 0.1037,
+      "step": 3924
+    },
+    {
+      "epoch": 2.0555264536406495,
+      "loss_partial": 0.05419921875,
+      "step": 3924
+    },
+    {
+      "epoch": 2.0555264536406495,
+      "loss_partial": 0.0546875,
+      "step": 3924
+    },
+    {
+      "epoch": 2.0560502881089575,
+      "grad_norm": 1.4826910495758057,
+      "learning_rate": 1.1258411971319436e-05,
+      "loss": 0.1031,
+      "step": 3925
+    },
+    {
+      "epoch": 2.0560502881089575,
+      "loss_partial": 0.060302734375,
+      "step": 3925
+    },
+    {
+      "epoch": 2.0560502881089575,
+      "loss_partial": 0.04931640625,
+      "step": 3925
+    },
+    {
+      "epoch": 2.0565741225772656,
+      "grad_norm": 1.7725979089736938,
+      "learning_rate": 1.1253876598062488e-05,
+      "loss": 0.1019,
+      "step": 3926
+    },
+    {
+      "epoch": 2.0565741225772656,
+      "loss_partial": 0.051025390625,
+      "step": 3926
+    },
+    {
+      "epoch": 2.0565741225772656,
+      "loss_partial": 0.046875,
+      "step": 3926
+    },
+    {
+      "epoch": 2.0570979570455736,
+      "grad_norm": 1.2787271738052368,
+      "learning_rate": 1.1249340962753084e-05,
+      "loss": 0.1036,
+      "step": 3927
+    },
+    {
+      "epoch": 2.0570979570455736,
+      "loss_partial": 0.047607421875,
+      "step": 3927
+    },
+    {
+      "epoch": 2.0570979570455736,
+      "loss_partial": 0.054443359375,
+      "step": 3927
+    },
+    {
+      "epoch": 2.0576217915138817,
+      "grad_norm": 1.757386326789856,
+      "learning_rate": 1.1244805066339146e-05,
+      "loss": 0.102,
+      "step": 3928
+    },
+    {
+      "epoch": 2.0576217915138817,
+      "loss_partial": 0.053955078125,
+      "step": 3928
+    },
+    {
+      "epoch": 2.0576217915138817,
+      "loss_partial": 0.052001953125,
+      "step": 3928
+    },
+    {
+      "epoch": 2.0581456259821898,
+      "grad_norm": 1.0655385255813599,
+      "learning_rate": 1.1240268909768646e-05,
+      "loss": 0.1044,
+      "step": 3929
+    },
+    {
+      "epoch": 2.0581456259821898,
+      "loss_partial": 0.048828125,
+      "step": 3929
+    },
+    {
+      "epoch": 2.0581456259821898,
+      "loss_partial": 0.0498046875,
+      "step": 3929
+    },
+    {
+      "epoch": 2.058669460450498,
+      "grad_norm": 1.6353110074996948,
+      "learning_rate": 1.123573249398961e-05,
+      "loss": 0.1031,
+      "step": 3930
+    },
+    {
+      "epoch": 2.058669460450498,
+      "loss_partial": 0.04736328125,
+      "step": 3930
+    },
+    {
+      "epoch": 2.058669460450498,
+      "loss_partial": 0.060791015625,
+      "step": 3930
+    },
+    {
+      "epoch": 2.059193294918806,
+      "grad_norm": 1.365593671798706,
+      "learning_rate": 1.1231195819950129e-05,
+      "loss": 0.1049,
+      "step": 3931
+    },
+    {
+      "epoch": 2.059193294918806,
+      "loss_partial": 0.05615234375,
+      "step": 3931
+    },
+    {
+      "epoch": 2.059193294918806,
+      "loss_partial": 0.064453125,
+      "step": 3931
+    },
+    {
+      "epoch": 2.0597171293871135,
+      "grad_norm": 1.4685893058776855,
+      "learning_rate": 1.1226658888598332e-05,
+      "loss": 0.1039,
+      "step": 3932
+    },
+    {
+      "epoch": 2.0597171293871135,
+      "loss_partial": 0.057373046875,
+      "step": 3932
+    },
+    {
+      "epoch": 2.0597171293871135,
+      "loss_partial": 0.06640625,
+      "step": 3932
+    },
+    {
+      "epoch": 2.0602409638554215,
+      "grad_norm": 1.517593502998352,
+      "learning_rate": 1.1222121700882413e-05,
+      "loss": 0.1024,
+      "step": 3933
+    },
+    {
+      "epoch": 2.0602409638554215,
+      "loss_partial": 0.0439453125,
+      "step": 3933
+    },
+    {
+      "epoch": 2.0602409638554215,
+      "loss_partial": 0.050048828125,
+      "step": 3933
+    },
+    {
+      "epoch": 2.0607647983237296,
+      "grad_norm": 1.803660273551941,
+      "learning_rate": 1.1217584257750619e-05,
+      "loss": 0.1044,
+      "step": 3934
+    },
+    {
+      "epoch": 2.0607647983237296,
+      "loss_partial": 0.06494140625,
+      "step": 3934
+    },
+    {
+      "epoch": 2.0607647983237296,
+      "loss_partial": 0.0517578125,
+      "step": 3934
+    },
+    {
+      "epoch": 2.0612886327920377,
+      "grad_norm": 1.5909326076507568,
+      "learning_rate": 1.1213046560151242e-05,
+      "loss": 0.1037,
+      "step": 3935
+    },
+    {
+      "epoch": 2.0612886327920377,
+      "loss_partial": 0.052001953125,
+      "step": 3935
+    },
+    {
+      "epoch": 2.0612886327920377,
+      "loss_partial": 0.042236328125,
+      "step": 3935
+    },
+    {
+      "epoch": 2.0618124672603457,
+      "grad_norm": 1.228933572769165,
+      "learning_rate": 1.1208508609032635e-05,
+      "loss": 0.1021,
+      "step": 3936
+    },
+    {
+      "epoch": 2.0618124672603457,
+      "loss_partial": 0.043212890625,
+      "step": 3936
+    },
+    {
+      "epoch": 2.0618124672603457,
+      "loss_partial": 0.045166015625,
+      "step": 3936
+    },
+    {
+      "epoch": 2.062336301728654,
+      "grad_norm": 1.3389172554016113,
+      "learning_rate": 1.1203970405343203e-05,
+      "loss": 0.0988,
+      "step": 3937
+    },
+    {
+      "epoch": 2.062336301728654,
+      "loss_partial": 0.053955078125,
+      "step": 3937
+    },
+    {
+      "epoch": 2.062336301728654,
+      "loss_partial": 0.0517578125,
+      "step": 3937
+    },
+    {
+      "epoch": 2.062860136196962,
+      "grad_norm": 1.4358570575714111,
+      "learning_rate": 1.11994319500314e-05,
+      "loss": 0.1063,
+      "step": 3938
+    },
+    {
+      "epoch": 2.062860136196962,
+      "loss_partial": 0.054931640625,
+      "step": 3938
+    },
+    {
+      "epoch": 2.062860136196962,
+      "loss_partial": 0.046142578125,
+      "step": 3938
+    },
+    {
+      "epoch": 2.06338397066527,
+      "grad_norm": 1.332529902458191,
+      "learning_rate": 1.1194893244045736e-05,
+      "loss": 0.1031,
+      "step": 3939
+    },
+    {
+      "epoch": 2.06338397066527,
+      "loss_partial": 0.0517578125,
+      "step": 3939
+    },
+    {
+      "epoch": 2.06338397066527,
+      "loss_partial": 0.0439453125,
+      "step": 3939
+    },
+    {
+      "epoch": 2.063907805133578,
+      "grad_norm": 1.3824410438537598,
+      "learning_rate": 1.1190354288334777e-05,
+      "loss": 0.1016,
+      "step": 3940
+    },
+    {
+      "epoch": 2.063907805133578,
+      "loss_partial": 0.058349609375,
+      "step": 3940
+    },
+    {
+      "epoch": 2.063907805133578,
+      "loss_partial": 0.056396484375,
+      "step": 3940
+    },
+    {
+      "epoch": 2.064431639601886,
+      "grad_norm": 1.6520940065383911,
+      "learning_rate": 1.1185815083847129e-05,
+      "loss": 0.1057,
+      "step": 3941
+    },
+    {
+      "epoch": 2.064431639601886,
+      "loss_partial": 0.0537109375,
+      "step": 3941
+    },
+    {
+      "epoch": 2.064431639601886,
+      "loss_partial": 0.051025390625,
+      "step": 3941
+    },
+    {
+      "epoch": 2.0649554740701936,
+      "grad_norm": 1.339263677597046,
+      "learning_rate": 1.1181275631531461e-05,
+      "loss": 0.1002,
+      "step": 3942
+    },
+    {
+      "epoch": 2.0649554740701936,
+      "loss_partial": 0.06494140625,
+      "step": 3942
+    },
+    {
+      "epoch": 2.0649554740701936,
+      "loss_partial": 0.0419921875,
+      "step": 3942
+    },
+    {
+      "epoch": 2.0654793085385017,
+      "grad_norm": 1.3653743267059326,
+      "learning_rate": 1.1176735932336494e-05,
+      "loss": 0.1006,
+      "step": 3943
+    },
+    {
+      "epoch": 2.0654793085385017,
+      "loss_partial": 0.051513671875,
+      "step": 3943
+    },
+    {
+      "epoch": 2.0654793085385017,
+      "loss_partial": 0.053466796875,
+      "step": 3943
+    },
+    {
+      "epoch": 2.0660031430068098,
+      "grad_norm": 2.088099718093872,
+      "learning_rate": 1.117219598721099e-05,
+      "loss": 0.1007,
+      "step": 3944
+    },
+    {
+      "epoch": 2.0660031430068098,
+      "loss_partial": 0.054931640625,
+      "step": 3944
+    },
+    {
+      "epoch": 2.0660031430068098,
+      "loss_partial": 0.052734375,
+      "step": 3944
+    },
+    {
+      "epoch": 2.066526977475118,
+      "grad_norm": 1.0438205003738403,
+      "learning_rate": 1.1167655797103775e-05,
+      "loss": 0.1068,
+      "step": 3945
+    },
+    {
+      "epoch": 2.066526977475118,
+      "loss_partial": 0.064453125,
+      "step": 3945
+    },
+    {
+      "epoch": 2.066526977475118,
+      "loss_partial": 0.0537109375,
+      "step": 3945
+    },
+    {
+      "epoch": 2.067050811943426,
+      "grad_norm": 1.4325330257415771,
+      "learning_rate": 1.1163115362963719e-05,
+      "loss": 0.1042,
+      "step": 3946
+    },
+    {
+      "epoch": 2.067050811943426,
+      "loss_partial": 0.055908203125,
+      "step": 3946
+    },
+    {
+      "epoch": 2.067050811943426,
+      "loss_partial": 0.05078125,
+      "step": 3946
+    },
+    {
+      "epoch": 2.067574646411734,
+      "grad_norm": 1.6043261289596558,
+      "learning_rate": 1.1158574685739747e-05,
+      "loss": 0.1037,
+      "step": 3947
+    },
+    {
+      "epoch": 2.067574646411734,
+      "loss_partial": 0.05810546875,
+      "step": 3947
+    },
+    {
+      "epoch": 2.067574646411734,
+      "loss_partial": 0.052001953125,
+      "step": 3947
+    },
+    {
+      "epoch": 2.068098480880042,
+      "grad_norm": 1.5853430032730103,
+      "learning_rate": 1.1154033766380827e-05,
+      "loss": 0.1024,
+      "step": 3948
+    },
+    {
+      "epoch": 2.068098480880042,
+      "loss_partial": 0.0595703125,
+      "step": 3948
+    },
+    {
+      "epoch": 2.068098480880042,
+      "loss_partial": 0.054443359375,
+      "step": 3948
+    },
+    {
+      "epoch": 2.06862231534835,
+      "grad_norm": 1.3049002885818481,
+      "learning_rate": 1.1149492605835988e-05,
+      "loss": 0.1018,
+      "step": 3949
+    },
+    {
+      "epoch": 2.06862231534835,
+      "loss_partial": 0.054443359375,
+      "step": 3949
+    },
+    {
+      "epoch": 2.06862231534835,
+      "loss_partial": 0.05419921875,
+      "step": 3949
+    },
+    {
+      "epoch": 2.069146149816658,
+      "grad_norm": 1.777056097984314,
+      "learning_rate": 1.11449512050543e-05,
+      "loss": 0.1066,
+      "step": 3950
+    },
+    {
+      "epoch": 2.069146149816658,
+      "loss_partial": 0.052734375,
+      "step": 3950
+    },
+    {
+      "epoch": 2.069146149816658,
+      "loss_partial": 0.044677734375,
+      "step": 3950
+    },
+    {
+      "epoch": 2.0696699842849657,
+      "grad_norm": 1.5397025346755981,
+      "learning_rate": 1.1140409564984893e-05,
+      "loss": 0.1048,
+      "step": 3951
+    },
+    {
+      "epoch": 2.0696699842849657,
+      "loss_partial": 0.05029296875,
+      "step": 3951
+    },
+    {
+      "epoch": 2.0696699842849657,
+      "loss_partial": 0.0537109375,
+      "step": 3951
+    },
+    {
+      "epoch": 2.070193818753274,
+      "grad_norm": 1.5018149614334106,
+      "learning_rate": 1.1135867686576937e-05,
+      "loss": 0.1047,
+      "step": 3952
+    },
+    {
+      "epoch": 2.070193818753274,
+      "loss_partial": 0.044921875,
+      "step": 3952
+    },
+    {
+      "epoch": 2.070193818753274,
+      "loss_partial": 0.057861328125,
+      "step": 3952
+    },
+    {
+      "epoch": 2.070717653221582,
+      "grad_norm": 1.2419787645339966,
+      "learning_rate": 1.1131325570779663e-05,
+      "loss": 0.1048,
+      "step": 3953
+    },
+    {
+      "epoch": 2.070717653221582,
+      "loss_partial": 0.05126953125,
+      "step": 3953
+    },
+    {
+      "epoch": 2.070717653221582,
+      "loss_partial": 0.055908203125,
+      "step": 3953
+    },
+    {
+      "epoch": 2.07124148768989,
+      "grad_norm": 1.9775426387786865,
+      "learning_rate": 1.1126783218542334e-05,
+      "loss": 0.1022,
+      "step": 3954
+    },
+    {
+      "epoch": 2.07124148768989,
+      "loss_partial": 0.04736328125,
+      "step": 3954
+    },
+    {
+      "epoch": 2.07124148768989,
+      "loss_partial": 0.046875,
+      "step": 3954
+    },
+    {
+      "epoch": 2.071765322158198,
+      "grad_norm": 1.739707112312317,
+      "learning_rate": 1.1122240630814285e-05,
+      "loss": 0.1032,
+      "step": 3955
+    },
+    {
+      "epoch": 2.071765322158198,
+      "loss_partial": 0.049560546875,
+      "step": 3955
+    },
+    {
+      "epoch": 2.071765322158198,
+      "loss_partial": 0.04638671875,
+      "step": 3955
+    },
+    {
+      "epoch": 2.072289156626506,
+      "grad_norm": 1.8995299339294434,
+      "learning_rate": 1.111769780854488e-05,
+      "loss": 0.103,
+      "step": 3956
+    },
+    {
+      "epoch": 2.072289156626506,
+      "loss_partial": 0.050537109375,
+      "step": 3956
+    },
+    {
+      "epoch": 2.072289156626506,
+      "loss_partial": 0.05615234375,
+      "step": 3956
+    },
+    {
+      "epoch": 2.072812991094814,
+      "grad_norm": 2.3786065578460693,
+      "learning_rate": 1.1113154752683548e-05,
+      "loss": 0.1057,
+      "step": 3957
+    },
+    {
+      "epoch": 2.072812991094814,
+      "loss_partial": 0.05517578125,
+      "step": 3957
+    },
+    {
+      "epoch": 2.072812991094814,
+      "loss_partial": 0.052734375,
+      "step": 3957
+    },
+    {
+      "epoch": 2.073336825563122,
+      "grad_norm": 1.8648487329483032,
+      "learning_rate": 1.1108611464179755e-05,
+      "loss": 0.1058,
+      "step": 3958
+    },
+    {
+      "epoch": 2.073336825563122,
+      "loss_partial": 0.049560546875,
+      "step": 3958
+    },
+    {
+      "epoch": 2.073336825563122,
+      "loss_partial": 0.042724609375,
+      "step": 3958
+    },
+    {
+      "epoch": 2.07386066003143,
+      "grad_norm": 1.670650601387024,
+      "learning_rate": 1.1104067943983023e-05,
+      "loss": 0.1031,
+      "step": 3959
+    },
+    {
+      "epoch": 2.07386066003143,
+      "loss_partial": 0.057373046875,
+      "step": 3959
+    },
+    {
+      "epoch": 2.07386066003143,
+      "loss_partial": 0.06298828125,
+      "step": 3959
+    },
+    {
+      "epoch": 2.0743844944997383,
+      "grad_norm": 1.927488088607788,
+      "learning_rate": 1.1099524193042915e-05,
+      "loss": 0.1021,
+      "step": 3960
+    },
+    {
+      "epoch": 2.0743844944997383,
+      "loss_partial": 0.05615234375,
+      "step": 3960
+    },
+    {
+      "epoch": 2.0743844944997383,
+      "loss_partial": 0.057861328125,
+      "step": 3960
+    },
+    {
+      "epoch": 2.074908328968046,
+      "grad_norm": 1.2698055505752563,
+      "learning_rate": 1.1094980212309052e-05,
+      "loss": 0.1033,
+      "step": 3961
+    },
+    {
+      "epoch": 2.074908328968046,
+      "loss_partial": 0.052490234375,
+      "step": 3961
+    },
+    {
+      "epoch": 2.074908328968046,
+      "loss_partial": 0.049560546875,
+      "step": 3961
+    },
+    {
+      "epoch": 2.075432163436354,
+      "grad_norm": 1.7752461433410645,
+      "learning_rate": 1.1090436002731093e-05,
+      "loss": 0.103,
+      "step": 3962
+    },
+    {
+      "epoch": 2.075432163436354,
+      "loss_partial": 0.04638671875,
+      "step": 3962
+    },
+    {
+      "epoch": 2.075432163436354,
+      "loss_partial": 0.06591796875,
+      "step": 3962
+    },
+    {
+      "epoch": 2.075955997904662,
+      "grad_norm": 1.50130033493042,
+      "learning_rate": 1.1085891565258757e-05,
+      "loss": 0.106,
+      "step": 3963
+    },
+    {
+      "epoch": 2.075955997904662,
+      "loss_partial": 0.06689453125,
+      "step": 3963
+    },
+    {
+      "epoch": 2.075955997904662,
+      "loss_partial": 0.052001953125,
+      "step": 3963
+    },
+    {
+      "epoch": 2.07647983237297,
+      "grad_norm": 1.3119391202926636,
+      "learning_rate": 1.1081346900841794e-05,
+      "loss": 0.1057,
+      "step": 3964
+    },
+    {
+      "epoch": 2.07647983237297,
+      "loss_partial": 0.050537109375,
+      "step": 3964
+    },
+    {
+      "epoch": 2.07647983237297,
+      "loss_partial": 0.05908203125,
+      "step": 3964
+    },
+    {
+      "epoch": 2.077003666841278,
+      "grad_norm": 1.6476655006408691,
+      "learning_rate": 1.107680201043002e-05,
+      "loss": 0.1016,
+      "step": 3965
+    },
+    {
+      "epoch": 2.077003666841278,
+      "loss_partial": 0.058349609375,
+      "step": 3965
+    },
+    {
+      "epoch": 2.077003666841278,
+      "loss_partial": 0.0556640625,
+      "step": 3965
+    },
+    {
+      "epoch": 2.077527501309586,
+      "grad_norm": 1.328023910522461,
+      "learning_rate": 1.1072256894973283e-05,
+      "loss": 0.104,
+      "step": 3966
+    },
+    {
+      "epoch": 2.077527501309586,
+      "loss_partial": 0.046875,
+      "step": 3966
+    },
+    {
+      "epoch": 2.077527501309586,
+      "loss_partial": 0.050048828125,
+      "step": 3966
+    },
+    {
+      "epoch": 2.0780513357778942,
+      "grad_norm": 1.4655464887619019,
+      "learning_rate": 1.1067711555421485e-05,
+      "loss": 0.1052,
+      "step": 3967
+    },
+    {
+      "epoch": 2.0780513357778942,
+      "loss_partial": 0.0546875,
+      "step": 3967
+    },
+    {
+      "epoch": 2.0780513357778942,
+      "loss_partial": 0.045654296875,
+      "step": 3967
+    },
+    {
+      "epoch": 2.0785751702462023,
+      "grad_norm": 1.148821473121643,
+      "learning_rate": 1.1063165992724575e-05,
+      "loss": 0.1034,
+      "step": 3968
+    },
+    {
+      "epoch": 2.0785751702462023,
+      "loss_partial": 0.050048828125,
+      "step": 3968
+    },
+    {
+      "epoch": 2.0785751702462023,
+      "loss_partial": 0.046630859375,
+      "step": 3968
+    },
+    {
+      "epoch": 2.0790990047145104,
+      "grad_norm": 1.879091501235962,
+      "learning_rate": 1.1058620207832547e-05,
+      "loss": 0.1018,
+      "step": 3969
+    },
+    {
+      "epoch": 2.0790990047145104,
+      "loss_partial": 0.0537109375,
+      "step": 3969
+    },
+    {
+      "epoch": 2.0790990047145104,
+      "loss_partial": 0.047607421875,
+      "step": 3969
+    },
+    {
+      "epoch": 2.0796228391828184,
+      "grad_norm": 1.7088193893432617,
+      "learning_rate": 1.105407420169544e-05,
+      "loss": 0.1008,
+      "step": 3970
+    },
+    {
+      "epoch": 2.0796228391828184,
+      "loss_partial": 0.04736328125,
+      "step": 3970
+    },
+    {
+      "epoch": 2.0796228391828184,
+      "loss_partial": 0.04931640625,
+      "step": 3970
+    },
+    {
+      "epoch": 2.080146673651126,
+      "grad_norm": 1.4290357828140259,
+      "learning_rate": 1.1049527975263348e-05,
+      "loss": 0.101,
+      "step": 3971
+    },
+    {
+      "epoch": 2.080146673651126,
+      "loss_partial": 0.051025390625,
+      "step": 3971
+    },
+    {
+      "epoch": 2.080146673651126,
+      "loss_partial": 0.052734375,
+      "step": 3971
+    },
+    {
+      "epoch": 2.080670508119434,
+      "grad_norm": 2.0424132347106934,
+      "learning_rate": 1.1044981529486395e-05,
+      "loss": 0.1016,
+      "step": 3972
+    },
+    {
+      "epoch": 2.080670508119434,
+      "loss_partial": 0.044677734375,
+      "step": 3972
+    },
+    {
+      "epoch": 2.080670508119434,
+      "loss_partial": 0.05859375,
+      "step": 3972
+    },
+    {
+      "epoch": 2.081194342587742,
+      "grad_norm": 1.6907830238342285,
+      "learning_rate": 1.1040434865314764e-05,
+      "loss": 0.1017,
+      "step": 3973
+    },
+    {
+      "epoch": 2.081194342587742,
+      "loss_partial": 0.05810546875,
+      "step": 3973
+    },
+    {
+      "epoch": 2.081194342587742,
+      "loss_partial": 0.04541015625,
+      "step": 3973
+    },
+    {
+      "epoch": 2.0817181770560502,
+      "grad_norm": 1.7788095474243164,
+      "learning_rate": 1.1035887983698683e-05,
+      "loss": 0.1002,
+      "step": 3974
+    },
+    {
+      "epoch": 2.0817181770560502,
+      "loss_partial": 0.04296875,
+      "step": 3974
+    },
+    {
+      "epoch": 2.0817181770560502,
+      "loss_partial": 0.051025390625,
+      "step": 3974
+    },
+    {
+      "epoch": 2.0822420115243583,
+      "grad_norm": 2.0631322860717773,
+      "learning_rate": 1.1031340885588415e-05,
+      "loss": 0.1005,
+      "step": 3975
+    },
+    {
+      "epoch": 2.0822420115243583,
+      "loss_partial": 0.0537109375,
+      "step": 3975
+    },
+    {
+      "epoch": 2.0822420115243583,
+      "loss_partial": 0.0615234375,
+      "step": 3975
+    },
+    {
+      "epoch": 2.0827658459926663,
+      "grad_norm": 1.660630226135254,
+      "learning_rate": 1.1026793571934283e-05,
+      "loss": 0.1043,
+      "step": 3976
+    },
+    {
+      "epoch": 2.0827658459926663,
+      "loss_partial": 0.0458984375,
+      "step": 3976
+    },
+    {
+      "epoch": 2.0827658459926663,
+      "loss_partial": 0.0546875,
+      "step": 3976
+    },
+    {
+      "epoch": 2.0832896804609744,
+      "grad_norm": 1.8546090126037598,
+      "learning_rate": 1.1022246043686644e-05,
+      "loss": 0.1015,
+      "step": 3977
+    },
+    {
+      "epoch": 2.0832896804609744,
+      "loss_partial": 0.048828125,
+      "step": 3977
+    },
+    {
+      "epoch": 2.0832896804609744,
+      "loss_partial": 0.04736328125,
+      "step": 3977
+    },
+    {
+      "epoch": 2.0838135149292825,
+      "grad_norm": 1.435342788696289,
+      "learning_rate": 1.1017698301795902e-05,
+      "loss": 0.1005,
+      "step": 3978
+    },
+    {
+      "epoch": 2.0838135149292825,
+      "loss_partial": 0.05419921875,
+      "step": 3978
+    },
+    {
+      "epoch": 2.0838135149292825,
+      "loss_partial": 0.05322265625,
+      "step": 3978
+    },
+    {
+      "epoch": 2.0843373493975905,
+      "grad_norm": 1.4227112531661987,
+      "learning_rate": 1.1013150347212506e-05,
+      "loss": 0.1026,
+      "step": 3979
+    },
+    {
+      "epoch": 2.0843373493975905,
+      "loss_partial": 0.0458984375,
+      "step": 3979
+    },
+    {
+      "epoch": 2.0843373493975905,
+      "loss_partial": 0.049072265625,
+      "step": 3979
+    },
+    {
+      "epoch": 2.0848611838658986,
+      "grad_norm": 1.6649980545043945,
+      "learning_rate": 1.1008602180886956e-05,
+      "loss": 0.0988,
+      "step": 3980
+    },
+    {
+      "epoch": 2.0848611838658986,
+      "loss_partial": 0.05322265625,
+      "step": 3980
+    },
+    {
+      "epoch": 2.0848611838658986,
+      "loss_partial": 0.06298828125,
+      "step": 3980
+    },
+    {
+      "epoch": 2.085385018334206,
+      "grad_norm": 1.6419166326522827,
+      "learning_rate": 1.1004053803769786e-05,
+      "loss": 0.1038,
+      "step": 3981
+    },
+    {
+      "epoch": 2.085385018334206,
+      "loss_partial": 0.0556640625,
+      "step": 3981
+    },
+    {
+      "epoch": 2.085385018334206,
+      "loss_partial": 0.048828125,
+      "step": 3981
+    },
+    {
+      "epoch": 2.0859088528025143,
+      "grad_norm": 2.339573860168457,
+      "learning_rate": 1.0999505216811578e-05,
+      "loss": 0.1002,
+      "step": 3982
+    },
+    {
+      "epoch": 2.0859088528025143,
+      "loss_partial": 0.043212890625,
+      "step": 3982
+    },
+    {
+      "epoch": 2.0859088528025143,
+      "loss_partial": 0.052001953125,
+      "step": 3982
+    },
+    {
+      "epoch": 2.0864326872708223,
+      "grad_norm": 2.0766680240631104,
+      "learning_rate": 1.0994956420962963e-05,
+      "loss": 0.1008,
+      "step": 3983
+    },
+    {
+      "epoch": 2.0864326872708223,
+      "loss_partial": 0.047607421875,
+      "step": 3983
+    },
+    {
+      "epoch": 2.0864326872708223,
+      "loss_partial": 0.044189453125,
+      "step": 3983
+    },
+    {
+      "epoch": 2.0869565217391304,
+      "grad_norm": 1.8107341527938843,
+      "learning_rate": 1.099040741717461e-05,
+      "loss": 0.1025,
+      "step": 3984
+    },
+    {
+      "epoch": 2.0869565217391304,
+      "loss_partial": 0.0615234375,
+      "step": 3984
+    },
+    {
+      "epoch": 2.0869565217391304,
+      "loss_partial": 0.045654296875,
+      "step": 3984
+    },
+    {
+      "epoch": 2.0874803562074384,
+      "grad_norm": 1.7081477642059326,
+      "learning_rate": 1.098585820639723e-05,
+      "loss": 0.1058,
+      "step": 3985
+    },
+    {
+      "epoch": 2.0874803562074384,
+      "loss_partial": 0.058837890625,
+      "step": 3985
+    },
+    {
+      "epoch": 2.0874803562074384,
+      "loss_partial": 0.046630859375,
+      "step": 3985
+    },
+    {
+      "epoch": 2.0880041906757465,
+      "grad_norm": 1.5865297317504883,
+      "learning_rate": 1.098130878958158e-05,
+      "loss": 0.1017,
+      "step": 3986
+    },
+    {
+      "epoch": 2.0880041906757465,
+      "loss_partial": 0.045166015625,
+      "step": 3986
+    },
+    {
+      "epoch": 2.0880041906757465,
+      "loss_partial": 0.04345703125,
+      "step": 3986
+    },
+    {
+      "epoch": 2.0885280251440546,
+      "grad_norm": 1.513489842414856,
+      "learning_rate": 1.0976759167678458e-05,
+      "loss": 0.1021,
+      "step": 3987
+    },
+    {
+      "epoch": 2.0885280251440546,
+      "loss_partial": 0.039306640625,
+      "step": 3987
+    },
+    {
+      "epoch": 2.0885280251440546,
+      "loss_partial": 0.045654296875,
+      "step": 3987
+    },
+    {
+      "epoch": 2.0890518596123626,
+      "grad_norm": 1.7077140808105469,
+      "learning_rate": 1.0972209341638716e-05,
+      "loss": 0.1029,
+      "step": 3988
+    },
+    {
+      "epoch": 2.0890518596123626,
+      "loss_partial": 0.050048828125,
+      "step": 3988
+    },
+    {
+      "epoch": 2.0890518596123626,
+      "loss_partial": 0.054931640625,
+      "step": 3988
+    },
+    {
+      "epoch": 2.0895756940806707,
+      "grad_norm": 1.6290191411972046,
+      "learning_rate": 1.096765931241323e-05,
+      "loss": 0.1031,
+      "step": 3989
+    },
+    {
+      "epoch": 2.0895756940806707,
+      "loss_partial": 0.043701171875,
+      "step": 3989
+    },
+    {
+      "epoch": 2.0895756940806707,
+      "loss_partial": 0.05126953125,
+      "step": 3989
+    },
+    {
+      "epoch": 2.0900995285489787,
+      "grad_norm": 1.9190478324890137,
+      "learning_rate": 1.0963109080952936e-05,
+      "loss": 0.1039,
+      "step": 3990
+    },
+    {
+      "epoch": 2.0900995285489787,
+      "loss_partial": 0.058837890625,
+      "step": 3990
+    },
+    {
+      "epoch": 2.0900995285489787,
+      "loss_partial": 0.046630859375,
+      "step": 3990
+    },
+    {
+      "epoch": 2.0906233630172864,
+      "grad_norm": 1.7715144157409668,
+      "learning_rate": 1.0958558648208794e-05,
+      "loss": 0.1049,
+      "step": 3991
+    },
+    {
+      "epoch": 2.0906233630172864,
+      "loss_partial": 0.052734375,
+      "step": 3991
+    },
+    {
+      "epoch": 2.0906233630172864,
+      "loss_partial": 0.051513671875,
+      "step": 3991
+    },
+    {
+      "epoch": 2.0911471974855944,
+      "grad_norm": 2.218928337097168,
+      "learning_rate": 1.0954008015131828e-05,
+      "loss": 0.1031,
+      "step": 3992
+    },
+    {
+      "epoch": 2.0911471974855944,
+      "loss_partial": 0.048095703125,
+      "step": 3992
+    },
+    {
+      "epoch": 2.0911471974855944,
+      "loss_partial": 0.042236328125,
+      "step": 3992
+    },
+    {
+      "epoch": 2.0916710319539025,
+      "grad_norm": 1.903926134109497,
+      "learning_rate": 1.094945718267308e-05,
+      "loss": 0.1012,
+      "step": 3993
+    },
+    {
+      "epoch": 2.0916710319539025,
+      "loss_partial": 0.058349609375,
+      "step": 3993
+    },
+    {
+      "epoch": 2.0916710319539025,
+      "loss_partial": 0.051025390625,
+      "step": 3993
+    },
+    {
+      "epoch": 2.0921948664222105,
+      "grad_norm": 2.0047760009765625,
+      "learning_rate": 1.0944906151783656e-05,
+      "loss": 0.1065,
+      "step": 3994
+    },
+    {
+      "epoch": 2.0921948664222105,
+      "loss_partial": 0.052978515625,
+      "step": 3994
+    },
+    {
+      "epoch": 2.0921948664222105,
+      "loss_partial": 0.051025390625,
+      "step": 3994
+    },
+    {
+      "epoch": 2.0927187008905186,
+      "grad_norm": 1.1942135095596313,
+      "learning_rate": 1.0940354923414687e-05,
+      "loss": 0.1034,
+      "step": 3995
+    },
+    {
+      "epoch": 2.0927187008905186,
+      "loss_partial": 0.05126953125,
+      "step": 3995
+    },
+    {
+      "epoch": 2.0927187008905186,
+      "loss_partial": 0.052490234375,
+      "step": 3995
+    },
+    {
+      "epoch": 2.0932425353588267,
+      "grad_norm": 2.068686008453369,
+      "learning_rate": 1.093580349851736e-05,
+      "loss": 0.1062,
+      "step": 3996
+    },
+    {
+      "epoch": 2.0932425353588267,
+      "loss_partial": 0.044677734375,
+      "step": 3996
+    },
+    {
+      "epoch": 2.0932425353588267,
+      "loss_partial": 0.056640625,
+      "step": 3996
+    },
+    {
+      "epoch": 2.0937663698271347,
+      "grad_norm": 1.5819171667099,
+      "learning_rate": 1.0931251878042882e-05,
+      "loss": 0.1076,
+      "step": 3997
+    },
+    {
+      "epoch": 2.0937663698271347,
+      "loss_partial": 0.0595703125,
+      "step": 3997
+    },
+    {
+      "epoch": 2.0937663698271347,
+      "loss_partial": 0.051513671875,
+      "step": 3997
+    },
+    {
+      "epoch": 2.0942902042954428,
+      "grad_norm": 1.629467248916626,
+      "learning_rate": 1.0926700062942522e-05,
+      "loss": 0.1035,
+      "step": 3998
+    },
+    {
+      "epoch": 2.0942902042954428,
+      "loss_partial": 0.053955078125,
+      "step": 3998
+    },
+    {
+      "epoch": 2.0942902042954428,
+      "loss_partial": 0.052001953125,
+      "step": 3998
+    },
+    {
+      "epoch": 2.094814038763751,
+      "grad_norm": 1.5901076793670654,
+      "learning_rate": 1.0922148054167579e-05,
+      "loss": 0.1012,
+      "step": 3999
+    },
+    {
+      "epoch": 2.094814038763751,
+      "loss_partial": 0.054931640625,
+      "step": 3999
+    },
+    {
+      "epoch": 2.094814038763751,
+      "loss_partial": 0.05029296875,
+      "step": 3999
+    },
+    {
+      "epoch": 2.095337873232059,
+      "grad_norm": 1.4114431142807007,
+      "learning_rate": 1.0917595852669396e-05,
+      "loss": 0.1057,
+      "step": 4000
+    },
+    {
+      "epoch": 2.095337873232059,
+      "loss_partial": 0.052978515625,
+      "step": 4000
+    },
+    {
+      "epoch": 2.095337873232059,
+      "loss_partial": 0.056884765625,
+      "step": 4000
+    },
+    {
+      "epoch": 2.0958617077003665,
+      "grad_norm": 1.3747546672821045,
+      "learning_rate": 1.091304345939935e-05,
+      "loss": 0.1014,
+      "step": 4001
+    },
+    {
+      "epoch": 2.0958617077003665,
+      "loss_partial": 0.055908203125,
+      "step": 4001
+    },
+    {
+      "epoch": 2.0958617077003665,
+      "loss_partial": 0.05029296875,
+      "step": 4001
+    },
+    {
+      "epoch": 2.0963855421686746,
+      "grad_norm": 1.2592564821243286,
+      "learning_rate": 1.0908490875308876e-05,
+      "loss": 0.1006,
+      "step": 4002
+    },
+    {
+      "epoch": 2.0963855421686746,
+      "loss_partial": 0.04443359375,
+      "step": 4002
+    },
+    {
+      "epoch": 2.0963855421686746,
+      "loss_partial": 0.0556640625,
+      "step": 4002
+    },
+    {
+      "epoch": 2.0969093766369826,
+      "grad_norm": 1.608677864074707,
+      "learning_rate": 1.0903938101349418e-05,
+      "loss": 0.0998,
+      "step": 4003
+    },
+    {
+      "epoch": 2.0969093766369826,
+      "loss_partial": 0.06494140625,
+      "step": 4003
+    },
+    {
+      "epoch": 2.0969093766369826,
+      "loss_partial": 0.057861328125,
+      "step": 4003
+    },
+    {
+      "epoch": 2.0974332111052907,
+      "grad_norm": 1.3664661645889282,
+      "learning_rate": 1.0899385138472494e-05,
+      "loss": 0.1033,
+      "step": 4004
+    },
+    {
+      "epoch": 2.0974332111052907,
+      "loss_partial": 0.059814453125,
+      "step": 4004
+    },
+    {
+      "epoch": 2.0974332111052907,
+      "loss_partial": 0.042724609375,
+      "step": 4004
+    },
+    {
+      "epoch": 2.0979570455735987,
+      "grad_norm": 1.8794127702713013,
+      "learning_rate": 1.089483198762963e-05,
+      "loss": 0.1009,
+      "step": 4005
+    },
+    {
+      "epoch": 2.0979570455735987,
+      "loss_partial": 0.04638671875,
+      "step": 4005
+    },
+    {
+      "epoch": 2.0979570455735987,
+      "loss_partial": 0.0673828125,
+      "step": 4005
+    },
+    {
+      "epoch": 2.098480880041907,
+      "grad_norm": 1.5713146924972534,
+      "learning_rate": 1.0890278649772421e-05,
+      "loss": 0.1047,
+      "step": 4006
+    },
+    {
+      "epoch": 2.098480880041907,
+      "loss_partial": 0.047119140625,
+      "step": 4006
+    },
+    {
+      "epoch": 2.098480880041907,
+      "loss_partial": 0.052734375,
+      "step": 4006
+    },
+    {
+      "epoch": 2.099004714510215,
+      "grad_norm": 1.702012538909912,
+      "learning_rate": 1.0885725125852474e-05,
+      "loss": 0.1017,
+      "step": 4007
+    },
+    {
+      "epoch": 2.099004714510215,
+      "loss_partial": 0.052490234375,
+      "step": 4007
+    },
+    {
+      "epoch": 2.099004714510215,
+      "loss_partial": 0.05029296875,
+      "step": 4007
+    },
+    {
+      "epoch": 2.099528548978523,
+      "grad_norm": 1.896368145942688,
+      "learning_rate": 1.0881171416821458e-05,
+      "loss": 0.1031,
+      "step": 4008
+    },
+    {
+      "epoch": 2.099528548978523,
+      "loss_partial": 0.04931640625,
+      "step": 4008
+    },
+    {
+      "epoch": 2.099528548978523,
+      "loss_partial": 0.04931640625,
+      "step": 4008
+    },
+    {
+      "epoch": 2.100052383446831,
+      "grad_norm": 1.6507755517959595,
+      "learning_rate": 1.0876617523631062e-05,
+      "loss": 0.0993,
+      "step": 4009
+    },
+    {
+      "epoch": 2.100052383446831,
+      "loss_partial": 0.04931640625,
+      "step": 4009
+    },
+    {
+      "epoch": 2.100052383446831,
+      "loss_partial": 0.0537109375,
+      "step": 4009
+    },
+    {
+      "epoch": 2.1005762179151386,
+      "grad_norm": 1.3731812238693237,
+      "learning_rate": 1.0872063447233026e-05,
+      "loss": 0.1035,
+      "step": 4010
+    },
+    {
+      "epoch": 2.1005762179151386,
+      "loss_partial": 0.0537109375,
+      "step": 4010
+    },
+    {
+      "epoch": 2.1005762179151386,
+      "loss_partial": 0.056640625,
+      "step": 4010
+    },
+    {
+      "epoch": 2.1011000523834467,
+      "grad_norm": 1.8273677825927734,
+      "learning_rate": 1.086750918857912e-05,
+      "loss": 0.103,
+      "step": 4011
+    },
+    {
+      "epoch": 2.1011000523834467,
+      "loss_partial": 0.05029296875,
+      "step": 4011
+    },
+    {
+      "epoch": 2.1011000523834467,
+      "loss_partial": 0.04931640625,
+      "step": 4011
+    },
+    {
+      "epoch": 2.1016238868517547,
+      "grad_norm": 1.5618139505386353,
+      "learning_rate": 1.086295474862116e-05,
+      "loss": 0.1004,
+      "step": 4012
+    },
+    {
+      "epoch": 2.1016238868517547,
+      "loss_partial": 0.042724609375,
+      "step": 4012
+    },
+    {
+      "epoch": 2.1016238868517547,
+      "loss_partial": 0.053466796875,
+      "step": 4012
+    },
+    {
+      "epoch": 2.102147721320063,
+      "grad_norm": 1.371507167816162,
+      "learning_rate": 1.0858400128310991e-05,
+      "loss": 0.1053,
+      "step": 4013
+    },
+    {
+      "epoch": 2.102147721320063,
+      "loss_partial": 0.05810546875,
+      "step": 4013
+    },
+    {
+      "epoch": 2.102147721320063,
+      "loss_partial": 0.052001953125,
+      "step": 4013
+    },
+    {
+      "epoch": 2.102671555788371,
+      "grad_norm": 1.6732739210128784,
+      "learning_rate": 1.0853845328600505e-05,
+      "loss": 0.1038,
+      "step": 4014
+    },
+    {
+      "epoch": 2.102671555788371,
+      "loss_partial": 0.049072265625,
+      "step": 4014
+    },
+    {
+      "epoch": 2.102671555788371,
+      "loss_partial": 0.051025390625,
+      "step": 4014
+    },
+    {
+      "epoch": 2.103195390256679,
+      "grad_norm": 1.4454927444458008,
+      "learning_rate": 1.0849290350441626e-05,
+      "loss": 0.1015,
+      "step": 4015
+    },
+    {
+      "epoch": 2.103195390256679,
+      "loss_partial": 0.0478515625,
+      "step": 4015
+    },
+    {
+      "epoch": 2.103195390256679,
+      "loss_partial": 0.060302734375,
+      "step": 4015
+    },
+    {
+      "epoch": 2.103719224724987,
+      "grad_norm": 1.5194976329803467,
+      "learning_rate": 1.0844735194786318e-05,
+      "loss": 0.1005,
+      "step": 4016
+    },
+    {
+      "epoch": 2.103719224724987,
+      "loss_partial": 0.0478515625,
+      "step": 4016
+    },
+    {
+      "epoch": 2.103719224724987,
+      "loss_partial": 0.047607421875,
+      "step": 4016
+    },
+    {
+      "epoch": 2.104243059193295,
+      "grad_norm": 1.6282069683074951,
+      "learning_rate": 1.0840179862586572e-05,
+      "loss": 0.1017,
+      "step": 4017
+    },
+    {
+      "epoch": 2.104243059193295,
+      "loss_partial": 0.04833984375,
+      "step": 4017
+    },
+    {
+      "epoch": 2.104243059193295,
+      "loss_partial": 0.06396484375,
+      "step": 4017
+    },
+    {
+      "epoch": 2.104766893661603,
+      "grad_norm": 1.6595613956451416,
+      "learning_rate": 1.0835624354794433e-05,
+      "loss": 0.1039,
+      "step": 4018
+    },
+    {
+      "epoch": 2.104766893661603,
+      "loss_partial": 0.0517578125,
+      "step": 4018
+    },
+    {
+      "epoch": 2.104766893661603,
+      "loss_partial": 0.044677734375,
+      "step": 4018
+    },
+    {
+      "epoch": 2.105290728129911,
+      "grad_norm": 2.9728281497955322,
+      "learning_rate": 1.0831068672361973e-05,
+      "loss": 0.1014,
+      "step": 4019
+    },
+    {
+      "epoch": 2.105290728129911,
+      "loss_partial": 0.052734375,
+      "step": 4019
+    },
+    {
+      "epoch": 2.105290728129911,
+      "loss_partial": 0.055908203125,
+      "step": 4019
+    },
+    {
+      "epoch": 2.1058145625982188,
+      "grad_norm": 1.6374229192733765,
+      "learning_rate": 1.0826512816241299e-05,
+      "loss": 0.1028,
+      "step": 4020
+    },
+    {
+      "epoch": 2.1058145625982188,
+      "loss_partial": 0.049072265625,
+      "step": 4020
+    },
+    {
+      "epoch": 2.1058145625982188,
+      "loss_partial": 0.056884765625,
+      "step": 4020
+    },
+    {
+      "epoch": 2.106338397066527,
+      "grad_norm": 2.270627975463867,
+      "learning_rate": 1.0821956787384557e-05,
+      "loss": 0.1011,
+      "step": 4021
+    },
+    {
+      "epoch": 2.106338397066527,
+      "loss_partial": 0.0458984375,
+      "step": 4021
+    },
+    {
+      "epoch": 2.106338397066527,
+      "loss_partial": 0.0556640625,
+      "step": 4021
+    },
+    {
+      "epoch": 2.106862231534835,
+      "grad_norm": 1.6833373308181763,
+      "learning_rate": 1.081740058674393e-05,
+      "loss": 0.1059,
+      "step": 4022
+    },
+    {
+      "epoch": 2.106862231534835,
+      "loss_partial": 0.048095703125,
+      "step": 4022
+    },
+    {
+      "epoch": 2.106862231534835,
+      "loss_partial": 0.05517578125,
+      "step": 4022
+    },
+    {
+      "epoch": 2.107386066003143,
+      "grad_norm": 1.610196590423584,
+      "learning_rate": 1.0812844215271632e-05,
+      "loss": 0.1039,
+      "step": 4023
+    },
+    {
+      "epoch": 2.107386066003143,
+      "loss_partial": 0.047119140625,
+      "step": 4023
+    },
+    {
+      "epoch": 2.107386066003143,
+      "loss_partial": 0.051513671875,
+      "step": 4023
+    },
+    {
+      "epoch": 2.107909900471451,
+      "grad_norm": 1.6112689971923828,
+      "learning_rate": 1.0808287673919922e-05,
+      "loss": 0.1023,
+      "step": 4024
+    },
+    {
+      "epoch": 2.107909900471451,
+      "loss_partial": 0.06298828125,
+      "step": 4024
+    },
+    {
+      "epoch": 2.107909900471451,
+      "loss_partial": 0.0478515625,
+      "step": 4024
+    },
+    {
+      "epoch": 2.108433734939759,
+      "grad_norm": 2.0769619941711426,
+      "learning_rate": 1.0803730963641086e-05,
+      "loss": 0.1032,
+      "step": 4025
+    },
+    {
+      "epoch": 2.108433734939759,
+      "loss_partial": 0.045654296875,
+      "step": 4025
+    },
+    {
+      "epoch": 2.108433734939759,
+      "loss_partial": 0.048828125,
+      "step": 4025
+    },
+    {
+      "epoch": 2.108957569408067,
+      "grad_norm": 1.928884744644165,
+      "learning_rate": 1.0799174085387449e-05,
+      "loss": 0.1019,
+      "step": 4026
+    },
+    {
+      "epoch": 2.108957569408067,
+      "loss_partial": 0.04833984375,
+      "step": 4026
+    },
+    {
+      "epoch": 2.108957569408067,
+      "loss_partial": 0.04296875,
+      "step": 4026
+    },
+    {
+      "epoch": 2.109481403876375,
+      "grad_norm": 1.5949244499206543,
+      "learning_rate": 1.079461704011137e-05,
+      "loss": 0.1045,
+      "step": 4027
+    },
+    {
+      "epoch": 2.109481403876375,
+      "loss_partial": 0.050537109375,
+      "step": 4027
+    },
+    {
+      "epoch": 2.109481403876375,
+      "loss_partial": 0.0556640625,
+      "step": 4027
+    },
+    {
+      "epoch": 2.1100052383446832,
+      "grad_norm": 2.418637275695801,
+      "learning_rate": 1.0790059828765245e-05,
+      "loss": 0.1022,
+      "step": 4028
+    },
+    {
+      "epoch": 2.1100052383446832,
+      "loss_partial": 0.05078125,
+      "step": 4028
+    },
+    {
+      "epoch": 2.1100052383446832,
+      "loss_partial": 0.046875,
+      "step": 4028
+    },
+    {
+      "epoch": 2.1105290728129913,
+      "grad_norm": 1.3223371505737305,
+      "learning_rate": 1.0785502452301497e-05,
+      "loss": 0.1021,
+      "step": 4029
+    },
+    {
+      "epoch": 2.1105290728129913,
+      "loss_partial": 0.051025390625,
+      "step": 4029
+    },
+    {
+      "epoch": 2.1105290728129913,
+      "loss_partial": 0.051513671875,
+      "step": 4029
+    },
+    {
+      "epoch": 2.111052907281299,
+      "grad_norm": 2.2350993156433105,
+      "learning_rate": 1.0780944911672597e-05,
+      "loss": 0.1054,
+      "step": 4030
+    },
+    {
+      "epoch": 2.111052907281299,
+      "loss_partial": 0.059326171875,
+      "step": 4030
+    },
+    {
+      "epoch": 2.111052907281299,
+      "loss_partial": 0.0458984375,
+      "step": 4030
+    },
+    {
+      "epoch": 2.111576741749607,
+      "grad_norm": 1.9570249319076538,
+      "learning_rate": 1.077638720783104e-05,
+      "loss": 0.1029,
+      "step": 4031
+    },
+    {
+      "epoch": 2.111576741749607,
+      "loss_partial": 0.047119140625,
+      "step": 4031
+    },
+    {
+      "epoch": 2.111576741749607,
+      "loss_partial": 0.05224609375,
+      "step": 4031
+    },
+    {
+      "epoch": 2.112100576217915,
+      "grad_norm": 2.0208208560943604,
+      "learning_rate": 1.0771829341729356e-05,
+      "loss": 0.1034,
+      "step": 4032
+    },
+    {
+      "epoch": 2.112100576217915,
+      "loss_partial": 0.058837890625,
+      "step": 4032
+    },
+    {
+      "epoch": 2.112100576217915,
+      "loss_partial": 0.045654296875,
+      "step": 4032
+    },
+    {
+      "epoch": 2.112624410686223,
+      "grad_norm": 1.488483190536499,
+      "learning_rate": 1.0767271314320115e-05,
+      "loss": 0.0989,
+      "step": 4033
+    },
+    {
+      "epoch": 2.112624410686223,
+      "loss_partial": 0.052001953125,
+      "step": 4033
+    },
+    {
+      "epoch": 2.112624410686223,
+      "loss_partial": 0.054931640625,
+      "step": 4033
+    },
+    {
+      "epoch": 2.113148245154531,
+      "grad_norm": 1.4714348316192627,
+      "learning_rate": 1.0762713126555911e-05,
+      "loss": 0.1042,
+      "step": 4034
+    },
+    {
+      "epoch": 2.113148245154531,
+      "loss_partial": 0.051025390625,
+      "step": 4034
+    },
+    {
+      "epoch": 2.113148245154531,
+      "loss_partial": 0.056396484375,
+      "step": 4034
+    },
+    {
+      "epoch": 2.113672079622839,
+      "grad_norm": 1.8882514238357544,
+      "learning_rate": 1.0758154779389383e-05,
+      "loss": 0.104,
+      "step": 4035
+    },
+    {
+      "epoch": 2.113672079622839,
+      "loss_partial": 0.060302734375,
+      "step": 4035
+    },
+    {
+      "epoch": 2.113672079622839,
+      "loss_partial": 0.0498046875,
+      "step": 4035
+    },
+    {
+      "epoch": 2.1141959140911473,
+      "grad_norm": 2.1787948608398438,
+      "learning_rate": 1.075359627377319e-05,
+      "loss": 0.106,
+      "step": 4036
+    },
+    {
+      "epoch": 2.1141959140911473,
+      "loss_partial": 0.048095703125,
+      "step": 4036
+    },
+    {
+      "epoch": 2.1141959140911473,
+      "loss_partial": 0.053466796875,
+      "step": 4036
+    },
+    {
+      "epoch": 2.1147197485594553,
+      "grad_norm": 1.9750850200653076,
+      "learning_rate": 1.0749037610660041e-05,
+      "loss": 0.1012,
+      "step": 4037
+    },
+    {
+      "epoch": 2.1147197485594553,
+      "loss_partial": 0.05224609375,
+      "step": 4037
+    },
+    {
+      "epoch": 2.1147197485594553,
+      "loss_partial": 0.0693359375,
+      "step": 4037
+    },
+    {
+      "epoch": 2.1152435830277634,
+      "grad_norm": 2.1511383056640625,
+      "learning_rate": 1.0744478791002656e-05,
+      "loss": 0.1045,
+      "step": 4038
+    },
+    {
+      "epoch": 2.1152435830277634,
+      "loss_partial": 0.04931640625,
+      "step": 4038
+    },
+    {
+      "epoch": 2.1152435830277634,
+      "loss_partial": 0.0478515625,
+      "step": 4038
+    },
+    {
+      "epoch": 2.1157674174960714,
+      "grad_norm": 1.8798941373825073,
+      "learning_rate": 1.0739919815753817e-05,
+      "loss": 0.1042,
+      "step": 4039
+    },
+    {
+      "epoch": 2.1157674174960714,
+      "loss_partial": 0.05810546875,
+      "step": 4039
+    },
+    {
+      "epoch": 2.1157674174960714,
+      "loss_partial": 0.0556640625,
+      "step": 4039
+    },
+    {
+      "epoch": 2.116291251964379,
+      "grad_norm": 1.6291728019714355,
+      "learning_rate": 1.0735360685866306e-05,
+      "loss": 0.107,
+      "step": 4040
+    },
+    {
+      "epoch": 2.116291251964379,
+      "loss_partial": 0.0576171875,
+      "step": 4040
+    },
+    {
+      "epoch": 2.116291251964379,
+      "loss_partial": 0.05078125,
+      "step": 4040
+    },
+    {
+      "epoch": 2.116815086432687,
+      "grad_norm": 1.9509419202804565,
+      "learning_rate": 1.073080140229296e-05,
+      "loss": 0.1036,
+      "step": 4041
+    },
+    {
+      "epoch": 2.116815086432687,
+      "loss_partial": 0.04931640625,
+      "step": 4041
+    },
+    {
+      "epoch": 2.116815086432687,
+      "loss_partial": 0.052978515625,
+      "step": 4041
+    },
+    {
+      "epoch": 2.117338920900995,
+      "grad_norm": 1.7140445709228516,
+      "learning_rate": 1.0726241965986639e-05,
+      "loss": 0.1046,
+      "step": 4042
+    },
+    {
+      "epoch": 2.117338920900995,
+      "loss_partial": 0.052001953125,
+      "step": 4042
+    },
+    {
+      "epoch": 2.117338920900995,
+      "loss_partial": 0.058837890625,
+      "step": 4042
+    },
+    {
+      "epoch": 2.1178627553693032,
+      "grad_norm": 1.7135467529296875,
+      "learning_rate": 1.072168237790024e-05,
+      "loss": 0.104,
+      "step": 4043
+    },
+    {
+      "epoch": 2.1178627553693032,
+      "loss_partial": 0.05224609375,
+      "step": 4043
+    },
+    {
+      "epoch": 2.1178627553693032,
+      "loss_partial": 0.047119140625,
+      "step": 4043
+    },
+    {
+      "epoch": 2.1183865898376113,
+      "grad_norm": 1.9594019651412964,
+      "learning_rate": 1.0717122638986683e-05,
+      "loss": 0.1037,
+      "step": 4044
+    },
+    {
+      "epoch": 2.1183865898376113,
+      "loss_partial": 0.03955078125,
+      "step": 4044
+    },
+    {
+      "epoch": 2.1183865898376113,
+      "loss_partial": 0.047607421875,
+      "step": 4044
+    },
+    {
+      "epoch": 2.1189104243059194,
+      "grad_norm": 1.7775791883468628,
+      "learning_rate": 1.0712562750198934e-05,
+      "loss": 0.103,
+      "step": 4045
+    },
+    {
+      "epoch": 2.1189104243059194,
+      "loss_partial": 0.0478515625,
+      "step": 4045
+    },
+    {
+      "epoch": 2.1189104243059194,
+      "loss_partial": 0.0458984375,
+      "step": 4045
+    },
+    {
+      "epoch": 2.1194342587742274,
+      "grad_norm": 2.2352051734924316,
+      "learning_rate": 1.0708002712489975e-05,
+      "loss": 0.1002,
+      "step": 4046
+    },
+    {
+      "epoch": 2.1194342587742274,
+      "loss_partial": 0.053466796875,
+      "step": 4046
+    },
+    {
+      "epoch": 2.1194342587742274,
+      "loss_partial": 0.051513671875,
+      "step": 4046
+    },
+    {
+      "epoch": 2.1199580932425355,
+      "grad_norm": 1.63973069190979,
+      "learning_rate": 1.0703442526812829e-05,
+      "loss": 0.1022,
+      "step": 4047
+    },
+    {
+      "epoch": 2.1199580932425355,
+      "loss_partial": 0.05322265625,
+      "step": 4047
+    },
+    {
+      "epoch": 2.1199580932425355,
+      "loss_partial": 0.060302734375,
+      "step": 4047
+    },
+    {
+      "epoch": 2.1204819277108435,
+      "grad_norm": 2.2510597705841064,
+      "learning_rate": 1.0698882194120542e-05,
+      "loss": 0.103,
+      "step": 4048
+    },
+    {
+      "epoch": 2.1204819277108435,
+      "loss_partial": 0.047607421875,
+      "step": 4048
+    },
+    {
+      "epoch": 2.1204819277108435,
+      "loss_partial": 0.051025390625,
+      "step": 4048
+    },
+    {
+      "epoch": 2.1210057621791516,
+      "grad_norm": 1.8912224769592285,
+      "learning_rate": 1.0694321715366198e-05,
+      "loss": 0.1022,
+      "step": 4049
+    },
+    {
+      "epoch": 2.1210057621791516,
+      "loss_partial": 0.046142578125,
+      "step": 4049
+    },
+    {
+      "epoch": 2.1210057621791516,
+      "loss_partial": 0.0576171875,
+      "step": 4049
+    },
+    {
+      "epoch": 2.121529596647459,
+      "grad_norm": 1.747047781944275,
+      "learning_rate": 1.0689761091502908e-05,
+      "loss": 0.1041,
+      "step": 4050
+    },
+    {
+      "epoch": 2.121529596647459,
+      "loss_partial": 0.048095703125,
+      "step": 4050
+    },
+    {
+      "epoch": 2.121529596647459,
+      "loss_partial": 0.047119140625,
+      "step": 4050
+    },
+    {
+      "epoch": 2.1220534311157673,
+      "grad_norm": 1.3788398504257202,
+      "learning_rate": 1.0685200323483818e-05,
+      "loss": 0.1008,
+      "step": 4051
+    },
+    {
+      "epoch": 2.1220534311157673,
+      "loss_partial": 0.046142578125,
+      "step": 4051
+    },
+    {
+      "epoch": 2.1220534311157673,
+      "loss_partial": 0.05126953125,
+      "step": 4051
+    },
+    {
+      "epoch": 2.1225772655840753,
+      "grad_norm": 2.39089298248291,
+      "learning_rate": 1.0680639412262097e-05,
+      "loss": 0.1024,
+      "step": 4052
+    },
+    {
+      "epoch": 2.1225772655840753,
+      "loss_partial": 0.05029296875,
+      "step": 4052
+    },
+    {
+      "epoch": 2.1225772655840753,
+      "loss_partial": 0.053955078125,
+      "step": 4052
+    },
+    {
+      "epoch": 2.1231011000523834,
+      "grad_norm": 1.5561953783035278,
+      "learning_rate": 1.0676078358790948e-05,
+      "loss": 0.101,
+      "step": 4053
+    },
+    {
+      "epoch": 2.1231011000523834,
+      "loss_partial": 0.0546875,
+      "step": 4053
+    },
+    {
+      "epoch": 2.1231011000523834,
+      "loss_partial": 0.054931640625,
+      "step": 4053
+    },
+    {
+      "epoch": 2.1236249345206915,
+      "grad_norm": 1.6892361640930176,
+      "learning_rate": 1.0671517164023601e-05,
+      "loss": 0.1037,
+      "step": 4054
+    },
+    {
+      "epoch": 2.1236249345206915,
+      "loss_partial": 0.05322265625,
+      "step": 4054
+    },
+    {
+      "epoch": 2.1236249345206915,
+      "loss_partial": 0.05517578125,
+      "step": 4054
+    },
+    {
+      "epoch": 2.1241487689889995,
+      "grad_norm": 1.3329702615737915,
+      "learning_rate": 1.0666955828913322e-05,
+      "loss": 0.107,
+      "step": 4055
+    },
+    {
+      "epoch": 2.1241487689889995,
+      "loss_partial": 0.052978515625,
+      "step": 4055
+    },
+    {
+      "epoch": 2.1241487689889995,
+      "loss_partial": 0.05078125,
+      "step": 4055
+    },
+    {
+      "epoch": 2.1246726034573076,
+      "grad_norm": 2.2349853515625,
+      "learning_rate": 1.0662394354413396e-05,
+      "loss": 0.1019,
+      "step": 4056
+    },
+    {
+      "epoch": 2.1246726034573076,
+      "loss_partial": 0.046142578125,
+      "step": 4056
+    },
+    {
+      "epoch": 2.1246726034573076,
+      "loss_partial": 0.0478515625,
+      "step": 4056
+    },
+    {
+      "epoch": 2.1251964379256156,
+      "grad_norm": 1.7405214309692383,
+      "learning_rate": 1.0657832741477147e-05,
+      "loss": 0.0988,
+      "step": 4057
+    },
+    {
+      "epoch": 2.1251964379256156,
+      "loss_partial": 0.053466796875,
+      "step": 4057
+    },
+    {
+      "epoch": 2.1251964379256156,
+      "loss_partial": 0.04833984375,
+      "step": 4057
+    },
+    {
+      "epoch": 2.1257202723939237,
+      "grad_norm": 2.1927943229675293,
+      "learning_rate": 1.0653270991057925e-05,
+      "loss": 0.1046,
+      "step": 4058
+    },
+    {
+      "epoch": 2.1257202723939237,
+      "loss_partial": 0.049560546875,
+      "step": 4058
+    },
+    {
+      "epoch": 2.1257202723939237,
+      "loss_partial": 0.05615234375,
+      "step": 4058
+    },
+    {
+      "epoch": 2.1262441068622318,
+      "grad_norm": 1.991890788078308,
+      "learning_rate": 1.0648709104109106e-05,
+      "loss": 0.105,
+      "step": 4059
+    },
+    {
+      "epoch": 2.1262441068622318,
+      "loss_partial": 0.056396484375,
+      "step": 4059
+    },
+    {
+      "epoch": 2.1262441068622318,
+      "loss_partial": 0.053466796875,
+      "step": 4059
+    },
+    {
+      "epoch": 2.1267679413305394,
+      "grad_norm": 1.8096978664398193,
+      "learning_rate": 1.0644147081584091e-05,
+      "loss": 0.1021,
+      "step": 4060
+    },
+    {
+      "epoch": 2.1267679413305394,
+      "loss_partial": 0.04833984375,
+      "step": 4060
+    },
+    {
+      "epoch": 2.1267679413305394,
+      "loss_partial": 0.0595703125,
+      "step": 4060
+    },
+    {
+      "epoch": 2.1272917757988474,
+      "grad_norm": 2.16140079498291,
+      "learning_rate": 1.0639584924436324e-05,
+      "loss": 0.1057,
+      "step": 4061
+    },
+    {
+      "epoch": 2.1272917757988474,
+      "loss_partial": 0.056884765625,
+      "step": 4061
+    },
+    {
+      "epoch": 2.1272917757988474,
+      "loss_partial": 0.046875,
+      "step": 4061
+    },
+    {
+      "epoch": 2.1278156102671555,
+      "grad_norm": 1.7979811429977417,
+      "learning_rate": 1.0635022633619258e-05,
+      "loss": 0.1028,
+      "step": 4062
+    },
+    {
+      "epoch": 2.1278156102671555,
+      "loss_partial": 0.045654296875,
+      "step": 4062
+    },
+    {
+      "epoch": 2.1278156102671555,
+      "loss_partial": 0.048828125,
+      "step": 4062
+    },
+    {
+      "epoch": 2.1283394447354635,
+      "grad_norm": 1.9312692880630493,
+      "learning_rate": 1.0630460210086394e-05,
+      "loss": 0.1003,
+      "step": 4063
+    },
+    {
+      "epoch": 2.1283394447354635,
+      "loss_partial": 0.04345703125,
+      "step": 4063
+    },
+    {
+      "epoch": 2.1283394447354635,
+      "loss_partial": 0.05029296875,
+      "step": 4063
+    },
+    {
+      "epoch": 2.1288632792037716,
+      "grad_norm": 1.5403679609298706,
+      "learning_rate": 1.0625897654791242e-05,
+      "loss": 0.1042,
+      "step": 4064
+    },
+    {
+      "epoch": 2.1288632792037716,
+      "loss_partial": 0.049560546875,
+      "step": 4064
+    },
+    {
+      "epoch": 2.1288632792037716,
+      "loss_partial": 0.050048828125,
+      "step": 4064
+    },
+    {
+      "epoch": 2.1293871136720797,
+      "grad_norm": 1.7772207260131836,
+      "learning_rate": 1.0621334968687352e-05,
+      "loss": 0.106,
+      "step": 4065
+    },
+    {
+      "epoch": 2.1293871136720797,
+      "loss_partial": 0.0625,
+      "step": 4065
+    },
+    {
+      "epoch": 2.1293871136720797,
+      "loss_partial": 0.05126953125,
+      "step": 4065
+    },
+    {
+      "epoch": 2.1299109481403877,
+      "grad_norm": 1.7155652046203613,
+      "learning_rate": 1.0616772152728292e-05,
+      "loss": 0.1052,
+      "step": 4066
+    },
+    {
+      "epoch": 2.1299109481403877,
+      "loss_partial": 0.059326171875,
+      "step": 4066
+    },
+    {
+      "epoch": 2.1299109481403877,
+      "loss_partial": 0.046142578125,
+      "step": 4066
+    },
+    {
+      "epoch": 2.130434782608696,
+      "grad_norm": 1.4373674392700195,
+      "learning_rate": 1.061220920786767e-05,
+      "loss": 0.1043,
+      "step": 4067
+    },
+    {
+      "epoch": 2.130434782608696,
+      "loss_partial": 0.053955078125,
+      "step": 4067
+    },
+    {
+      "epoch": 2.130434782608696,
+      "loss_partial": 0.05810546875,
+      "step": 4067
+    },
+    {
+      "epoch": 2.130958617077004,
+      "grad_norm": 1.7783482074737549,
+      "learning_rate": 1.0607646135059106e-05,
+      "loss": 0.1066,
+      "step": 4068
+    },
+    {
+      "epoch": 2.130958617077004,
+      "loss_partial": 0.044921875,
+      "step": 4068
+    },
+    {
+      "epoch": 2.130958617077004,
+      "loss_partial": 0.050048828125,
+      "step": 4068
+    },
+    {
+      "epoch": 2.1314824515453115,
+      "grad_norm": 1.4015753269195557,
+      "learning_rate": 1.0603082935256261e-05,
+      "loss": 0.1034,
+      "step": 4069
+    },
+    {
+      "epoch": 2.1314824515453115,
+      "loss_partial": 0.046875,
+      "step": 4069
+    },
+    {
+      "epoch": 2.1314824515453115,
+      "loss_partial": 0.047119140625,
+      "step": 4069
+    },
+    {
+      "epoch": 2.1320062860136195,
+      "grad_norm": 2.172870635986328,
+      "learning_rate": 1.059851960941281e-05,
+      "loss": 0.1018,
+      "step": 4070
+    },
+    {
+      "epoch": 2.1320062860136195,
+      "loss_partial": 0.04296875,
+      "step": 4070
+    },
+    {
+      "epoch": 2.1320062860136195,
+      "loss_partial": 0.054443359375,
+      "step": 4070
+    },
+    {
+      "epoch": 2.1325301204819276,
+      "grad_norm": 1.5321199893951416,
+      "learning_rate": 1.0593956158482463e-05,
+      "loss": 0.102,
+      "step": 4071
+    },
+    {
+      "epoch": 2.1325301204819276,
+      "loss_partial": 0.046875,
+      "step": 4071
+    },
+    {
+      "epoch": 2.1325301204819276,
+      "loss_partial": 0.049560546875,
+      "step": 4071
+    },
+    {
+      "epoch": 2.1330539549502356,
+      "grad_norm": 1.745668649673462,
+      "learning_rate": 1.0589392583418949e-05,
+      "loss": 0.1042,
+      "step": 4072
+    },
+    {
+      "epoch": 2.1330539549502356,
+      "loss_partial": 0.040283203125,
+      "step": 4072
+    },
+    {
+      "epoch": 2.1330539549502356,
+      "loss_partial": 0.06298828125,
+      "step": 4072
+    },
+    {
+      "epoch": 2.1335777894185437,
+      "grad_norm": 1.1800267696380615,
+      "learning_rate": 1.0584828885176032e-05,
+      "loss": 0.098,
+      "step": 4073
+    },
+    {
+      "epoch": 2.1335777894185437,
+      "loss_partial": 0.05419921875,
+      "step": 4073
+    },
+    {
+      "epoch": 2.1335777894185437,
+      "loss_partial": 0.046142578125,
+      "step": 4073
+    },
+    {
+      "epoch": 2.1341016238868518,
+      "grad_norm": 1.4997552633285522,
+      "learning_rate": 1.0580265064707493e-05,
+      "loss": 0.1018,
+      "step": 4074
+    },
+    {
+      "epoch": 2.1341016238868518,
+      "loss_partial": 0.051025390625,
+      "step": 4074
+    },
+    {
+      "epoch": 2.1341016238868518,
+      "loss_partial": 0.047607421875,
+      "step": 4074
+    },
+    {
+      "epoch": 2.13462545835516,
+      "grad_norm": 1.680398941040039,
+      "learning_rate": 1.0575701122967146e-05,
+      "loss": 0.1005,
+      "step": 4075
+    },
+    {
+      "epoch": 2.13462545835516,
+      "loss_partial": 0.043701171875,
+      "step": 4075
+    },
+    {
+      "epoch": 2.13462545835516,
+      "loss_partial": 0.0576171875,
+      "step": 4075
+    },
+    {
+      "epoch": 2.135149292823468,
+      "grad_norm": 1.8576648235321045,
+      "learning_rate": 1.0571137060908822e-05,
+      "loss": 0.1036,
+      "step": 4076
+    },
+    {
+      "epoch": 2.135149292823468,
+      "loss_partial": 0.057373046875,
+      "step": 4076
+    },
+    {
+      "epoch": 2.135149292823468,
+      "loss_partial": 0.054443359375,
+      "step": 4076
+    },
+    {
+      "epoch": 2.135673127291776,
+      "grad_norm": 1.9267114400863647,
+      "learning_rate": 1.0566572879486388e-05,
+      "loss": 0.1005,
+      "step": 4077
+    },
+    {
+      "epoch": 2.135673127291776,
+      "loss_partial": 0.048095703125,
+      "step": 4077
+    },
+    {
+      "epoch": 2.135673127291776,
+      "loss_partial": 0.04638671875,
+      "step": 4077
+    },
+    {
+      "epoch": 2.136196961760084,
+      "grad_norm": 1.4628067016601562,
+      "learning_rate": 1.056200857965372e-05,
+      "loss": 0.1039,
+      "step": 4078
+    },
+    {
+      "epoch": 2.136196961760084,
+      "loss_partial": 0.053955078125,
+      "step": 4078
+    },
+    {
+      "epoch": 2.136196961760084,
+      "loss_partial": 0.059814453125,
+      "step": 4078
+    },
+    {
+      "epoch": 2.136720796228392,
+      "grad_norm": 1.9948612451553345,
+      "learning_rate": 1.0557444162364739e-05,
+      "loss": 0.1017,
+      "step": 4079
+    },
+    {
+      "epoch": 2.136720796228392,
+      "loss_partial": 0.0478515625,
+      "step": 4079
+    },
+    {
+      "epoch": 2.136720796228392,
+      "loss_partial": 0.050048828125,
+      "step": 4079
+    },
+    {
+      "epoch": 2.1372446306966997,
+      "grad_norm": 1.3657382726669312,
+      "learning_rate": 1.0552879628573371e-05,
+      "loss": 0.0994,
+      "step": 4080
+    },
+    {
+      "epoch": 2.1372446306966997,
+      "loss_partial": 0.05517578125,
+      "step": 4080
+    },
+    {
+      "epoch": 2.1372446306966997,
+      "loss_partial": 0.049072265625,
+      "step": 4080
+    },
+    {
+      "epoch": 2.1377684651650077,
+      "grad_norm": 1.8395854234695435,
+      "learning_rate": 1.0548314979233583e-05,
+      "loss": 0.1003,
+      "step": 4081
+    },
+    {
+      "epoch": 2.1377684651650077,
+      "loss_partial": 0.048583984375,
+      "step": 4081
+    },
+    {
+      "epoch": 2.1377684651650077,
+      "loss_partial": 0.051025390625,
+      "step": 4081
+    },
+    {
+      "epoch": 2.138292299633316,
+      "grad_norm": 2.1096835136413574,
+      "learning_rate": 1.0543750215299353e-05,
+      "loss": 0.1008,
+      "step": 4082
+    },
+    {
+      "epoch": 2.138292299633316,
+      "loss_partial": 0.056884765625,
+      "step": 4082
+    },
+    {
+      "epoch": 2.138292299633316,
+      "loss_partial": 0.050048828125,
+      "step": 4082
+    },
+    {
+      "epoch": 2.138816134101624,
+      "grad_norm": 1.84113347530365,
+      "learning_rate": 1.0539185337724694e-05,
+      "loss": 0.1025,
+      "step": 4083
+    },
+    {
+      "epoch": 2.138816134101624,
+      "loss_partial": 0.0693359375,
+      "step": 4083
+    },
+    {
+      "epoch": 2.138816134101624,
+      "loss_partial": 0.0546875,
+      "step": 4083
+    },
+    {
+      "epoch": 2.139339968569932,
+      "grad_norm": 2.3664608001708984,
+      "learning_rate": 1.0534620347463629e-05,
+      "loss": 0.1049,
+      "step": 4084
+    },
+    {
+      "epoch": 2.139339968569932,
+      "loss_partial": 0.06298828125,
+      "step": 4084
+    },
+    {
+      "epoch": 2.139339968569932,
+      "loss_partial": 0.05078125,
+      "step": 4084
+    },
+    {
+      "epoch": 2.13986380303824,
+      "grad_norm": 2.0166187286376953,
+      "learning_rate": 1.0530055245470219e-05,
+      "loss": 0.1033,
+      "step": 4085
+    },
+    {
+      "epoch": 2.13986380303824,
+      "loss_partial": 0.059814453125,
+      "step": 4085
+    },
+    {
+      "epoch": 2.13986380303824,
+      "loss_partial": 0.04638671875,
+      "step": 4085
+    },
+    {
+      "epoch": 2.140387637506548,
+      "grad_norm": 1.6381560564041138,
+      "learning_rate": 1.052549003269854e-05,
+      "loss": 0.1021,
+      "step": 4086
+    },
+    {
+      "epoch": 2.140387637506548,
+      "loss_partial": 0.05810546875,
+      "step": 4086
+    },
+    {
+      "epoch": 2.140387637506548,
+      "loss_partial": 0.0576171875,
+      "step": 4086
+    },
+    {
+      "epoch": 2.140911471974856,
+      "grad_norm": 2.1849756240844727,
+      "learning_rate": 1.0520924710102696e-05,
+      "loss": 0.1024,
+      "step": 4087
+    },
+    {
+      "epoch": 2.140911471974856,
+      "loss_partial": 0.052001953125,
+      "step": 4087
+    },
+    {
+      "epoch": 2.140911471974856,
+      "loss_partial": 0.054443359375,
+      "step": 4087
+    },
+    {
+      "epoch": 2.141435306443164,
+      "grad_norm": 1.6909451484680176,
+      "learning_rate": 1.0516359278636807e-05,
+      "loss": 0.0995,
+      "step": 4088
+    },
+    {
+      "epoch": 2.141435306443164,
+      "loss_partial": 0.048828125,
+      "step": 4088
+    },
+    {
+      "epoch": 2.141435306443164,
+      "loss_partial": 0.0478515625,
+      "step": 4088
+    },
+    {
+      "epoch": 2.1419591409114718,
+      "grad_norm": 1.709808588027954,
+      "learning_rate": 1.0511793739255022e-05,
+      "loss": 0.1017,
+      "step": 4089
+    },
+    {
+      "epoch": 2.1419591409114718,
+      "loss_partial": 0.0478515625,
+      "step": 4089
+    },
+    {
+      "epoch": 2.1419591409114718,
+      "loss_partial": 0.052001953125,
+      "step": 4089
+    },
+    {
+      "epoch": 2.14248297537978,
+      "grad_norm": 1.6155524253845215,
+      "learning_rate": 1.0507228092911511e-05,
+      "loss": 0.1041,
+      "step": 4090
+    },
+    {
+      "epoch": 2.14248297537978,
+      "loss_partial": 0.056396484375,
+      "step": 4090
+    },
+    {
+      "epoch": 2.14248297537978,
+      "loss_partial": 0.04931640625,
+      "step": 4090
+    },
+    {
+      "epoch": 2.143006809848088,
+      "grad_norm": 1.867705225944519,
+      "learning_rate": 1.0502662340560464e-05,
+      "loss": 0.1023,
+      "step": 4091
+    },
+    {
+      "epoch": 2.143006809848088,
+      "loss_partial": 0.047119140625,
+      "step": 4091
+    },
+    {
+      "epoch": 2.143006809848088,
+      "loss_partial": 0.052734375,
+      "step": 4091
+    },
+    {
+      "epoch": 2.143530644316396,
+      "grad_norm": 2.754697322845459,
+      "learning_rate": 1.0498096483156097e-05,
+      "loss": 0.1059,
+      "step": 4092
+    },
+    {
+      "epoch": 2.143530644316396,
+      "loss_partial": 0.059814453125,
+      "step": 4092
+    },
+    {
+      "epoch": 2.143530644316396,
+      "loss_partial": 0.0478515625,
+      "step": 4092
+    },
+    {
+      "epoch": 2.144054478784704,
+      "grad_norm": 2.3086984157562256,
+      "learning_rate": 1.0493530521652644e-05,
+      "loss": 0.1057,
+      "step": 4093
+    },
+    {
+      "epoch": 2.144054478784704,
+      "loss_partial": 0.04736328125,
+      "step": 4093
+    },
+    {
+      "epoch": 2.144054478784704,
+      "loss_partial": 0.04931640625,
+      "step": 4093
+    },
+    {
+      "epoch": 2.144578313253012,
+      "grad_norm": 2.1367228031158447,
+      "learning_rate": 1.0488964457004367e-05,
+      "loss": 0.0988,
+      "step": 4094
+    },
+    {
+      "epoch": 2.144578313253012,
+      "loss_partial": 0.055419921875,
+      "step": 4094
+    },
+    {
+      "epoch": 2.144578313253012,
+      "loss_partial": 0.0751953125,
+      "step": 4094
+    },
+    {
+      "epoch": 2.14510214772132,
+      "grad_norm": 1.886370062828064,
+      "learning_rate": 1.0484398290165541e-05,
+      "loss": 0.1042,
+      "step": 4095
+    },
+    {
+      "epoch": 2.14510214772132,
+      "loss_partial": 0.053955078125,
+      "step": 4095
+    },
+    {
+      "epoch": 2.14510214772132,
+      "loss_partial": 0.048583984375,
+      "step": 4095
+    },
+    {
+      "epoch": 2.145625982189628,
+      "grad_norm": 1.7025706768035889,
+      "learning_rate": 1.0479832022090465e-05,
+      "loss": 0.1035,
+      "step": 4096
+    },
+    {
+      "epoch": 2.145625982189628,
+      "loss_partial": 0.055908203125,
+      "step": 4096
+    },
+    {
+      "epoch": 2.145625982189628,
+      "loss_partial": 0.05517578125,
+      "step": 4096
+    },
+    {
+      "epoch": 2.1461498166579362,
+      "grad_norm": 2.417754888534546,
+      "learning_rate": 1.0475265653733466e-05,
+      "loss": 0.1024,
+      "step": 4097
+    },
+    {
+      "epoch": 2.1461498166579362,
+      "loss_partial": 0.056640625,
+      "step": 4097
+    },
+    {
+      "epoch": 2.1461498166579362,
+      "loss_partial": 0.050048828125,
+      "step": 4097
+    },
+    {
+      "epoch": 2.1466736511262443,
+      "grad_norm": 1.63076913356781,
+      "learning_rate": 1.0470699186048885e-05,
+      "loss": 0.1077,
+      "step": 4098
+    },
+    {
+      "epoch": 2.1466736511262443,
+      "loss_partial": 0.048583984375,
+      "step": 4098
+    },
+    {
+      "epoch": 2.1466736511262443,
+      "loss_partial": 0.0498046875,
+      "step": 4098
+    },
+    {
+      "epoch": 2.147197485594552,
+      "grad_norm": 2.2057034969329834,
+      "learning_rate": 1.046613261999108e-05,
+      "loss": 0.1029,
+      "step": 4099
+    },
+    {
+      "epoch": 2.147197485594552,
+      "loss_partial": 0.046630859375,
+      "step": 4099
+    },
+    {
+      "epoch": 2.147197485594552,
+      "loss_partial": 0.0576171875,
+      "step": 4099
+    },
+    {
+      "epoch": 2.14772132006286,
+      "grad_norm": 2.0025794506073,
+      "learning_rate": 1.0461565956514445e-05,
+      "loss": 0.1052,
+      "step": 4100
+    },
+    {
+      "epoch": 2.14772132006286,
+      "loss_partial": 0.047119140625,
+      "step": 4100
+    },
+    {
+      "epoch": 2.14772132006286,
+      "loss_partial": 0.04833984375,
+      "step": 4100
+    },
+    {
+      "epoch": 2.148245154531168,
+      "grad_norm": 2.1091887950897217,
+      "learning_rate": 1.0456999196573378e-05,
+      "loss": 0.1042,
+      "step": 4101
+    },
+    {
+      "epoch": 2.148245154531168,
+      "loss_partial": 0.059814453125,
+      "step": 4101
+    },
+    {
+      "epoch": 2.148245154531168,
+      "loss_partial": 0.0576171875,
+      "step": 4101
+    },
+    {
+      "epoch": 2.148768988999476,
+      "grad_norm": 1.7094686031341553,
+      "learning_rate": 1.0452432341122307e-05,
+      "loss": 0.1065,
+      "step": 4102
+    },
+    {
+      "epoch": 2.148768988999476,
+      "loss_partial": 0.048828125,
+      "step": 4102
+    },
+    {
+      "epoch": 2.148768988999476,
+      "loss_partial": 0.04736328125,
+      "step": 4102
+    },
+    {
+      "epoch": 2.149292823467784,
+      "grad_norm": 1.622968316078186,
+      "learning_rate": 1.044786539111567e-05,
+      "loss": 0.1037,
+      "step": 4103
+    },
+    {
+      "epoch": 2.149292823467784,
+      "loss_partial": 0.05322265625,
+      "step": 4103
+    },
+    {
+      "epoch": 2.149292823467784,
+      "loss_partial": 0.045654296875,
+      "step": 4103
+    },
+    {
+      "epoch": 2.1498166579360922,
+      "grad_norm": 1.7944066524505615,
+      "learning_rate": 1.0443298347507938e-05,
+      "loss": 0.1001,
+      "step": 4104
+    },
+    {
+      "epoch": 2.1498166579360922,
+      "loss_partial": 0.045166015625,
+      "step": 4104
+    },
+    {
+      "epoch": 2.1498166579360922,
+      "loss_partial": 0.0498046875,
+      "step": 4104
+    },
+    {
+      "epoch": 2.1503404924044003,
+      "grad_norm": 1.710695743560791,
+      "learning_rate": 1.0438731211253592e-05,
+      "loss": 0.0995,
+      "step": 4105
+    },
+    {
+      "epoch": 2.1503404924044003,
+      "loss_partial": 0.049072265625,
+      "step": 4105
+    },
+    {
+      "epoch": 2.1503404924044003,
+      "loss_partial": 0.0537109375,
+      "step": 4105
+    },
+    {
+      "epoch": 2.1508643268727083,
+      "grad_norm": 1.844683289527893,
+      "learning_rate": 1.0434163983307138e-05,
+      "loss": 0.1004,
+      "step": 4106
+    },
+    {
+      "epoch": 2.1508643268727083,
+      "loss_partial": 0.048583984375,
+      "step": 4106
+    },
+    {
+      "epoch": 2.1508643268727083,
+      "loss_partial": 0.04638671875,
+      "step": 4106
+    },
+    {
+      "epoch": 2.1513881613410164,
+      "grad_norm": 1.6547383069992065,
+      "learning_rate": 1.0429596664623094e-05,
+      "loss": 0.0984,
+      "step": 4107
+    },
+    {
+      "epoch": 2.1513881613410164,
+      "loss_partial": 0.05224609375,
+      "step": 4107
+    },
+    {
+      "epoch": 2.1513881613410164,
+      "loss_partial": 0.044189453125,
+      "step": 4107
+    },
+    {
+      "epoch": 2.151911995809324,
+      "grad_norm": 2.068185329437256,
+      "learning_rate": 1.0425029256156004e-05,
+      "loss": 0.1026,
+      "step": 4108
+    },
+    {
+      "epoch": 2.151911995809324,
+      "loss_partial": 0.04736328125,
+      "step": 4108
+    },
+    {
+      "epoch": 2.151911995809324,
+      "loss_partial": 0.053466796875,
+      "step": 4108
+    },
+    {
+      "epoch": 2.152435830277632,
+      "grad_norm": 1.8018879890441895,
+      "learning_rate": 1.0420461758860427e-05,
+      "loss": 0.1057,
+      "step": 4109
+    },
+    {
+      "epoch": 2.152435830277632,
+      "loss_partial": 0.048583984375,
+      "step": 4109
+    },
+    {
+      "epoch": 2.152435830277632,
+      "loss_partial": 0.049560546875,
+      "step": 4109
+    },
+    {
+      "epoch": 2.15295966474594,
+      "grad_norm": 1.9713505506515503,
+      "learning_rate": 1.0415894173690946e-05,
+      "loss": 0.1035,
+      "step": 4110
+    },
+    {
+      "epoch": 2.15295966474594,
+      "loss_partial": 0.05126953125,
+      "step": 4110
+    },
+    {
+      "epoch": 2.15295966474594,
+      "loss_partial": 0.046875,
+      "step": 4110
+    },
+    {
+      "epoch": 2.153483499214248,
+      "grad_norm": 1.8041311502456665,
+      "learning_rate": 1.041132650160215e-05,
+      "loss": 0.1043,
+      "step": 4111
+    },
+    {
+      "epoch": 2.153483499214248,
+      "loss_partial": 0.04443359375,
+      "step": 4111
+    },
+    {
+      "epoch": 2.153483499214248,
+      "loss_partial": 0.05615234375,
+      "step": 4111
+    },
+    {
+      "epoch": 2.1540073336825563,
+      "grad_norm": 1.8635635375976562,
+      "learning_rate": 1.0406758743548662e-05,
+      "loss": 0.1068,
+      "step": 4112
+    },
+    {
+      "epoch": 2.1540073336825563,
+      "loss_partial": 0.051025390625,
+      "step": 4112
+    },
+    {
+      "epoch": 2.1540073336825563,
+      "loss_partial": 0.06103515625,
+      "step": 4112
+    },
+    {
+      "epoch": 2.1545311681508643,
+      "grad_norm": 2.577803373336792,
+      "learning_rate": 1.0402190900485112e-05,
+      "loss": 0.1032,
+      "step": 4113
+    },
+    {
+      "epoch": 2.1545311681508643,
+      "loss_partial": 0.054443359375,
+      "step": 4113
+    },
+    {
+      "epoch": 2.1545311681508643,
+      "loss_partial": 0.044677734375,
+      "step": 4113
+    },
+    {
+      "epoch": 2.1550550026191724,
+      "grad_norm": 1.9618034362792969,
+      "learning_rate": 1.039762297336615e-05,
+      "loss": 0.1015,
+      "step": 4114
+    },
+    {
+      "epoch": 2.1550550026191724,
+      "loss_partial": 0.050048828125,
+      "step": 4114
+    },
+    {
+      "epoch": 2.1550550026191724,
+      "loss_partial": 0.043701171875,
+      "step": 4114
+    },
+    {
+      "epoch": 2.1555788370874804,
+      "grad_norm": 1.9370859861373901,
+      "learning_rate": 1.0393054963146443e-05,
+      "loss": 0.0989,
+      "step": 4115
+    },
+    {
+      "epoch": 2.1555788370874804,
+      "loss_partial": 0.056884765625,
+      "step": 4115
+    },
+    {
+      "epoch": 2.1555788370874804,
+      "loss_partial": 0.05712890625,
+      "step": 4115
+    },
+    {
+      "epoch": 2.1561026715557885,
+      "grad_norm": 2.051772117614746,
+      "learning_rate": 1.0388486870780684e-05,
+      "loss": 0.1007,
+      "step": 4116
+    },
+    {
+      "epoch": 2.1561026715557885,
+      "loss_partial": 0.046630859375,
+      "step": 4116
+    },
+    {
+      "epoch": 2.1561026715557885,
+      "loss_partial": 0.059326171875,
+      "step": 4116
+    },
+    {
+      "epoch": 2.1566265060240966,
+      "grad_norm": 1.9208698272705078,
+      "learning_rate": 1.0383918697223564e-05,
+      "loss": 0.1022,
+      "step": 4117
+    },
+    {
+      "epoch": 2.1566265060240966,
+      "loss_partial": 0.055908203125,
+      "step": 4117
+    },
+    {
+      "epoch": 2.1566265060240966,
+      "loss_partial": 0.047119140625,
+      "step": 4117
+    },
+    {
+      "epoch": 2.1571503404924046,
+      "grad_norm": 2.225302219390869,
+      "learning_rate": 1.0379350443429816e-05,
+      "loss": 0.1036,
+      "step": 4118
+    },
+    {
+      "epoch": 2.1571503404924046,
+      "loss_partial": 0.04638671875,
+      "step": 4118
+    },
+    {
+      "epoch": 2.1571503404924046,
+      "loss_partial": 0.056396484375,
+      "step": 4118
+    },
+    {
+      "epoch": 2.1576741749607122,
+      "grad_norm": 2.13185715675354,
+      "learning_rate": 1.0374782110354171e-05,
+      "loss": 0.1065,
+      "step": 4119
+    },
+    {
+      "epoch": 2.1576741749607122,
+      "loss_partial": 0.053955078125,
+      "step": 4119
+    },
+    {
+      "epoch": 2.1576741749607122,
+      "loss_partial": 0.05126953125,
+      "step": 4119
+    },
+    {
+      "epoch": 2.1581980094290203,
+      "grad_norm": 2.024218797683716,
+      "learning_rate": 1.0370213698951383e-05,
+      "loss": 0.1022,
+      "step": 4120
+    },
+    {
+      "epoch": 2.1581980094290203,
+      "loss_partial": 0.052734375,
+      "step": 4120
+    },
+    {
+      "epoch": 2.1581980094290203,
+      "loss_partial": 0.046630859375,
+      "step": 4120
+    },
+    {
+      "epoch": 2.1587218438973284,
+      "grad_norm": 1.8886231184005737,
+      "learning_rate": 1.0365645210176219e-05,
+      "loss": 0.1016,
+      "step": 4121
+    },
+    {
+      "epoch": 2.1587218438973284,
+      "loss_partial": 0.04736328125,
+      "step": 4121
+    },
+    {
+      "epoch": 2.1587218438973284,
+      "loss_partial": 0.0498046875,
+      "step": 4121
+    },
+    {
+      "epoch": 2.1592456783656364,
+      "grad_norm": 1.5422173738479614,
+      "learning_rate": 1.0361076644983467e-05,
+      "loss": 0.1009,
+      "step": 4122
+    },
+    {
+      "epoch": 2.1592456783656364,
+      "loss_partial": 0.04443359375,
+      "step": 4122
+    },
+    {
+      "epoch": 2.1592456783656364,
+      "loss_partial": 0.05126953125,
+      "step": 4122
+    },
+    {
+      "epoch": 2.1597695128339445,
+      "grad_norm": 1.516809105873108,
+      "learning_rate": 1.035650800432793e-05,
+      "loss": 0.1014,
+      "step": 4123
+    },
+    {
+      "epoch": 2.1597695128339445,
+      "loss_partial": 0.057373046875,
+      "step": 4123
+    },
+    {
+      "epoch": 2.1597695128339445,
+      "loss_partial": 0.0546875,
+      "step": 4123
+    },
+    {
+      "epoch": 2.1602933473022525,
+      "grad_norm": 1.5033122301101685,
+      "learning_rate": 1.0351939289164426e-05,
+      "loss": 0.1011,
+      "step": 4124
+    },
+    {
+      "epoch": 2.1602933473022525,
+      "loss_partial": 0.051513671875,
+      "step": 4124
+    },
+    {
+      "epoch": 2.1602933473022525,
+      "loss_partial": 0.059326171875,
+      "step": 4124
+    },
+    {
+      "epoch": 2.1608171817705606,
+      "grad_norm": 1.1838656663894653,
+      "learning_rate": 1.0347370500447787e-05,
+      "loss": 0.1042,
+      "step": 4125
+    },
+    {
+      "epoch": 2.1608171817705606,
+      "loss_partial": 0.0546875,
+      "step": 4125
+    },
+    {
+      "epoch": 2.1608171817705606,
+      "loss_partial": 0.06298828125,
+      "step": 4125
+    },
+    {
+      "epoch": 2.1613410162388687,
+      "grad_norm": 1.4949301481246948,
+      "learning_rate": 1.034280163913286e-05,
+      "loss": 0.101,
+      "step": 4126
+    },
+    {
+      "epoch": 2.1613410162388687,
+      "loss_partial": 0.056640625,
+      "step": 4126
+    },
+    {
+      "epoch": 2.1613410162388687,
+      "loss_partial": 0.044921875,
+      "step": 4126
+    },
+    {
+      "epoch": 2.1618648507071767,
+      "grad_norm": 1.52484130859375,
+      "learning_rate": 1.0338232706174509e-05,
+      "loss": 0.1021,
+      "step": 4127
+    },
+    {
+      "epoch": 2.1618648507071767,
+      "loss_partial": 0.04931640625,
+      "step": 4127
+    },
+    {
+      "epoch": 2.1618648507071767,
+      "loss_partial": 0.048095703125,
+      "step": 4127
+    },
+    {
+      "epoch": 2.1623886851754843,
+      "grad_norm": 1.885177731513977,
+      "learning_rate": 1.0333663702527616e-05,
+      "loss": 0.0989,
+      "step": 4128
+    },
+    {
+      "epoch": 2.1623886851754843,
+      "loss_partial": 0.055908203125,
+      "step": 4128
+    },
+    {
+      "epoch": 2.1623886851754843,
+      "loss_partial": 0.048583984375,
+      "step": 4128
+    },
+    {
+      "epoch": 2.1629125196437924,
+      "grad_norm": 1.455143690109253,
+      "learning_rate": 1.0329094629147072e-05,
+      "loss": 0.1026,
+      "step": 4129
+    },
+    {
+      "epoch": 2.1629125196437924,
+      "loss_partial": 0.048095703125,
+      "step": 4129
+    },
+    {
+      "epoch": 2.1629125196437924,
+      "loss_partial": 0.05029296875,
+      "step": 4129
+    },
+    {
+      "epoch": 2.1634363541121004,
+      "grad_norm": 1.6273397207260132,
+      "learning_rate": 1.0324525486987784e-05,
+      "loss": 0.1035,
+      "step": 4130
+    },
+    {
+      "epoch": 2.1634363541121004,
+      "loss_partial": 0.05224609375,
+      "step": 4130
+    },
+    {
+      "epoch": 2.1634363541121004,
+      "loss_partial": 0.053466796875,
+      "step": 4130
+    },
+    {
+      "epoch": 2.1639601885804085,
+      "grad_norm": 1.3130093812942505,
+      "learning_rate": 1.0319956277004679e-05,
+      "loss": 0.1032,
+      "step": 4131
+    },
+    {
+      "epoch": 2.1639601885804085,
+      "loss_partial": 0.045654296875,
+      "step": 4131
+    },
+    {
+      "epoch": 2.1639601885804085,
+      "loss_partial": 0.043701171875,
+      "step": 4131
+    },
+    {
+      "epoch": 2.1644840230487166,
+      "grad_norm": 1.491435170173645,
+      "learning_rate": 1.031538700015269e-05,
+      "loss": 0.1031,
+      "step": 4132
+    },
+    {
+      "epoch": 2.1644840230487166,
+      "loss_partial": 0.0615234375,
+      "step": 4132
+    },
+    {
+      "epoch": 2.1644840230487166,
+      "loss_partial": 0.053955078125,
+      "step": 4132
+    },
+    {
+      "epoch": 2.1650078575170246,
+      "grad_norm": 1.689961314201355,
+      "learning_rate": 1.0310817657386768e-05,
+      "loss": 0.1051,
+      "step": 4133
+    },
+    {
+      "epoch": 2.1650078575170246,
+      "loss_partial": 0.053955078125,
+      "step": 4133
+    },
+    {
+      "epoch": 2.1650078575170246,
+      "loss_partial": 0.05517578125,
+      "step": 4133
+    },
+    {
+      "epoch": 2.1655316919853327,
+      "grad_norm": 1.5102492570877075,
+      "learning_rate": 1.0306248249661876e-05,
+      "loss": 0.1043,
+      "step": 4134
+    },
+    {
+      "epoch": 2.1655316919853327,
+      "loss_partial": 0.045654296875,
+      "step": 4134
+    },
+    {
+      "epoch": 2.1655316919853327,
+      "loss_partial": 0.043701171875,
+      "step": 4134
+    },
+    {
+      "epoch": 2.1660555264536407,
+      "grad_norm": 1.786487102508545,
+      "learning_rate": 1.0301678777932994e-05,
+      "loss": 0.104,
+      "step": 4135
+    },
+    {
+      "epoch": 2.1660555264536407,
+      "loss_partial": 0.04736328125,
+      "step": 4135
+    },
+    {
+      "epoch": 2.1660555264536407,
+      "loss_partial": 0.046875,
+      "step": 4135
+    },
+    {
+      "epoch": 2.166579360921949,
+      "grad_norm": 1.5625617504119873,
+      "learning_rate": 1.0297109243155112e-05,
+      "loss": 0.1042,
+      "step": 4136
+    },
+    {
+      "epoch": 2.166579360921949,
+      "loss_partial": 0.04833984375,
+      "step": 4136
+    },
+    {
+      "epoch": 2.166579360921949,
+      "loss_partial": 0.0458984375,
+      "step": 4136
+    },
+    {
+      "epoch": 2.167103195390257,
+      "grad_norm": 1.8912571668624878,
+      "learning_rate": 1.0292539646283239e-05,
+      "loss": 0.1031,
+      "step": 4137
+    },
+    {
+      "epoch": 2.167103195390257,
+      "loss_partial": 0.05029296875,
+      "step": 4137
+    },
+    {
+      "epoch": 2.167103195390257,
+      "loss_partial": 0.046875,
+      "step": 4137
+    },
+    {
+      "epoch": 2.167627029858565,
+      "grad_norm": 2.04268479347229,
+      "learning_rate": 1.0287969988272388e-05,
+      "loss": 0.1017,
+      "step": 4138
+    },
+    {
+      "epoch": 2.167627029858565,
+      "loss_partial": 0.047607421875,
+      "step": 4138
+    },
+    {
+      "epoch": 2.167627029858565,
+      "loss_partial": 0.059814453125,
+      "step": 4138
+    },
+    {
+      "epoch": 2.1681508643268725,
+      "grad_norm": 1.7593014240264893,
+      "learning_rate": 1.0283400270077587e-05,
+      "loss": 0.0997,
+      "step": 4139
+    },
+    {
+      "epoch": 2.1681508643268725,
+      "loss_partial": 0.0380859375,
+      "step": 4139
+    },
+    {
+      "epoch": 2.1681508643268725,
+      "loss_partial": 0.052490234375,
+      "step": 4139
+    },
+    {
+      "epoch": 2.1686746987951806,
+      "grad_norm": 1.2966252565383911,
+      "learning_rate": 1.0278830492653886e-05,
+      "loss": 0.1023,
+      "step": 4140
+    },
+    {
+      "epoch": 2.1686746987951806,
+      "loss_partial": 0.05517578125,
+      "step": 4140
+    },
+    {
+      "epoch": 2.1686746987951806,
+      "loss_partial": 0.052001953125,
+      "step": 4140
+    },
+    {
+      "epoch": 2.1691985332634887,
+      "grad_norm": 1.5664851665496826,
+      "learning_rate": 1.0274260656956331e-05,
+      "loss": 0.0996,
+      "step": 4141
+    },
+    {
+      "epoch": 2.1691985332634887,
+      "loss_partial": 0.052001953125,
+      "step": 4141
+    },
+    {
+      "epoch": 2.1691985332634887,
+      "loss_partial": 0.051025390625,
+      "step": 4141
+    },
+    {
+      "epoch": 2.1697223677317967,
+      "grad_norm": 1.2041106224060059,
+      "learning_rate": 1.026969076394e-05,
+      "loss": 0.1036,
+      "step": 4142
+    },
+    {
+      "epoch": 2.1697223677317967,
+      "loss_partial": 0.049560546875,
+      "step": 4142
+    },
+    {
+      "epoch": 2.1697223677317967,
+      "loss_partial": 0.0419921875,
+      "step": 4142
+    },
+    {
+      "epoch": 2.170246202200105,
+      "grad_norm": 1.3477574586868286,
+      "learning_rate": 1.0265120814559966e-05,
+      "loss": 0.1026,
+      "step": 4143
+    },
+    {
+      "epoch": 2.170246202200105,
+      "loss_partial": 0.048583984375,
+      "step": 4143
+    },
+    {
+      "epoch": 2.170246202200105,
+      "loss_partial": 0.0615234375,
+      "step": 4143
+    },
+    {
+      "epoch": 2.170770036668413,
+      "grad_norm": 1.2197036743164062,
+      "learning_rate": 1.0260550809771322e-05,
+      "loss": 0.1039,
+      "step": 4144
+    },
+    {
+      "epoch": 2.170770036668413,
+      "loss_partial": 0.053955078125,
+      "step": 4144
+    },
+    {
+      "epoch": 2.170770036668413,
+      "loss_partial": 0.0556640625,
+      "step": 4144
+    },
+    {
+      "epoch": 2.171293871136721,
+      "grad_norm": 1.2471767663955688,
+      "learning_rate": 1.0255980750529169e-05,
+      "loss": 0.0991,
+      "step": 4145
+    },
+    {
+      "epoch": 2.171293871136721,
+      "loss_partial": 0.059326171875,
+      "step": 4145
+    },
+    {
+      "epoch": 2.171293871136721,
+      "loss_partial": 0.0439453125,
+      "step": 4145
+    },
+    {
+      "epoch": 2.171817705605029,
+      "grad_norm": 1.333691954612732,
+      "learning_rate": 1.0251410637788618e-05,
+      "loss": 0.1031,
+      "step": 4146
+    },
+    {
+      "epoch": 2.171817705605029,
+      "loss_partial": 0.056884765625,
+      "step": 4146
+    },
+    {
+      "epoch": 2.171817705605029,
+      "loss_partial": 0.05322265625,
+      "step": 4146
+    },
+    {
+      "epoch": 2.172341540073337,
+      "grad_norm": 1.3683232069015503,
+      "learning_rate": 1.0246840472504805e-05,
+      "loss": 0.1006,
+      "step": 4147
+    },
+    {
+      "epoch": 2.172341540073337,
+      "loss_partial": 0.04931640625,
+      "step": 4147
+    },
+    {
+      "epoch": 2.172341540073337,
+      "loss_partial": 0.06396484375,
+      "step": 4147
+    },
+    {
+      "epoch": 2.1728653745416446,
+      "grad_norm": 1.1487563848495483,
+      "learning_rate": 1.0242270255632854e-05,
+      "loss": 0.1015,
+      "step": 4148
+    },
+    {
+      "epoch": 2.1728653745416446,
+      "loss_partial": 0.046875,
+      "step": 4148
+    },
+    {
+      "epoch": 2.1728653745416446,
+      "loss_partial": 0.056884765625,
+      "step": 4148
+    },
+    {
+      "epoch": 2.1733892090099527,
+      "grad_norm": 1.7154943943023682,
+      "learning_rate": 1.023769998812792e-05,
+      "loss": 0.106,
+      "step": 4149
+    },
+    {
+      "epoch": 2.1733892090099527,
+      "loss_partial": 0.051025390625,
+      "step": 4149
+    },
+    {
+      "epoch": 2.1733892090099527,
+      "loss_partial": 0.04736328125,
+      "step": 4149
+    },
+    {
+      "epoch": 2.1739130434782608,
+      "grad_norm": 1.5217820405960083,
+      "learning_rate": 1.0233129670945161e-05,
+      "loss": 0.0998,
+      "step": 4150
+    },
+    {
+      "epoch": 2.1739130434782608,
+      "loss_partial": 0.04248046875,
+      "step": 4150
+    },
+    {
+      "epoch": 2.1739130434782608,
+      "loss_partial": 0.04052734375,
+      "step": 4150
+    },
+    {
+      "epoch": 2.174436877946569,
+      "grad_norm": 1.8495161533355713,
+      "learning_rate": 1.0228559305039741e-05,
+      "loss": 0.1078,
+      "step": 4151
+    },
+    {
+      "epoch": 2.174436877946569,
+      "loss_partial": 0.056396484375,
+      "step": 4151
+    },
+    {
+      "epoch": 2.174436877946569,
+      "loss_partial": 0.043212890625,
+      "step": 4151
+    },
+    {
+      "epoch": 2.174960712414877,
+      "grad_norm": 1.3380447626113892,
+      "learning_rate": 1.0223988891366836e-05,
+      "loss": 0.1038,
+      "step": 4152
+    },
+    {
+      "epoch": 2.174960712414877,
+      "loss_partial": 0.06103515625,
+      "step": 4152
+    },
+    {
+      "epoch": 2.174960712414877,
+      "loss_partial": 0.054931640625,
+      "step": 4152
+    },
+    {
+      "epoch": 2.175484546883185,
+      "grad_norm": 1.6905590295791626,
+      "learning_rate": 1.021941843088164e-05,
+      "loss": 0.1024,
+      "step": 4153
+    },
+    {
+      "epoch": 2.175484546883185,
+      "loss_partial": 0.046875,
+      "step": 4153
+    },
+    {
+      "epoch": 2.175484546883185,
+      "loss_partial": 0.0537109375,
+      "step": 4153
+    },
+    {
+      "epoch": 2.176008381351493,
+      "grad_norm": 1.459219217300415,
+      "learning_rate": 1.0214847924539348e-05,
+      "loss": 0.1016,
+      "step": 4154
+    },
+    {
+      "epoch": 2.176008381351493,
+      "loss_partial": 0.0458984375,
+      "step": 4154
+    },
+    {
+      "epoch": 2.176008381351493,
+      "loss_partial": 0.048095703125,
+      "step": 4154
+    },
+    {
+      "epoch": 2.176532215819801,
+      "grad_norm": 1.4688098430633545,
+      "learning_rate": 1.0210277373295172e-05,
+      "loss": 0.1031,
+      "step": 4155
+    },
+    {
+      "epoch": 2.176532215819801,
+      "loss_partial": 0.044921875,
+      "step": 4155
+    },
+    {
+      "epoch": 2.176532215819801,
+      "loss_partial": 0.052734375,
+      "step": 4155
+    },
+    {
+      "epoch": 2.177056050288109,
+      "grad_norm": 1.2336399555206299,
+      "learning_rate": 1.0205706778104322e-05,
+      "loss": 0.1009,
+      "step": 4156
+    },
+    {
+      "epoch": 2.177056050288109,
+      "loss_partial": 0.045166015625,
+      "step": 4156
+    },
+    {
+      "epoch": 2.177056050288109,
+      "loss_partial": 0.043701171875,
+      "step": 4156
+    },
+    {
+      "epoch": 2.177579884756417,
+      "grad_norm": 1.5236401557922363,
+      "learning_rate": 1.020113613992203e-05,
+      "loss": 0.1037,
+      "step": 4157
+    },
+    {
+      "epoch": 2.177579884756417,
+      "loss_partial": 0.0556640625,
+      "step": 4157
+    },
+    {
+      "epoch": 2.177579884756417,
+      "loss_partial": 0.0517578125,
+      "step": 4157
+    },
+    {
+      "epoch": 2.178103719224725,
+      "grad_norm": 1.7284578084945679,
+      "learning_rate": 1.0196565459703524e-05,
+      "loss": 0.1025,
+      "step": 4158
+    },
+    {
+      "epoch": 2.178103719224725,
+      "loss_partial": 0.046142578125,
+      "step": 4158
+    },
+    {
+      "epoch": 2.178103719224725,
+      "loss_partial": 0.051513671875,
+      "step": 4158
+    },
+    {
+      "epoch": 2.178627553693033,
+      "grad_norm": 1.633273959159851,
+      "learning_rate": 1.0191994738404057e-05,
+      "loss": 0.1038,
+      "step": 4159
+    },
+    {
+      "epoch": 2.178627553693033,
+      "loss_partial": 0.053466796875,
+      "step": 4159
+    },
+    {
+      "epoch": 2.178627553693033,
+      "loss_partial": 0.0537109375,
+      "step": 4159
+    },
+    {
+      "epoch": 2.179151388161341,
+      "grad_norm": 1.7244657278060913,
+      "learning_rate": 1.0187423976978873e-05,
+      "loss": 0.1009,
+      "step": 4160
+    },
+    {
+      "epoch": 2.179151388161341,
+      "loss_partial": 0.045166015625,
+      "step": 4160
+    },
+    {
+      "epoch": 2.179151388161341,
+      "loss_partial": 0.056640625,
+      "step": 4160
+    },
+    {
+      "epoch": 2.179675222629649,
+      "grad_norm": 1.6531956195831299,
+      "learning_rate": 1.018285317638324e-05,
+      "loss": 0.1009,
+      "step": 4161
+    },
+    {
+      "epoch": 2.179675222629649,
+      "loss_partial": 0.0498046875,
+      "step": 4161
+    },
+    {
+      "epoch": 2.179675222629649,
+      "loss_partial": 0.050537109375,
+      "step": 4161
+    },
+    {
+      "epoch": 2.180199057097957,
+      "grad_norm": 1.5959746837615967,
+      "learning_rate": 1.0178282337572424e-05,
+      "loss": 0.1031,
+      "step": 4162
+    },
+    {
+      "epoch": 2.180199057097957,
+      "loss_partial": 0.0537109375,
+      "step": 4162
+    },
+    {
+      "epoch": 2.180199057097957,
+      "loss_partial": 0.044921875,
+      "step": 4162
+    },
+    {
+      "epoch": 2.180722891566265,
+      "grad_norm": 1.197356939315796,
+      "learning_rate": 1.0173711461501702e-05,
+      "loss": 0.1033,
+      "step": 4163
+    },
+    {
+      "epoch": 2.180722891566265,
+      "loss_partial": 0.05078125,
+      "step": 4163
+    },
+    {
+      "epoch": 2.180722891566265,
+      "loss_partial": 0.053466796875,
+      "step": 4163
+    },
+    {
+      "epoch": 2.181246726034573,
+      "grad_norm": 1.0022531747817993,
+      "learning_rate": 1.0169140549126355e-05,
+      "loss": 0.1046,
+      "step": 4164
+    },
+    {
+      "epoch": 2.181246726034573,
+      "loss_partial": 0.048828125,
+      "step": 4164
+    },
+    {
+      "epoch": 2.181246726034573,
+      "loss_partial": 0.049072265625,
+      "step": 4164
+    },
+    {
+      "epoch": 2.181770560502881,
+      "grad_norm": 1.2660574913024902,
+      "learning_rate": 1.0164569601401684e-05,
+      "loss": 0.1013,
+      "step": 4165
+    },
+    {
+      "epoch": 2.181770560502881,
+      "loss_partial": 0.05419921875,
+      "step": 4165
+    },
+    {
+      "epoch": 2.181770560502881,
+      "loss_partial": 0.04638671875,
+      "step": 4165
+    },
+    {
+      "epoch": 2.1822943949711893,
+      "grad_norm": 1.4133031368255615,
+      "learning_rate": 1.015999861928298e-05,
+      "loss": 0.104,
+      "step": 4166
+    },
+    {
+      "epoch": 2.1822943949711893,
+      "loss_partial": 0.042724609375,
+      "step": 4166
+    },
+    {
+      "epoch": 2.1822943949711893,
+      "loss_partial": 0.053955078125,
+      "step": 4166
+    },
+    {
+      "epoch": 2.1828182294394973,
+      "grad_norm": 1.8802087306976318,
+      "learning_rate": 1.0155427603725555e-05,
+      "loss": 0.1004,
+      "step": 4167
+    },
+    {
+      "epoch": 2.1828182294394973,
+      "loss_partial": 0.049072265625,
+      "step": 4167
+    },
+    {
+      "epoch": 2.1828182294394973,
+      "loss_partial": 0.052001953125,
+      "step": 4167
+    },
+    {
+      "epoch": 2.183342063907805,
+      "grad_norm": 1.4965193271636963,
+      "learning_rate": 1.0150856555684724e-05,
+      "loss": 0.1,
+      "step": 4168
+    },
+    {
+      "epoch": 2.183342063907805,
+      "loss_partial": 0.045654296875,
+      "step": 4168
+    },
+    {
+      "epoch": 2.183342063907805,
+      "loss_partial": 0.058349609375,
+      "step": 4168
+    },
+    {
+      "epoch": 2.183865898376113,
+      "grad_norm": 1.9810736179351807,
+      "learning_rate": 1.0146285476115805e-05,
+      "loss": 0.1018,
+      "step": 4169
+    },
+    {
+      "epoch": 2.183865898376113,
+      "loss_partial": 0.046875,
+      "step": 4169
+    },
+    {
+      "epoch": 2.183865898376113,
+      "loss_partial": 0.05712890625,
+      "step": 4169
+    },
+    {
+      "epoch": 2.184389732844421,
+      "grad_norm": 1.4623470306396484,
+      "learning_rate": 1.0141714365974125e-05,
+      "loss": 0.1085,
+      "step": 4170
+    },
+    {
+      "epoch": 2.184389732844421,
+      "loss_partial": 0.06396484375,
+      "step": 4170
+    },
+    {
+      "epoch": 2.184389732844421,
+      "loss_partial": 0.05712890625,
+      "step": 4170
+    },
+    {
+      "epoch": 2.184913567312729,
+      "grad_norm": 1.730088710784912,
+      "learning_rate": 1.013714322621502e-05,
+      "loss": 0.1064,
+      "step": 4171
+    },
+    {
+      "epoch": 2.184913567312729,
+      "loss_partial": 0.0498046875,
+      "step": 4171
+    },
+    {
+      "epoch": 2.184913567312729,
+      "loss_partial": 0.04638671875,
+      "step": 4171
+    },
+    {
+      "epoch": 2.185437401781037,
+      "grad_norm": 1.700749158859253,
+      "learning_rate": 1.0132572057793826e-05,
+      "loss": 0.104,
+      "step": 4172
+    },
+    {
+      "epoch": 2.185437401781037,
+      "loss_partial": 0.05078125,
+      "step": 4172
+    },
+    {
+      "epoch": 2.185437401781037,
+      "loss_partial": 0.0478515625,
+      "step": 4172
+    },
+    {
+      "epoch": 2.1859612362493452,
+      "grad_norm": 1.6663877964019775,
+      "learning_rate": 1.0128000861665894e-05,
+      "loss": 0.1013,
+      "step": 4173
+    },
+    {
+      "epoch": 2.1859612362493452,
+      "loss_partial": 0.04541015625,
+      "step": 4173
+    },
+    {
+      "epoch": 2.1859612362493452,
+      "loss_partial": 0.052490234375,
+      "step": 4173
+    },
+    {
+      "epoch": 2.1864850707176533,
+      "grad_norm": 1.7678511142730713,
+      "learning_rate": 1.0123429638786575e-05,
+      "loss": 0.101,
+      "step": 4174
+    },
+    {
+      "epoch": 2.1864850707176533,
+      "loss_partial": 0.04052734375,
+      "step": 4174
+    },
+    {
+      "epoch": 2.1864850707176533,
+      "loss_partial": 0.05712890625,
+      "step": 4174
+    },
+    {
+      "epoch": 2.1870089051859614,
+      "grad_norm": 1.9411550760269165,
+      "learning_rate": 1.0118858390111223e-05,
+      "loss": 0.1056,
+      "step": 4175
+    },
+    {
+      "epoch": 2.1870089051859614,
+      "loss_partial": 0.04638671875,
+      "step": 4175
+    },
+    {
+      "epoch": 2.1870089051859614,
+      "loss_partial": 0.05322265625,
+      "step": 4175
+    },
+    {
+      "epoch": 2.1875327396542694,
+      "grad_norm": 1.293156623840332,
+      "learning_rate": 1.01142871165952e-05,
+      "loss": 0.1042,
+      "step": 4176
+    },
+    {
+      "epoch": 2.1875327396542694,
+      "loss_partial": 0.048095703125,
+      "step": 4176
+    },
+    {
+      "epoch": 2.1875327396542694,
+      "loss_partial": 0.058837890625,
+      "step": 4176
+    },
+    {
+      "epoch": 2.1880565741225775,
+      "grad_norm": 1.3686038255691528,
+      "learning_rate": 1.0109715819193881e-05,
+      "loss": 0.0994,
+      "step": 4177
+    },
+    {
+      "epoch": 2.1880565741225775,
+      "loss_partial": 0.048828125,
+      "step": 4177
+    },
+    {
+      "epoch": 2.1880565741225775,
+      "loss_partial": 0.0546875,
+      "step": 4177
+    },
+    {
+      "epoch": 2.188580408590885,
+      "grad_norm": 1.5656650066375732,
+      "learning_rate": 1.010514449886263e-05,
+      "loss": 0.103,
+      "step": 4178
+    },
+    {
+      "epoch": 2.188580408590885,
+      "loss_partial": 0.0498046875,
+      "step": 4178
+    },
+    {
+      "epoch": 2.188580408590885,
+      "loss_partial": 0.05322265625,
+      "step": 4178
+    },
+    {
+      "epoch": 2.189104243059193,
+      "grad_norm": 1.3511844873428345,
+      "learning_rate": 1.0100573156556833e-05,
+      "loss": 0.1013,
+      "step": 4179
+    },
+    {
+      "epoch": 2.189104243059193,
+      "loss_partial": 0.053466796875,
+      "step": 4179
+    },
+    {
+      "epoch": 2.189104243059193,
+      "loss_partial": 0.056884765625,
+      "step": 4179
+    },
+    {
+      "epoch": 2.189628077527501,
+      "grad_norm": 1.130759358406067,
+      "learning_rate": 1.0096001793231866e-05,
+      "loss": 0.1007,
+      "step": 4180
+    },
+    {
+      "epoch": 2.189628077527501,
+      "loss_partial": 0.06005859375,
+      "step": 4180
+    },
+    {
+      "epoch": 2.189628077527501,
+      "loss_partial": 0.0556640625,
+      "step": 4180
+    },
+    {
+      "epoch": 2.1901519119958093,
+      "grad_norm": 1.2209633588790894,
+      "learning_rate": 1.0091430409843119e-05,
+      "loss": 0.1069,
+      "step": 4181
+    },
+    {
+      "epoch": 2.1901519119958093,
+      "loss_partial": 0.044921875,
+      "step": 4181
+    },
+    {
+      "epoch": 2.1901519119958093,
+      "loss_partial": 0.060791015625,
+      "step": 4181
+    },
+    {
+      "epoch": 2.1906757464641173,
+      "grad_norm": 1.452519178390503,
+      "learning_rate": 1.0086859007345979e-05,
+      "loss": 0.1012,
+      "step": 4182
+    },
+    {
+      "epoch": 2.1906757464641173,
+      "loss_partial": 0.045166015625,
+      "step": 4182
+    },
+    {
+      "epoch": 2.1906757464641173,
+      "loss_partial": 0.0478515625,
+      "step": 4182
+    },
+    {
+      "epoch": 2.1911995809324254,
+      "grad_norm": 1.4211674928665161,
+      "learning_rate": 1.0082287586695847e-05,
+      "loss": 0.0996,
+      "step": 4183
+    },
+    {
+      "epoch": 2.1911995809324254,
+      "loss_partial": 0.04443359375,
+      "step": 4183
+    },
+    {
+      "epoch": 2.1911995809324254,
+      "loss_partial": 0.054443359375,
+      "step": 4183
+    },
+    {
+      "epoch": 2.1917234154007335,
+      "grad_norm": 1.7200795412063599,
+      "learning_rate": 1.0077716148848116e-05,
+      "loss": 0.1054,
+      "step": 4184
+    },
+    {
+      "epoch": 2.1917234154007335,
+      "loss_partial": 0.047607421875,
+      "step": 4184
+    },
+    {
+      "epoch": 2.1917234154007335,
+      "loss_partial": 0.05859375,
+      "step": 4184
+    },
+    {
+      "epoch": 2.1922472498690415,
+      "grad_norm": 1.6367071866989136,
+      "learning_rate": 1.0073144694758191e-05,
+      "loss": 0.1069,
+      "step": 4185
+    },
+    {
+      "epoch": 2.1922472498690415,
+      "loss_partial": 0.052490234375,
+      "step": 4185
+    },
+    {
+      "epoch": 2.1922472498690415,
+      "loss_partial": 0.057373046875,
+      "step": 4185
+    },
+    {
+      "epoch": 2.1927710843373496,
+      "grad_norm": 1.6790261268615723,
+      "learning_rate": 1.0068573225381478e-05,
+      "loss": 0.1023,
+      "step": 4186
+    },
+    {
+      "epoch": 2.1927710843373496,
+      "loss_partial": 0.041259765625,
+      "step": 4186
+    },
+    {
+      "epoch": 2.1927710843373496,
+      "loss_partial": 0.052490234375,
+      "step": 4186
+    },
+    {
+      "epoch": 2.193294918805657,
+      "grad_norm": 1.702471375465393,
+      "learning_rate": 1.0064001741673384e-05,
+      "loss": 0.1027,
+      "step": 4187
+    },
+    {
+      "epoch": 2.193294918805657,
+      "loss_partial": 0.0439453125,
+      "step": 4187
+    },
+    {
+      "epoch": 2.193294918805657,
+      "loss_partial": 0.046142578125,
+      "step": 4187
+    },
+    {
+      "epoch": 2.1938187532739652,
+      "grad_norm": 1.4961748123168945,
+      "learning_rate": 1.0059430244589321e-05,
+      "loss": 0.1021,
+      "step": 4188
+    },
+    {
+      "epoch": 2.1938187532739652,
+      "loss_partial": 0.0419921875,
+      "step": 4188
+    },
+    {
+      "epoch": 2.1938187532739652,
+      "loss_partial": 0.044921875,
+      "step": 4188
+    },
+    {
+      "epoch": 2.1943425877422733,
+      "grad_norm": 1.250600814819336,
+      "learning_rate": 1.0054858735084705e-05,
+      "loss": 0.0989,
+      "step": 4189
+    },
+    {
+      "epoch": 2.1943425877422733,
+      "loss_partial": 0.0576171875,
+      "step": 4189
+    },
+    {
+      "epoch": 2.1943425877422733,
+      "loss_partial": 0.046875,
+      "step": 4189
+    },
+    {
+      "epoch": 2.1948664222105814,
+      "grad_norm": 1.8705341815948486,
+      "learning_rate": 1.0050287214114954e-05,
+      "loss": 0.1048,
+      "step": 4190
+    },
+    {
+      "epoch": 2.1948664222105814,
+      "loss_partial": 0.045654296875,
+      "step": 4190
+    },
+    {
+      "epoch": 2.1948664222105814,
+      "loss_partial": 0.052490234375,
+      "step": 4190
+    },
+    {
+      "epoch": 2.1953902566788894,
+      "grad_norm": 1.4862220287322998,
+      "learning_rate": 1.0045715682635484e-05,
+      "loss": 0.1003,
+      "step": 4191
+    },
+    {
+      "epoch": 2.1953902566788894,
+      "loss_partial": 0.0458984375,
+      "step": 4191
+    },
+    {
+      "epoch": 2.1953902566788894,
+      "loss_partial": 0.044189453125,
+      "step": 4191
+    },
+    {
+      "epoch": 2.1959140911471975,
+      "grad_norm": 1.7467539310455322,
+      "learning_rate": 1.0041144141601721e-05,
+      "loss": 0.1035,
+      "step": 4192
+    },
+    {
+      "epoch": 2.1959140911471975,
+      "loss_partial": 0.047119140625,
+      "step": 4192
+    },
+    {
+      "epoch": 2.1959140911471975,
+      "loss_partial": 0.0576171875,
+      "step": 4192
+    },
+    {
+      "epoch": 2.1964379256155055,
+      "grad_norm": 1.2998385429382324,
+      "learning_rate": 1.0036572591969087e-05,
+      "loss": 0.0993,
+      "step": 4193
+    },
+    {
+      "epoch": 2.1964379256155055,
+      "loss_partial": 0.052490234375,
+      "step": 4193
+    },
+    {
+      "epoch": 2.1964379256155055,
+      "loss_partial": 0.0478515625,
+      "step": 4193
+    },
+    {
+      "epoch": 2.1969617600838136,
+      "grad_norm": 1.7096006870269775,
+      "learning_rate": 1.0032001034693003e-05,
+      "loss": 0.1025,
+      "step": 4194
+    },
+    {
+      "epoch": 2.1969617600838136,
+      "loss_partial": 0.051513671875,
+      "step": 4194
+    },
+    {
+      "epoch": 2.1969617600838136,
+      "loss_partial": 0.045166015625,
+      "step": 4194
+    },
+    {
+      "epoch": 2.1974855945521217,
+      "grad_norm": 1.2881338596343994,
+      "learning_rate": 1.0027429470728907e-05,
+      "loss": 0.1046,
+      "step": 4195
+    },
+    {
+      "epoch": 2.1974855945521217,
+      "loss_partial": 0.05029296875,
+      "step": 4195
+    },
+    {
+      "epoch": 2.1974855945521217,
+      "loss_partial": 0.044921875,
+      "step": 4195
+    },
+    {
+      "epoch": 2.1980094290204297,
+      "grad_norm": 2.125941038131714,
+      "learning_rate": 1.0022857901032217e-05,
+      "loss": 0.1037,
+      "step": 4196
+    },
+    {
+      "epoch": 2.1980094290204297,
+      "loss_partial": 0.051513671875,
+      "step": 4196
+    },
+    {
+      "epoch": 2.1980094290204297,
+      "loss_partial": 0.0439453125,
+      "step": 4196
+    },
+    {
+      "epoch": 2.198533263488738,
+      "grad_norm": 1.6261475086212158,
+      "learning_rate": 1.001828632655837e-05,
+      "loss": 0.1009,
+      "step": 4197
+    },
+    {
+      "epoch": 2.198533263488738,
+      "loss_partial": 0.05615234375,
+      "step": 4197
+    },
+    {
+      "epoch": 2.198533263488738,
+      "loss_partial": 0.048583984375,
+      "step": 4197
+    },
+    {
+      "epoch": 2.1990570979570454,
+      "grad_norm": 2.4941916465759277,
+      "learning_rate": 1.001371474826279e-05,
+      "loss": 0.1011,
+      "step": 4198
+    },
+    {
+      "epoch": 2.1990570979570454,
+      "loss_partial": 0.049560546875,
+      "step": 4198
+    },
+    {
+      "epoch": 2.1990570979570454,
+      "loss_partial": 0.0576171875,
+      "step": 4198
+    },
+    {
+      "epoch": 2.1995809324253535,
+      "grad_norm": 1.5362756252288818,
+      "learning_rate": 1.0009143167100916e-05,
+      "loss": 0.1026,
+      "step": 4199
+    },
+    {
+      "epoch": 2.1995809324253535,
+      "loss_partial": 0.056884765625,
+      "step": 4199
+    },
+    {
+      "epoch": 2.1995809324253535,
+      "loss_partial": 0.053955078125,
+      "step": 4199
+    },
+    {
+      "epoch": 2.2001047668936615,
+      "grad_norm": 1.9084261655807495,
+      "learning_rate": 1.0004571584028173e-05,
+      "loss": 0.102,
+      "step": 4200
+    },
+    {
+      "epoch": 2.2001047668936615,
+      "loss_partial": 0.05126953125,
+      "step": 4200
+    },
+    {
+      "epoch": 2.2001047668936615,
+      "loss_partial": 0.047119140625,
+      "step": 4200
+    },
+    {
+      "epoch": 2.2006286013619696,
+      "grad_norm": 1.5518667697906494,
+      "learning_rate": 1e-05,
+      "loss": 0.107,
+      "step": 4201
+    },
+    {
+      "epoch": 2.2006286013619696,
+      "loss_partial": 0.047607421875,
+      "step": 4201
+    },
+    {
+      "epoch": 2.2006286013619696,
+      "loss_partial": 0.04638671875,
+      "step": 4201
+    },
+    {
+      "epoch": 2.2011524358302776,
+      "grad_norm": 1.4603468179702759,
+      "learning_rate": 9.995428415971828e-06,
+      "loss": 0.1008,
+      "step": 4202
+    },
+    {
+      "epoch": 2.2011524358302776,
+      "loss_partial": 0.054931640625,
+      "step": 4202
+    },
+    {
+      "epoch": 2.2011524358302776,
+      "loss_partial": 0.047119140625,
+      "step": 4202
+    },
+    {
+      "epoch": 2.2016762702985857,
+      "grad_norm": 1.9416978359222412,
+      "learning_rate": 9.990856832899089e-06,
+      "loss": 0.1044,
+      "step": 4203
+    },
+    {
+      "epoch": 2.2016762702985857,
+      "loss_partial": 0.053466796875,
+      "step": 4203
+    },
+    {
+      "epoch": 2.2016762702985857,
+      "loss_partial": 0.050537109375,
+      "step": 4203
+    },
+    {
+      "epoch": 2.2022001047668938,
+      "grad_norm": 1.5304337739944458,
+      "learning_rate": 9.986285251737211e-06,
+      "loss": 0.1074,
+      "step": 4204
+    },
+    {
+      "epoch": 2.2022001047668938,
+      "loss_partial": 0.0556640625,
+      "step": 4204
+    },
+    {
+      "epoch": 2.2022001047668938,
+      "loss_partial": 0.044921875,
+      "step": 4204
+    },
+    {
+      "epoch": 2.202723939235202,
+      "grad_norm": 1.9996459484100342,
+      "learning_rate": 9.981713673441634e-06,
+      "loss": 0.1045,
+      "step": 4205
+    },
+    {
+      "epoch": 2.202723939235202,
+      "loss_partial": 0.056640625,
+      "step": 4205
+    },
+    {
+      "epoch": 2.202723939235202,
+      "loss_partial": 0.048583984375,
+      "step": 4205
+    },
+    {
+      "epoch": 2.20324777370351,
+      "grad_norm": 1.41926908493042,
+      "learning_rate": 9.977142098967787e-06,
+      "loss": 0.1053,
+      "step": 4206
+    },
+    {
+      "epoch": 2.20324777370351,
+      "loss_partial": 0.054931640625,
+      "step": 4206
+    },
+    {
+      "epoch": 2.20324777370351,
+      "loss_partial": 0.056640625,
+      "step": 4206
+    },
+    {
+      "epoch": 2.2037716081718175,
+      "grad_norm": 1.7515085935592651,
+      "learning_rate": 9.972570529271098e-06,
+      "loss": 0.1008,
+      "step": 4207
+    },
+    {
+      "epoch": 2.2037716081718175,
+      "loss_partial": 0.047607421875,
+      "step": 4207
+    },
+    {
+      "epoch": 2.2037716081718175,
+      "loss_partial": 0.05078125,
+      "step": 4207
+    },
+    {
+      "epoch": 2.2042954426401256,
+      "grad_norm": 1.277738094329834,
+      "learning_rate": 9.967998965306997e-06,
+      "loss": 0.1031,
+      "step": 4208
+    },
+    {
+      "epoch": 2.2042954426401256,
+      "loss_partial": 0.052490234375,
+      "step": 4208
+    },
+    {
+      "epoch": 2.2042954426401256,
+      "loss_partial": 0.054931640625,
+      "step": 4208
+    },
+    {
+      "epoch": 2.2048192771084336,
+      "grad_norm": 2.136545181274414,
+      "learning_rate": 9.963427408030916e-06,
+      "loss": 0.1021,
+      "step": 4209
+    },
+    {
+      "epoch": 2.2048192771084336,
+      "loss_partial": 0.043212890625,
+      "step": 4209
+    },
+    {
+      "epoch": 2.2048192771084336,
+      "loss_partial": 0.046630859375,
+      "step": 4209
+    },
+    {
+      "epoch": 2.2053431115767417,
+      "grad_norm": 1.811760663986206,
+      "learning_rate": 9.958855858398282e-06,
+      "loss": 0.102,
+      "step": 4210
+    },
+    {
+      "epoch": 2.2053431115767417,
+      "loss_partial": 0.06298828125,
+      "step": 4210
+    },
+    {
+      "epoch": 2.2053431115767417,
+      "loss_partial": 0.054443359375,
+      "step": 4210
+    },
+    {
+      "epoch": 2.2058669460450497,
+      "grad_norm": 1.7451505661010742,
+      "learning_rate": 9.95428431736452e-06,
+      "loss": 0.1046,
+      "step": 4211
+    },
+    {
+      "epoch": 2.2058669460450497,
+      "loss_partial": 0.052734375,
+      "step": 4211
+    },
+    {
+      "epoch": 2.2058669460450497,
+      "loss_partial": 0.044189453125,
+      "step": 4211
+    },
+    {
+      "epoch": 2.206390780513358,
+      "grad_norm": 1.7915705442428589,
+      "learning_rate": 9.949712785885048e-06,
+      "loss": 0.1027,
+      "step": 4212
+    },
+    {
+      "epoch": 2.206390780513358,
+      "loss_partial": 0.050537109375,
+      "step": 4212
+    },
+    {
+      "epoch": 2.206390780513358,
+      "loss_partial": 0.050537109375,
+      "step": 4212
+    },
+    {
+      "epoch": 2.206914614981666,
+      "grad_norm": 2.1435928344726562,
+      "learning_rate": 9.945141264915297e-06,
+      "loss": 0.1033,
+      "step": 4213
+    },
+    {
+      "epoch": 2.206914614981666,
+      "loss_partial": 0.06005859375,
+      "step": 4213
+    },
+    {
+      "epoch": 2.206914614981666,
+      "loss_partial": 0.043701171875,
+      "step": 4213
+    },
+    {
+      "epoch": 2.207438449449974,
+      "grad_norm": 1.6082404851913452,
+      "learning_rate": 9.940569755410682e-06,
+      "loss": 0.1031,
+      "step": 4214
+    },
+    {
+      "epoch": 2.207438449449974,
+      "loss_partial": 0.04638671875,
+      "step": 4214
+    },
+    {
+      "epoch": 2.207438449449974,
+      "loss_partial": 0.0517578125,
+      "step": 4214
+    },
+    {
+      "epoch": 2.207962283918282,
+      "grad_norm": 2.077068328857422,
+      "learning_rate": 9.935998258326621e-06,
+      "loss": 0.1003,
+      "step": 4215
+    },
+    {
+      "epoch": 2.207962283918282,
+      "loss_partial": 0.051513671875,
+      "step": 4215
+    },
+    {
+      "epoch": 2.207962283918282,
+      "loss_partial": 0.050048828125,
+      "step": 4215
+    },
+    {
+      "epoch": 2.20848611838659,
+      "grad_norm": 1.7939000129699707,
+      "learning_rate": 9.931426774618526e-06,
+      "loss": 0.1011,
+      "step": 4216
+    },
+    {
+      "epoch": 2.20848611838659,
+      "loss_partial": 0.05322265625,
+      "step": 4216
+    },
+    {
+      "epoch": 2.20848611838659,
+      "loss_partial": 0.052734375,
+      "step": 4216
+    },
+    {
+      "epoch": 2.2090099528548977,
+      "grad_norm": 1.5068491697311401,
+      "learning_rate": 9.926855305241812e-06,
+      "loss": 0.1074,
+      "step": 4217
+    },
+    {
+      "epoch": 2.2090099528548977,
+      "loss_partial": 0.041259765625,
+      "step": 4217
+    },
+    {
+      "epoch": 2.2090099528548977,
+      "loss_partial": 0.045654296875,
+      "step": 4217
+    },
+    {
+      "epoch": 2.2095337873232057,
+      "grad_norm": 1.7598544359207153,
+      "learning_rate": 9.922283851151889e-06,
+      "loss": 0.1021,
+      "step": 4218
+    },
+    {
+      "epoch": 2.2095337873232057,
+      "loss_partial": 0.04296875,
+      "step": 4218
+    },
+    {
+      "epoch": 2.2095337873232057,
+      "loss_partial": 0.041259765625,
+      "step": 4218
+    },
+    {
+      "epoch": 2.2100576217915138,
+      "grad_norm": 1.736751914024353,
+      "learning_rate": 9.91771241330416e-06,
+      "loss": 0.101,
+      "step": 4219
+    },
+    {
+      "epoch": 2.2100576217915138,
+      "loss_partial": 0.050537109375,
+      "step": 4219
+    },
+    {
+      "epoch": 2.2100576217915138,
+      "loss_partial": 0.048828125,
+      "step": 4219
+    },
+    {
+      "epoch": 2.210581456259822,
+      "grad_norm": 1.7537963390350342,
+      "learning_rate": 9.913140992654021e-06,
+      "loss": 0.1029,
+      "step": 4220
+    },
+    {
+      "epoch": 2.210581456259822,
+      "loss_partial": 0.0498046875,
+      "step": 4220
+    },
+    {
+      "epoch": 2.210581456259822,
+      "loss_partial": 0.055908203125,
+      "step": 4220
+    },
+    {
+      "epoch": 2.21110529072813,
+      "grad_norm": 2.648059368133545,
+      "learning_rate": 9.908569590156883e-06,
+      "loss": 0.0999,
+      "step": 4221
+    },
+    {
+      "epoch": 2.21110529072813,
+      "loss_partial": 0.0498046875,
+      "step": 4221
+    },
+    {
+      "epoch": 2.21110529072813,
+      "loss_partial": 0.05029296875,
+      "step": 4221
+    },
+    {
+      "epoch": 2.211629125196438,
+      "grad_norm": 1.5873527526855469,
+      "learning_rate": 9.903998206768137e-06,
+      "loss": 0.1057,
+      "step": 4222
+    },
+    {
+      "epoch": 2.211629125196438,
+      "loss_partial": 0.052001953125,
+      "step": 4222
+    },
+    {
+      "epoch": 2.211629125196438,
+      "loss_partial": 0.051513671875,
+      "step": 4222
+    },
+    {
+      "epoch": 2.212152959664746,
+      "grad_norm": 2.3928439617156982,
+      "learning_rate": 9.89942684344317e-06,
+      "loss": 0.1045,
+      "step": 4223
+    },
+    {
+      "epoch": 2.212152959664746,
+      "loss_partial": 0.05029296875,
+      "step": 4223
+    },
+    {
+      "epoch": 2.212152959664746,
+      "loss_partial": 0.05126953125,
+      "step": 4223
+    },
+    {
+      "epoch": 2.212676794133054,
+      "grad_norm": 1.8606536388397217,
+      "learning_rate": 9.894855501137371e-06,
+      "loss": 0.1039,
+      "step": 4224
+    },
+    {
+      "epoch": 2.212676794133054,
+      "loss_partial": 0.053466796875,
+      "step": 4224
+    },
+    {
+      "epoch": 2.212676794133054,
+      "loss_partial": 0.04931640625,
+      "step": 4224
+    },
+    {
+      "epoch": 2.213200628601362,
+      "grad_norm": 1.574575424194336,
+      "learning_rate": 9.890284180806122e-06,
+      "loss": 0.1011,
+      "step": 4225
+    },
+    {
+      "epoch": 2.213200628601362,
+      "loss_partial": 0.052734375,
+      "step": 4225
+    },
+    {
+      "epoch": 2.213200628601362,
+      "loss_partial": 0.049560546875,
+      "step": 4225
+    },
+    {
+      "epoch": 2.21372446306967,
+      "grad_norm": 1.6959481239318848,
+      "learning_rate": 9.885712883404802e-06,
+      "loss": 0.0999,
+      "step": 4226
+    },
+    {
+      "epoch": 2.21372446306967,
+      "loss_partial": 0.0546875,
+      "step": 4226
+    },
+    {
+      "epoch": 2.21372446306967,
+      "loss_partial": 0.05029296875,
+      "step": 4226
+    },
+    {
+      "epoch": 2.214248297537978,
+      "grad_norm": 1.7416411638259888,
+      "learning_rate": 9.881141609888782e-06,
+      "loss": 0.1014,
+      "step": 4227
+    },
+    {
+      "epoch": 2.214248297537978,
+      "loss_partial": 0.046142578125,
+      "step": 4227
+    },
+    {
+      "epoch": 2.214248297537978,
+      "loss_partial": 0.03662109375,
+      "step": 4227
+    },
+    {
+      "epoch": 2.214772132006286,
+      "grad_norm": 1.6641134023666382,
+      "learning_rate": 9.876570361213429e-06,
+      "loss": 0.1017,
+      "step": 4228
+    },
+    {
+      "epoch": 2.214772132006286,
+      "loss_partial": 0.044189453125,
+      "step": 4228
+    },
+    {
+      "epoch": 2.214772132006286,
+      "loss_partial": 0.053955078125,
+      "step": 4228
+    },
+    {
+      "epoch": 2.215295966474594,
+      "grad_norm": 2.1777162551879883,
+      "learning_rate": 9.87199913833411e-06,
+      "loss": 0.1003,
+      "step": 4229
+    },
+    {
+      "epoch": 2.215295966474594,
+      "loss_partial": 0.04931640625,
+      "step": 4229
+    },
+    {
+      "epoch": 2.215295966474594,
+      "loss_partial": 0.061279296875,
+      "step": 4229
+    },
+    {
+      "epoch": 2.215819800942902,
+      "grad_norm": 1.8411035537719727,
+      "learning_rate": 9.867427942206178e-06,
+      "loss": 0.1057,
+      "step": 4230
+    },
+    {
+      "epoch": 2.215819800942902,
+      "loss_partial": 0.04638671875,
+      "step": 4230
+    },
+    {
+      "epoch": 2.215819800942902,
+      "loss_partial": 0.049560546875,
+      "step": 4230
+    },
+    {
+      "epoch": 2.21634363541121,
+      "grad_norm": 2.0156900882720947,
+      "learning_rate": 9.862856773784985e-06,
+      "loss": 0.1056,
+      "step": 4231
+    },
+    {
+      "epoch": 2.21634363541121,
+      "loss_partial": 0.047119140625,
+      "step": 4231
+    },
+    {
+      "epoch": 2.21634363541121,
+      "loss_partial": 0.04345703125,
+      "step": 4231
+    },
+    {
+      "epoch": 2.216867469879518,
+      "grad_norm": 1.8921122550964355,
+      "learning_rate": 9.858285634025876e-06,
+      "loss": 0.1002,
+      "step": 4232
+    },
+    {
+      "epoch": 2.216867469879518,
+      "loss_partial": 0.044189453125,
+      "step": 4232
+    },
+    {
+      "epoch": 2.216867469879518,
+      "loss_partial": 0.0654296875,
+      "step": 4232
+    },
+    {
+      "epoch": 2.217391304347826,
+      "grad_norm": 1.7969988584518433,
+      "learning_rate": 9.853714523884196e-06,
+      "loss": 0.1074,
+      "step": 4233
+    },
+    {
+      "epoch": 2.217391304347826,
+      "loss_partial": 0.050537109375,
+      "step": 4233
+    },
+    {
+      "epoch": 2.217391304347826,
+      "loss_partial": 0.051025390625,
+      "step": 4233
+    },
+    {
+      "epoch": 2.2179151388161342,
+      "grad_norm": 1.4230087995529175,
+      "learning_rate": 9.849143444315279e-06,
+      "loss": 0.1009,
+      "step": 4234
+    },
+    {
+      "epoch": 2.2179151388161342,
+      "loss_partial": 0.05078125,
+      "step": 4234
+    },
+    {
+      "epoch": 2.2179151388161342,
+      "loss_partial": 0.05859375,
+      "step": 4234
+    },
+    {
+      "epoch": 2.2184389732844423,
+      "grad_norm": 1.3830294609069824,
+      "learning_rate": 9.844572396274446e-06,
+      "loss": 0.103,
+      "step": 4235
+    },
+    {
+      "epoch": 2.2184389732844423,
+      "loss_partial": 0.0673828125,
+      "step": 4235
+    },
+    {
+      "epoch": 2.2184389732844423,
+      "loss_partial": 0.058349609375,
+      "step": 4235
+    },
+    {
+      "epoch": 2.2189628077527503,
+      "grad_norm": 1.8377314805984497,
+      "learning_rate": 9.840001380717022e-06,
+      "loss": 0.1046,
+      "step": 4236
+    },
+    {
+      "epoch": 2.2189628077527503,
+      "loss_partial": 0.048583984375,
+      "step": 4236
+    },
+    {
+      "epoch": 2.2189628077527503,
+      "loss_partial": 0.04638671875,
+      "step": 4236
+    },
+    {
+      "epoch": 2.219486642221058,
+      "grad_norm": 1.7613424062728882,
+      "learning_rate": 9.835430398598319e-06,
+      "loss": 0.1082,
+      "step": 4237
+    },
+    {
+      "epoch": 2.219486642221058,
+      "loss_partial": 0.05224609375,
+      "step": 4237
+    },
+    {
+      "epoch": 2.219486642221058,
+      "loss_partial": 0.055908203125,
+      "step": 4237
+    },
+    {
+      "epoch": 2.220010476689366,
+      "grad_norm": 1.9182441234588623,
+      "learning_rate": 9.830859450873646e-06,
+      "loss": 0.1048,
+      "step": 4238
+    },
+    {
+      "epoch": 2.220010476689366,
+      "loss_partial": 0.051513671875,
+      "step": 4238
+    },
+    {
+      "epoch": 2.220010476689366,
+      "loss_partial": 0.047607421875,
+      "step": 4238
+    },
+    {
+      "epoch": 2.220534311157674,
+      "grad_norm": 1.3446003198623657,
+      "learning_rate": 9.826288538498303e-06,
+      "loss": 0.1002,
+      "step": 4239
+    },
+    {
+      "epoch": 2.220534311157674,
+      "loss_partial": 0.05908203125,
+      "step": 4239
+    },
+    {
+      "epoch": 2.220534311157674,
+      "loss_partial": 0.046142578125,
+      "step": 4239
+    },
+    {
+      "epoch": 2.221058145625982,
+      "grad_norm": 1.9064276218414307,
+      "learning_rate": 9.821717662427579e-06,
+      "loss": 0.1046,
+      "step": 4240
+    },
+    {
+      "epoch": 2.221058145625982,
+      "loss_partial": 0.055908203125,
+      "step": 4240
+    },
+    {
+      "epoch": 2.221058145625982,
+      "loss_partial": 0.046142578125,
+      "step": 4240
+    },
+    {
+      "epoch": 2.22158198009429,
+      "grad_norm": 1.4228261709213257,
+      "learning_rate": 9.817146823616763e-06,
+      "loss": 0.1017,
+      "step": 4241
+    },
+    {
+      "epoch": 2.22158198009429,
+      "loss_partial": 0.054443359375,
+      "step": 4241
+    },
+    {
+      "epoch": 2.22158198009429,
+      "loss_partial": 0.048828125,
+      "step": 4241
+    },
+    {
+      "epoch": 2.2221058145625983,
+      "grad_norm": 2.0816683769226074,
+      "learning_rate": 9.812576023021128e-06,
+      "loss": 0.1035,
+      "step": 4242
+    },
+    {
+      "epoch": 2.2221058145625983,
+      "loss_partial": 0.05029296875,
+      "step": 4242
+    },
+    {
+      "epoch": 2.2221058145625983,
+      "loss_partial": 0.04931640625,
+      "step": 4242
+    },
+    {
+      "epoch": 2.2226296490309063,
+      "grad_norm": 1.6847883462905884,
+      "learning_rate": 9.808005261595948e-06,
+      "loss": 0.1007,
+      "step": 4243
+    },
+    {
+      "epoch": 2.2226296490309063,
+      "loss_partial": 0.0458984375,
+      "step": 4243
+    },
+    {
+      "epoch": 2.2226296490309063,
+      "loss_partial": 0.049072265625,
+      "step": 4243
+    },
+    {
+      "epoch": 2.2231534834992144,
+      "grad_norm": 1.7656292915344238,
+      "learning_rate": 9.803434540296475e-06,
+      "loss": 0.0963,
+      "step": 4244
+    },
+    {
+      "epoch": 2.2231534834992144,
+      "loss_partial": 0.048095703125,
+      "step": 4244
+    },
+    {
+      "epoch": 2.2231534834992144,
+      "loss_partial": 0.055419921875,
+      "step": 4244
+    },
+    {
+      "epoch": 2.2236773179675224,
+      "grad_norm": 2.0851595401763916,
+      "learning_rate": 9.798863860077972e-06,
+      "loss": 0.1055,
+      "step": 4245
+    },
+    {
+      "epoch": 2.2236773179675224,
+      "loss_partial": 0.051025390625,
+      "step": 4245
+    },
+    {
+      "epoch": 2.2236773179675224,
+      "loss_partial": 0.0458984375,
+      "step": 4245
+    },
+    {
+      "epoch": 2.22420115243583,
+      "grad_norm": 1.452027678489685,
+      "learning_rate": 9.79429322189568e-06,
+      "loss": 0.1069,
+      "step": 4246
+    },
+    {
+      "epoch": 2.22420115243583,
+      "loss_partial": 0.0478515625,
+      "step": 4246
+    },
+    {
+      "epoch": 2.22420115243583,
+      "loss_partial": 0.0615234375,
+      "step": 4246
+    },
+    {
+      "epoch": 2.224724986904138,
+      "grad_norm": 1.3154654502868652,
+      "learning_rate": 9.789722626704832e-06,
+      "loss": 0.1008,
+      "step": 4247
+    },
+    {
+      "epoch": 2.224724986904138,
+      "loss_partial": 0.054443359375,
+      "step": 4247
+    },
+    {
+      "epoch": 2.224724986904138,
+      "loss_partial": 0.06396484375,
+      "step": 4247
+    },
+    {
+      "epoch": 2.225248821372446,
+      "grad_norm": 1.0848314762115479,
+      "learning_rate": 9.785152075460652e-06,
+      "loss": 0.107,
+      "step": 4248
+    },
+    {
+      "epoch": 2.225248821372446,
+      "loss_partial": 0.04150390625,
+      "step": 4248
+    },
+    {
+      "epoch": 2.225248821372446,
+      "loss_partial": 0.053955078125,
+      "step": 4248
+    },
+    {
+      "epoch": 2.2257726558407542,
+      "grad_norm": 1.4409408569335938,
+      "learning_rate": 9.780581569118361e-06,
+      "loss": 0.098,
+      "step": 4249
+    },
+    {
+      "epoch": 2.2257726558407542,
+      "loss_partial": 0.04638671875,
+      "step": 4249
+    },
+    {
+      "epoch": 2.2257726558407542,
+      "loss_partial": 0.0478515625,
+      "step": 4249
+    },
+    {
+      "epoch": 2.2262964903090623,
+      "grad_norm": 1.4773473739624023,
+      "learning_rate": 9.776011108633167e-06,
+      "loss": 0.0998,
+      "step": 4250
+    },
+    {
+      "epoch": 2.2262964903090623,
+      "loss_partial": 0.052490234375,
+      "step": 4250
+    },
+    {
+      "epoch": 2.2262964903090623,
+      "loss_partial": 0.0458984375,
+      "step": 4250
+    },
+    {
+      "epoch": 2.2268203247773704,
+      "grad_norm": 1.6622670888900757,
+      "learning_rate": 9.771440694960264e-06,
+      "loss": 0.1005,
+      "step": 4251
+    },
+    {
+      "epoch": 2.2268203247773704,
+      "loss_partial": 0.05078125,
+      "step": 4251
+    },
+    {
+      "epoch": 2.2268203247773704,
+      "loss_partial": 0.0478515625,
+      "step": 4251
+    },
+    {
+      "epoch": 2.2273441592456784,
+      "grad_norm": 1.269511342048645,
+      "learning_rate": 9.766870329054844e-06,
+      "loss": 0.1003,
+      "step": 4252
+    },
+    {
+      "epoch": 2.2273441592456784,
+      "loss_partial": 0.052490234375,
+      "step": 4252
+    },
+    {
+      "epoch": 2.2273441592456784,
+      "loss_partial": 0.049072265625,
+      "step": 4252
+    },
+    {
+      "epoch": 2.2278679937139865,
+      "grad_norm": 1.2537087202072144,
+      "learning_rate": 9.762300011872083e-06,
+      "loss": 0.103,
+      "step": 4253
+    },
+    {
+      "epoch": 2.2278679937139865,
+      "loss_partial": 0.04736328125,
+      "step": 4253
+    },
+    {
+      "epoch": 2.2278679937139865,
+      "loss_partial": 0.0498046875,
+      "step": 4253
+    },
+    {
+      "epoch": 2.2283918281822945,
+      "grad_norm": 1.6222795248031616,
+      "learning_rate": 9.757729744367149e-06,
+      "loss": 0.0983,
+      "step": 4254
+    },
+    {
+      "epoch": 2.2283918281822945,
+      "loss_partial": 0.054443359375,
+      "step": 4254
+    },
+    {
+      "epoch": 2.2283918281822945,
+      "loss_partial": 0.062255859375,
+      "step": 4254
+    },
+    {
+      "epoch": 2.2289156626506026,
+      "grad_norm": 1.5473613739013672,
+      "learning_rate": 9.753159527495202e-06,
+      "loss": 0.1079,
+      "step": 4255
+    },
+    {
+      "epoch": 2.2289156626506026,
+      "loss_partial": 0.05908203125,
+      "step": 4255
+    },
+    {
+      "epoch": 2.2289156626506026,
+      "loss_partial": 0.05029296875,
+      "step": 4255
+    },
+    {
+      "epoch": 2.2294394971189107,
+      "grad_norm": 1.1632676124572754,
+      "learning_rate": 9.748589362211382e-06,
+      "loss": 0.1031,
+      "step": 4256
+    },
+    {
+      "epoch": 2.2294394971189107,
+      "loss_partial": 0.0517578125,
+      "step": 4256
+    },
+    {
+      "epoch": 2.2294394971189107,
+      "loss_partial": 0.052490234375,
+      "step": 4256
+    },
+    {
+      "epoch": 2.2299633315872183,
+      "grad_norm": 1.5409921407699585,
+      "learning_rate": 9.744019249470834e-06,
+      "loss": 0.108,
+      "step": 4257
+    },
+    {
+      "epoch": 2.2299633315872183,
+      "loss_partial": 0.05029296875,
+      "step": 4257
+    },
+    {
+      "epoch": 2.2299633315872183,
+      "loss_partial": 0.045166015625,
+      "step": 4257
+    },
+    {
+      "epoch": 2.2304871660555263,
+      "grad_norm": 1.2226970195770264,
+      "learning_rate": 9.739449190228681e-06,
+      "loss": 0.1,
+      "step": 4258
+    },
+    {
+      "epoch": 2.2304871660555263,
+      "loss_partial": 0.052490234375,
+      "step": 4258
+    },
+    {
+      "epoch": 2.2304871660555263,
+      "loss_partial": 0.0478515625,
+      "step": 4258
+    },
+    {
+      "epoch": 2.2310110005238344,
+      "grad_norm": 1.57381272315979,
+      "learning_rate": 9.734879185440037e-06,
+      "loss": 0.1026,
+      "step": 4259
+    },
+    {
+      "epoch": 2.2310110005238344,
+      "loss_partial": 0.054443359375,
+      "step": 4259
+    },
+    {
+      "epoch": 2.2310110005238344,
+      "loss_partial": 0.046630859375,
+      "step": 4259
+    },
+    {
+      "epoch": 2.2315348349921424,
+      "grad_norm": 1.468201756477356,
+      "learning_rate": 9.730309236060002e-06,
+      "loss": 0.1042,
+      "step": 4260
+    },
+    {
+      "epoch": 2.2315348349921424,
+      "loss_partial": 0.05810546875,
+      "step": 4260
+    },
+    {
+      "epoch": 2.2315348349921424,
+      "loss_partial": 0.046142578125,
+      "step": 4260
+    },
+    {
+      "epoch": 2.2320586694604505,
+      "grad_norm": 1.1426441669464111,
+      "learning_rate": 9.725739343043669e-06,
+      "loss": 0.1017,
+      "step": 4261
+    },
+    {
+      "epoch": 2.2320586694604505,
+      "loss_partial": 0.044677734375,
+      "step": 4261
+    },
+    {
+      "epoch": 2.2320586694604505,
+      "loss_partial": 0.06640625,
+      "step": 4261
+    },
+    {
+      "epoch": 2.2325825039287586,
+      "grad_norm": 1.5745383501052856,
+      "learning_rate": 9.721169507346118e-06,
+      "loss": 0.106,
+      "step": 4262
+    },
+    {
+      "epoch": 2.2325825039287586,
+      "loss_partial": 0.047119140625,
+      "step": 4262
+    },
+    {
+      "epoch": 2.2325825039287586,
+      "loss_partial": 0.050048828125,
+      "step": 4262
+    },
+    {
+      "epoch": 2.2331063383970666,
+      "grad_norm": 1.496314287185669,
+      "learning_rate": 9.716599729922416e-06,
+      "loss": 0.1002,
+      "step": 4263
+    },
+    {
+      "epoch": 2.2331063383970666,
+      "loss_partial": 0.05078125,
+      "step": 4263
+    },
+    {
+      "epoch": 2.2331063383970666,
+      "loss_partial": 0.0556640625,
+      "step": 4263
+    },
+    {
+      "epoch": 2.2336301728653747,
+      "grad_norm": 1.2780050039291382,
+      "learning_rate": 9.712030011727618e-06,
+      "loss": 0.1046,
+      "step": 4264
+    },
+    {
+      "epoch": 2.2336301728653747,
+      "loss_partial": 0.0556640625,
+      "step": 4264
+    },
+    {
+      "epoch": 2.2336301728653747,
+      "loss_partial": 0.049072265625,
+      "step": 4264
+    },
+    {
+      "epoch": 2.2341540073336827,
+      "grad_norm": 1.5265984535217285,
+      "learning_rate": 9.707460353716765e-06,
+      "loss": 0.1026,
+      "step": 4265
+    },
+    {
+      "epoch": 2.2341540073336827,
+      "loss_partial": 0.039794921875,
+      "step": 4265
+    },
+    {
+      "epoch": 2.2341540073336827,
+      "loss_partial": 0.055908203125,
+      "step": 4265
+    },
+    {
+      "epoch": 2.2346778418019904,
+      "grad_norm": 1.5688835382461548,
+      "learning_rate": 9.70289075684489e-06,
+      "loss": 0.1021,
+      "step": 4266
+    },
+    {
+      "epoch": 2.2346778418019904,
+      "loss_partial": 0.049560546875,
+      "step": 4266
+    },
+    {
+      "epoch": 2.2346778418019904,
+      "loss_partial": 0.05322265625,
+      "step": 4266
+    },
+    {
+      "epoch": 2.2352016762702984,
+      "grad_norm": 1.2929191589355469,
+      "learning_rate": 9.69832122206701e-06,
+      "loss": 0.1037,
+      "step": 4267
+    },
+    {
+      "epoch": 2.2352016762702984,
+      "loss_partial": 0.052001953125,
+      "step": 4267
+    },
+    {
+      "epoch": 2.2352016762702984,
+      "loss_partial": 0.0615234375,
+      "step": 4267
+    },
+    {
+      "epoch": 2.2357255107386065,
+      "grad_norm": 1.7161024808883667,
+      "learning_rate": 9.693751750338129e-06,
+      "loss": 0.1071,
+      "step": 4268
+    },
+    {
+      "epoch": 2.2357255107386065,
+      "loss_partial": 0.045654296875,
+      "step": 4268
+    },
+    {
+      "epoch": 2.2357255107386065,
+      "loss_partial": 0.05078125,
+      "step": 4268
+    },
+    {
+      "epoch": 2.2362493452069145,
+      "grad_norm": 1.428708553314209,
+      "learning_rate": 9.689182342613234e-06,
+      "loss": 0.1035,
+      "step": 4269
+    },
+    {
+      "epoch": 2.2362493452069145,
+      "loss_partial": 0.054443359375,
+      "step": 4269
+    },
+    {
+      "epoch": 2.2362493452069145,
+      "loss_partial": 0.05419921875,
+      "step": 4269
+    },
+    {
+      "epoch": 2.2367731796752226,
+      "grad_norm": 2.087587594985962,
+      "learning_rate": 9.684612999847313e-06,
+      "loss": 0.1048,
+      "step": 4270
+    },
+    {
+      "epoch": 2.2367731796752226,
+      "loss_partial": 0.04736328125,
+      "step": 4270
+    },
+    {
+      "epoch": 2.2367731796752226,
+      "loss_partial": 0.051513671875,
+      "step": 4270
+    },
+    {
+      "epoch": 2.2372970141435307,
+      "grad_norm": 1.2033672332763672,
+      "learning_rate": 9.680043722995324e-06,
+      "loss": 0.1006,
+      "step": 4271
+    },
+    {
+      "epoch": 2.2372970141435307,
+      "loss_partial": 0.045654296875,
+      "step": 4271
+    },
+    {
+      "epoch": 2.2372970141435307,
+      "loss_partial": 0.058349609375,
+      "step": 4271
+    },
+    {
+      "epoch": 2.2378208486118387,
+      "grad_norm": 1.8276952505111694,
+      "learning_rate": 9.675474513012218e-06,
+      "loss": 0.1039,
+      "step": 4272
+    },
+    {
+      "epoch": 2.2378208486118387,
+      "loss_partial": 0.046630859375,
+      "step": 4272
+    },
+    {
+      "epoch": 2.2378208486118387,
+      "loss_partial": 0.043212890625,
+      "step": 4272
+    },
+    {
+      "epoch": 2.238344683080147,
+      "grad_norm": 1.48714280128479,
+      "learning_rate": 9.67090537085293e-06,
+      "loss": 0.1027,
+      "step": 4273
+    },
+    {
+      "epoch": 2.238344683080147,
+      "loss_partial": 0.05078125,
+      "step": 4273
+    },
+    {
+      "epoch": 2.238344683080147,
+      "loss_partial": 0.047607421875,
+      "step": 4273
+    },
+    {
+      "epoch": 2.238868517548455,
+      "grad_norm": 1.7037454843521118,
+      "learning_rate": 9.666336297472387e-06,
+      "loss": 0.1038,
+      "step": 4274
+    },
+    {
+      "epoch": 2.238868517548455,
+      "loss_partial": 0.052978515625,
+      "step": 4274
+    },
+    {
+      "epoch": 2.238868517548455,
+      "loss_partial": 0.052734375,
+      "step": 4274
+    },
+    {
+      "epoch": 2.239392352016763,
+      "grad_norm": 1.56212317943573,
+      "learning_rate": 9.661767293825493e-06,
+      "loss": 0.0997,
+      "step": 4275
+    },
+    {
+      "epoch": 2.239392352016763,
+      "loss_partial": 0.045166015625,
+      "step": 4275
+    },
+    {
+      "epoch": 2.239392352016763,
+      "loss_partial": 0.0478515625,
+      "step": 4275
+    },
+    {
+      "epoch": 2.2399161864850705,
+      "grad_norm": 1.7166409492492676,
+      "learning_rate": 9.657198360867145e-06,
+      "loss": 0.1003,
+      "step": 4276
+    },
+    {
+      "epoch": 2.2399161864850705,
+      "loss_partial": 0.05029296875,
+      "step": 4276
+    },
+    {
+      "epoch": 2.2399161864850705,
+      "loss_partial": 0.05224609375,
+      "step": 4276
+    },
+    {
+      "epoch": 2.2404400209533786,
+      "grad_norm": 1.4628703594207764,
+      "learning_rate": 9.652629499552216e-06,
+      "loss": 0.1011,
+      "step": 4277
+    },
+    {
+      "epoch": 2.2404400209533786,
+      "loss_partial": 0.052734375,
+      "step": 4277
+    },
+    {
+      "epoch": 2.2404400209533786,
+      "loss_partial": 0.04833984375,
+      "step": 4277
+    },
+    {
+      "epoch": 2.2409638554216866,
+      "grad_norm": 1.6617333889007568,
+      "learning_rate": 9.648060710835578e-06,
+      "loss": 0.1032,
+      "step": 4278
+    },
+    {
+      "epoch": 2.2409638554216866,
+      "loss_partial": 0.046142578125,
+      "step": 4278
+    },
+    {
+      "epoch": 2.2409638554216866,
+      "loss_partial": 0.054443359375,
+      "step": 4278
+    },
+    {
+      "epoch": 2.2414876898899947,
+      "grad_norm": 1.9579455852508545,
+      "learning_rate": 9.643491995672073e-06,
+      "loss": 0.1023,
+      "step": 4279
+    },
+    {
+      "epoch": 2.2414876898899947,
+      "loss_partial": 0.0458984375,
+      "step": 4279
+    },
+    {
+      "epoch": 2.2414876898899947,
+      "loss_partial": 0.059326171875,
+      "step": 4279
+    },
+    {
+      "epoch": 2.2420115243583028,
+      "grad_norm": 1.8604017496109009,
+      "learning_rate": 9.638923355016536e-06,
+      "loss": 0.1037,
+      "step": 4280
+    },
+    {
+      "epoch": 2.2420115243583028,
+      "loss_partial": 0.05029296875,
+      "step": 4280
+    },
+    {
+      "epoch": 2.2420115243583028,
+      "loss_partial": 0.053466796875,
+      "step": 4280
+    },
+    {
+      "epoch": 2.242535358826611,
+      "grad_norm": 1.8512917757034302,
+      "learning_rate": 9.634354789823783e-06,
+      "loss": 0.103,
+      "step": 4281
+    },
+    {
+      "epoch": 2.242535358826611,
+      "loss_partial": 0.04833984375,
+      "step": 4281
+    },
+    {
+      "epoch": 2.242535358826611,
+      "loss_partial": 0.0615234375,
+      "step": 4281
+    },
+    {
+      "epoch": 2.243059193294919,
+      "grad_norm": 1.5535701513290405,
+      "learning_rate": 9.629786301048619e-06,
+      "loss": 0.1021,
+      "step": 4282
+    },
+    {
+      "epoch": 2.243059193294919,
+      "loss_partial": 0.056884765625,
+      "step": 4282
+    },
+    {
+      "epoch": 2.243059193294919,
+      "loss_partial": 0.054931640625,
+      "step": 4282
+    },
+    {
+      "epoch": 2.243583027763227,
+      "grad_norm": 1.8543908596038818,
+      "learning_rate": 9.62521788964583e-06,
+      "loss": 0.1013,
+      "step": 4283
+    },
+    {
+      "epoch": 2.243583027763227,
+      "loss_partial": 0.046630859375,
+      "step": 4283
+    },
+    {
+      "epoch": 2.243583027763227,
+      "loss_partial": 0.049072265625,
+      "step": 4283
+    },
+    {
+      "epoch": 2.244106862231535,
+      "grad_norm": 1.514361023902893,
+      "learning_rate": 9.620649556570186e-06,
+      "loss": 0.102,
+      "step": 4284
+    },
+    {
+      "epoch": 2.244106862231535,
+      "loss_partial": 0.05712890625,
+      "step": 4284
+    },
+    {
+      "epoch": 2.244106862231535,
+      "loss_partial": 0.045166015625,
+      "step": 4284
+    },
+    {
+      "epoch": 2.244630696699843,
+      "grad_norm": 1.4289398193359375,
+      "learning_rate": 9.616081302776436e-06,
+      "loss": 0.1034,
+      "step": 4285
+    },
+    {
+      "epoch": 2.244630696699843,
+      "loss_partial": 0.04833984375,
+      "step": 4285
+    },
+    {
+      "epoch": 2.244630696699843,
+      "loss_partial": 0.0439453125,
+      "step": 4285
+    },
+    {
+      "epoch": 2.2451545311681507,
+      "grad_norm": 1.848132848739624,
+      "learning_rate": 9.611513129219321e-06,
+      "loss": 0.0988,
+      "step": 4286
+    },
+    {
+      "epoch": 2.2451545311681507,
+      "loss_partial": 0.051025390625,
+      "step": 4286
+    },
+    {
+      "epoch": 2.2451545311681507,
+      "loss_partial": 0.05419921875,
+      "step": 4286
+    },
+    {
+      "epoch": 2.2456783656364587,
+      "grad_norm": 1.6531524658203125,
+      "learning_rate": 9.60694503685356e-06,
+      "loss": 0.103,
+      "step": 4287
+    },
+    {
+      "epoch": 2.2456783656364587,
+      "loss_partial": 0.05224609375,
+      "step": 4287
+    },
+    {
+      "epoch": 2.2456783656364587,
+      "loss_partial": 0.053955078125,
+      "step": 4287
+    },
+    {
+      "epoch": 2.246202200104767,
+      "grad_norm": 1.8191628456115723,
+      "learning_rate": 9.602377026633855e-06,
+      "loss": 0.1009,
+      "step": 4288
+    },
+    {
+      "epoch": 2.246202200104767,
+      "loss_partial": 0.04736328125,
+      "step": 4288
+    },
+    {
+      "epoch": 2.246202200104767,
+      "loss_partial": 0.049072265625,
+      "step": 4288
+    },
+    {
+      "epoch": 2.246726034573075,
+      "grad_norm": 1.7086400985717773,
+      "learning_rate": 9.597809099514892e-06,
+      "loss": 0.1027,
+      "step": 4289
+    },
+    {
+      "epoch": 2.246726034573075,
+      "loss_partial": 0.044189453125,
+      "step": 4289
+    },
+    {
+      "epoch": 2.246726034573075,
+      "loss_partial": 0.049072265625,
+      "step": 4289
+    },
+    {
+      "epoch": 2.247249869041383,
+      "grad_norm": 2.10903000831604,
+      "learning_rate": 9.593241256451341e-06,
+      "loss": 0.1033,
+      "step": 4290
+    },
+    {
+      "epoch": 2.247249869041383,
+      "loss_partial": 0.05078125,
+      "step": 4290
+    },
+    {
+      "epoch": 2.247249869041383,
+      "loss_partial": 0.05419921875,
+      "step": 4290
+    },
+    {
+      "epoch": 2.247773703509691,
+      "grad_norm": 1.5713423490524292,
+      "learning_rate": 9.588673498397852e-06,
+      "loss": 0.1048,
+      "step": 4291
+    },
+    {
+      "epoch": 2.247773703509691,
+      "loss_partial": 0.044189453125,
+      "step": 4291
+    },
+    {
+      "epoch": 2.247773703509691,
+      "loss_partial": 0.043701171875,
+      "step": 4291
+    },
+    {
+      "epoch": 2.248297537977999,
+      "grad_norm": 2.036210775375366,
+      "learning_rate": 9.584105826309059e-06,
+      "loss": 0.105,
+      "step": 4292
+    },
+    {
+      "epoch": 2.248297537977999,
+      "loss_partial": 0.04248046875,
+      "step": 4292
+    },
+    {
+      "epoch": 2.248297537977999,
+      "loss_partial": 0.052734375,
+      "step": 4292
+    },
+    {
+      "epoch": 2.248821372446307,
+      "grad_norm": 2.0108368396759033,
+      "learning_rate": 9.579538241139573e-06,
+      "loss": 0.1023,
+      "step": 4293
+    },
+    {
+      "epoch": 2.248821372446307,
+      "loss_partial": 0.053466796875,
+      "step": 4293
+    },
+    {
+      "epoch": 2.248821372446307,
+      "loss_partial": 0.05810546875,
+      "step": 4293
+    },
+    {
+      "epoch": 2.249345206914615,
+      "grad_norm": 1.6197173595428467,
+      "learning_rate": 9.574970743843997e-06,
+      "loss": 0.1036,
+      "step": 4294
+    },
+    {
+      "epoch": 2.249345206914615,
+      "loss_partial": 0.047607421875,
+      "step": 4294
+    },
+    {
+      "epoch": 2.249345206914615,
+      "loss_partial": 0.059326171875,
+      "step": 4294
+    },
+    {
+      "epoch": 2.249869041382923,
+      "grad_norm": 1.6124897003173828,
+      "learning_rate": 9.570403335376909e-06,
+      "loss": 0.0998,
+      "step": 4295
+    },
+    {
+      "epoch": 2.249869041382923,
+      "loss_partial": 0.052490234375,
+      "step": 4295
+    },
+    {
+      "epoch": 2.249869041382923,
+      "loss_partial": 0.04833984375,
+      "step": 4295
+    },
+    {
+      "epoch": 2.250392875851231,
+      "grad_norm": 1.472089171409607,
+      "learning_rate": 9.565836016692867e-06,
+      "loss": 0.1017,
+      "step": 4296
+    },
+    {
+      "epoch": 2.250392875851231,
+      "loss_partial": 0.04736328125,
+      "step": 4296
+    },
+    {
+      "epoch": 2.250392875851231,
+      "loss_partial": 0.05078125,
+      "step": 4296
+    },
+    {
+      "epoch": 2.250916710319539,
+      "grad_norm": 1.7796118259429932,
+      "learning_rate": 9.56126878874641e-06,
+      "loss": 0.1063,
+      "step": 4297
+    },
+    {
+      "epoch": 2.250916710319539,
+      "loss_partial": 0.0537109375,
+      "step": 4297
+    },
+    {
+      "epoch": 2.250916710319539,
+      "loss_partial": 0.04443359375,
+      "step": 4297
+    },
+    {
+      "epoch": 2.251440544787847,
+      "grad_norm": 1.5813599824905396,
+      "learning_rate": 9.556701652492063e-06,
+      "loss": 0.1057,
+      "step": 4298
+    },
+    {
+      "epoch": 2.251440544787847,
+      "loss_partial": 0.047607421875,
+      "step": 4298
+    },
+    {
+      "epoch": 2.251440544787847,
+      "loss_partial": 0.056640625,
+      "step": 4298
+    },
+    {
+      "epoch": 2.251964379256155,
+      "grad_norm": 2.0501110553741455,
+      "learning_rate": 9.552134608884333e-06,
+      "loss": 0.1021,
+      "step": 4299
+    },
+    {
+      "epoch": 2.251964379256155,
+      "loss_partial": 0.044921875,
+      "step": 4299
+    },
+    {
+      "epoch": 2.251964379256155,
+      "loss_partial": 0.0576171875,
+      "step": 4299
+    },
+    {
+      "epoch": 2.252488213724463,
+      "grad_norm": 2.0881588459014893,
+      "learning_rate": 9.547567658877698e-06,
+      "loss": 0.1077,
+      "step": 4300
+    },
+    {
+      "epoch": 2.252488213724463,
+      "loss_partial": 0.0439453125,
+      "step": 4300
+    },
+    {
+      "epoch": 2.252488213724463,
+      "loss_partial": 0.052001953125,
+      "step": 4300
+    },
+    {
+      "epoch": 2.253012048192771,
+      "grad_norm": 2.291334629058838,
+      "learning_rate": 9.543000803426622e-06,
+      "loss": 0.1019,
+      "step": 4301
+    },
+    {
+      "epoch": 2.253012048192771,
+      "loss_partial": 0.054931640625,
+      "step": 4301
+    },
+    {
+      "epoch": 2.253012048192771,
+      "loss_partial": 0.0517578125,
+      "step": 4301
+    },
+    {
+      "epoch": 2.253535882661079,
+      "grad_norm": 1.3388279676437378,
+      "learning_rate": 9.538434043485558e-06,
+      "loss": 0.1059,
+      "step": 4302
+    },
+    {
+      "epoch": 2.253535882661079,
+      "loss_partial": 0.04541015625,
+      "step": 4302
+    },
+    {
+      "epoch": 2.253535882661079,
+      "loss_partial": 0.049072265625,
+      "step": 4302
+    },
+    {
+      "epoch": 2.2540597171293872,
+      "grad_norm": 1.556854009628296,
+      "learning_rate": 9.533867380008923e-06,
+      "loss": 0.1034,
+      "step": 4303
+    },
+    {
+      "epoch": 2.2540597171293872,
+      "loss_partial": 0.05615234375,
+      "step": 4303
+    },
+    {
+      "epoch": 2.2540597171293872,
+      "loss_partial": 0.053955078125,
+      "step": 4303
+    },
+    {
+      "epoch": 2.2545835515976953,
+      "grad_norm": 2.064364433288574,
+      "learning_rate": 9.52930081395112e-06,
+      "loss": 0.1049,
+      "step": 4304
+    },
+    {
+      "epoch": 2.2545835515976953,
+      "loss_partial": 0.0537109375,
+      "step": 4304
+    },
+    {
+      "epoch": 2.2545835515976953,
+      "loss_partial": 0.052490234375,
+      "step": 4304
+    },
+    {
+      "epoch": 2.255107386066003,
+      "grad_norm": 2.458923816680908,
+      "learning_rate": 9.524734346266535e-06,
+      "loss": 0.1034,
+      "step": 4305
+    },
+    {
+      "epoch": 2.255107386066003,
+      "loss_partial": 0.045166015625,
+      "step": 4305
+    },
+    {
+      "epoch": 2.255107386066003,
+      "loss_partial": 0.049072265625,
+      "step": 4305
+    },
+    {
+      "epoch": 2.255631220534311,
+      "grad_norm": 2.1752748489379883,
+      "learning_rate": 9.520167977909536e-06,
+      "loss": 0.0992,
+      "step": 4306
+    },
+    {
+      "epoch": 2.255631220534311,
+      "loss_partial": 0.052490234375,
+      "step": 4306
+    },
+    {
+      "epoch": 2.255631220534311,
+      "loss_partial": 0.051025390625,
+      "step": 4306
+    },
+    {
+      "epoch": 2.256155055002619,
+      "grad_norm": 1.9594647884368896,
+      "learning_rate": 9.515601709834464e-06,
+      "loss": 0.0993,
+      "step": 4307
+    },
+    {
+      "epoch": 2.256155055002619,
+      "loss_partial": 0.04736328125,
+      "step": 4307
+    },
+    {
+      "epoch": 2.256155055002619,
+      "loss_partial": 0.0478515625,
+      "step": 4307
+    },
+    {
+      "epoch": 2.256678889470927,
+      "grad_norm": 2.081861734390259,
+      "learning_rate": 9.511035542995638e-06,
+      "loss": 0.0985,
+      "step": 4308
+    },
+    {
+      "epoch": 2.256678889470927,
+      "loss_partial": 0.04833984375,
+      "step": 4308
+    },
+    {
+      "epoch": 2.256678889470927,
+      "loss_partial": 0.046142578125,
+      "step": 4308
+    },
+    {
+      "epoch": 2.257202723939235,
+      "grad_norm": 1.864859700202942,
+      "learning_rate": 9.506469478347357e-06,
+      "loss": 0.1018,
+      "step": 4309
+    },
+    {
+      "epoch": 2.257202723939235,
+      "loss_partial": 0.05859375,
+      "step": 4309
+    },
+    {
+      "epoch": 2.257202723939235,
+      "loss_partial": 0.049072265625,
+      "step": 4309
+    },
+    {
+      "epoch": 2.257726558407543,
+      "grad_norm": 1.9281914234161377,
+      "learning_rate": 9.501903516843905e-06,
+      "loss": 0.1026,
+      "step": 4310
+    },
+    {
+      "epoch": 2.257726558407543,
+      "loss_partial": 0.04833984375,
+      "step": 4310
+    },
+    {
+      "epoch": 2.257726558407543,
+      "loss_partial": 0.058837890625,
+      "step": 4310
+    },
+    {
+      "epoch": 2.2582503928758513,
+      "grad_norm": 1.901439905166626,
+      "learning_rate": 9.497337659439538e-06,
+      "loss": 0.102,
+      "step": 4311
+    },
+    {
+      "epoch": 2.2582503928758513,
+      "loss_partial": 0.049560546875,
+      "step": 4311
+    },
+    {
+      "epoch": 2.2582503928758513,
+      "loss_partial": 0.06396484375,
+      "step": 4311
+    },
+    {
+      "epoch": 2.2587742273441593,
+      "grad_norm": 1.802533745765686,
+      "learning_rate": 9.492771907088492e-06,
+      "loss": 0.1022,
+      "step": 4312
+    },
+    {
+      "epoch": 2.2587742273441593,
+      "loss_partial": 0.052001953125,
+      "step": 4312
+    },
+    {
+      "epoch": 2.2587742273441593,
+      "loss_partial": 0.047119140625,
+      "step": 4312
+    },
+    {
+      "epoch": 2.2592980618124674,
+      "grad_norm": 1.9472684860229492,
+      "learning_rate": 9.488206260744978e-06,
+      "loss": 0.1049,
+      "step": 4313
+    },
+    {
+      "epoch": 2.2592980618124674,
+      "loss_partial": 0.058837890625,
+      "step": 4313
+    },
+    {
+      "epoch": 2.2592980618124674,
+      "loss_partial": 0.05126953125,
+      "step": 4313
+    },
+    {
+      "epoch": 2.2598218962807755,
+      "grad_norm": 1.4930506944656372,
+      "learning_rate": 9.483640721363195e-06,
+      "loss": 0.1038,
+      "step": 4314
+    },
+    {
+      "epoch": 2.2598218962807755,
+      "loss_partial": 0.046142578125,
+      "step": 4314
+    },
+    {
+      "epoch": 2.2598218962807755,
+      "loss_partial": 0.040771484375,
+      "step": 4314
+    },
+    {
+      "epoch": 2.2603457307490835,
+      "grad_norm": 1.352145791053772,
+      "learning_rate": 9.479075289897308e-06,
+      "loss": 0.0997,
+      "step": 4315
+    },
+    {
+      "epoch": 2.2603457307490835,
+      "loss_partial": 0.04638671875,
+      "step": 4315
+    },
+    {
+      "epoch": 2.2603457307490835,
+      "loss_partial": 0.048095703125,
+      "step": 4315
+    },
+    {
+      "epoch": 2.260869565217391,
+      "grad_norm": 1.4549720287322998,
+      "learning_rate": 9.474509967301463e-06,
+      "loss": 0.1029,
+      "step": 4316
+    },
+    {
+      "epoch": 2.260869565217391,
+      "loss_partial": 0.06494140625,
+      "step": 4316
+    },
+    {
+      "epoch": 2.260869565217391,
+      "loss_partial": 0.058349609375,
+      "step": 4316
+    },
+    {
+      "epoch": 2.261393399685699,
+      "grad_norm": 1.526616096496582,
+      "learning_rate": 9.469944754529784e-06,
+      "loss": 0.1038,
+      "step": 4317
+    },
+    {
+      "epoch": 2.261393399685699,
+      "loss_partial": 0.049072265625,
+      "step": 4317
+    },
+    {
+      "epoch": 2.261393399685699,
+      "loss_partial": 0.042724609375,
+      "step": 4317
+    },
+    {
+      "epoch": 2.2619172341540073,
+      "grad_norm": 1.373612403869629,
+      "learning_rate": 9.465379652536371e-06,
+      "loss": 0.0994,
+      "step": 4318
+    },
+    {
+      "epoch": 2.2619172341540073,
+      "loss_partial": 0.053466796875,
+      "step": 4318
+    },
+    {
+      "epoch": 2.2619172341540073,
+      "loss_partial": 0.047607421875,
+      "step": 4318
+    },
+    {
+      "epoch": 2.2624410686223153,
+      "grad_norm": 1.297000765800476,
+      "learning_rate": 9.46081466227531e-06,
+      "loss": 0.1017,
+      "step": 4319
+    },
+    {
+      "epoch": 2.2624410686223153,
+      "loss_partial": 0.047607421875,
+      "step": 4319
+    },
+    {
+      "epoch": 2.2624410686223153,
+      "loss_partial": 0.0537109375,
+      "step": 4319
+    },
+    {
+      "epoch": 2.2629649030906234,
+      "grad_norm": 1.5841641426086426,
+      "learning_rate": 9.456249784700648e-06,
+      "loss": 0.1024,
+      "step": 4320
+    },
+    {
+      "epoch": 2.2629649030906234,
+      "loss_partial": 0.049560546875,
+      "step": 4320
+    },
+    {
+      "epoch": 2.2629649030906234,
+      "loss_partial": 0.0517578125,
+      "step": 4320
+    },
+    {
+      "epoch": 2.2634887375589314,
+      "grad_norm": 1.5563751459121704,
+      "learning_rate": 9.45168502076642e-06,
+      "loss": 0.1037,
+      "step": 4321
+    },
+    {
+      "epoch": 2.2634887375589314,
+      "loss_partial": 0.057861328125,
+      "step": 4321
+    },
+    {
+      "epoch": 2.2634887375589314,
+      "loss_partial": 0.049072265625,
+      "step": 4321
+    },
+    {
+      "epoch": 2.2640125720272395,
+      "grad_norm": 1.6051068305969238,
+      "learning_rate": 9.44712037142663e-06,
+      "loss": 0.0994,
+      "step": 4322
+    },
+    {
+      "epoch": 2.2640125720272395,
+      "loss_partial": 0.048828125,
+      "step": 4322
+    },
+    {
+      "epoch": 2.2640125720272395,
+      "loss_partial": 0.06591796875,
+      "step": 4322
+    },
+    {
+      "epoch": 2.2645364064955475,
+      "grad_norm": 1.5114511251449585,
+      "learning_rate": 9.442555837635264e-06,
+      "loss": 0.1003,
+      "step": 4323
+    },
+    {
+      "epoch": 2.2645364064955475,
+      "loss_partial": 0.04833984375,
+      "step": 4323
+    },
+    {
+      "epoch": 2.2645364064955475,
+      "loss_partial": 0.05029296875,
+      "step": 4323
+    },
+    {
+      "epoch": 2.2650602409638556,
+      "grad_norm": 1.563490629196167,
+      "learning_rate": 9.437991420346282e-06,
+      "loss": 0.1016,
+      "step": 4324
+    },
+    {
+      "epoch": 2.2650602409638556,
+      "loss_partial": 0.050048828125,
+      "step": 4324
+    },
+    {
+      "epoch": 2.2650602409638556,
+      "loss_partial": 0.046630859375,
+      "step": 4324
+    },
+    {
+      "epoch": 2.2655840754321632,
+      "grad_norm": 1.5977861881256104,
+      "learning_rate": 9.433427120513619e-06,
+      "loss": 0.1022,
+      "step": 4325
+    },
+    {
+      "epoch": 2.2655840754321632,
+      "loss_partial": 0.052490234375,
+      "step": 4325
+    },
+    {
+      "epoch": 2.2655840754321632,
+      "loss_partial": 0.05078125,
+      "step": 4325
+    },
+    {
+      "epoch": 2.2661079099004713,
+      "grad_norm": 1.6332244873046875,
+      "learning_rate": 9.428862939091178e-06,
+      "loss": 0.0992,
+      "step": 4326
+    },
+    {
+      "epoch": 2.2661079099004713,
+      "loss_partial": 0.043701171875,
+      "step": 4326
+    },
+    {
+      "epoch": 2.2661079099004713,
+      "loss_partial": 0.052001953125,
+      "step": 4326
+    },
+    {
+      "epoch": 2.2666317443687793,
+      "grad_norm": 1.0592817068099976,
+      "learning_rate": 9.424298877032857e-06,
+      "loss": 0.1028,
+      "step": 4327
+    },
+    {
+      "epoch": 2.2666317443687793,
+      "loss_partial": 0.04345703125,
+      "step": 4327
+    },
+    {
+      "epoch": 2.2666317443687793,
+      "loss_partial": 0.04638671875,
+      "step": 4327
+    },
+    {
+      "epoch": 2.2671555788370874,
+      "grad_norm": 1.753961205482483,
+      "learning_rate": 9.41973493529251e-06,
+      "loss": 0.1054,
+      "step": 4328
+    },
+    {
+      "epoch": 2.2671555788370874,
+      "loss_partial": 0.06201171875,
+      "step": 4328
+    },
+    {
+      "epoch": 2.2671555788370874,
+      "loss_partial": 0.0537109375,
+      "step": 4328
+    },
+    {
+      "epoch": 2.2676794133053955,
+      "grad_norm": 1.1360411643981934,
+      "learning_rate": 9.415171114823973e-06,
+      "loss": 0.1014,
+      "step": 4329
+    },
+    {
+      "epoch": 2.2676794133053955,
+      "loss_partial": 0.0595703125,
+      "step": 4329
+    },
+    {
+      "epoch": 2.2676794133053955,
+      "loss_partial": 0.046630859375,
+      "step": 4329
+    },
+    {
+      "epoch": 2.2682032477737035,
+      "grad_norm": 1.5966156721115112,
+      "learning_rate": 9.410607416581051e-06,
+      "loss": 0.1006,
+      "step": 4330
+    },
+    {
+      "epoch": 2.2682032477737035,
+      "loss_partial": 0.056640625,
+      "step": 4330
+    },
+    {
+      "epoch": 2.2682032477737035,
+      "loss_partial": 0.045166015625,
+      "step": 4330
+    },
+    {
+      "epoch": 2.2687270822420116,
+      "grad_norm": 1.3343446254730225,
+      "learning_rate": 9.40604384151754e-06,
+      "loss": 0.1004,
+      "step": 4331
+    },
+    {
+      "epoch": 2.2687270822420116,
+      "loss_partial": 0.046875,
+      "step": 4331
+    },
+    {
+      "epoch": 2.2687270822420116,
+      "loss_partial": 0.0478515625,
+      "step": 4331
+    },
+    {
+      "epoch": 2.2692509167103196,
+      "grad_norm": 2.7350356578826904,
+      "learning_rate": 9.401480390587192e-06,
+      "loss": 0.102,
+      "step": 4332
+    },
+    {
+      "epoch": 2.2692509167103196,
+      "loss_partial": 0.056396484375,
+      "step": 4332
+    },
+    {
+      "epoch": 2.2692509167103196,
+      "loss_partial": 0.056884765625,
+      "step": 4332
+    },
+    {
+      "epoch": 2.2697747511786277,
+      "grad_norm": 1.4432604312896729,
+      "learning_rate": 9.396917064743742e-06,
+      "loss": 0.1023,
+      "step": 4333
+    },
+    {
+      "epoch": 2.2697747511786277,
+      "loss_partial": 0.052978515625,
+      "step": 4333
+    },
+    {
+      "epoch": 2.2697747511786277,
+      "loss_partial": 0.050048828125,
+      "step": 4333
+    },
+    {
+      "epoch": 2.2702985856469358,
+      "grad_norm": 1.6224708557128906,
+      "learning_rate": 9.392353864940894e-06,
+      "loss": 0.1002,
+      "step": 4334
+    },
+    {
+      "epoch": 2.2702985856469358,
+      "loss_partial": 0.046875,
+      "step": 4334
+    },
+    {
+      "epoch": 2.2702985856469358,
+      "loss_partial": 0.048583984375,
+      "step": 4334
+    },
+    {
+      "epoch": 2.270822420115244,
+      "grad_norm": 2.325427532196045,
+      "learning_rate": 9.387790792132333e-06,
+      "loss": 0.1032,
+      "step": 4335
+    },
+    {
+      "epoch": 2.270822420115244,
+      "loss_partial": 0.053955078125,
+      "step": 4335
+    },
+    {
+      "epoch": 2.270822420115244,
+      "loss_partial": 0.054931640625,
+      "step": 4335
+    },
+    {
+      "epoch": 2.2713462545835514,
+      "grad_norm": 1.5488948822021484,
+      "learning_rate": 9.38322784727171e-06,
+      "loss": 0.1023,
+      "step": 4336
+    },
+    {
+      "epoch": 2.2713462545835514,
+      "loss_partial": 0.046875,
+      "step": 4336
+    },
+    {
+      "epoch": 2.2713462545835514,
+      "loss_partial": 0.05322265625,
+      "step": 4336
+    },
+    {
+      "epoch": 2.2718700890518595,
+      "grad_norm": 1.6380606889724731,
+      "learning_rate": 9.378665031312653e-06,
+      "loss": 0.1002,
+      "step": 4337
+    },
+    {
+      "epoch": 2.2718700890518595,
+      "loss_partial": 0.042724609375,
+      "step": 4337
+    },
+    {
+      "epoch": 2.2718700890518595,
+      "loss_partial": 0.047607421875,
+      "step": 4337
+    },
+    {
+      "epoch": 2.2723939235201676,
+      "grad_norm": 1.5896540880203247,
+      "learning_rate": 9.374102345208758e-06,
+      "loss": 0.1026,
+      "step": 4338
+    },
+    {
+      "epoch": 2.2723939235201676,
+      "loss_partial": 0.05078125,
+      "step": 4338
+    },
+    {
+      "epoch": 2.2723939235201676,
+      "loss_partial": 0.051513671875,
+      "step": 4338
+    },
+    {
+      "epoch": 2.2729177579884756,
+      "grad_norm": 1.7393896579742432,
+      "learning_rate": 9.36953978991361e-06,
+      "loss": 0.1,
+      "step": 4339
+    },
+    {
+      "epoch": 2.2729177579884756,
+      "loss_partial": 0.0517578125,
+      "step": 4339
+    },
+    {
+      "epoch": 2.2729177579884756,
+      "loss_partial": 0.04541015625,
+      "step": 4339
+    },
+    {
+      "epoch": 2.2734415924567837,
+      "grad_norm": 1.2259700298309326,
+      "learning_rate": 9.364977366380743e-06,
+      "loss": 0.1035,
+      "step": 4340
+    },
+    {
+      "epoch": 2.2734415924567837,
+      "loss_partial": 0.04296875,
+      "step": 4340
+    },
+    {
+      "epoch": 2.2734415924567837,
+      "loss_partial": 0.046630859375,
+      "step": 4340
+    },
+    {
+      "epoch": 2.2739654269250917,
+      "grad_norm": 1.5393977165222168,
+      "learning_rate": 9.360415075563681e-06,
+      "loss": 0.1044,
+      "step": 4341
+    },
+    {
+      "epoch": 2.2739654269250917,
+      "loss_partial": 0.046875,
+      "step": 4341
+    },
+    {
+      "epoch": 2.2739654269250917,
+      "loss_partial": 0.046142578125,
+      "step": 4341
+    },
+    {
+      "epoch": 2.2744892613934,
+      "grad_norm": 1.4034793376922607,
+      "learning_rate": 9.355852918415909e-06,
+      "loss": 0.0986,
+      "step": 4342
+    },
+    {
+      "epoch": 2.2744892613934,
+      "loss_partial": 0.0419921875,
+      "step": 4342
+    },
+    {
+      "epoch": 2.2744892613934,
+      "loss_partial": 0.044677734375,
+      "step": 4342
+    },
+    {
+      "epoch": 2.275013095861708,
+      "grad_norm": 1.2441768646240234,
+      "learning_rate": 9.351290895890898e-06,
+      "loss": 0.1033,
+      "step": 4343
+    },
+    {
+      "epoch": 2.275013095861708,
+      "loss_partial": 0.0517578125,
+      "step": 4343
+    },
+    {
+      "epoch": 2.275013095861708,
+      "loss_partial": 0.051513671875,
+      "step": 4343
+    },
+    {
+      "epoch": 2.2755369303300155,
+      "grad_norm": 1.5250575542449951,
+      "learning_rate": 9.346729008942077e-06,
+      "loss": 0.0989,
+      "step": 4344
+    },
+    {
+      "epoch": 2.2755369303300155,
+      "loss_partial": 0.044921875,
+      "step": 4344
+    },
+    {
+      "epoch": 2.2755369303300155,
+      "loss_partial": 0.05126953125,
+      "step": 4344
+    },
+    {
+      "epoch": 2.2760607647983235,
+      "grad_norm": 1.289361596107483,
+      "learning_rate": 9.342167258522855e-06,
+      "loss": 0.1047,
+      "step": 4345
+    },
+    {
+      "epoch": 2.2760607647983235,
+      "loss_partial": 0.048095703125,
+      "step": 4345
+    },
+    {
+      "epoch": 2.2760607647983235,
+      "loss_partial": 0.045166015625,
+      "step": 4345
+    },
+    {
+      "epoch": 2.2765845992666316,
+      "grad_norm": 1.3066424131393433,
+      "learning_rate": 9.337605645586605e-06,
+      "loss": 0.1027,
+      "step": 4346
+    },
+    {
+      "epoch": 2.2765845992666316,
+      "loss_partial": 0.03662109375,
+      "step": 4346
+    },
+    {
+      "epoch": 2.2765845992666316,
+      "loss_partial": 0.056640625,
+      "step": 4346
+    },
+    {
+      "epoch": 2.2771084337349397,
+      "grad_norm": 1.1473042964935303,
+      "learning_rate": 9.333044171086682e-06,
+      "loss": 0.097,
+      "step": 4347
+    },
+    {
+      "epoch": 2.2771084337349397,
+      "loss_partial": 0.049560546875,
+      "step": 4347
+    },
+    {
+      "epoch": 2.2771084337349397,
+      "loss_partial": 0.04833984375,
+      "step": 4347
+    },
+    {
+      "epoch": 2.2776322682032477,
+      "grad_norm": 1.8957451581954956,
+      "learning_rate": 9.3284828359764e-06,
+      "loss": 0.1018,
+      "step": 4348
+    },
+    {
+      "epoch": 2.2776322682032477,
+      "loss_partial": 0.0537109375,
+      "step": 4348
+    },
+    {
+      "epoch": 2.2776322682032477,
+      "loss_partial": 0.0546875,
+      "step": 4348
+    },
+    {
+      "epoch": 2.2781561026715558,
+      "grad_norm": 1.8033424615859985,
+      "learning_rate": 9.323921641209057e-06,
+      "loss": 0.1029,
+      "step": 4349
+    },
+    {
+      "epoch": 2.2781561026715558,
+      "loss_partial": 0.04931640625,
+      "step": 4349
+    },
+    {
+      "epoch": 2.2781561026715558,
+      "loss_partial": 0.06640625,
+      "step": 4349
+    },
+    {
+      "epoch": 2.278679937139864,
+      "grad_norm": 1.5073124170303345,
+      "learning_rate": 9.319360587737903e-06,
+      "loss": 0.1028,
+      "step": 4350
+    },
+    {
+      "epoch": 2.278679937139864,
+      "loss_partial": 0.05810546875,
+      "step": 4350
+    },
+    {
+      "epoch": 2.278679937139864,
+      "loss_partial": 0.061279296875,
+      "step": 4350
+    },
+    {
+      "epoch": 2.279203771608172,
+      "grad_norm": 2.023768186569214,
+      "learning_rate": 9.314799676516184e-06,
+      "loss": 0.1053,
+      "step": 4351
+    },
+    {
+      "epoch": 2.279203771608172,
+      "loss_partial": 0.0478515625,
+      "step": 4351
+    },
+    {
+      "epoch": 2.279203771608172,
+      "loss_partial": 0.044921875,
+      "step": 4351
+    },
+    {
+      "epoch": 2.27972760607648,
+      "grad_norm": 1.7857836484909058,
+      "learning_rate": 9.310238908497095e-06,
+      "loss": 0.102,
+      "step": 4352
+    },
+    {
+      "epoch": 2.27972760607648,
+      "loss_partial": 0.047607421875,
+      "step": 4352
+    },
+    {
+      "epoch": 2.27972760607648,
+      "loss_partial": 0.051025390625,
+      "step": 4352
+    },
+    {
+      "epoch": 2.280251440544788,
+      "grad_norm": 1.5700697898864746,
+      "learning_rate": 9.305678284633809e-06,
+      "loss": 0.1075,
+      "step": 4353
+    },
+    {
+      "epoch": 2.280251440544788,
+      "loss_partial": 0.046630859375,
+      "step": 4353
+    },
+    {
+      "epoch": 2.280251440544788,
+      "loss_partial": 0.047607421875,
+      "step": 4353
+    },
+    {
+      "epoch": 2.280775275013096,
+      "grad_norm": 1.313978910446167,
+      "learning_rate": 9.30111780587946e-06,
+      "loss": 0.0995,
+      "step": 4354
+    },
+    {
+      "epoch": 2.280775275013096,
+      "loss_partial": 0.05126953125,
+      "step": 4354
+    },
+    {
+      "epoch": 2.280775275013096,
+      "loss_partial": 0.053466796875,
+      "step": 4354
+    },
+    {
+      "epoch": 2.2812991094814037,
+      "grad_norm": 1.892934799194336,
+      "learning_rate": 9.296557473187174e-06,
+      "loss": 0.0963,
+      "step": 4355
+    },
+    {
+      "epoch": 2.2812991094814037,
+      "loss_partial": 0.04736328125,
+      "step": 4355
+    },
+    {
+      "epoch": 2.2812991094814037,
+      "loss_partial": 0.049072265625,
+      "step": 4355
+    },
+    {
+      "epoch": 2.2818229439497117,
+      "grad_norm": 1.7675940990447998,
+      "learning_rate": 9.291997287510027e-06,
+      "loss": 0.1036,
+      "step": 4356
+    },
+    {
+      "epoch": 2.2818229439497117,
+      "loss_partial": 0.050537109375,
+      "step": 4356
+    },
+    {
+      "epoch": 2.2818229439497117,
+      "loss_partial": 0.048828125,
+      "step": 4356
+    },
+    {
+      "epoch": 2.28234677841802,
+      "grad_norm": 1.464134931564331,
+      "learning_rate": 9.28743724980107e-06,
+      "loss": 0.1032,
+      "step": 4357
+    },
+    {
+      "epoch": 2.28234677841802,
+      "loss_partial": 0.04052734375,
+      "step": 4357
+    },
+    {
+      "epoch": 2.28234677841802,
+      "loss_partial": 0.05908203125,
+      "step": 4357
+    },
+    {
+      "epoch": 2.282870612886328,
+      "grad_norm": 1.5158867835998535,
+      "learning_rate": 9.282877361013315e-06,
+      "loss": 0.1011,
+      "step": 4358
+    },
+    {
+      "epoch": 2.282870612886328,
+      "loss_partial": 0.060546875,
+      "step": 4358
+    },
+    {
+      "epoch": 2.282870612886328,
+      "loss_partial": 0.052001953125,
+      "step": 4358
+    },
+    {
+      "epoch": 2.283394447354636,
+      "grad_norm": 2.1857593059539795,
+      "learning_rate": 9.278317622099762e-06,
+      "loss": 0.1007,
+      "step": 4359
+    },
+    {
+      "epoch": 2.283394447354636,
+      "loss_partial": 0.052978515625,
+      "step": 4359
+    },
+    {
+      "epoch": 2.283394447354636,
+      "loss_partial": 0.048095703125,
+      "step": 4359
+    },
+    {
+      "epoch": 2.283918281822944,
+      "grad_norm": 1.6733468770980835,
+      "learning_rate": 9.273758034013365e-06,
+      "loss": 0.0998,
+      "step": 4360
+    },
+    {
+      "epoch": 2.283918281822944,
+      "loss_partial": 0.053466796875,
+      "step": 4360
+    },
+    {
+      "epoch": 2.283918281822944,
+      "loss_partial": 0.056640625,
+      "step": 4360
+    },
+    {
+      "epoch": 2.284442116291252,
+      "grad_norm": 1.861789345741272,
+      "learning_rate": 9.269198597707045e-06,
+      "loss": 0.1007,
+      "step": 4361
+    },
+    {
+      "epoch": 2.284442116291252,
+      "loss_partial": 0.05078125,
+      "step": 4361
+    },
+    {
+      "epoch": 2.284442116291252,
+      "loss_partial": 0.046630859375,
+      "step": 4361
+    },
+    {
+      "epoch": 2.28496595075956,
+      "grad_norm": 1.8918828964233398,
+      "learning_rate": 9.264639314133696e-06,
+      "loss": 0.1066,
+      "step": 4362
+    },
+    {
+      "epoch": 2.28496595075956,
+      "loss_partial": 0.046630859375,
+      "step": 4362
+    },
+    {
+      "epoch": 2.28496595075956,
+      "loss_partial": 0.050048828125,
+      "step": 4362
+    },
+    {
+      "epoch": 2.285489785227868,
+      "grad_norm": 1.4049466848373413,
+      "learning_rate": 9.260080184246187e-06,
+      "loss": 0.1011,
+      "step": 4363
+    },
+    {
+      "epoch": 2.285489785227868,
+      "loss_partial": 0.052001953125,
+      "step": 4363
+    },
+    {
+      "epoch": 2.285489785227868,
+      "loss_partial": 0.0478515625,
+      "step": 4363
+    },
+    {
+      "epoch": 2.286013619696176,
+      "grad_norm": 1.636338472366333,
+      "learning_rate": 9.255521208997345e-06,
+      "loss": 0.1044,
+      "step": 4364
+    },
+    {
+      "epoch": 2.286013619696176,
+      "loss_partial": 0.0546875,
+      "step": 4364
+    },
+    {
+      "epoch": 2.286013619696176,
+      "loss_partial": 0.060546875,
+      "step": 4364
+    },
+    {
+      "epoch": 2.286537454164484,
+      "grad_norm": 1.406287670135498,
+      "learning_rate": 9.250962389339966e-06,
+      "loss": 0.0995,
+      "step": 4365
+    },
+    {
+      "epoch": 2.286537454164484,
+      "loss_partial": 0.05078125,
+      "step": 4365
+    },
+    {
+      "epoch": 2.286537454164484,
+      "loss_partial": 0.048828125,
+      "step": 4365
+    },
+    {
+      "epoch": 2.287061288632792,
+      "grad_norm": 1.858933448791504,
+      "learning_rate": 9.24640372622681e-06,
+      "loss": 0.1038,
+      "step": 4366
+    },
+    {
+      "epoch": 2.287061288632792,
+      "loss_partial": 0.0546875,
+      "step": 4366
+    },
+    {
+      "epoch": 2.287061288632792,
+      "loss_partial": 0.04833984375,
+      "step": 4366
+    },
+    {
+      "epoch": 2.2875851231011,
+      "grad_norm": 1.8667733669281006,
+      "learning_rate": 9.24184522061062e-06,
+      "loss": 0.1012,
+      "step": 4367
+    },
+    {
+      "epoch": 2.2875851231011,
+      "loss_partial": 0.051025390625,
+      "step": 4367
+    },
+    {
+      "epoch": 2.2875851231011,
+      "loss_partial": 0.0478515625,
+      "step": 4367
+    },
+    {
+      "epoch": 2.288108957569408,
+      "grad_norm": 1.6312483549118042,
+      "learning_rate": 9.23728687344409e-06,
+      "loss": 0.1012,
+      "step": 4368
+    },
+    {
+      "epoch": 2.288108957569408,
+      "loss_partial": 0.046630859375,
+      "step": 4368
+    },
+    {
+      "epoch": 2.288108957569408,
+      "loss_partial": 0.05126953125,
+      "step": 4368
+    },
+    {
+      "epoch": 2.288632792037716,
+      "grad_norm": 1.3998781442642212,
+      "learning_rate": 9.23272868567989e-06,
+      "loss": 0.1013,
+      "step": 4369
+    },
+    {
+      "epoch": 2.288632792037716,
+      "loss_partial": 0.04931640625,
+      "step": 4369
+    },
+    {
+      "epoch": 2.288632792037716,
+      "loss_partial": 0.054931640625,
+      "step": 4369
+    },
+    {
+      "epoch": 2.289156626506024,
+      "grad_norm": 1.5627121925354004,
+      "learning_rate": 9.228170658270646e-06,
+      "loss": 0.1023,
+      "step": 4370
+    },
+    {
+      "epoch": 2.289156626506024,
+      "loss_partial": 0.047119140625,
+      "step": 4370
+    },
+    {
+      "epoch": 2.289156626506024,
+      "loss_partial": 0.052001953125,
+      "step": 4370
+    },
+    {
+      "epoch": 2.289680460974332,
+      "grad_norm": 1.4131550788879395,
+      "learning_rate": 9.223612792168963e-06,
+      "loss": 0.1012,
+      "step": 4371
+    },
+    {
+      "epoch": 2.289680460974332,
+      "loss_partial": 0.0498046875,
+      "step": 4371
+    },
+    {
+      "epoch": 2.289680460974332,
+      "loss_partial": 0.05078125,
+      "step": 4371
+    },
+    {
+      "epoch": 2.2902042954426403,
+      "grad_norm": 1.7114694118499756,
+      "learning_rate": 9.219055088327406e-06,
+      "loss": 0.0997,
+      "step": 4372
+    },
+    {
+      "epoch": 2.2902042954426403,
+      "loss_partial": 0.046630859375,
+      "step": 4372
+    },
+    {
+      "epoch": 2.2902042954426403,
+      "loss_partial": 0.04736328125,
+      "step": 4372
+    },
+    {
+      "epoch": 2.2907281299109483,
+      "grad_norm": 1.394004225730896,
+      "learning_rate": 9.214497547698505e-06,
+      "loss": 0.1016,
+      "step": 4373
+    },
+    {
+      "epoch": 2.2907281299109483,
+      "loss_partial": 0.046875,
+      "step": 4373
+    },
+    {
+      "epoch": 2.2907281299109483,
+      "loss_partial": 0.04296875,
+      "step": 4373
+    },
+    {
+      "epoch": 2.2912519643792564,
+      "grad_norm": 1.2628740072250366,
+      "learning_rate": 9.209940171234762e-06,
+      "loss": 0.0992,
+      "step": 4374
+    },
+    {
+      "epoch": 2.2912519643792564,
+      "loss_partial": 0.057861328125,
+      "step": 4374
+    },
+    {
+      "epoch": 2.2912519643792564,
+      "loss_partial": 0.04736328125,
+      "step": 4374
+    },
+    {
+      "epoch": 2.291775798847564,
+      "grad_norm": 1.4698386192321777,
+      "learning_rate": 9.205382959888631e-06,
+      "loss": 0.1048,
+      "step": 4375
+    },
+    {
+      "epoch": 2.291775798847564,
+      "loss_partial": 0.0537109375,
+      "step": 4375
+    },
+    {
+      "epoch": 2.291775798847564,
+      "loss_partial": 0.0517578125,
+      "step": 4375
+    },
+    {
+      "epoch": 2.292299633315872,
+      "grad_norm": 1.7915985584259033,
+      "learning_rate": 9.200825914612555e-06,
+      "loss": 0.1036,
+      "step": 4376
+    },
+    {
+      "epoch": 2.292299633315872,
+      "loss_partial": 0.051513671875,
+      "step": 4376
+    },
+    {
+      "epoch": 2.292299633315872,
+      "loss_partial": 0.05322265625,
+      "step": 4376
+    },
+    {
+      "epoch": 2.29282346778418,
+      "grad_norm": 1.7928398847579956,
+      "learning_rate": 9.196269036358919e-06,
+      "loss": 0.1,
+      "step": 4377
+    },
+    {
+      "epoch": 2.29282346778418,
+      "loss_partial": 0.05126953125,
+      "step": 4377
+    },
+    {
+      "epoch": 2.29282346778418,
+      "loss_partial": 0.055908203125,
+      "step": 4377
+    },
+    {
+      "epoch": 2.293347302252488,
+      "grad_norm": 1.895228624343872,
+      "learning_rate": 9.191712326080082e-06,
+      "loss": 0.0995,
+      "step": 4378
+    },
+    {
+      "epoch": 2.293347302252488,
+      "loss_partial": 0.04736328125,
+      "step": 4378
+    },
+    {
+      "epoch": 2.293347302252488,
+      "loss_partial": 0.044189453125,
+      "step": 4378
+    },
+    {
+      "epoch": 2.2938711367207962,
+      "grad_norm": 1.7119561433792114,
+      "learning_rate": 9.187155784728369e-06,
+      "loss": 0.103,
+      "step": 4379
+    },
+    {
+      "epoch": 2.2938711367207962,
+      "loss_partial": 0.0595703125,
+      "step": 4379
+    },
+    {
+      "epoch": 2.2938711367207962,
+      "loss_partial": 0.046630859375,
+      "step": 4379
+    },
+    {
+      "epoch": 2.2943949711891043,
+      "grad_norm": 1.8054412603378296,
+      "learning_rate": 9.182599413256074e-06,
+      "loss": 0.1013,
+      "step": 4380
+    },
+    {
+      "epoch": 2.2943949711891043,
+      "loss_partial": 0.049560546875,
+      "step": 4380
+    },
+    {
+      "epoch": 2.2943949711891043,
+      "loss_partial": 0.043212890625,
+      "step": 4380
+    },
+    {
+      "epoch": 2.2949188056574124,
+      "grad_norm": 1.678789496421814,
+      "learning_rate": 9.178043212615446e-06,
+      "loss": 0.1029,
+      "step": 4381
+    },
+    {
+      "epoch": 2.2949188056574124,
+      "loss_partial": 0.046875,
+      "step": 4381
+    },
+    {
+      "epoch": 2.2949188056574124,
+      "loss_partial": 0.05029296875,
+      "step": 4381
+    },
+    {
+      "epoch": 2.2954426401257204,
+      "grad_norm": 1.8732959032058716,
+      "learning_rate": 9.173487183758704e-06,
+      "loss": 0.1038,
+      "step": 4382
+    },
+    {
+      "epoch": 2.2954426401257204,
+      "loss_partial": 0.047119140625,
+      "step": 4382
+    },
+    {
+      "epoch": 2.2954426401257204,
+      "loss_partial": 0.04443359375,
+      "step": 4382
+    },
+    {
+      "epoch": 2.2959664745940285,
+      "grad_norm": 1.6826986074447632,
+      "learning_rate": 9.168931327638029e-06,
+      "loss": 0.1015,
+      "step": 4383
+    },
+    {
+      "epoch": 2.2959664745940285,
+      "loss_partial": 0.04931640625,
+      "step": 4383
+    },
+    {
+      "epoch": 2.2959664745940285,
+      "loss_partial": 0.048095703125,
+      "step": 4383
+    },
+    {
+      "epoch": 2.296490309062336,
+      "grad_norm": 1.741658329963684,
+      "learning_rate": 9.164375645205568e-06,
+      "loss": 0.1019,
+      "step": 4384
+    },
+    {
+      "epoch": 2.296490309062336,
+      "loss_partial": 0.047607421875,
+      "step": 4384
+    },
+    {
+      "epoch": 2.296490309062336,
+      "loss_partial": 0.0498046875,
+      "step": 4384
+    },
+    {
+      "epoch": 2.297014143530644,
+      "grad_norm": 1.7116453647613525,
+      "learning_rate": 9.15982013741343e-06,
+      "loss": 0.0995,
+      "step": 4385
+    },
+    {
+      "epoch": 2.297014143530644,
+      "loss_partial": 0.049072265625,
+      "step": 4385
+    },
+    {
+      "epoch": 2.297014143530644,
+      "loss_partial": 0.05126953125,
+      "step": 4385
+    },
+    {
+      "epoch": 2.297537977998952,
+      "grad_norm": 1.6430343389511108,
+      "learning_rate": 9.155264805213689e-06,
+      "loss": 0.101,
+      "step": 4386
+    },
+    {
+      "epoch": 2.297537977998952,
+      "loss_partial": 0.0576171875,
+      "step": 4386
+    },
+    {
+      "epoch": 2.297537977998952,
+      "loss_partial": 0.05224609375,
+      "step": 4386
+    },
+    {
+      "epoch": 2.2980618124672603,
+      "grad_norm": 1.9955682754516602,
+      "learning_rate": 9.150709649558374e-06,
+      "loss": 0.1009,
+      "step": 4387
+    },
+    {
+      "epoch": 2.2980618124672603,
+      "loss_partial": 0.05419921875,
+      "step": 4387
+    },
+    {
+      "epoch": 2.2980618124672603,
+      "loss_partial": 0.054443359375,
+      "step": 4387
+    },
+    {
+      "epoch": 2.2985856469355683,
+      "grad_norm": 1.7616199254989624,
+      "learning_rate": 9.146154671399498e-06,
+      "loss": 0.1019,
+      "step": 4388
+    },
+    {
+      "epoch": 2.2985856469355683,
+      "loss_partial": 0.0400390625,
+      "step": 4388
+    },
+    {
+      "epoch": 2.2985856469355683,
+      "loss_partial": 0.056884765625,
+      "step": 4388
+    },
+    {
+      "epoch": 2.2991094814038764,
+      "grad_norm": 2.0831336975097656,
+      "learning_rate": 9.141599871689014e-06,
+      "loss": 0.1066,
+      "step": 4389
+    },
+    {
+      "epoch": 2.2991094814038764,
+      "loss_partial": 0.060302734375,
+      "step": 4389
+    },
+    {
+      "epoch": 2.2991094814038764,
+      "loss_partial": 0.05078125,
+      "step": 4389
+    },
+    {
+      "epoch": 2.2996333158721844,
+      "grad_norm": 1.6596599817276,
+      "learning_rate": 9.137045251378847e-06,
+      "loss": 0.1038,
+      "step": 4390
+    },
+    {
+      "epoch": 2.2996333158721844,
+      "loss_partial": 0.0595703125,
+      "step": 4390
+    },
+    {
+      "epoch": 2.2996333158721844,
+      "loss_partial": 0.042236328125,
+      "step": 4390
+    },
+    {
+      "epoch": 2.3001571503404925,
+      "grad_norm": 1.8690167665481567,
+      "learning_rate": 9.132490811420882e-06,
+      "loss": 0.0981,
+      "step": 4391
+    },
+    {
+      "epoch": 2.3001571503404925,
+      "loss_partial": 0.055908203125,
+      "step": 4391
+    },
+    {
+      "epoch": 2.3001571503404925,
+      "loss_partial": 0.05078125,
+      "step": 4391
+    },
+    {
+      "epoch": 2.3006809848088006,
+      "grad_norm": 1.3176156282424927,
+      "learning_rate": 9.127936552766979e-06,
+      "loss": 0.1018,
+      "step": 4392
+    },
+    {
+      "epoch": 2.3006809848088006,
+      "loss_partial": 0.05908203125,
+      "step": 4392
+    },
+    {
+      "epoch": 2.3006809848088006,
+      "loss_partial": 0.044677734375,
+      "step": 4392
+    },
+    {
+      "epoch": 2.3012048192771086,
+      "grad_norm": 1.298062801361084,
+      "learning_rate": 9.123382476368941e-06,
+      "loss": 0.107,
+      "step": 4393
+    },
+    {
+      "epoch": 2.3012048192771086,
+      "loss_partial": 0.043212890625,
+      "step": 4393
+    },
+    {
+      "epoch": 2.3012048192771086,
+      "loss_partial": 0.05810546875,
+      "step": 4393
+    },
+    {
+      "epoch": 2.3017286537454167,
+      "grad_norm": 1.945356011390686,
+      "learning_rate": 9.118828583178547e-06,
+      "loss": 0.1052,
+      "step": 4394
+    },
+    {
+      "epoch": 2.3017286537454167,
+      "loss_partial": 0.0498046875,
+      "step": 4394
+    },
+    {
+      "epoch": 2.3017286537454167,
+      "loss_partial": 0.046630859375,
+      "step": 4394
+    },
+    {
+      "epoch": 2.3022524882137243,
+      "grad_norm": 1.4133086204528809,
+      "learning_rate": 9.114274874147526e-06,
+      "loss": 0.1003,
+      "step": 4395
+    },
+    {
+      "epoch": 2.3022524882137243,
+      "loss_partial": 0.060791015625,
+      "step": 4395
+    },
+    {
+      "epoch": 2.3022524882137243,
+      "loss_partial": 0.052978515625,
+      "step": 4395
+    },
+    {
+      "epoch": 2.3027763226820324,
+      "grad_norm": 1.878559947013855,
+      "learning_rate": 9.109721350227582e-06,
+      "loss": 0.1029,
+      "step": 4396
+    },
+    {
+      "epoch": 2.3027763226820324,
+      "loss_partial": 0.0634765625,
+      "step": 4396
+    },
+    {
+      "epoch": 2.3027763226820324,
+      "loss_partial": 0.0478515625,
+      "step": 4396
+    },
+    {
+      "epoch": 2.3033001571503404,
+      "grad_norm": 1.6925745010375977,
+      "learning_rate": 9.105168012370372e-06,
+      "loss": 0.1071,
+      "step": 4397
+    },
+    {
+      "epoch": 2.3033001571503404,
+      "loss_partial": 0.055908203125,
+      "step": 4397
+    },
+    {
+      "epoch": 2.3033001571503404,
+      "loss_partial": 0.04248046875,
+      "step": 4397
+    },
+    {
+      "epoch": 2.3038239916186485,
+      "grad_norm": 1.414062738418579,
+      "learning_rate": 9.100614861527513e-06,
+      "loss": 0.1009,
+      "step": 4398
+    },
+    {
+      "epoch": 2.3038239916186485,
+      "loss_partial": 0.04296875,
+      "step": 4398
+    },
+    {
+      "epoch": 2.3038239916186485,
+      "loss_partial": 0.050048828125,
+      "step": 4398
+    },
+    {
+      "epoch": 2.3043478260869565,
+      "grad_norm": 1.3341288566589355,
+      "learning_rate": 9.09606189865058e-06,
+      "loss": 0.1013,
+      "step": 4399
+    },
+    {
+      "epoch": 2.3043478260869565,
+      "loss_partial": 0.047607421875,
+      "step": 4399
+    },
+    {
+      "epoch": 2.3043478260869565,
+      "loss_partial": 0.04931640625,
+      "step": 4399
+    },
+    {
+      "epoch": 2.3048716605552646,
+      "grad_norm": 1.5521283149719238,
+      "learning_rate": 9.091509124691128e-06,
+      "loss": 0.1027,
+      "step": 4400
+    },
+    {
+      "epoch": 2.3048716605552646,
+      "loss_partial": 0.045654296875,
+      "step": 4400
+    },
+    {
+      "epoch": 2.3048716605552646,
+      "loss_partial": 0.049560546875,
+      "step": 4400
+    },
+    {
+      "epoch": 2.3053954950235727,
+      "grad_norm": 1.3544154167175293,
+      "learning_rate": 9.086956540600651e-06,
+      "loss": 0.1004,
+      "step": 4401
+    },
+    {
+      "epoch": 2.3053954950235727,
+      "loss_partial": 0.047119140625,
+      "step": 4401
+    },
+    {
+      "epoch": 2.3053954950235727,
+      "loss_partial": 0.046630859375,
+      "step": 4401
+    },
+    {
+      "epoch": 2.3059193294918807,
+      "grad_norm": 1.6969293355941772,
+      "learning_rate": 9.082404147330611e-06,
+      "loss": 0.101,
+      "step": 4402
+    },
+    {
+      "epoch": 2.3059193294918807,
+      "loss_partial": 0.044677734375,
+      "step": 4402
+    },
+    {
+      "epoch": 2.3059193294918807,
+      "loss_partial": 0.044189453125,
+      "step": 4402
+    },
+    {
+      "epoch": 2.3064431639601883,
+      "grad_norm": 1.4591940641403198,
+      "learning_rate": 9.077851945832423e-06,
+      "loss": 0.1002,
+      "step": 4403
+    },
+    {
+      "epoch": 2.3064431639601883,
+      "loss_partial": 0.052734375,
+      "step": 4403
+    },
+    {
+      "epoch": 2.3064431639601883,
+      "loss_partial": 0.053466796875,
+      "step": 4403
+    },
+    {
+      "epoch": 2.3069669984284964,
+      "grad_norm": 2.072890520095825,
+      "learning_rate": 9.073299937057481e-06,
+      "loss": 0.1033,
+      "step": 4404
+    },
+    {
+      "epoch": 2.3069669984284964,
+      "loss_partial": 0.05322265625,
+      "step": 4404
+    },
+    {
+      "epoch": 2.3069669984284964,
+      "loss_partial": 0.057861328125,
+      "step": 4404
+    },
+    {
+      "epoch": 2.3074908328968045,
+      "grad_norm": 1.7209339141845703,
+      "learning_rate": 9.068748121957123e-06,
+      "loss": 0.1022,
+      "step": 4405
+    },
+    {
+      "epoch": 2.3074908328968045,
+      "loss_partial": 0.0419921875,
+      "step": 4405
+    },
+    {
+      "epoch": 2.3074908328968045,
+      "loss_partial": 0.04833984375,
+      "step": 4405
+    },
+    {
+      "epoch": 2.3080146673651125,
+      "grad_norm": 1.8180209398269653,
+      "learning_rate": 9.064196501482647e-06,
+      "loss": 0.1057,
+      "step": 4406
+    },
+    {
+      "epoch": 2.3080146673651125,
+      "loss_partial": 0.0517578125,
+      "step": 4406
+    },
+    {
+      "epoch": 2.3080146673651125,
+      "loss_partial": 0.061767578125,
+      "step": 4406
+    },
+    {
+      "epoch": 2.3085385018334206,
+      "grad_norm": 1.3093042373657227,
+      "learning_rate": 9.059645076585314e-06,
+      "loss": 0.1065,
+      "step": 4407
+    },
+    {
+      "epoch": 2.3085385018334206,
+      "loss_partial": 0.05615234375,
+      "step": 4407
+    },
+    {
+      "epoch": 2.3085385018334206,
+      "loss_partial": 0.04638671875,
+      "step": 4407
+    },
+    {
+      "epoch": 2.3090623363017286,
+      "grad_norm": 2.116434335708618,
+      "learning_rate": 9.055093848216347e-06,
+      "loss": 0.104,
+      "step": 4408
+    },
+    {
+      "epoch": 2.3090623363017286,
+      "loss_partial": 0.045166015625,
+      "step": 4408
+    },
+    {
+      "epoch": 2.3090623363017286,
+      "loss_partial": 0.0576171875,
+      "step": 4408
+    },
+    {
+      "epoch": 2.3095861707700367,
+      "grad_norm": 1.3030552864074707,
+      "learning_rate": 9.050542817326923e-06,
+      "loss": 0.1031,
+      "step": 4409
+    },
+    {
+      "epoch": 2.3095861707700367,
+      "loss_partial": 0.04541015625,
+      "step": 4409
+    },
+    {
+      "epoch": 2.3095861707700367,
+      "loss_partial": 0.04638671875,
+      "step": 4409
+    },
+    {
+      "epoch": 2.3101100052383448,
+      "grad_norm": 1.6172674894332886,
+      "learning_rate": 9.045991984868179e-06,
+      "loss": 0.0978,
+      "step": 4410
+    },
+    {
+      "epoch": 2.3101100052383448,
+      "loss_partial": 0.055908203125,
+      "step": 4410
+    },
+    {
+      "epoch": 2.3101100052383448,
+      "loss_partial": 0.052001953125,
+      "step": 4410
+    },
+    {
+      "epoch": 2.310633839706653,
+      "grad_norm": 1.3851871490478516,
+      "learning_rate": 9.041441351791206e-06,
+      "loss": 0.103,
+      "step": 4411
+    },
+    {
+      "epoch": 2.310633839706653,
+      "loss_partial": 0.044921875,
+      "step": 4411
+    },
+    {
+      "epoch": 2.310633839706653,
+      "loss_partial": 0.05224609375,
+      "step": 4411
+    },
+    {
+      "epoch": 2.311157674174961,
+      "grad_norm": 1.5216597318649292,
+      "learning_rate": 9.036890919047068e-06,
+      "loss": 0.1006,
+      "step": 4412
+    },
+    {
+      "epoch": 2.311157674174961,
+      "loss_partial": 0.052734375,
+      "step": 4412
+    },
+    {
+      "epoch": 2.311157674174961,
+      "loss_partial": 0.056640625,
+      "step": 4412
+    },
+    {
+      "epoch": 2.311681508643269,
+      "grad_norm": 1.6991819143295288,
+      "learning_rate": 9.032340687586772e-06,
+      "loss": 0.1011,
+      "step": 4413
+    },
+    {
+      "epoch": 2.311681508643269,
+      "loss_partial": 0.05224609375,
+      "step": 4413
+    },
+    {
+      "epoch": 2.311681508643269,
+      "loss_partial": 0.060302734375,
+      "step": 4413
+    },
+    {
+      "epoch": 2.3122053431115765,
+      "grad_norm": 1.742643117904663,
+      "learning_rate": 9.02779065836129e-06,
+      "loss": 0.1023,
+      "step": 4414
+    },
+    {
+      "epoch": 2.3122053431115765,
+      "loss_partial": 0.0439453125,
+      "step": 4414
+    },
+    {
+      "epoch": 2.3122053431115765,
+      "loss_partial": 0.0546875,
+      "step": 4414
+    },
+    {
+      "epoch": 2.3127291775798846,
+      "grad_norm": 1.8212794065475464,
+      "learning_rate": 9.023240832321541e-06,
+      "loss": 0.1005,
+      "step": 4415
+    },
+    {
+      "epoch": 2.3127291775798846,
+      "loss_partial": 0.04638671875,
+      "step": 4415
+    },
+    {
+      "epoch": 2.3127291775798846,
+      "loss_partial": 0.04833984375,
+      "step": 4415
+    },
+    {
+      "epoch": 2.3132530120481927,
+      "grad_norm": 1.3424451351165771,
+      "learning_rate": 9.018691210418423e-06,
+      "loss": 0.1003,
+      "step": 4416
+    },
+    {
+      "epoch": 2.3132530120481927,
+      "loss_partial": 0.042724609375,
+      "step": 4416
+    },
+    {
+      "epoch": 2.3132530120481927,
+      "loss_partial": 0.054443359375,
+      "step": 4416
+    },
+    {
+      "epoch": 2.3137768465165007,
+      "grad_norm": 1.4070229530334473,
+      "learning_rate": 9.014141793602776e-06,
+      "loss": 0.1013,
+      "step": 4417
+    },
+    {
+      "epoch": 2.3137768465165007,
+      "loss_partial": 0.048828125,
+      "step": 4417
+    },
+    {
+      "epoch": 2.3137768465165007,
+      "loss_partial": 0.0458984375,
+      "step": 4417
+    },
+    {
+      "epoch": 2.314300680984809,
+      "grad_norm": 1.1354115009307861,
+      "learning_rate": 9.009592582825394e-06,
+      "loss": 0.103,
+      "step": 4418
+    },
+    {
+      "epoch": 2.314300680984809,
+      "loss_partial": 0.06103515625,
+      "step": 4418
+    },
+    {
+      "epoch": 2.314300680984809,
+      "loss_partial": 0.052734375,
+      "step": 4418
+    },
+    {
+      "epoch": 2.314824515453117,
+      "grad_norm": 1.3299485445022583,
+      "learning_rate": 9.005043579037039e-06,
+      "loss": 0.1073,
+      "step": 4419
+    },
+    {
+      "epoch": 2.314824515453117,
+      "loss_partial": 0.04931640625,
+      "step": 4419
+    },
+    {
+      "epoch": 2.314824515453117,
+      "loss_partial": 0.044921875,
+      "step": 4419
+    },
+    {
+      "epoch": 2.315348349921425,
+      "grad_norm": 1.7325536012649536,
+      "learning_rate": 9.000494783188423e-06,
+      "loss": 0.1022,
+      "step": 4420
+    },
+    {
+      "epoch": 2.315348349921425,
+      "loss_partial": 0.057373046875,
+      "step": 4420
+    },
+    {
+      "epoch": 2.315348349921425,
+      "loss_partial": 0.057373046875,
+      "step": 4420
+    },
+    {
+      "epoch": 2.315872184389733,
+      "grad_norm": 1.553101897239685,
+      "learning_rate": 8.995946196230218e-06,
+      "loss": 0.1039,
+      "step": 4421
+    },
+    {
+      "epoch": 2.315872184389733,
+      "loss_partial": 0.06396484375,
+      "step": 4421
+    },
+    {
+      "epoch": 2.315872184389733,
+      "loss_partial": 0.0458984375,
+      "step": 4421
+    },
+    {
+      "epoch": 2.316396018858041,
+      "grad_norm": 1.5206489562988281,
+      "learning_rate": 8.991397819113049e-06,
+      "loss": 0.1018,
+      "step": 4422
+    },
+    {
+      "epoch": 2.316396018858041,
+      "loss_partial": 0.05419921875,
+      "step": 4422
+    },
+    {
+      "epoch": 2.316396018858041,
+      "loss_partial": 0.048095703125,
+      "step": 4422
+    },
+    {
+      "epoch": 2.3169198533263486,
+      "grad_norm": 1.433853030204773,
+      "learning_rate": 8.986849652787494e-06,
+      "loss": 0.0998,
+      "step": 4423
+    },
+    {
+      "epoch": 2.3169198533263486,
+      "loss_partial": 0.04931640625,
+      "step": 4423
+    },
+    {
+      "epoch": 2.3169198533263486,
+      "loss_partial": 0.04833984375,
+      "step": 4423
+    },
+    {
+      "epoch": 2.3174436877946567,
+      "grad_norm": 1.9311739206314087,
+      "learning_rate": 8.982301698204101e-06,
+      "loss": 0.0985,
+      "step": 4424
+    },
+    {
+      "epoch": 2.3174436877946567,
+      "loss_partial": 0.055419921875,
+      "step": 4424
+    },
+    {
+      "epoch": 2.3174436877946567,
+      "loss_partial": 0.0458984375,
+      "step": 4424
+    },
+    {
+      "epoch": 2.3179675222629648,
+      "grad_norm": 2.0194098949432373,
+      "learning_rate": 8.977753956313359e-06,
+      "loss": 0.1012,
+      "step": 4425
+    },
+    {
+      "epoch": 2.3179675222629648,
+      "loss_partial": 0.056640625,
+      "step": 4425
+    },
+    {
+      "epoch": 2.3179675222629648,
+      "loss_partial": 0.052490234375,
+      "step": 4425
+    },
+    {
+      "epoch": 2.318491356731273,
+      "grad_norm": 2.119060516357422,
+      "learning_rate": 8.973206428065722e-06,
+      "loss": 0.1028,
+      "step": 4426
+    },
+    {
+      "epoch": 2.318491356731273,
+      "loss_partial": 0.04638671875,
+      "step": 4426
+    },
+    {
+      "epoch": 2.318491356731273,
+      "loss_partial": 0.06201171875,
+      "step": 4426
+    },
+    {
+      "epoch": 2.319015191199581,
+      "grad_norm": 1.8248361349105835,
+      "learning_rate": 8.968659114411585e-06,
+      "loss": 0.1052,
+      "step": 4427
+    },
+    {
+      "epoch": 2.319015191199581,
+      "loss_partial": 0.0556640625,
+      "step": 4427
+    },
+    {
+      "epoch": 2.319015191199581,
+      "loss_partial": 0.05126953125,
+      "step": 4427
+    },
+    {
+      "epoch": 2.319539025667889,
+      "grad_norm": 1.7289921045303345,
+      "learning_rate": 8.96411201630132e-06,
+      "loss": 0.1032,
+      "step": 4428
+    },
+    {
+      "epoch": 2.319539025667889,
+      "loss_partial": 0.057861328125,
+      "step": 4428
+    },
+    {
+      "epoch": 2.319539025667889,
+      "loss_partial": 0.055908203125,
+      "step": 4428
+    },
+    {
+      "epoch": 2.320062860136197,
+      "grad_norm": 1.5377099514007568,
+      "learning_rate": 8.95956513468524e-06,
+      "loss": 0.0993,
+      "step": 4429
+    },
+    {
+      "epoch": 2.320062860136197,
+      "loss_partial": 0.042724609375,
+      "step": 4429
+    },
+    {
+      "epoch": 2.320062860136197,
+      "loss_partial": 0.05322265625,
+      "step": 4429
+    },
+    {
+      "epoch": 2.320586694604505,
+      "grad_norm": 1.5346843004226685,
+      "learning_rate": 8.95501847051361e-06,
+      "loss": 0.1023,
+      "step": 4430
+    },
+    {
+      "epoch": 2.320586694604505,
+      "loss_partial": 0.050537109375,
+      "step": 4430
+    },
+    {
+      "epoch": 2.320586694604505,
+      "loss_partial": 0.0419921875,
+      "step": 4430
+    },
+    {
+      "epoch": 2.321110529072813,
+      "grad_norm": 1.2926831245422363,
+      "learning_rate": 8.950472024736659e-06,
+      "loss": 0.1012,
+      "step": 4431
+    },
+    {
+      "epoch": 2.321110529072813,
+      "loss_partial": 0.053466796875,
+      "step": 4431
+    },
+    {
+      "epoch": 2.321110529072813,
+      "loss_partial": 0.055908203125,
+      "step": 4431
+    },
+    {
+      "epoch": 2.321634363541121,
+      "grad_norm": 1.3011127710342407,
+      "learning_rate": 8.945925798304562e-06,
+      "loss": 0.105,
+      "step": 4432
+    },
+    {
+      "epoch": 2.321634363541121,
+      "loss_partial": 0.046875,
+      "step": 4432
+    },
+    {
+      "epoch": 2.321634363541121,
+      "loss_partial": 0.056396484375,
+      "step": 4432
+    },
+    {
+      "epoch": 2.3221581980094292,
+      "grad_norm": 1.4563583135604858,
+      "learning_rate": 8.941379792167457e-06,
+      "loss": 0.1033,
+      "step": 4433
+    },
+    {
+      "epoch": 2.3221581980094292,
+      "loss_partial": 0.053466796875,
+      "step": 4433
+    },
+    {
+      "epoch": 2.3221581980094292,
+      "loss_partial": 0.041259765625,
+      "step": 4433
+    },
+    {
+      "epoch": 2.322682032477737,
+      "grad_norm": 1.2868551015853882,
+      "learning_rate": 8.936834007275429e-06,
+      "loss": 0.1015,
+      "step": 4434
+    },
+    {
+      "epoch": 2.322682032477737,
+      "loss_partial": 0.047119140625,
+      "step": 4434
+    },
+    {
+      "epoch": 2.322682032477737,
+      "loss_partial": 0.057373046875,
+      "step": 4434
+    },
+    {
+      "epoch": 2.323205866946045,
+      "grad_norm": 1.6194247007369995,
+      "learning_rate": 8.932288444578518e-06,
+      "loss": 0.1039,
+      "step": 4435
+    },
+    {
+      "epoch": 2.323205866946045,
+      "loss_partial": 0.041748046875,
+      "step": 4435
+    },
+    {
+      "epoch": 2.323205866946045,
+      "loss_partial": 0.054931640625,
+      "step": 4435
+    },
+    {
+      "epoch": 2.323729701414353,
+      "grad_norm": 1.224615216255188,
+      "learning_rate": 8.927743105026717e-06,
+      "loss": 0.1031,
+      "step": 4436
+    },
+    {
+      "epoch": 2.323729701414353,
+      "loss_partial": 0.049560546875,
+      "step": 4436
+    },
+    {
+      "epoch": 2.323729701414353,
+      "loss_partial": 0.053955078125,
+      "step": 4436
+    },
+    {
+      "epoch": 2.324253535882661,
+      "grad_norm": 1.7025086879730225,
+      "learning_rate": 8.923197989569981e-06,
+      "loss": 0.1004,
+      "step": 4437
+    },
+    {
+      "epoch": 2.324253535882661,
+      "loss_partial": 0.05322265625,
+      "step": 4437
+    },
+    {
+      "epoch": 2.324253535882661,
+      "loss_partial": 0.048095703125,
+      "step": 4437
+    },
+    {
+      "epoch": 2.324777370350969,
+      "grad_norm": 1.2463889122009277,
+      "learning_rate": 8.918653099158208e-06,
+      "loss": 0.0996,
+      "step": 4438
+    },
+    {
+      "epoch": 2.324777370350969,
+      "loss_partial": 0.050048828125,
+      "step": 4438
+    },
+    {
+      "epoch": 2.324777370350969,
+      "loss_partial": 0.053466796875,
+      "step": 4438
+    },
+    {
+      "epoch": 2.325301204819277,
+      "grad_norm": 1.7190848588943481,
+      "learning_rate": 8.91410843474125e-06,
+      "loss": 0.1031,
+      "step": 4439
+    },
+    {
+      "epoch": 2.325301204819277,
+      "loss_partial": 0.04736328125,
+      "step": 4439
+    },
+    {
+      "epoch": 2.325301204819277,
+      "loss_partial": 0.052978515625,
+      "step": 4439
+    },
+    {
+      "epoch": 2.325825039287585,
+      "grad_norm": 1.344190239906311,
+      "learning_rate": 8.909563997268907e-06,
+      "loss": 0.1018,
+      "step": 4440
+    },
+    {
+      "epoch": 2.325825039287585,
+      "loss_partial": 0.068359375,
+      "step": 4440
+    },
+    {
+      "epoch": 2.325825039287585,
+      "loss_partial": 0.052490234375,
+      "step": 4440
+    },
+    {
+      "epoch": 2.3263488737558933,
+      "grad_norm": 1.4745410680770874,
+      "learning_rate": 8.905019787690951e-06,
+      "loss": 0.1039,
+      "step": 4441
+    },
+    {
+      "epoch": 2.3263488737558933,
+      "loss_partial": 0.0546875,
+      "step": 4441
+    },
+    {
+      "epoch": 2.3263488737558933,
+      "loss_partial": 0.0478515625,
+      "step": 4441
+    },
+    {
+      "epoch": 2.3268727082242013,
+      "grad_norm": 1.145776629447937,
+      "learning_rate": 8.900475806957087e-06,
+      "loss": 0.0996,
+      "step": 4442
+    },
+    {
+      "epoch": 2.3268727082242013,
+      "loss_partial": 0.049560546875,
+      "step": 4442
+    },
+    {
+      "epoch": 2.3268727082242013,
+      "loss_partial": 0.0537109375,
+      "step": 4442
+    },
+    {
+      "epoch": 2.327396542692509,
+      "grad_norm": 1.5996785163879395,
+      "learning_rate": 8.895932056016982e-06,
+      "loss": 0.0976,
+      "step": 4443
+    },
+    {
+      "epoch": 2.327396542692509,
+      "loss_partial": 0.052734375,
+      "step": 4443
+    },
+    {
+      "epoch": 2.327396542692509,
+      "loss_partial": 0.04931640625,
+      "step": 4443
+    },
+    {
+      "epoch": 2.327920377160817,
+      "grad_norm": 1.406530499458313,
+      "learning_rate": 8.891388535820247e-06,
+      "loss": 0.0982,
+      "step": 4444
+    },
+    {
+      "epoch": 2.327920377160817,
+      "loss_partial": 0.056396484375,
+      "step": 4444
+    },
+    {
+      "epoch": 2.327920377160817,
+      "loss_partial": 0.048583984375,
+      "step": 4444
+    },
+    {
+      "epoch": 2.328444211629125,
+      "grad_norm": 1.7459743022918701,
+      "learning_rate": 8.886845247316455e-06,
+      "loss": 0.1048,
+      "step": 4445
+    },
+    {
+      "epoch": 2.328444211629125,
+      "loss_partial": 0.055908203125,
+      "step": 4445
+    },
+    {
+      "epoch": 2.328444211629125,
+      "loss_partial": 0.055908203125,
+      "step": 4445
+    },
+    {
+      "epoch": 2.328968046097433,
+      "grad_norm": 1.652561068534851,
+      "learning_rate": 8.882302191455121e-06,
+      "loss": 0.1028,
+      "step": 4446
+    },
+    {
+      "epoch": 2.328968046097433,
+      "loss_partial": 0.052978515625,
+      "step": 4446
+    },
+    {
+      "epoch": 2.328968046097433,
+      "loss_partial": 0.051025390625,
+      "step": 4446
+    },
+    {
+      "epoch": 2.329491880565741,
+      "grad_norm": 1.5962274074554443,
+      "learning_rate": 8.87775936918572e-06,
+      "loss": 0.1014,
+      "step": 4447
+    },
+    {
+      "epoch": 2.329491880565741,
+      "loss_partial": 0.048828125,
+      "step": 4447
+    },
+    {
+      "epoch": 2.329491880565741,
+      "loss_partial": 0.052734375,
+      "step": 4447
+    },
+    {
+      "epoch": 2.3300157150340493,
+      "grad_norm": 1.9747391939163208,
+      "learning_rate": 8.873216781457666e-06,
+      "loss": 0.1012,
+      "step": 4448
+    },
+    {
+      "epoch": 2.3300157150340493,
+      "loss_partial": 0.0419921875,
+      "step": 4448
+    },
+    {
+      "epoch": 2.3300157150340493,
+      "loss_partial": 0.05126953125,
+      "step": 4448
+    },
+    {
+      "epoch": 2.3305395495023573,
+      "grad_norm": 1.9122653007507324,
+      "learning_rate": 8.86867442922034e-06,
+      "loss": 0.1006,
+      "step": 4449
+    },
+    {
+      "epoch": 2.3305395495023573,
+      "loss_partial": 0.0654296875,
+      "step": 4449
+    },
+    {
+      "epoch": 2.3305395495023573,
+      "loss_partial": 0.0576171875,
+      "step": 4449
+    },
+    {
+      "epoch": 2.3310633839706654,
+      "grad_norm": 1.4297784566879272,
+      "learning_rate": 8.864132313423066e-06,
+      "loss": 0.1067,
+      "step": 4450
+    },
+    {
+      "epoch": 2.3310633839706654,
+      "loss_partial": 0.051025390625,
+      "step": 4450
+    },
+    {
+      "epoch": 2.3310633839706654,
+      "loss_partial": 0.043212890625,
+      "step": 4450
+    },
+    {
+      "epoch": 2.3315872184389734,
+      "grad_norm": 2.011317253112793,
+      "learning_rate": 8.859590435015112e-06,
+      "loss": 0.098,
+      "step": 4451
+    },
+    {
+      "epoch": 2.3315872184389734,
+      "loss_partial": 0.048583984375,
+      "step": 4451
+    },
+    {
+      "epoch": 2.3315872184389734,
+      "loss_partial": 0.06201171875,
+      "step": 4451
+    },
+    {
+      "epoch": 2.3321110529072815,
+      "grad_norm": 1.4178022146224976,
+      "learning_rate": 8.8550487949457e-06,
+      "loss": 0.1005,
+      "step": 4452
+    },
+    {
+      "epoch": 2.3321110529072815,
+      "loss_partial": 0.03857421875,
+      "step": 4452
+    },
+    {
+      "epoch": 2.3321110529072815,
+      "loss_partial": 0.052978515625,
+      "step": 4452
+    },
+    {
+      "epoch": 2.3326348873755895,
+      "grad_norm": 1.8215268850326538,
+      "learning_rate": 8.850507394164016e-06,
+      "loss": 0.1032,
+      "step": 4453
+    },
+    {
+      "epoch": 2.3326348873755895,
+      "loss_partial": 0.051025390625,
+      "step": 4453
+    },
+    {
+      "epoch": 2.3326348873755895,
+      "loss_partial": 0.05615234375,
+      "step": 4453
+    },
+    {
+      "epoch": 2.333158721843897,
+      "grad_norm": 1.4268394708633423,
+      "learning_rate": 8.845966233619175e-06,
+      "loss": 0.1044,
+      "step": 4454
+    },
+    {
+      "epoch": 2.333158721843897,
+      "loss_partial": 0.037109375,
+      "step": 4454
+    },
+    {
+      "epoch": 2.333158721843897,
+      "loss_partial": 0.05322265625,
+      "step": 4454
+    },
+    {
+      "epoch": 2.3336825563122052,
+      "grad_norm": 2.519617795944214,
+      "learning_rate": 8.841425314260257e-06,
+      "loss": 0.1033,
+      "step": 4455
+    },
+    {
+      "epoch": 2.3336825563122052,
+      "loss_partial": 0.04736328125,
+      "step": 4455
+    },
+    {
+      "epoch": 2.3336825563122052,
+      "loss_partial": 0.052734375,
+      "step": 4455
+    },
+    {
+      "epoch": 2.3342063907805133,
+      "grad_norm": 1.9496541023254395,
+      "learning_rate": 8.836884637036281e-06,
+      "loss": 0.1032,
+      "step": 4456
+    },
+    {
+      "epoch": 2.3342063907805133,
+      "loss_partial": 0.051025390625,
+      "step": 4456
+    },
+    {
+      "epoch": 2.3342063907805133,
+      "loss_partial": 0.05126953125,
+      "step": 4456
+    },
+    {
+      "epoch": 2.3347302252488213,
+      "grad_norm": 1.8401938676834106,
+      "learning_rate": 8.832344202896226e-06,
+      "loss": 0.1017,
+      "step": 4457
+    },
+    {
+      "epoch": 2.3347302252488213,
+      "loss_partial": 0.056884765625,
+      "step": 4457
+    },
+    {
+      "epoch": 2.3347302252488213,
+      "loss_partial": 0.0537109375,
+      "step": 4457
+    },
+    {
+      "epoch": 2.3352540597171294,
+      "grad_norm": 1.8421911001205444,
+      "learning_rate": 8.827804012789013e-06,
+      "loss": 0.106,
+      "step": 4458
+    },
+    {
+      "epoch": 2.3352540597171294,
+      "loss_partial": 0.05712890625,
+      "step": 4458
+    },
+    {
+      "epoch": 2.3352540597171294,
+      "loss_partial": 0.04638671875,
+      "step": 4458
+    },
+    {
+      "epoch": 2.3357778941854375,
+      "grad_norm": 1.5475691556930542,
+      "learning_rate": 8.823264067663511e-06,
+      "loss": 0.1008,
+      "step": 4459
+    },
+    {
+      "epoch": 2.3357778941854375,
+      "loss_partial": 0.048828125,
+      "step": 4459
+    },
+    {
+      "epoch": 2.3357778941854375,
+      "loss_partial": 0.046630859375,
+      "step": 4459
+    },
+    {
+      "epoch": 2.3363017286537455,
+      "grad_norm": 1.8891441822052002,
+      "learning_rate": 8.818724368468539e-06,
+      "loss": 0.1006,
+      "step": 4460
+    },
+    {
+      "epoch": 2.3363017286537455,
+      "loss_partial": 0.056396484375,
+      "step": 4460
+    },
+    {
+      "epoch": 2.3363017286537455,
+      "loss_partial": 0.044921875,
+      "step": 4460
+    },
+    {
+      "epoch": 2.3368255631220536,
+      "grad_norm": 1.5982508659362793,
+      "learning_rate": 8.814184916152873e-06,
+      "loss": 0.102,
+      "step": 4461
+    },
+    {
+      "epoch": 2.3368255631220536,
+      "loss_partial": 0.0458984375,
+      "step": 4461
+    },
+    {
+      "epoch": 2.3368255631220536,
+      "loss_partial": 0.07177734375,
+      "step": 4461
+    },
+    {
+      "epoch": 2.337349397590361,
+      "grad_norm": 2.042365074157715,
+      "learning_rate": 8.809645711665226e-06,
+      "loss": 0.1024,
+      "step": 4462
+    },
+    {
+      "epoch": 2.337349397590361,
+      "loss_partial": 0.053466796875,
+      "step": 4462
+    },
+    {
+      "epoch": 2.337349397590361,
+      "loss_partial": 0.04638671875,
+      "step": 4462
+    },
+    {
+      "epoch": 2.3378732320586693,
+      "grad_norm": 1.748291015625,
+      "learning_rate": 8.805106755954269e-06,
+      "loss": 0.1019,
+      "step": 4463
+    },
+    {
+      "epoch": 2.3378732320586693,
+      "loss_partial": 0.051025390625,
+      "step": 4463
+    },
+    {
+      "epoch": 2.3378732320586693,
+      "loss_partial": 0.04638671875,
+      "step": 4463
+    },
+    {
+      "epoch": 2.3383970665269773,
+      "grad_norm": 1.2839500904083252,
+      "learning_rate": 8.800568049968602e-06,
+      "loss": 0.1049,
+      "step": 4464
+    },
+    {
+      "epoch": 2.3383970665269773,
+      "loss_partial": 0.052490234375,
+      "step": 4464
+    },
+    {
+      "epoch": 2.3383970665269773,
+      "loss_partial": 0.04931640625,
+      "step": 4464
+    },
+    {
+      "epoch": 2.3389209009952854,
+      "grad_norm": 2.230727195739746,
+      "learning_rate": 8.7960295946568e-06,
+      "loss": 0.1034,
+      "step": 4465
+    },
+    {
+      "epoch": 2.3389209009952854,
+      "loss_partial": 0.042724609375,
+      "step": 4465
+    },
+    {
+      "epoch": 2.3389209009952854,
+      "loss_partial": 0.05322265625,
+      "step": 4465
+    },
+    {
+      "epoch": 2.3394447354635934,
+      "grad_norm": 1.1785823106765747,
+      "learning_rate": 8.791491390967368e-06,
+      "loss": 0.1028,
+      "step": 4466
+    },
+    {
+      "epoch": 2.3394447354635934,
+      "loss_partial": 0.04833984375,
+      "step": 4466
+    },
+    {
+      "epoch": 2.3394447354635934,
+      "loss_partial": 0.064453125,
+      "step": 4466
+    },
+    {
+      "epoch": 2.3399685699319015,
+      "grad_norm": 1.923329472541809,
+      "learning_rate": 8.786953439848762e-06,
+      "loss": 0.1066,
+      "step": 4467
+    },
+    {
+      "epoch": 2.3399685699319015,
+      "loss_partial": 0.04638671875,
+      "step": 4467
+    },
+    {
+      "epoch": 2.3399685699319015,
+      "loss_partial": 0.043212890625,
+      "step": 4467
+    },
+    {
+      "epoch": 2.3404924044002096,
+      "grad_norm": 1.5378915071487427,
+      "learning_rate": 8.782415742249383e-06,
+      "loss": 0.1019,
+      "step": 4468
+    },
+    {
+      "epoch": 2.3404924044002096,
+      "loss_partial": 0.059326171875,
+      "step": 4468
+    },
+    {
+      "epoch": 2.3404924044002096,
+      "loss_partial": 0.044189453125,
+      "step": 4468
+    },
+    {
+      "epoch": 2.3410162388685176,
+      "grad_norm": 1.8043971061706543,
+      "learning_rate": 8.777878299117589e-06,
+      "loss": 0.1034,
+      "step": 4469
+    },
+    {
+      "epoch": 2.3410162388685176,
+      "loss_partial": 0.049072265625,
+      "step": 4469
+    },
+    {
+      "epoch": 2.3410162388685176,
+      "loss_partial": 0.0556640625,
+      "step": 4469
+    },
+    {
+      "epoch": 2.3415400733368257,
+      "grad_norm": 1.545474886894226,
+      "learning_rate": 8.773341111401672e-06,
+      "loss": 0.1021,
+      "step": 4470
+    },
+    {
+      "epoch": 2.3415400733368257,
+      "loss_partial": 0.061767578125,
+      "step": 4470
+    },
+    {
+      "epoch": 2.3415400733368257,
+      "loss_partial": 0.045654296875,
+      "step": 4470
+    },
+    {
+      "epoch": 2.3420639078051337,
+      "grad_norm": 1.9379117488861084,
+      "learning_rate": 8.768804180049878e-06,
+      "loss": 0.1035,
+      "step": 4471
+    },
+    {
+      "epoch": 2.3420639078051337,
+      "loss_partial": 0.044189453125,
+      "step": 4471
+    },
+    {
+      "epoch": 2.3420639078051337,
+      "loss_partial": 0.050048828125,
+      "step": 4471
+    },
+    {
+      "epoch": 2.342587742273442,
+      "grad_norm": 1.9209550619125366,
+      "learning_rate": 8.76426750601039e-06,
+      "loss": 0.106,
+      "step": 4472
+    },
+    {
+      "epoch": 2.342587742273442,
+      "loss_partial": 0.053466796875,
+      "step": 4472
+    },
+    {
+      "epoch": 2.342587742273442,
+      "loss_partial": 0.050537109375,
+      "step": 4472
+    },
+    {
+      "epoch": 2.3431115767417494,
+      "grad_norm": 1.5488996505737305,
+      "learning_rate": 8.759731090231357e-06,
+      "loss": 0.1011,
+      "step": 4473
+    },
+    {
+      "epoch": 2.3431115767417494,
+      "loss_partial": 0.0458984375,
+      "step": 4473
+    },
+    {
+      "epoch": 2.3431115767417494,
+      "loss_partial": 0.05712890625,
+      "step": 4473
+    },
+    {
+      "epoch": 2.3436354112100575,
+      "grad_norm": 1.8284780979156494,
+      "learning_rate": 8.755194933660857e-06,
+      "loss": 0.1031,
+      "step": 4474
+    },
+    {
+      "epoch": 2.3436354112100575,
+      "loss_partial": 0.05859375,
+      "step": 4474
+    },
+    {
+      "epoch": 2.3436354112100575,
+      "loss_partial": 0.05126953125,
+      "step": 4474
+    },
+    {
+      "epoch": 2.3441592456783655,
+      "grad_norm": 1.7574565410614014,
+      "learning_rate": 8.75065903724692e-06,
+      "loss": 0.102,
+      "step": 4475
+    },
+    {
+      "epoch": 2.3441592456783655,
+      "loss_partial": 0.0556640625,
+      "step": 4475
+    },
+    {
+      "epoch": 2.3441592456783655,
+      "loss_partial": 0.050048828125,
+      "step": 4475
+    },
+    {
+      "epoch": 2.3446830801466736,
+      "grad_norm": 1.8951871395111084,
+      "learning_rate": 8.746123401937514e-06,
+      "loss": 0.101,
+      "step": 4476
+    },
+    {
+      "epoch": 2.3446830801466736,
+      "loss_partial": 0.0498046875,
+      "step": 4476
+    },
+    {
+      "epoch": 2.3446830801466736,
+      "loss_partial": 0.055419921875,
+      "step": 4476
+    },
+    {
+      "epoch": 2.3452069146149817,
+      "grad_norm": 1.4653795957565308,
+      "learning_rate": 8.741588028680566e-06,
+      "loss": 0.1008,
+      "step": 4477
+    },
+    {
+      "epoch": 2.3452069146149817,
+      "loss_partial": 0.050048828125,
+      "step": 4477
+    },
+    {
+      "epoch": 2.3452069146149817,
+      "loss_partial": 0.0517578125,
+      "step": 4477
+    },
+    {
+      "epoch": 2.3457307490832897,
+      "grad_norm": 2.6122961044311523,
+      "learning_rate": 8.737052918423939e-06,
+      "loss": 0.1019,
+      "step": 4478
+    },
+    {
+      "epoch": 2.3457307490832897,
+      "loss_partial": 0.0458984375,
+      "step": 4478
+    },
+    {
+      "epoch": 2.3457307490832897,
+      "loss_partial": 0.039794921875,
+      "step": 4478
+    },
+    {
+      "epoch": 2.3462545835515978,
+      "grad_norm": 1.786685585975647,
+      "learning_rate": 8.73251807211544e-06,
+      "loss": 0.0994,
+      "step": 4479
+    },
+    {
+      "epoch": 2.3462545835515978,
+      "loss_partial": 0.04443359375,
+      "step": 4479
+    },
+    {
+      "epoch": 2.3462545835515978,
+      "loss_partial": 0.051025390625,
+      "step": 4479
+    },
+    {
+      "epoch": 2.346778418019906,
+      "grad_norm": 2.0257086753845215,
+      "learning_rate": 8.727983490702826e-06,
+      "loss": 0.1013,
+      "step": 4480
+    },
+    {
+      "epoch": 2.346778418019906,
+      "loss_partial": 0.044189453125,
+      "step": 4480
+    },
+    {
+      "epoch": 2.346778418019906,
+      "loss_partial": 0.05322265625,
+      "step": 4480
+    },
+    {
+      "epoch": 2.347302252488214,
+      "grad_norm": 1.8929990530014038,
+      "learning_rate": 8.723449175133799e-06,
+      "loss": 0.1038,
+      "step": 4481
+    },
+    {
+      "epoch": 2.347302252488214,
+      "loss_partial": 0.05419921875,
+      "step": 4481
+    },
+    {
+      "epoch": 2.347302252488214,
+      "loss_partial": 0.052001953125,
+      "step": 4481
+    },
+    {
+      "epoch": 2.3478260869565215,
+      "grad_norm": 1.8394705057144165,
+      "learning_rate": 8.718915126355999e-06,
+      "loss": 0.1058,
+      "step": 4482
+    },
+    {
+      "epoch": 2.3478260869565215,
+      "loss_partial": 0.052001953125,
+      "step": 4482
+    },
+    {
+      "epoch": 2.3478260869565215,
+      "loss_partial": 0.06103515625,
+      "step": 4482
+    },
+    {
+      "epoch": 2.3483499214248296,
+      "grad_norm": 1.8875155448913574,
+      "learning_rate": 8.714381345317015e-06,
+      "loss": 0.1024,
+      "step": 4483
+    },
+    {
+      "epoch": 2.3483499214248296,
+      "loss_partial": 0.0498046875,
+      "step": 4483
+    },
+    {
+      "epoch": 2.3483499214248296,
+      "loss_partial": 0.0478515625,
+      "step": 4483
+    },
+    {
+      "epoch": 2.3488737558931376,
+      "grad_norm": 2.0078558921813965,
+      "learning_rate": 8.709847832964378e-06,
+      "loss": 0.103,
+      "step": 4484
+    },
+    {
+      "epoch": 2.3488737558931376,
+      "loss_partial": 0.044921875,
+      "step": 4484
+    },
+    {
+      "epoch": 2.3488737558931376,
+      "loss_partial": 0.047119140625,
+      "step": 4484
+    },
+    {
+      "epoch": 2.3493975903614457,
+      "grad_norm": 1.894219994544983,
+      "learning_rate": 8.705314590245568e-06,
+      "loss": 0.1002,
+      "step": 4485
+    },
+    {
+      "epoch": 2.3493975903614457,
+      "loss_partial": 0.048095703125,
+      "step": 4485
+    },
+    {
+      "epoch": 2.3493975903614457,
+      "loss_partial": 0.05517578125,
+      "step": 4485
+    },
+    {
+      "epoch": 2.3499214248297537,
+      "grad_norm": 1.52084219455719,
+      "learning_rate": 8.700781618108003e-06,
+      "loss": 0.0984,
+      "step": 4486
+    },
+    {
+      "epoch": 2.3499214248297537,
+      "loss_partial": 0.049560546875,
+      "step": 4486
+    },
+    {
+      "epoch": 2.3499214248297537,
+      "loss_partial": 0.059814453125,
+      "step": 4486
+    },
+    {
+      "epoch": 2.350445259298062,
+      "grad_norm": 1.637711524963379,
+      "learning_rate": 8.696248917499049e-06,
+      "loss": 0.1054,
+      "step": 4487
+    },
+    {
+      "epoch": 2.350445259298062,
+      "loss_partial": 0.048828125,
+      "step": 4487
+    },
+    {
+      "epoch": 2.350445259298062,
+      "loss_partial": 0.05615234375,
+      "step": 4487
+    },
+    {
+      "epoch": 2.35096909376637,
+      "grad_norm": 1.7618331909179688,
+      "learning_rate": 8.691716489366006e-06,
+      "loss": 0.1057,
+      "step": 4488
+    },
+    {
+      "epoch": 2.35096909376637,
+      "loss_partial": 0.07421875,
+      "step": 4488
+    },
+    {
+      "epoch": 2.35096909376637,
+      "loss_partial": 0.049072265625,
+      "step": 4488
+    },
+    {
+      "epoch": 2.351492928234678,
+      "grad_norm": 1.7915499210357666,
+      "learning_rate": 8.687184334656122e-06,
+      "loss": 0.1075,
+      "step": 4489
+    },
+    {
+      "epoch": 2.351492928234678,
+      "loss_partial": 0.049072265625,
+      "step": 4489
+    },
+    {
+      "epoch": 2.351492928234678,
+      "loss_partial": 0.053955078125,
+      "step": 4489
+    },
+    {
+      "epoch": 2.352016762702986,
+      "grad_norm": 1.709713339805603,
+      "learning_rate": 8.682652454316598e-06,
+      "loss": 0.1043,
+      "step": 4490
+    },
+    {
+      "epoch": 2.352016762702986,
+      "loss_partial": 0.053466796875,
+      "step": 4490
+    },
+    {
+      "epoch": 2.352016762702986,
+      "loss_partial": 0.05126953125,
+      "step": 4490
+    },
+    {
+      "epoch": 2.352540597171294,
+      "grad_norm": 1.4580925703048706,
+      "learning_rate": 8.678120849294566e-06,
+      "loss": 0.1012,
+      "step": 4491
+    },
+    {
+      "epoch": 2.352540597171294,
+      "loss_partial": 0.048828125,
+      "step": 4491
+    },
+    {
+      "epoch": 2.352540597171294,
+      "loss_partial": 0.043701171875,
+      "step": 4491
+    },
+    {
+      "epoch": 2.353064431639602,
+      "grad_norm": 1.7233365774154663,
+      "learning_rate": 8.673589520537102e-06,
+      "loss": 0.1024,
+      "step": 4492
+    },
+    {
+      "epoch": 2.353064431639602,
+      "loss_partial": 0.0537109375,
+      "step": 4492
+    },
+    {
+      "epoch": 2.353064431639602,
+      "loss_partial": 0.048583984375,
+      "step": 4492
+    },
+    {
+      "epoch": 2.3535882661079097,
+      "grad_norm": 1.7322160005569458,
+      "learning_rate": 8.669058468991224e-06,
+      "loss": 0.1025,
+      "step": 4493
+    },
+    {
+      "epoch": 2.3535882661079097,
+      "loss_partial": 0.0537109375,
+      "step": 4493
+    },
+    {
+      "epoch": 2.3535882661079097,
+      "loss_partial": 0.06005859375,
+      "step": 4493
+    },
+    {
+      "epoch": 2.354112100576218,
+      "grad_norm": 2.704944610595703,
+      "learning_rate": 8.664527695603897e-06,
+      "loss": 0.1061,
+      "step": 4494
+    },
+    {
+      "epoch": 2.354112100576218,
+      "loss_partial": 0.0537109375,
+      "step": 4494
+    },
+    {
+      "epoch": 2.354112100576218,
+      "loss_partial": 0.04443359375,
+      "step": 4494
+    },
+    {
+      "epoch": 2.354635935044526,
+      "grad_norm": 1.1287200450897217,
+      "learning_rate": 8.659997201322024e-06,
+      "loss": 0.1025,
+      "step": 4495
+    },
+    {
+      "epoch": 2.354635935044526,
+      "loss_partial": 0.047607421875,
+      "step": 4495
+    },
+    {
+      "epoch": 2.354635935044526,
+      "loss_partial": 0.05615234375,
+      "step": 4495
+    },
+    {
+      "epoch": 2.355159769512834,
+      "grad_norm": 2.3049583435058594,
+      "learning_rate": 8.65546698709245e-06,
+      "loss": 0.1058,
+      "step": 4496
+    },
+    {
+      "epoch": 2.355159769512834,
+      "loss_partial": 0.052734375,
+      "step": 4496
+    },
+    {
+      "epoch": 2.355159769512834,
+      "loss_partial": 0.06298828125,
+      "step": 4496
+    },
+    {
+      "epoch": 2.355683603981142,
+      "grad_norm": 2.0003669261932373,
+      "learning_rate": 8.650937053861954e-06,
+      "loss": 0.103,
+      "step": 4497
+    },
+    {
+      "epoch": 2.355683603981142,
+      "loss_partial": 0.0498046875,
+      "step": 4497
+    },
+    {
+      "epoch": 2.355683603981142,
+      "loss_partial": 0.056640625,
+      "step": 4497
+    },
+    {
+      "epoch": 2.35620743844945,
+      "grad_norm": 1.4333171844482422,
+      "learning_rate": 8.646407402577279e-06,
+      "loss": 0.1024,
+      "step": 4498
+    },
+    {
+      "epoch": 2.35620743844945,
+      "loss_partial": 0.056396484375,
+      "step": 4498
+    },
+    {
+      "epoch": 2.35620743844945,
+      "loss_partial": 0.04638671875,
+      "step": 4498
+    },
+    {
+      "epoch": 2.356731272917758,
+      "grad_norm": 1.508885145187378,
+      "learning_rate": 8.641878034185085e-06,
+      "loss": 0.1018,
+      "step": 4499
+    },
+    {
+      "epoch": 2.356731272917758,
+      "loss_partial": 0.056884765625,
+      "step": 4499
+    },
+    {
+      "epoch": 2.356731272917758,
+      "loss_partial": 0.0458984375,
+      "step": 4499
+    },
+    {
+      "epoch": 2.357255107386066,
+      "grad_norm": 2.286069631576538,
+      "learning_rate": 8.637348949631986e-06,
+      "loss": 0.1028,
+      "step": 4500
+    },
+    {
+      "epoch": 2.357255107386066,
+      "loss_partial": 0.044677734375,
+      "step": 4500
+    },
+    {
+      "epoch": 2.357255107386066,
+      "loss_partial": 0.0576171875,
+      "step": 4500
+    },
+    {
+      "epoch": 2.357778941854374,
+      "grad_norm": 1.2658271789550781,
+      "learning_rate": 8.632820149864524e-06,
+      "loss": 0.1007,
+      "step": 4501
+    },
+    {
+      "epoch": 2.357778941854374,
+      "loss_partial": 0.052490234375,
+      "step": 4501
+    },
+    {
+      "epoch": 2.357778941854374,
+      "loss_partial": 0.049560546875,
+      "step": 4501
+    },
+    {
+      "epoch": 2.358302776322682,
+      "grad_norm": 2.401089668273926,
+      "learning_rate": 8.6282916358292e-06,
+      "loss": 0.1039,
+      "step": 4502
+    },
+    {
+      "epoch": 2.358302776322682,
+      "loss_partial": 0.0498046875,
+      "step": 4502
+    },
+    {
+      "epoch": 2.358302776322682,
+      "loss_partial": 0.05224609375,
+      "step": 4502
+    },
+    {
+      "epoch": 2.35882661079099,
+      "grad_norm": 1.2499477863311768,
+      "learning_rate": 8.623763408472443e-06,
+      "loss": 0.0989,
+      "step": 4503
+    },
+    {
+      "epoch": 2.35882661079099,
+      "loss_partial": 0.05126953125,
+      "step": 4503
+    },
+    {
+      "epoch": 2.35882661079099,
+      "loss_partial": 0.056396484375,
+      "step": 4503
+    },
+    {
+      "epoch": 2.359350445259298,
+      "grad_norm": 2.3595468997955322,
+      "learning_rate": 8.619235468740621e-06,
+      "loss": 0.1024,
+      "step": 4504
+    },
+    {
+      "epoch": 2.359350445259298,
+      "loss_partial": 0.047119140625,
+      "step": 4504
+    },
+    {
+      "epoch": 2.359350445259298,
+      "loss_partial": 0.052734375,
+      "step": 4504
+    },
+    {
+      "epoch": 2.359874279727606,
+      "grad_norm": 1.5207692384719849,
+      "learning_rate": 8.614707817580046e-06,
+      "loss": 0.1036,
+      "step": 4505
+    },
+    {
+      "epoch": 2.359874279727606,
+      "loss_partial": 0.049072265625,
+      "step": 4505
+    },
+    {
+      "epoch": 2.359874279727606,
+      "loss_partial": 0.049560546875,
+      "step": 4505
+    },
+    {
+      "epoch": 2.360398114195914,
+      "grad_norm": 2.071824073791504,
+      "learning_rate": 8.610180455936974e-06,
+      "loss": 0.1059,
+      "step": 4506
+    },
+    {
+      "epoch": 2.360398114195914,
+      "loss_partial": 0.051513671875,
+      "step": 4506
+    },
+    {
+      "epoch": 2.360398114195914,
+      "loss_partial": 0.050537109375,
+      "step": 4506
+    },
+    {
+      "epoch": 2.360921948664222,
+      "grad_norm": 1.4176607131958008,
+      "learning_rate": 8.605653384757592e-06,
+      "loss": 0.0996,
+      "step": 4507
+    },
+    {
+      "epoch": 2.360921948664222,
+      "loss_partial": 0.048828125,
+      "step": 4507
+    },
+    {
+      "epoch": 2.360921948664222,
+      "loss_partial": 0.056396484375,
+      "step": 4507
+    },
+    {
+      "epoch": 2.36144578313253,
+      "grad_norm": 1.5385297536849976,
+      "learning_rate": 8.60112660498803e-06,
+      "loss": 0.1049,
+      "step": 4508
+    },
+    {
+      "epoch": 2.36144578313253,
+      "loss_partial": 0.055908203125,
+      "step": 4508
+    },
+    {
+      "epoch": 2.36144578313253,
+      "loss_partial": 0.057861328125,
+      "step": 4508
+    },
+    {
+      "epoch": 2.3619696176008382,
+      "grad_norm": 1.251641869544983,
+      "learning_rate": 8.596600117574353e-06,
+      "loss": 0.1008,
+      "step": 4509
+    },
+    {
+      "epoch": 2.3619696176008382,
+      "loss_partial": 0.046630859375,
+      "step": 4509
+    },
+    {
+      "epoch": 2.3619696176008382,
+      "loss_partial": 0.043212890625,
+      "step": 4509
+    },
+    {
+      "epoch": 2.3624934520691463,
+      "grad_norm": 1.6160829067230225,
+      "learning_rate": 8.592073923462578e-06,
+      "loss": 0.1005,
+      "step": 4510
+    },
+    {
+      "epoch": 2.3624934520691463,
+      "loss_partial": 0.04931640625,
+      "step": 4510
+    },
+    {
+      "epoch": 2.3624934520691463,
+      "loss_partial": 0.051513671875,
+      "step": 4510
+    },
+    {
+      "epoch": 2.3630172865374544,
+      "grad_norm": 1.2653508186340332,
+      "learning_rate": 8.587548023598646e-06,
+      "loss": 0.0984,
+      "step": 4511
+    },
+    {
+      "epoch": 2.3630172865374544,
+      "loss_partial": 0.05810546875,
+      "step": 4511
+    },
+    {
+      "epoch": 2.3630172865374544,
+      "loss_partial": 0.056640625,
+      "step": 4511
+    },
+    {
+      "epoch": 2.3635411210057624,
+      "grad_norm": 1.4921979904174805,
+      "learning_rate": 8.583022418928444e-06,
+      "loss": 0.1016,
+      "step": 4512
+    },
+    {
+      "epoch": 2.3635411210057624,
+      "loss_partial": 0.0458984375,
+      "step": 4512
+    },
+    {
+      "epoch": 2.3635411210057624,
+      "loss_partial": 0.036376953125,
+      "step": 4512
+    },
+    {
+      "epoch": 2.36406495547407,
+      "grad_norm": 1.4313846826553345,
+      "learning_rate": 8.57849711039779e-06,
+      "loss": 0.0997,
+      "step": 4513
+    },
+    {
+      "epoch": 2.36406495547407,
+      "loss_partial": 0.047607421875,
+      "step": 4513
+    },
+    {
+      "epoch": 2.36406495547407,
+      "loss_partial": 0.05029296875,
+      "step": 4513
+    },
+    {
+      "epoch": 2.364588789942378,
+      "grad_norm": 1.5028691291809082,
+      "learning_rate": 8.573972098952452e-06,
+      "loss": 0.0992,
+      "step": 4514
+    },
+    {
+      "epoch": 2.364588789942378,
+      "loss_partial": 0.05029296875,
+      "step": 4514
+    },
+    {
+      "epoch": 2.364588789942378,
+      "loss_partial": 0.0439453125,
+      "step": 4514
+    },
+    {
+      "epoch": 2.365112624410686,
+      "grad_norm": 1.4575612545013428,
+      "learning_rate": 8.569447385538128e-06,
+      "loss": 0.0982,
+      "step": 4515
+    },
+    {
+      "epoch": 2.365112624410686,
+      "loss_partial": 0.043701171875,
+      "step": 4515
+    },
+    {
+      "epoch": 2.365112624410686,
+      "loss_partial": 0.05078125,
+      "step": 4515
+    },
+    {
+      "epoch": 2.365636458878994,
+      "grad_norm": 1.5960453748703003,
+      "learning_rate": 8.564922971100453e-06,
+      "loss": 0.1051,
+      "step": 4516
+    },
+    {
+      "epoch": 2.365636458878994,
+      "loss_partial": 0.056640625,
+      "step": 4516
+    },
+    {
+      "epoch": 2.365636458878994,
+      "loss_partial": 0.057861328125,
+      "step": 4516
+    },
+    {
+      "epoch": 2.3661602933473023,
+      "grad_norm": 1.687628984451294,
+      "learning_rate": 8.560398856585002e-06,
+      "loss": 0.104,
+      "step": 4517
+    },
+    {
+      "epoch": 2.3661602933473023,
+      "loss_partial": 0.0478515625,
+      "step": 4517
+    },
+    {
+      "epoch": 2.3661602933473023,
+      "loss_partial": 0.05126953125,
+      "step": 4517
+    },
+    {
+      "epoch": 2.3666841278156103,
+      "grad_norm": 1.4938985109329224,
+      "learning_rate": 8.555875042937288e-06,
+      "loss": 0.1039,
+      "step": 4518
+    },
+    {
+      "epoch": 2.3666841278156103,
+      "loss_partial": 0.052001953125,
+      "step": 4518
+    },
+    {
+      "epoch": 2.3666841278156103,
+      "loss_partial": 0.052001953125,
+      "step": 4518
+    },
+    {
+      "epoch": 2.3672079622839184,
+      "grad_norm": 1.7583786249160767,
+      "learning_rate": 8.551351531102762e-06,
+      "loss": 0.1049,
+      "step": 4519
+    },
+    {
+      "epoch": 2.3672079622839184,
+      "loss_partial": 0.049072265625,
+      "step": 4519
+    },
+    {
+      "epoch": 2.3672079622839184,
+      "loss_partial": 0.053466796875,
+      "step": 4519
+    },
+    {
+      "epoch": 2.3677317967522264,
+      "grad_norm": 1.2027814388275146,
+      "learning_rate": 8.546828322026806e-06,
+      "loss": 0.0999,
+      "step": 4520
+    },
+    {
+      "epoch": 2.3677317967522264,
+      "loss_partial": 0.04638671875,
+      "step": 4520
+    },
+    {
+      "epoch": 2.3677317967522264,
+      "loss_partial": 0.052734375,
+      "step": 4520
+    },
+    {
+      "epoch": 2.368255631220534,
+      "grad_norm": 1.327979326248169,
+      "learning_rate": 8.542305416654741e-06,
+      "loss": 0.1016,
+      "step": 4521
+    },
+    {
+      "epoch": 2.368255631220534,
+      "loss_partial": 0.048828125,
+      "step": 4521
+    },
+    {
+      "epoch": 2.368255631220534,
+      "loss_partial": 0.05126953125,
+      "step": 4521
+    },
+    {
+      "epoch": 2.368779465688842,
+      "grad_norm": 1.3883562088012695,
+      "learning_rate": 8.537782815931834e-06,
+      "loss": 0.1017,
+      "step": 4522
+    },
+    {
+      "epoch": 2.368779465688842,
+      "loss_partial": 0.04443359375,
+      "step": 4522
+    },
+    {
+      "epoch": 2.368779465688842,
+      "loss_partial": 0.056396484375,
+      "step": 4522
+    },
+    {
+      "epoch": 2.36930330015715,
+      "grad_norm": 1.1816929578781128,
+      "learning_rate": 8.533260520803277e-06,
+      "loss": 0.1041,
+      "step": 4523
+    },
+    {
+      "epoch": 2.36930330015715,
+      "loss_partial": 0.041748046875,
+      "step": 4523
+    },
+    {
+      "epoch": 2.36930330015715,
+      "loss_partial": 0.056396484375,
+      "step": 4523
+    },
+    {
+      "epoch": 2.3698271346254582,
+      "grad_norm": 1.732053518295288,
+      "learning_rate": 8.528738532214202e-06,
+      "loss": 0.1007,
+      "step": 4524
+    },
+    {
+      "epoch": 2.3698271346254582,
+      "loss_partial": 0.04931640625,
+      "step": 4524
+    },
+    {
+      "epoch": 2.3698271346254582,
+      "loss_partial": 0.05126953125,
+      "step": 4524
+    },
+    {
+      "epoch": 2.3703509690937663,
+      "grad_norm": 1.2407256364822388,
+      "learning_rate": 8.524216851109672e-06,
+      "loss": 0.1034,
+      "step": 4525
+    },
+    {
+      "epoch": 2.3703509690937663,
+      "loss_partial": 0.0517578125,
+      "step": 4525
+    },
+    {
+      "epoch": 2.3703509690937663,
+      "loss_partial": 0.05126953125,
+      "step": 4525
+    },
+    {
+      "epoch": 2.3708748035620744,
+      "grad_norm": 1.8476130962371826,
+      "learning_rate": 8.519695478434697e-06,
+      "loss": 0.1042,
+      "step": 4526
+    },
+    {
+      "epoch": 2.3708748035620744,
+      "loss_partial": 0.05810546875,
+      "step": 4526
+    },
+    {
+      "epoch": 2.3708748035620744,
+      "loss_partial": 0.047119140625,
+      "step": 4526
+    },
+    {
+      "epoch": 2.3713986380303824,
+      "grad_norm": 1.5558186769485474,
+      "learning_rate": 8.515174415134212e-06,
+      "loss": 0.1024,
+      "step": 4527
+    },
+    {
+      "epoch": 2.3713986380303824,
+      "loss_partial": 0.056396484375,
+      "step": 4527
+    },
+    {
+      "epoch": 2.3713986380303824,
+      "loss_partial": 0.052001953125,
+      "step": 4527
+    },
+    {
+      "epoch": 2.3719224724986905,
+      "grad_norm": 1.377243995666504,
+      "learning_rate": 8.510653662153093e-06,
+      "loss": 0.0997,
+      "step": 4528
+    },
+    {
+      "epoch": 2.3719224724986905,
+      "loss_partial": 0.052490234375,
+      "step": 4528
+    },
+    {
+      "epoch": 2.3719224724986905,
+      "loss_partial": 0.05322265625,
+      "step": 4528
+    },
+    {
+      "epoch": 2.3724463069669985,
+      "grad_norm": 1.4360464811325073,
+      "learning_rate": 8.506133220436147e-06,
+      "loss": 0.1004,
+      "step": 4529
+    },
+    {
+      "epoch": 2.3724463069669985,
+      "loss_partial": 0.051025390625,
+      "step": 4529
+    },
+    {
+      "epoch": 2.3724463069669985,
+      "loss_partial": 0.053955078125,
+      "step": 4529
+    },
+    {
+      "epoch": 2.3729701414353066,
+      "grad_norm": 1.5705136060714722,
+      "learning_rate": 8.501613090928122e-06,
+      "loss": 0.099,
+      "step": 4530
+    },
+    {
+      "epoch": 2.3729701414353066,
+      "loss_partial": 0.044677734375,
+      "step": 4530
+    },
+    {
+      "epoch": 2.3729701414353066,
+      "loss_partial": 0.058349609375,
+      "step": 4530
+    },
+    {
+      "epoch": 2.3734939759036147,
+      "grad_norm": 1.5565390586853027,
+      "learning_rate": 8.497093274573695e-06,
+      "loss": 0.1021,
+      "step": 4531
+    },
+    {
+      "epoch": 2.3734939759036147,
+      "loss_partial": 0.044677734375,
+      "step": 4531
+    },
+    {
+      "epoch": 2.3734939759036147,
+      "loss_partial": 0.0517578125,
+      "step": 4531
+    },
+    {
+      "epoch": 2.3740178103719227,
+      "grad_norm": 1.2570598125457764,
+      "learning_rate": 8.49257377231748e-06,
+      "loss": 0.1022,
+      "step": 4532
+    },
+    {
+      "epoch": 2.3740178103719227,
+      "loss_partial": 0.055908203125,
+      "step": 4532
+    },
+    {
+      "epoch": 2.3740178103719227,
+      "loss_partial": 0.046630859375,
+      "step": 4532
+    },
+    {
+      "epoch": 2.3745416448402303,
+      "grad_norm": 1.9581624269485474,
+      "learning_rate": 8.48805458510402e-06,
+      "loss": 0.1038,
+      "step": 4533
+    },
+    {
+      "epoch": 2.3745416448402303,
+      "loss_partial": 0.04345703125,
+      "step": 4533
+    },
+    {
+      "epoch": 2.3745416448402303,
+      "loss_partial": 0.042236328125,
+      "step": 4533
+    },
+    {
+      "epoch": 2.3750654793085384,
+      "grad_norm": 1.0174909830093384,
+      "learning_rate": 8.483535713877808e-06,
+      "loss": 0.1031,
+      "step": 4534
+    },
+    {
+      "epoch": 2.3750654793085384,
+      "loss_partial": 0.04443359375,
+      "step": 4534
+    },
+    {
+      "epoch": 2.3750654793085384,
+      "loss_partial": 0.0478515625,
+      "step": 4534
+    },
+    {
+      "epoch": 2.3755893137768465,
+      "grad_norm": 1.7862029075622559,
+      "learning_rate": 8.479017159583253e-06,
+      "loss": 0.0992,
+      "step": 4535
+    },
+    {
+      "epoch": 2.3755893137768465,
+      "loss_partial": 0.06787109375,
+      "step": 4535
+    },
+    {
+      "epoch": 2.3755893137768465,
+      "loss_partial": 0.040283203125,
+      "step": 4535
+    },
+    {
+      "epoch": 2.3761131482451545,
+      "grad_norm": 1.4391883611679077,
+      "learning_rate": 8.474498923164707e-06,
+      "loss": 0.1035,
+      "step": 4536
+    },
+    {
+      "epoch": 2.3761131482451545,
+      "loss_partial": 0.044921875,
+      "step": 4536
+    },
+    {
+      "epoch": 2.3761131482451545,
+      "loss_partial": 0.055419921875,
+      "step": 4536
+    },
+    {
+      "epoch": 2.3766369827134626,
+      "grad_norm": 1.3820863962173462,
+      "learning_rate": 8.46998100556645e-06,
+      "loss": 0.1009,
+      "step": 4537
+    },
+    {
+      "epoch": 2.3766369827134626,
+      "loss_partial": 0.045654296875,
+      "step": 4537
+    },
+    {
+      "epoch": 2.3766369827134626,
+      "loss_partial": 0.04931640625,
+      "step": 4537
+    },
+    {
+      "epoch": 2.3771608171817706,
+      "grad_norm": 1.8979923725128174,
+      "learning_rate": 8.465463407732705e-06,
+      "loss": 0.1024,
+      "step": 4538
+    },
+    {
+      "epoch": 2.3771608171817706,
+      "loss_partial": 0.043212890625,
+      "step": 4538
+    },
+    {
+      "epoch": 2.3771608171817706,
+      "loss_partial": 0.044189453125,
+      "step": 4538
+    },
+    {
+      "epoch": 2.3776846516500787,
+      "grad_norm": 1.6248363256454468,
+      "learning_rate": 8.460946130607617e-06,
+      "loss": 0.0985,
+      "step": 4539
+    },
+    {
+      "epoch": 2.3776846516500787,
+      "loss_partial": 0.044677734375,
+      "step": 4539
+    },
+    {
+      "epoch": 2.3776846516500787,
+      "loss_partial": 0.052490234375,
+      "step": 4539
+    },
+    {
+      "epoch": 2.3782084861183868,
+      "grad_norm": 2.2039520740509033,
+      "learning_rate": 8.45642917513527e-06,
+      "loss": 0.1011,
+      "step": 4540
+    },
+    {
+      "epoch": 2.3782084861183868,
+      "loss_partial": 0.048828125,
+      "step": 4540
+    },
+    {
+      "epoch": 2.3782084861183868,
+      "loss_partial": 0.0498046875,
+      "step": 4540
+    },
+    {
+      "epoch": 2.3787323205866944,
+      "grad_norm": 1.5382323265075684,
+      "learning_rate": 8.451912542259679e-06,
+      "loss": 0.1034,
+      "step": 4541
+    },
+    {
+      "epoch": 2.3787323205866944,
+      "loss_partial": 0.051025390625,
+      "step": 4541
+    },
+    {
+      "epoch": 2.3787323205866944,
+      "loss_partial": 0.060791015625,
+      "step": 4541
+    },
+    {
+      "epoch": 2.3792561550550024,
+      "grad_norm": 1.809592843055725,
+      "learning_rate": 8.447396232924796e-06,
+      "loss": 0.101,
+      "step": 4542
+    },
+    {
+      "epoch": 2.3792561550550024,
+      "loss_partial": 0.051025390625,
+      "step": 4542
+    },
+    {
+      "epoch": 2.3792561550550024,
+      "loss_partial": 0.050048828125,
+      "step": 4542
+    },
+    {
+      "epoch": 2.3797799895233105,
+      "grad_norm": 1.5060185194015503,
+      "learning_rate": 8.442880248074499e-06,
+      "loss": 0.1042,
+      "step": 4543
+    },
+    {
+      "epoch": 2.3797799895233105,
+      "loss_partial": 0.045166015625,
+      "step": 4543
+    },
+    {
+      "epoch": 2.3797799895233105,
+      "loss_partial": 0.047119140625,
+      "step": 4543
+    },
+    {
+      "epoch": 2.3803038239916186,
+      "grad_norm": 1.3639713525772095,
+      "learning_rate": 8.438364588652601e-06,
+      "loss": 0.1005,
+      "step": 4544
+    },
+    {
+      "epoch": 2.3803038239916186,
+      "loss_partial": 0.0546875,
+      "step": 4544
+    },
+    {
+      "epoch": 2.3803038239916186,
+      "loss_partial": 0.04833984375,
+      "step": 4544
+    },
+    {
+      "epoch": 2.3808276584599266,
+      "grad_norm": 1.7546781301498413,
+      "learning_rate": 8.433849255602847e-06,
+      "loss": 0.1015,
+      "step": 4545
+    },
+    {
+      "epoch": 2.3808276584599266,
+      "loss_partial": 0.051025390625,
+      "step": 4545
+    },
+    {
+      "epoch": 2.3808276584599266,
+      "loss_partial": 0.0458984375,
+      "step": 4545
+    },
+    {
+      "epoch": 2.3813514929282347,
+      "grad_norm": 1.532900094985962,
+      "learning_rate": 8.429334249868911e-06,
+      "loss": 0.1064,
+      "step": 4546
+    },
+    {
+      "epoch": 2.3813514929282347,
+      "loss_partial": 0.04736328125,
+      "step": 4546
+    },
+    {
+      "epoch": 2.3813514929282347,
+      "loss_partial": 0.051025390625,
+      "step": 4546
+    },
+    {
+      "epoch": 2.3818753273965427,
+      "grad_norm": 1.67107355594635,
+      "learning_rate": 8.424819572394406e-06,
+      "loss": 0.1043,
+      "step": 4547
+    },
+    {
+      "epoch": 2.3818753273965427,
+      "loss_partial": 0.05029296875,
+      "step": 4547
+    },
+    {
+      "epoch": 2.3818753273965427,
+      "loss_partial": 0.0439453125,
+      "step": 4547
+    },
+    {
+      "epoch": 2.382399161864851,
+      "grad_norm": 1.901638150215149,
+      "learning_rate": 8.420305224122871e-06,
+      "loss": 0.1022,
+      "step": 4548
+    },
+    {
+      "epoch": 2.382399161864851,
+      "loss_partial": 0.04443359375,
+      "step": 4548
+    },
+    {
+      "epoch": 2.382399161864851,
+      "loss_partial": 0.048583984375,
+      "step": 4548
+    },
+    {
+      "epoch": 2.382922996333159,
+      "grad_norm": 1.541877269744873,
+      "learning_rate": 8.415791205997778e-06,
+      "loss": 0.0995,
+      "step": 4549
+    },
+    {
+      "epoch": 2.382922996333159,
+      "loss_partial": 0.053955078125,
+      "step": 4549
+    },
+    {
+      "epoch": 2.382922996333159,
+      "loss_partial": 0.0458984375,
+      "step": 4549
+    },
+    {
+      "epoch": 2.383446830801467,
+      "grad_norm": 2.4670040607452393,
+      "learning_rate": 8.41127751896252e-06,
+      "loss": 0.101,
+      "step": 4550
+    },
+    {
+      "epoch": 2.383446830801467,
+      "loss_partial": 0.053466796875,
+      "step": 4550
+    },
+    {
+      "epoch": 2.383446830801467,
+      "loss_partial": 0.052978515625,
+      "step": 4550
+    },
+    {
+      "epoch": 2.383970665269775,
+      "grad_norm": 1.5291987657546997,
+      "learning_rate": 8.406764163960439e-06,
+      "loss": 0.1021,
+      "step": 4551
+    },
+    {
+      "epoch": 2.383970665269775,
+      "loss_partial": 0.052978515625,
+      "step": 4551
+    },
+    {
+      "epoch": 2.383970665269775,
+      "loss_partial": 0.06787109375,
+      "step": 4551
+    },
+    {
+      "epoch": 2.3844944997380826,
+      "grad_norm": 2.164578914642334,
+      "learning_rate": 8.402251141934797e-06,
+      "loss": 0.1054,
+      "step": 4552
+    },
+    {
+      "epoch": 2.3844944997380826,
+      "loss_partial": 0.045166015625,
+      "step": 4552
+    },
+    {
+      "epoch": 2.3844944997380826,
+      "loss_partial": 0.0517578125,
+      "step": 4552
+    },
+    {
+      "epoch": 2.3850183342063906,
+      "grad_norm": 1.2257922887802124,
+      "learning_rate": 8.397738453828786e-06,
+      "loss": 0.1039,
+      "step": 4553
+    },
+    {
+      "epoch": 2.3850183342063906,
+      "loss_partial": 0.0556640625,
+      "step": 4553
+    },
+    {
+      "epoch": 2.3850183342063906,
+      "loss_partial": 0.04443359375,
+      "step": 4553
+    },
+    {
+      "epoch": 2.3855421686746987,
+      "grad_norm": 1.5078755617141724,
+      "learning_rate": 8.393226100585527e-06,
+      "loss": 0.0965,
+      "step": 4554
+    },
+    {
+      "epoch": 2.3855421686746987,
+      "loss_partial": 0.054443359375,
+      "step": 4554
+    },
+    {
+      "epoch": 2.3855421686746987,
+      "loss_partial": 0.048583984375,
+      "step": 4554
+    },
+    {
+      "epoch": 2.3860660031430068,
+      "grad_norm": 1.3112468719482422,
+      "learning_rate": 8.38871408314808e-06,
+      "loss": 0.1015,
+      "step": 4555
+    },
+    {
+      "epoch": 2.3860660031430068,
+      "loss_partial": 0.06005859375,
+      "step": 4555
+    },
+    {
+      "epoch": 2.3860660031430068,
+      "loss_partial": 0.04296875,
+      "step": 4555
+    },
+    {
+      "epoch": 2.386589837611315,
+      "grad_norm": 1.997292399406433,
+      "learning_rate": 8.384202402459425e-06,
+      "loss": 0.1025,
+      "step": 4556
+    },
+    {
+      "epoch": 2.386589837611315,
+      "loss_partial": 0.0703125,
+      "step": 4556
+    },
+    {
+      "epoch": 2.386589837611315,
+      "loss_partial": 0.055908203125,
+      "step": 4556
+    },
+    {
+      "epoch": 2.387113672079623,
+      "grad_norm": 1.5822666883468628,
+      "learning_rate": 8.379691059462478e-06,
+      "loss": 0.1027,
+      "step": 4557
+    },
+    {
+      "epoch": 2.387113672079623,
+      "loss_partial": 0.048095703125,
+      "step": 4557
+    },
+    {
+      "epoch": 2.387113672079623,
+      "loss_partial": 0.060791015625,
+      "step": 4557
+    },
+    {
+      "epoch": 2.387637506547931,
+      "grad_norm": 1.5548020601272583,
+      "learning_rate": 8.375180055100074e-06,
+      "loss": 0.1027,
+      "step": 4558
+    },
+    {
+      "epoch": 2.387637506547931,
+      "loss_partial": 0.05517578125,
+      "step": 4558
+    },
+    {
+      "epoch": 2.387637506547931,
+      "loss_partial": 0.04931640625,
+      "step": 4558
+    },
+    {
+      "epoch": 2.388161341016239,
+      "grad_norm": 1.9508726596832275,
+      "learning_rate": 8.370669390314994e-06,
+      "loss": 0.1053,
+      "step": 4559
+    },
+    {
+      "epoch": 2.388161341016239,
+      "loss_partial": 0.045166015625,
+      "step": 4559
+    },
+    {
+      "epoch": 2.388161341016239,
+      "loss_partial": 0.056640625,
+      "step": 4559
+    },
+    {
+      "epoch": 2.388685175484547,
+      "grad_norm": 1.540851354598999,
+      "learning_rate": 8.366159066049938e-06,
+      "loss": 0.1049,
+      "step": 4560
+    },
+    {
+      "epoch": 2.388685175484547,
+      "loss_partial": 0.05810546875,
+      "step": 4560
+    },
+    {
+      "epoch": 2.388685175484547,
+      "loss_partial": 0.045654296875,
+      "step": 4560
+    },
+    {
+      "epoch": 2.3892090099528547,
+      "grad_norm": 1.5650265216827393,
+      "learning_rate": 8.361649083247533e-06,
+      "loss": 0.1047,
+      "step": 4561
+    },
+    {
+      "epoch": 2.3892090099528547,
+      "loss_partial": 0.0546875,
+      "step": 4561
+    },
+    {
+      "epoch": 2.3892090099528547,
+      "loss_partial": 0.049560546875,
+      "step": 4561
+    },
+    {
+      "epoch": 2.3897328444211627,
+      "grad_norm": 1.6228653192520142,
+      "learning_rate": 8.357139442850336e-06,
+      "loss": 0.101,
+      "step": 4562
+    },
+    {
+      "epoch": 2.3897328444211627,
+      "loss_partial": 0.047607421875,
+      "step": 4562
+    },
+    {
+      "epoch": 2.3897328444211627,
+      "loss_partial": 0.050537109375,
+      "step": 4562
+    },
+    {
+      "epoch": 2.390256678889471,
+      "grad_norm": 1.5274356603622437,
+      "learning_rate": 8.352630145800839e-06,
+      "loss": 0.1023,
+      "step": 4563
+    },
+    {
+      "epoch": 2.390256678889471,
+      "loss_partial": 0.0595703125,
+      "step": 4563
+    },
+    {
+      "epoch": 2.390256678889471,
+      "loss_partial": 0.050048828125,
+      "step": 4563
+    },
+    {
+      "epoch": 2.390780513357779,
+      "grad_norm": 1.2581512928009033,
+      "learning_rate": 8.348121193041454e-06,
+      "loss": 0.1056,
+      "step": 4564
+    },
+    {
+      "epoch": 2.390780513357779,
+      "loss_partial": 0.04931640625,
+      "step": 4564
+    },
+    {
+      "epoch": 2.390780513357779,
+      "loss_partial": 0.05322265625,
+      "step": 4564
+    },
+    {
+      "epoch": 2.391304347826087,
+      "grad_norm": 1.1760646104812622,
+      "learning_rate": 8.343612585514525e-06,
+      "loss": 0.1027,
+      "step": 4565
+    },
+    {
+      "epoch": 2.391304347826087,
+      "loss_partial": 0.05126953125,
+      "step": 4565
+    },
+    {
+      "epoch": 2.391304347826087,
+      "loss_partial": 0.05126953125,
+      "step": 4565
+    },
+    {
+      "epoch": 2.391828182294395,
+      "grad_norm": 1.036668300628662,
+      "learning_rate": 8.339104324162322e-06,
+      "loss": 0.0982,
+      "step": 4566
+    },
+    {
+      "epoch": 2.391828182294395,
+      "loss_partial": 0.049072265625,
+      "step": 4566
+    },
+    {
+      "epoch": 2.391828182294395,
+      "loss_partial": 0.06298828125,
+      "step": 4566
+    },
+    {
+      "epoch": 2.392352016762703,
+      "grad_norm": 1.9395500421524048,
+      "learning_rate": 8.334596409927045e-06,
+      "loss": 0.1055,
+      "step": 4567
+    },
+    {
+      "epoch": 2.392352016762703,
+      "loss_partial": 0.048828125,
+      "step": 4567
+    },
+    {
+      "epoch": 2.392352016762703,
+      "loss_partial": 0.045654296875,
+      "step": 4567
+    },
+    {
+      "epoch": 2.392875851231011,
+      "grad_norm": 1.4850499629974365,
+      "learning_rate": 8.33008884375082e-06,
+      "loss": 0.1027,
+      "step": 4568
+    },
+    {
+      "epoch": 2.392875851231011,
+      "loss_partial": 0.05419921875,
+      "step": 4568
+    },
+    {
+      "epoch": 2.392875851231011,
+      "loss_partial": 0.056884765625,
+      "step": 4568
+    },
+    {
+      "epoch": 2.393399685699319,
+      "grad_norm": 1.6983470916748047,
+      "learning_rate": 8.3255816265757e-06,
+      "loss": 0.1017,
+      "step": 4569
+    },
+    {
+      "epoch": 2.393399685699319,
+      "loss_partial": 0.046142578125,
+      "step": 4569
+    },
+    {
+      "epoch": 2.393399685699319,
+      "loss_partial": 0.0576171875,
+      "step": 4569
+    },
+    {
+      "epoch": 2.393923520167627,
+      "grad_norm": 2.116151809692383,
+      "learning_rate": 8.321074759343663e-06,
+      "loss": 0.101,
+      "step": 4570
+    },
+    {
+      "epoch": 2.393923520167627,
+      "loss_partial": 0.042236328125,
+      "step": 4570
+    },
+    {
+      "epoch": 2.393923520167627,
+      "loss_partial": 0.0439453125,
+      "step": 4570
+    },
+    {
+      "epoch": 2.3944473546359353,
+      "grad_norm": 1.12233304977417,
+      "learning_rate": 8.31656824299662e-06,
+      "loss": 0.1042,
+      "step": 4571
+    },
+    {
+      "epoch": 2.3944473546359353,
+      "loss_partial": 0.052734375,
+      "step": 4571
+    },
+    {
+      "epoch": 2.3944473546359353,
+      "loss_partial": 0.04541015625,
+      "step": 4571
+    },
+    {
+      "epoch": 2.394971189104243,
+      "grad_norm": 1.5937165021896362,
+      "learning_rate": 8.312062078476409e-06,
+      "loss": 0.1031,
+      "step": 4572
+    },
+    {
+      "epoch": 2.394971189104243,
+      "loss_partial": 0.0546875,
+      "step": 4572
+    },
+    {
+      "epoch": 2.394971189104243,
+      "loss_partial": 0.0546875,
+      "step": 4572
+    },
+    {
+      "epoch": 2.395495023572551,
+      "grad_norm": 1.2550387382507324,
+      "learning_rate": 8.307556266724783e-06,
+      "loss": 0.1045,
+      "step": 4573
+    },
+    {
+      "epoch": 2.395495023572551,
+      "loss_partial": 0.06591796875,
+      "step": 4573
+    },
+    {
+      "epoch": 2.395495023572551,
+      "loss_partial": 0.045166015625,
+      "step": 4573
+    },
+    {
+      "epoch": 2.396018858040859,
+      "grad_norm": 1.427921175956726,
+      "learning_rate": 8.303050808683429e-06,
+      "loss": 0.103,
+      "step": 4574
+    },
+    {
+      "epoch": 2.396018858040859,
+      "loss_partial": 0.049072265625,
+      "step": 4574
+    },
+    {
+      "epoch": 2.396018858040859,
+      "loss_partial": 0.048095703125,
+      "step": 4574
+    },
+    {
+      "epoch": 2.396542692509167,
+      "grad_norm": 1.7291399240493774,
+      "learning_rate": 8.298545705293966e-06,
+      "loss": 0.1026,
+      "step": 4575
+    },
+    {
+      "epoch": 2.396542692509167,
+      "loss_partial": 0.045654296875,
+      "step": 4575
+    },
+    {
+      "epoch": 2.396542692509167,
+      "loss_partial": 0.04248046875,
+      "step": 4575
+    },
+    {
+      "epoch": 2.397066526977475,
+      "grad_norm": 1.5479285717010498,
+      "learning_rate": 8.294040957497927e-06,
+      "loss": 0.1037,
+      "step": 4576
+    },
+    {
+      "epoch": 2.397066526977475,
+      "loss_partial": 0.0458984375,
+      "step": 4576
+    },
+    {
+      "epoch": 2.397066526977475,
+      "loss_partial": 0.041748046875,
+      "step": 4576
+    },
+    {
+      "epoch": 2.397590361445783,
+      "grad_norm": 1.6285167932510376,
+      "learning_rate": 8.28953656623678e-06,
+      "loss": 0.1033,
+      "step": 4577
+    },
+    {
+      "epoch": 2.397590361445783,
+      "loss_partial": 0.047607421875,
+      "step": 4577
+    },
+    {
+      "epoch": 2.397590361445783,
+      "loss_partial": 0.052978515625,
+      "step": 4577
+    },
+    {
+      "epoch": 2.3981141959140913,
+      "grad_norm": 1.9155828952789307,
+      "learning_rate": 8.285032532451911e-06,
+      "loss": 0.1029,
+      "step": 4578
+    },
+    {
+      "epoch": 2.3981141959140913,
+      "loss_partial": 0.050048828125,
+      "step": 4578
+    },
+    {
+      "epoch": 2.3981141959140913,
+      "loss_partial": 0.04052734375,
+      "step": 4578
+    },
+    {
+      "epoch": 2.3986380303823993,
+      "grad_norm": 1.7953470945358276,
+      "learning_rate": 8.28052885708464e-06,
+      "loss": 0.1043,
+      "step": 4579
+    },
+    {
+      "epoch": 2.3986380303823993,
+      "loss_partial": 0.048583984375,
+      "step": 4579
+    },
+    {
+      "epoch": 2.3986380303823993,
+      "loss_partial": 0.054443359375,
+      "step": 4579
+    },
+    {
+      "epoch": 2.3991618648507074,
+      "grad_norm": 1.7209317684173584,
+      "learning_rate": 8.276025541076205e-06,
+      "loss": 0.1023,
+      "step": 4580
+    },
+    {
+      "epoch": 2.3991618648507074,
+      "loss_partial": 0.04736328125,
+      "step": 4580
+    },
+    {
+      "epoch": 2.3991618648507074,
+      "loss_partial": 0.0400390625,
+      "step": 4580
+    },
+    {
+      "epoch": 2.399685699319015,
+      "grad_norm": 1.7985467910766602,
+      "learning_rate": 8.271522585367772e-06,
+      "loss": 0.0992,
+      "step": 4581
+    },
+    {
+      "epoch": 2.399685699319015,
+      "loss_partial": 0.047607421875,
+      "step": 4581
+    },
+    {
+      "epoch": 2.399685699319015,
+      "loss_partial": 0.058349609375,
+      "step": 4581
+    },
+    {
+      "epoch": 2.400209533787323,
+      "grad_norm": 1.9204000234603882,
+      "learning_rate": 8.267019990900425e-06,
+      "loss": 0.0989,
+      "step": 4582
+    },
+    {
+      "epoch": 2.400209533787323,
+      "loss_partial": 0.043701171875,
+      "step": 4582
+    },
+    {
+      "epoch": 2.400209533787323,
+      "loss_partial": 0.05712890625,
+      "step": 4582
+    },
+    {
+      "epoch": 2.400733368255631,
+      "grad_norm": 1.5268243551254272,
+      "learning_rate": 8.262517758615187e-06,
+      "loss": 0.1065,
+      "step": 4583
+    },
+    {
+      "epoch": 2.400733368255631,
+      "loss_partial": 0.054931640625,
+      "step": 4583
+    },
+    {
+      "epoch": 2.400733368255631,
+      "loss_partial": 0.04443359375,
+      "step": 4583
+    },
+    {
+      "epoch": 2.401257202723939,
+      "grad_norm": 1.6169960498809814,
+      "learning_rate": 8.258015889452993e-06,
+      "loss": 0.1026,
+      "step": 4584
+    },
+    {
+      "epoch": 2.401257202723939,
+      "loss_partial": 0.05419921875,
+      "step": 4584
+    },
+    {
+      "epoch": 2.401257202723939,
+      "loss_partial": 0.05224609375,
+      "step": 4584
+    },
+    {
+      "epoch": 2.4017810371922472,
+      "grad_norm": 1.7670745849609375,
+      "learning_rate": 8.253514384354706e-06,
+      "loss": 0.1025,
+      "step": 4585
+    },
+    {
+      "epoch": 2.4017810371922472,
+      "loss_partial": 0.048828125,
+      "step": 4585
+    },
+    {
+      "epoch": 2.4017810371922472,
+      "loss_partial": 0.0517578125,
+      "step": 4585
+    },
+    {
+      "epoch": 2.4023048716605553,
+      "grad_norm": 1.5542443990707397,
+      "learning_rate": 8.249013244261113e-06,
+      "loss": 0.0984,
+      "step": 4586
+    },
+    {
+      "epoch": 2.4023048716605553,
+      "loss_partial": 0.05908203125,
+      "step": 4586
+    },
+    {
+      "epoch": 2.4023048716605553,
+      "loss_partial": 0.0556640625,
+      "step": 4586
+    },
+    {
+      "epoch": 2.4028287061288633,
+      "grad_norm": 1.6890146732330322,
+      "learning_rate": 8.244512470112925e-06,
+      "loss": 0.1045,
+      "step": 4587
+    },
+    {
+      "epoch": 2.4028287061288633,
+      "loss_partial": 0.054931640625,
+      "step": 4587
+    },
+    {
+      "epoch": 2.4028287061288633,
+      "loss_partial": 0.049560546875,
+      "step": 4587
+    },
+    {
+      "epoch": 2.4033525405971714,
+      "grad_norm": 1.5427454710006714,
+      "learning_rate": 8.240012062850773e-06,
+      "loss": 0.1032,
+      "step": 4588
+    },
+    {
+      "epoch": 2.4033525405971714,
+      "loss_partial": 0.041748046875,
+      "step": 4588
+    },
+    {
+      "epoch": 2.4033525405971714,
+      "loss_partial": 0.048828125,
+      "step": 4588
+    },
+    {
+      "epoch": 2.4038763750654795,
+      "grad_norm": 1.57435142993927,
+      "learning_rate": 8.23551202341522e-06,
+      "loss": 0.1023,
+      "step": 4589
+    },
+    {
+      "epoch": 2.4038763750654795,
+      "loss_partial": 0.051025390625,
+      "step": 4589
+    },
+    {
+      "epoch": 2.4038763750654795,
+      "loss_partial": 0.052734375,
+      "step": 4589
+    },
+    {
+      "epoch": 2.4044002095337875,
+      "grad_norm": 1.6965874433517456,
+      "learning_rate": 8.23101235274674e-06,
+      "loss": 0.1058,
+      "step": 4590
+    },
+    {
+      "epoch": 2.4044002095337875,
+      "loss_partial": 0.056640625,
+      "step": 4590
+    },
+    {
+      "epoch": 2.4044002095337875,
+      "loss_partial": 0.0498046875,
+      "step": 4590
+    },
+    {
+      "epoch": 2.4049240440020956,
+      "grad_norm": 1.6706842184066772,
+      "learning_rate": 8.226513051785738e-06,
+      "loss": 0.1017,
+      "step": 4591
+    },
+    {
+      "epoch": 2.4049240440020956,
+      "loss_partial": 0.044189453125,
+      "step": 4591
+    },
+    {
+      "epoch": 2.4049240440020956,
+      "loss_partial": 0.056396484375,
+      "step": 4591
+    },
+    {
+      "epoch": 2.405447878470403,
+      "grad_norm": 1.626432180404663,
+      "learning_rate": 8.222014121472545e-06,
+      "loss": 0.1049,
+      "step": 4592
+    },
+    {
+      "epoch": 2.405447878470403,
+      "loss_partial": 0.052001953125,
+      "step": 4592
+    },
+    {
+      "epoch": 2.405447878470403,
+      "loss_partial": 0.052001953125,
+      "step": 4592
+    },
+    {
+      "epoch": 2.4059717129387113,
+      "grad_norm": 1.3622524738311768,
+      "learning_rate": 8.217515562747404e-06,
+      "loss": 0.0959,
+      "step": 4593
+    },
+    {
+      "epoch": 2.4059717129387113,
+      "loss_partial": 0.048095703125,
+      "step": 4593
+    },
+    {
+      "epoch": 2.4059717129387113,
+      "loss_partial": 0.04443359375,
+      "step": 4593
+    },
+    {
+      "epoch": 2.4064955474070193,
+      "grad_norm": 1.5072429180145264,
+      "learning_rate": 8.213017376550484e-06,
+      "loss": 0.1022,
+      "step": 4594
+    },
+    {
+      "epoch": 2.4064955474070193,
+      "loss_partial": 0.059814453125,
+      "step": 4594
+    },
+    {
+      "epoch": 2.4064955474070193,
+      "loss_partial": 0.05517578125,
+      "step": 4594
+    },
+    {
+      "epoch": 2.4070193818753274,
+      "grad_norm": 1.5903606414794922,
+      "learning_rate": 8.208519563821886e-06,
+      "loss": 0.1056,
+      "step": 4595
+    },
+    {
+      "epoch": 2.4070193818753274,
+      "loss_partial": 0.05322265625,
+      "step": 4595
+    },
+    {
+      "epoch": 2.4070193818753274,
+      "loss_partial": 0.06396484375,
+      "step": 4595
+    },
+    {
+      "epoch": 2.4075432163436354,
+      "grad_norm": 1.6229327917099,
+      "learning_rate": 8.204022125501622e-06,
+      "loss": 0.1048,
+      "step": 4596
+    },
+    {
+      "epoch": 2.4075432163436354,
+      "loss_partial": 0.051513671875,
+      "step": 4596
+    },
+    {
+      "epoch": 2.4075432163436354,
+      "loss_partial": 0.056884765625,
+      "step": 4596
+    },
+    {
+      "epoch": 2.4080670508119435,
+      "grad_norm": 1.4374393224716187,
+      "learning_rate": 8.199525062529626e-06,
+      "loss": 0.0995,
+      "step": 4597
+    },
+    {
+      "epoch": 2.4080670508119435,
+      "loss_partial": 0.049560546875,
+      "step": 4597
+    },
+    {
+      "epoch": 2.4080670508119435,
+      "loss_partial": 0.056640625,
+      "step": 4597
+    },
+    {
+      "epoch": 2.4085908852802516,
+      "grad_norm": 2.0431392192840576,
+      "learning_rate": 8.195028375845758e-06,
+      "loss": 0.1011,
+      "step": 4598
+    },
+    {
+      "epoch": 2.4085908852802516,
+      "loss_partial": 0.045166015625,
+      "step": 4598
+    },
+    {
+      "epoch": 2.4085908852802516,
+      "loss_partial": 0.05126953125,
+      "step": 4598
+    },
+    {
+      "epoch": 2.4091147197485596,
+      "grad_norm": 1.4535167217254639,
+      "learning_rate": 8.190532066389798e-06,
+      "loss": 0.0999,
+      "step": 4599
+    },
+    {
+      "epoch": 2.4091147197485596,
+      "loss_partial": 0.053466796875,
+      "step": 4599
+    },
+    {
+      "epoch": 2.4091147197485596,
+      "loss_partial": 0.047607421875,
+      "step": 4599
+    },
+    {
+      "epoch": 2.4096385542168672,
+      "grad_norm": 1.52773118019104,
+      "learning_rate": 8.186036135101448e-06,
+      "loss": 0.1053,
+      "step": 4600
+    },
+    {
+      "epoch": 2.4096385542168672,
+      "loss_partial": 0.048095703125,
+      "step": 4600
+    },
+    {
+      "epoch": 2.4096385542168672,
+      "loss_partial": 0.048095703125,
+      "step": 4600
+    },
+    {
+      "epoch": 2.4101623886851753,
+      "grad_norm": 1.4589564800262451,
+      "learning_rate": 8.181540582920328e-06,
+      "loss": 0.102,
+      "step": 4601
+    },
+    {
+      "epoch": 2.4101623886851753,
+      "loss_partial": 0.052978515625,
+      "step": 4601
+    },
+    {
+      "epoch": 2.4101623886851753,
+      "loss_partial": 0.047607421875,
+      "step": 4601
+    },
+    {
+      "epoch": 2.4106862231534834,
+      "grad_norm": 1.8838711977005005,
+      "learning_rate": 8.177045410785981e-06,
+      "loss": 0.1017,
+      "step": 4602
+    },
+    {
+      "epoch": 2.4106862231534834,
+      "loss_partial": 0.060791015625,
+      "step": 4602
+    },
+    {
+      "epoch": 2.4106862231534834,
+      "loss_partial": 0.04541015625,
+      "step": 4602
+    },
+    {
+      "epoch": 2.4112100576217914,
+      "grad_norm": 1.3490629196166992,
+      "learning_rate": 8.17255061963787e-06,
+      "loss": 0.1008,
+      "step": 4603
+    },
+    {
+      "epoch": 2.4112100576217914,
+      "loss_partial": 0.06298828125,
+      "step": 4603
+    },
+    {
+      "epoch": 2.4112100576217914,
+      "loss_partial": 0.046875,
+      "step": 4603
+    },
+    {
+      "epoch": 2.4117338920900995,
+      "grad_norm": 1.4034428596496582,
+      "learning_rate": 8.168056210415378e-06,
+      "loss": 0.1007,
+      "step": 4604
+    },
+    {
+      "epoch": 2.4117338920900995,
+      "loss_partial": 0.0576171875,
+      "step": 4604
+    },
+    {
+      "epoch": 2.4117338920900995,
+      "loss_partial": 0.06396484375,
+      "step": 4604
+    },
+    {
+      "epoch": 2.4122577265584075,
+      "grad_norm": 1.7453097105026245,
+      "learning_rate": 8.163562184057812e-06,
+      "loss": 0.1046,
+      "step": 4605
+    },
+    {
+      "epoch": 2.4122577265584075,
+      "loss_partial": 0.04736328125,
+      "step": 4605
+    },
+    {
+      "epoch": 2.4122577265584075,
+      "loss_partial": 0.05224609375,
+      "step": 4605
+    },
+    {
+      "epoch": 2.4127815610267156,
+      "grad_norm": 1.7129429578781128,
+      "learning_rate": 8.159068541504393e-06,
+      "loss": 0.1012,
+      "step": 4606
+    },
+    {
+      "epoch": 2.4127815610267156,
+      "loss_partial": 0.0439453125,
+      "step": 4606
+    },
+    {
+      "epoch": 2.4127815610267156,
+      "loss_partial": 0.053955078125,
+      "step": 4606
+    },
+    {
+      "epoch": 2.4133053954950237,
+      "grad_norm": 1.3808057308197021,
+      "learning_rate": 8.15457528369426e-06,
+      "loss": 0.0993,
+      "step": 4607
+    },
+    {
+      "epoch": 2.4133053954950237,
+      "loss_partial": 0.05078125,
+      "step": 4607
+    },
+    {
+      "epoch": 2.4133053954950237,
+      "loss_partial": 0.049560546875,
+      "step": 4607
+    },
+    {
+      "epoch": 2.4138292299633317,
+      "grad_norm": 1.4450087547302246,
+      "learning_rate": 8.150082411566485e-06,
+      "loss": 0.1021,
+      "step": 4608
+    },
+    {
+      "epoch": 2.4138292299633317,
+      "loss_partial": 0.053955078125,
+      "step": 4608
+    },
+    {
+      "epoch": 2.4138292299633317,
+      "loss_partial": 0.0498046875,
+      "step": 4608
+    },
+    {
+      "epoch": 2.4143530644316398,
+      "grad_norm": 1.283553957939148,
+      "learning_rate": 8.145589926060048e-06,
+      "loss": 0.1025,
+      "step": 4609
+    },
+    {
+      "epoch": 2.4143530644316398,
+      "loss_partial": 0.042236328125,
+      "step": 4609
+    },
+    {
+      "epoch": 2.4143530644316398,
+      "loss_partial": 0.062255859375,
+      "step": 4609
+    },
+    {
+      "epoch": 2.414876898899948,
+      "grad_norm": 1.4798803329467773,
+      "learning_rate": 8.141097828113847e-06,
+      "loss": 0.1024,
+      "step": 4610
+    },
+    {
+      "epoch": 2.414876898899948,
+      "loss_partial": 0.051025390625,
+      "step": 4610
+    },
+    {
+      "epoch": 2.414876898899948,
+      "loss_partial": 0.05712890625,
+      "step": 4610
+    },
+    {
+      "epoch": 2.4154007333682554,
+      "grad_norm": 1.35856032371521,
+      "learning_rate": 8.136606118666703e-06,
+      "loss": 0.1003,
+      "step": 4611
+    },
+    {
+      "epoch": 2.4154007333682554,
+      "loss_partial": 0.046630859375,
+      "step": 4611
+    },
+    {
+      "epoch": 2.4154007333682554,
+      "loss_partial": 0.04736328125,
+      "step": 4611
+    },
+    {
+      "epoch": 2.4159245678365635,
+      "grad_norm": 1.9903404712677002,
+      "learning_rate": 8.13211479865736e-06,
+      "loss": 0.1026,
+      "step": 4612
+    },
+    {
+      "epoch": 2.4159245678365635,
+      "loss_partial": 0.052001953125,
+      "step": 4612
+    },
+    {
+      "epoch": 2.4159245678365635,
+      "loss_partial": 0.0537109375,
+      "step": 4612
+    },
+    {
+      "epoch": 2.4164484023048716,
+      "grad_norm": 1.628983736038208,
+      "learning_rate": 8.127623869024473e-06,
+      "loss": 0.1053,
+      "step": 4613
+    },
+    {
+      "epoch": 2.4164484023048716,
+      "loss_partial": 0.046875,
+      "step": 4613
+    },
+    {
+      "epoch": 2.4164484023048716,
+      "loss_partial": 0.05078125,
+      "step": 4613
+    },
+    {
+      "epoch": 2.4169722367731796,
+      "grad_norm": 1.4322704076766968,
+      "learning_rate": 8.12313333070662e-06,
+      "loss": 0.1016,
+      "step": 4614
+    },
+    {
+      "epoch": 2.4169722367731796,
+      "loss_partial": 0.054931640625,
+      "step": 4614
+    },
+    {
+      "epoch": 2.4169722367731796,
+      "loss_partial": 0.048095703125,
+      "step": 4614
+    },
+    {
+      "epoch": 2.4174960712414877,
+      "grad_norm": 1.399267315864563,
+      "learning_rate": 8.118643184642291e-06,
+      "loss": 0.1028,
+      "step": 4615
+    },
+    {
+      "epoch": 2.4174960712414877,
+      "loss_partial": 0.053466796875,
+      "step": 4615
+    },
+    {
+      "epoch": 2.4174960712414877,
+      "loss_partial": 0.042236328125,
+      "step": 4615
+    },
+    {
+      "epoch": 2.4180199057097957,
+      "grad_norm": 1.45967435836792,
+      "learning_rate": 8.114153431769906e-06,
+      "loss": 0.0992,
+      "step": 4616
+    },
+    {
+      "epoch": 2.4180199057097957,
+      "loss_partial": 0.04443359375,
+      "step": 4616
+    },
+    {
+      "epoch": 2.4180199057097957,
+      "loss_partial": 0.047607421875,
+      "step": 4616
+    },
+    {
+      "epoch": 2.418543740178104,
+      "grad_norm": 1.524221658706665,
+      "learning_rate": 8.109664073027792e-06,
+      "loss": 0.0984,
+      "step": 4617
+    },
+    {
+      "epoch": 2.418543740178104,
+      "loss_partial": 0.051025390625,
+      "step": 4617
+    },
+    {
+      "epoch": 2.418543740178104,
+      "loss_partial": 0.045166015625,
+      "step": 4617
+    },
+    {
+      "epoch": 2.419067574646412,
+      "grad_norm": 1.7816376686096191,
+      "learning_rate": 8.105175109354198e-06,
+      "loss": 0.0992,
+      "step": 4618
+    },
+    {
+      "epoch": 2.419067574646412,
+      "loss_partial": 0.045654296875,
+      "step": 4618
+    },
+    {
+      "epoch": 2.419067574646412,
+      "loss_partial": 0.0478515625,
+      "step": 4618
+    },
+    {
+      "epoch": 2.41959140911472,
+      "grad_norm": 1.351569652557373,
+      "learning_rate": 8.100686541687282e-06,
+      "loss": 0.0982,
+      "step": 4619
+    },
+    {
+      "epoch": 2.41959140911472,
+      "loss_partial": 0.058349609375,
+      "step": 4619
+    },
+    {
+      "epoch": 2.41959140911472,
+      "loss_partial": 0.045166015625,
+      "step": 4619
+    },
+    {
+      "epoch": 2.4201152435830275,
+      "grad_norm": 1.879224419593811,
+      "learning_rate": 8.09619837096514e-06,
+      "loss": 0.1037,
+      "step": 4620
+    },
+    {
+      "epoch": 2.4201152435830275,
+      "loss_partial": 0.047119140625,
+      "step": 4620
+    },
+    {
+      "epoch": 2.4201152435830275,
+      "loss_partial": 0.04248046875,
+      "step": 4620
+    },
+    {
+      "epoch": 2.4206390780513356,
+      "grad_norm": 1.4045010805130005,
+      "learning_rate": 8.091710598125766e-06,
+      "loss": 0.1005,
+      "step": 4621
+    },
+    {
+      "epoch": 2.4206390780513356,
+      "loss_partial": 0.050048828125,
+      "step": 4621
+    },
+    {
+      "epoch": 2.4206390780513356,
+      "loss_partial": 0.05322265625,
+      "step": 4621
+    },
+    {
+      "epoch": 2.4211629125196437,
+      "grad_norm": 2.003256320953369,
+      "learning_rate": 8.087223224107075e-06,
+      "loss": 0.1029,
+      "step": 4622
+    },
+    {
+      "epoch": 2.4211629125196437,
+      "loss_partial": 0.051513671875,
+      "step": 4622
+    },
+    {
+      "epoch": 2.4211629125196437,
+      "loss_partial": 0.06494140625,
+      "step": 4622
+    },
+    {
+      "epoch": 2.4216867469879517,
+      "grad_norm": 1.682249903678894,
+      "learning_rate": 8.082736249846901e-06,
+      "loss": 0.1027,
+      "step": 4623
+    },
+    {
+      "epoch": 2.4216867469879517,
+      "loss_partial": 0.03955078125,
+      "step": 4623
+    },
+    {
+      "epoch": 2.4216867469879517,
+      "loss_partial": 0.05224609375,
+      "step": 4623
+    },
+    {
+      "epoch": 2.42221058145626,
+      "grad_norm": 2.41636323928833,
+      "learning_rate": 8.078249676282996e-06,
+      "loss": 0.1018,
+      "step": 4624
+    },
+    {
+      "epoch": 2.42221058145626,
+      "loss_partial": 0.053955078125,
+      "step": 4624
+    },
+    {
+      "epoch": 2.42221058145626,
+      "loss_partial": 0.053466796875,
+      "step": 4624
+    },
+    {
+      "epoch": 2.422734415924568,
+      "grad_norm": 1.4909499883651733,
+      "learning_rate": 8.073763504353025e-06,
+      "loss": 0.0984,
+      "step": 4625
+    },
+    {
+      "epoch": 2.422734415924568,
+      "loss_partial": 0.047119140625,
+      "step": 4625
+    },
+    {
+      "epoch": 2.422734415924568,
+      "loss_partial": 0.04931640625,
+      "step": 4625
+    },
+    {
+      "epoch": 2.423258250392876,
+      "grad_norm": 1.5788904428482056,
+      "learning_rate": 8.06927773499457e-06,
+      "loss": 0.1038,
+      "step": 4626
+    },
+    {
+      "epoch": 2.423258250392876,
+      "loss_partial": 0.048828125,
+      "step": 4626
+    },
+    {
+      "epoch": 2.423258250392876,
+      "loss_partial": 0.050048828125,
+      "step": 4626
+    },
+    {
+      "epoch": 2.423782084861184,
+      "grad_norm": 1.6734942197799683,
+      "learning_rate": 8.064792369145128e-06,
+      "loss": 0.1019,
+      "step": 4627
+    },
+    {
+      "epoch": 2.423782084861184,
+      "loss_partial": 0.0517578125,
+      "step": 4627
+    },
+    {
+      "epoch": 2.423782084861184,
+      "loss_partial": 0.046142578125,
+      "step": 4627
+    },
+    {
+      "epoch": 2.424305919329492,
+      "grad_norm": 1.6173300743103027,
+      "learning_rate": 8.060307407742117e-06,
+      "loss": 0.1045,
+      "step": 4628
+    },
+    {
+      "epoch": 2.424305919329492,
+      "loss_partial": 0.046142578125,
+      "step": 4628
+    },
+    {
+      "epoch": 2.424305919329492,
+      "loss_partial": 0.056884765625,
+      "step": 4628
+    },
+    {
+      "epoch": 2.4248297537978,
+      "grad_norm": 1.6856797933578491,
+      "learning_rate": 8.05582285172286e-06,
+      "loss": 0.103,
+      "step": 4629
+    },
+    {
+      "epoch": 2.4248297537978,
+      "loss_partial": 0.044677734375,
+      "step": 4629
+    },
+    {
+      "epoch": 2.4248297537978,
+      "loss_partial": 0.058837890625,
+      "step": 4629
+    },
+    {
+      "epoch": 2.425353588266108,
+      "grad_norm": 1.4851436614990234,
+      "learning_rate": 8.051338702024607e-06,
+      "loss": 0.1029,
+      "step": 4630
+    },
+    {
+      "epoch": 2.425353588266108,
+      "loss_partial": 0.0498046875,
+      "step": 4630
+    },
+    {
+      "epoch": 2.425353588266108,
+      "loss_partial": 0.060302734375,
+      "step": 4630
+    },
+    {
+      "epoch": 2.4258774227344158,
+      "grad_norm": 1.9378924369812012,
+      "learning_rate": 8.046854959584508e-06,
+      "loss": 0.1024,
+      "step": 4631
+    },
+    {
+      "epoch": 2.4258774227344158,
+      "loss_partial": 0.05615234375,
+      "step": 4631
+    },
+    {
+      "epoch": 2.4258774227344158,
+      "loss_partial": 0.047119140625,
+      "step": 4631
+    },
+    {
+      "epoch": 2.426401257202724,
+      "grad_norm": 1.386507511138916,
+      "learning_rate": 8.042371625339653e-06,
+      "loss": 0.1043,
+      "step": 4632
+    },
+    {
+      "epoch": 2.426401257202724,
+      "loss_partial": 0.051513671875,
+      "step": 4632
+    },
+    {
+      "epoch": 2.426401257202724,
+      "loss_partial": 0.058837890625,
+      "step": 4632
+    },
+    {
+      "epoch": 2.426925091671032,
+      "grad_norm": 1.1942282915115356,
+      "learning_rate": 8.03788870022702e-06,
+      "loss": 0.0999,
+      "step": 4633
+    },
+    {
+      "epoch": 2.426925091671032,
+      "loss_partial": 0.050048828125,
+      "step": 4633
+    },
+    {
+      "epoch": 2.426925091671032,
+      "loss_partial": 0.053466796875,
+      "step": 4633
+    },
+    {
+      "epoch": 2.42744892613934,
+      "grad_norm": 1.7009904384613037,
+      "learning_rate": 8.033406185183515e-06,
+      "loss": 0.101,
+      "step": 4634
+    },
+    {
+      "epoch": 2.42744892613934,
+      "loss_partial": 0.0537109375,
+      "step": 4634
+    },
+    {
+      "epoch": 2.42744892613934,
+      "loss_partial": 0.045654296875,
+      "step": 4634
+    },
+    {
+      "epoch": 2.427972760607648,
+      "grad_norm": 1.667488932609558,
+      "learning_rate": 8.028924081145956e-06,
+      "loss": 0.1019,
+      "step": 4635
+    },
+    {
+      "epoch": 2.427972760607648,
+      "loss_partial": 0.0517578125,
+      "step": 4635
+    },
+    {
+      "epoch": 2.427972760607648,
+      "loss_partial": 0.0439453125,
+      "step": 4635
+    },
+    {
+      "epoch": 2.428496595075956,
+      "grad_norm": 1.5001356601715088,
+      "learning_rate": 8.024442389051076e-06,
+      "loss": 0.1049,
+      "step": 4636
+    },
+    {
+      "epoch": 2.428496595075956,
+      "loss_partial": 0.0458984375,
+      "step": 4636
+    },
+    {
+      "epoch": 2.428496595075956,
+      "loss_partial": 0.04833984375,
+      "step": 4636
+    },
+    {
+      "epoch": 2.429020429544264,
+      "grad_norm": 2.0552000999450684,
+      "learning_rate": 8.01996110983552e-06,
+      "loss": 0.1026,
+      "step": 4637
+    },
+    {
+      "epoch": 2.429020429544264,
+      "loss_partial": 0.045166015625,
+      "step": 4637
+    },
+    {
+      "epoch": 2.429020429544264,
+      "loss_partial": 0.051025390625,
+      "step": 4637
+    },
+    {
+      "epoch": 2.429544264012572,
+      "grad_norm": 1.5544261932373047,
+      "learning_rate": 8.01548024443585e-06,
+      "loss": 0.1019,
+      "step": 4638
+    },
+    {
+      "epoch": 2.429544264012572,
+      "loss_partial": 0.04638671875,
+      "step": 4638
+    },
+    {
+      "epoch": 2.429544264012572,
+      "loss_partial": 0.05419921875,
+      "step": 4638
+    },
+    {
+      "epoch": 2.4300680984808802,
+      "grad_norm": 2.1263229846954346,
+      "learning_rate": 8.010999793788531e-06,
+      "loss": 0.1039,
+      "step": 4639
+    },
+    {
+      "epoch": 2.4300680984808802,
+      "loss_partial": 0.06982421875,
+      "step": 4639
+    },
+    {
+      "epoch": 2.4300680984808802,
+      "loss_partial": 0.049072265625,
+      "step": 4639
+    },
+    {
+      "epoch": 2.430591932949188,
+      "grad_norm": 1.4542148113250732,
+      "learning_rate": 8.00651975882996e-06,
+      "loss": 0.1053,
+      "step": 4640
+    },
+    {
+      "epoch": 2.430591932949188,
+      "loss_partial": 0.0498046875,
+      "step": 4640
+    },
+    {
+      "epoch": 2.430591932949188,
+      "loss_partial": 0.05224609375,
+      "step": 4640
+    },
+    {
+      "epoch": 2.431115767417496,
+      "grad_norm": 1.7912260293960571,
+      "learning_rate": 8.002040140496433e-06,
+      "loss": 0.1026,
+      "step": 4641
+    },
+    {
+      "epoch": 2.431115767417496,
+      "loss_partial": 0.059326171875,
+      "step": 4641
+    },
+    {
+      "epoch": 2.431115767417496,
+      "loss_partial": 0.0498046875,
+      "step": 4641
+    },
+    {
+      "epoch": 2.431639601885804,
+      "grad_norm": 1.4728065729141235,
+      "learning_rate": 7.99756093972416e-06,
+      "loss": 0.0997,
+      "step": 4642
+    },
+    {
+      "epoch": 2.431639601885804,
+      "loss_partial": 0.061767578125,
+      "step": 4642
+    },
+    {
+      "epoch": 2.431639601885804,
+      "loss_partial": 0.04931640625,
+      "step": 4642
+    },
+    {
+      "epoch": 2.432163436354112,
+      "grad_norm": 1.624577522277832,
+      "learning_rate": 7.993082157449265e-06,
+      "loss": 0.103,
+      "step": 4643
+    },
+    {
+      "epoch": 2.432163436354112,
+      "loss_partial": 0.052490234375,
+      "step": 4643
+    },
+    {
+      "epoch": 2.432163436354112,
+      "loss_partial": 0.041259765625,
+      "step": 4643
+    },
+    {
+      "epoch": 2.43268727082242,
+      "grad_norm": 1.4253058433532715,
+      "learning_rate": 7.988603794607793e-06,
+      "loss": 0.1007,
+      "step": 4644
+    },
+    {
+      "epoch": 2.43268727082242,
+      "loss_partial": 0.053466796875,
+      "step": 4644
+    },
+    {
+      "epoch": 2.43268727082242,
+      "loss_partial": 0.050537109375,
+      "step": 4644
+    },
+    {
+      "epoch": 2.433211105290728,
+      "grad_norm": 1.824504017829895,
+      "learning_rate": 7.984125852135688e-06,
+      "loss": 0.1014,
+      "step": 4645
+    },
+    {
+      "epoch": 2.433211105290728,
+      "loss_partial": 0.05517578125,
+      "step": 4645
+    },
+    {
+      "epoch": 2.433211105290728,
+      "loss_partial": 0.0615234375,
+      "step": 4645
+    },
+    {
+      "epoch": 2.433734939759036,
+      "grad_norm": 1.340914011001587,
+      "learning_rate": 7.979648330968814e-06,
+      "loss": 0.1037,
+      "step": 4646
+    },
+    {
+      "epoch": 2.433734939759036,
+      "loss_partial": 0.061279296875,
+      "step": 4646
+    },
+    {
+      "epoch": 2.433734939759036,
+      "loss_partial": 0.052978515625,
+      "step": 4646
+    },
+    {
+      "epoch": 2.4342587742273443,
+      "grad_norm": 1.4484132528305054,
+      "learning_rate": 7.975171232042945e-06,
+      "loss": 0.1088,
+      "step": 4647
+    },
+    {
+      "epoch": 2.4342587742273443,
+      "loss_partial": 0.06005859375,
+      "step": 4647
+    },
+    {
+      "epoch": 2.4342587742273443,
+      "loss_partial": 0.04736328125,
+      "step": 4647
+    },
+    {
+      "epoch": 2.4347826086956523,
+      "grad_norm": 1.1163967847824097,
+      "learning_rate": 7.97069455629377e-06,
+      "loss": 0.1065,
+      "step": 4648
+    },
+    {
+      "epoch": 2.4347826086956523,
+      "loss_partial": 0.04833984375,
+      "step": 4648
+    },
+    {
+      "epoch": 2.4347826086956523,
+      "loss_partial": 0.051513671875,
+      "step": 4648
+    },
+    {
+      "epoch": 2.4353064431639604,
+      "grad_norm": 1.0413405895233154,
+      "learning_rate": 7.966218304656881e-06,
+      "loss": 0.0994,
+      "step": 4649
+    },
+    {
+      "epoch": 2.4353064431639604,
+      "loss_partial": 0.05078125,
+      "step": 4649
+    },
+    {
+      "epoch": 2.4353064431639604,
+      "loss_partial": 0.047119140625,
+      "step": 4649
+    },
+    {
+      "epoch": 2.4358302776322684,
+      "grad_norm": 1.7764990329742432,
+      "learning_rate": 7.961742478067793e-06,
+      "loss": 0.0965,
+      "step": 4650
+    },
+    {
+      "epoch": 2.4358302776322684,
+      "loss_partial": 0.05029296875,
+      "step": 4650
+    },
+    {
+      "epoch": 2.4358302776322684,
+      "loss_partial": 0.05322265625,
+      "step": 4650
+    },
+    {
+      "epoch": 2.436354112100576,
+      "grad_norm": 1.4523913860321045,
+      "learning_rate": 7.95726707746192e-06,
+      "loss": 0.1006,
+      "step": 4651
+    },
+    {
+      "epoch": 2.436354112100576,
+      "loss_partial": 0.057373046875,
+      "step": 4651
+    },
+    {
+      "epoch": 2.436354112100576,
+      "loss_partial": 0.0556640625,
+      "step": 4651
+    },
+    {
+      "epoch": 2.436877946568884,
+      "grad_norm": 1.3930267095565796,
+      "learning_rate": 7.952792103774597e-06,
+      "loss": 0.1033,
+      "step": 4652
+    },
+    {
+      "epoch": 2.436877946568884,
+      "loss_partial": 0.04833984375,
+      "step": 4652
+    },
+    {
+      "epoch": 2.436877946568884,
+      "loss_partial": 0.060791015625,
+      "step": 4652
+    },
+    {
+      "epoch": 2.437401781037192,
+      "grad_norm": 1.4885213375091553,
+      "learning_rate": 7.948317557941064e-06,
+      "loss": 0.1001,
+      "step": 4653
+    },
+    {
+      "epoch": 2.437401781037192,
+      "loss_partial": 0.06298828125,
+      "step": 4653
+    },
+    {
+      "epoch": 2.437401781037192,
+      "loss_partial": 0.050048828125,
+      "step": 4653
+    },
+    {
+      "epoch": 2.4379256155055002,
+      "grad_norm": 1.4266597032546997,
+      "learning_rate": 7.943843440896474e-06,
+      "loss": 0.1024,
+      "step": 4654
+    },
+    {
+      "epoch": 2.4379256155055002,
+      "loss_partial": 0.04638671875,
+      "step": 4654
+    },
+    {
+      "epoch": 2.4379256155055002,
+      "loss_partial": 0.047607421875,
+      "step": 4654
+    },
+    {
+      "epoch": 2.4384494499738083,
+      "grad_norm": 2.1052756309509277,
+      "learning_rate": 7.939369753575888e-06,
+      "loss": 0.0986,
+      "step": 4655
+    },
+    {
+      "epoch": 2.4384494499738083,
+      "loss_partial": 0.05322265625,
+      "step": 4655
+    },
+    {
+      "epoch": 2.4384494499738083,
+      "loss_partial": 0.052001953125,
+      "step": 4655
+    },
+    {
+      "epoch": 2.4389732844421164,
+      "grad_norm": 1.4995962381362915,
+      "learning_rate": 7.934896496914284e-06,
+      "loss": 0.0999,
+      "step": 4656
+    },
+    {
+      "epoch": 2.4389732844421164,
+      "loss_partial": 0.047607421875,
+      "step": 4656
+    },
+    {
+      "epoch": 2.4389732844421164,
+      "loss_partial": 0.04833984375,
+      "step": 4656
+    },
+    {
+      "epoch": 2.4394971189104244,
+      "grad_norm": 2.285701036453247,
+      "learning_rate": 7.930423671846543e-06,
+      "loss": 0.0992,
+      "step": 4657
+    },
+    {
+      "epoch": 2.4394971189104244,
+      "loss_partial": 0.049072265625,
+      "step": 4657
+    },
+    {
+      "epoch": 2.4394971189104244,
+      "loss_partial": 0.052978515625,
+      "step": 4657
+    },
+    {
+      "epoch": 2.4400209533787325,
+      "grad_norm": 1.5395584106445312,
+      "learning_rate": 7.925951279307455e-06,
+      "loss": 0.1021,
+      "step": 4658
+    },
+    {
+      "epoch": 2.4400209533787325,
+      "loss_partial": 0.051025390625,
+      "step": 4658
+    },
+    {
+      "epoch": 2.4400209533787325,
+      "loss_partial": 0.040771484375,
+      "step": 4658
+    },
+    {
+      "epoch": 2.44054478784704,
+      "grad_norm": 2.134248733520508,
+      "learning_rate": 7.921479320231723e-06,
+      "loss": 0.1035,
+      "step": 4659
+    },
+    {
+      "epoch": 2.44054478784704,
+      "loss_partial": 0.0498046875,
+      "step": 4659
+    },
+    {
+      "epoch": 2.44054478784704,
+      "loss_partial": 0.051513671875,
+      "step": 4659
+    },
+    {
+      "epoch": 2.441068622315348,
+      "grad_norm": 1.6199105978012085,
+      "learning_rate": 7.917007795553962e-06,
+      "loss": 0.1007,
+      "step": 4660
+    },
+    {
+      "epoch": 2.441068622315348,
+      "loss_partial": 0.058349609375,
+      "step": 4660
+    },
+    {
+      "epoch": 2.441068622315348,
+      "loss_partial": 0.057373046875,
+      "step": 4660
+    },
+    {
+      "epoch": 2.441592456783656,
+      "grad_norm": 1.6613349914550781,
+      "learning_rate": 7.912536706208691e-06,
+      "loss": 0.109,
+      "step": 4661
+    },
+    {
+      "epoch": 2.441592456783656,
+      "loss_partial": 0.047607421875,
+      "step": 4661
+    },
+    {
+      "epoch": 2.441592456783656,
+      "loss_partial": 0.04345703125,
+      "step": 4661
+    },
+    {
+      "epoch": 2.4421162912519643,
+      "grad_norm": 1.1881539821624756,
+      "learning_rate": 7.908066053130338e-06,
+      "loss": 0.1001,
+      "step": 4662
+    },
+    {
+      "epoch": 2.4421162912519643,
+      "loss_partial": 0.0498046875,
+      "step": 4662
+    },
+    {
+      "epoch": 2.4421162912519643,
+      "loss_partial": 0.048583984375,
+      "step": 4662
+    },
+    {
+      "epoch": 2.4426401257202723,
+      "grad_norm": 1.4437260627746582,
+      "learning_rate": 7.903595837253247e-06,
+      "loss": 0.1016,
+      "step": 4663
+    },
+    {
+      "epoch": 2.4426401257202723,
+      "loss_partial": 0.0498046875,
+      "step": 4663
+    },
+    {
+      "epoch": 2.4426401257202723,
+      "loss_partial": 0.046630859375,
+      "step": 4663
+    },
+    {
+      "epoch": 2.4431639601885804,
+      "grad_norm": 1.63087797164917,
+      "learning_rate": 7.899126059511659e-06,
+      "loss": 0.1012,
+      "step": 4664
+    },
+    {
+      "epoch": 2.4431639601885804,
+      "loss_partial": 0.052734375,
+      "step": 4664
+    },
+    {
+      "epoch": 2.4431639601885804,
+      "loss_partial": 0.0400390625,
+      "step": 4664
+    },
+    {
+      "epoch": 2.4436877946568885,
+      "grad_norm": 1.5756067037582397,
+      "learning_rate": 7.894656720839733e-06,
+      "loss": 0.1027,
+      "step": 4665
+    },
+    {
+      "epoch": 2.4436877946568885,
+      "loss_partial": 0.050537109375,
+      "step": 4665
+    },
+    {
+      "epoch": 2.4436877946568885,
+      "loss_partial": 0.047607421875,
+      "step": 4665
+    },
+    {
+      "epoch": 2.4442116291251965,
+      "grad_norm": 1.8180103302001953,
+      "learning_rate": 7.890187822171536e-06,
+      "loss": 0.1034,
+      "step": 4666
+    },
+    {
+      "epoch": 2.4442116291251965,
+      "loss_partial": 0.04541015625,
+      "step": 4666
+    },
+    {
+      "epoch": 2.4442116291251965,
+      "loss_partial": 0.045654296875,
+      "step": 4666
+    },
+    {
+      "epoch": 2.4447354635935046,
+      "grad_norm": 1.4601832628250122,
+      "learning_rate": 7.885719364441035e-06,
+      "loss": 0.0995,
+      "step": 4667
+    },
+    {
+      "epoch": 2.4447354635935046,
+      "loss_partial": 0.05322265625,
+      "step": 4667
+    },
+    {
+      "epoch": 2.4447354635935046,
+      "loss_partial": 0.046142578125,
+      "step": 4667
+    },
+    {
+      "epoch": 2.4452592980618126,
+      "grad_norm": 2.253138303756714,
+      "learning_rate": 7.881251348582111e-06,
+      "loss": 0.1024,
+      "step": 4668
+    },
+    {
+      "epoch": 2.4452592980618126,
+      "loss_partial": 0.060302734375,
+      "step": 4668
+    },
+    {
+      "epoch": 2.4452592980618126,
+      "loss_partial": 0.04736328125,
+      "step": 4668
+    },
+    {
+      "epoch": 2.4457831325301207,
+      "grad_norm": 1.8601857423782349,
+      "learning_rate": 7.876783775528556e-06,
+      "loss": 0.104,
+      "step": 4669
+    },
+    {
+      "epoch": 2.4457831325301207,
+      "loss_partial": 0.0458984375,
+      "step": 4669
+    },
+    {
+      "epoch": 2.4457831325301207,
+      "loss_partial": 0.04931640625,
+      "step": 4669
+    },
+    {
+      "epoch": 2.4463069669984283,
+      "grad_norm": 1.82535982131958,
+      "learning_rate": 7.872316646214063e-06,
+      "loss": 0.1063,
+      "step": 4670
+    },
+    {
+      "epoch": 2.4463069669984283,
+      "loss_partial": 0.054443359375,
+      "step": 4670
+    },
+    {
+      "epoch": 2.4463069669984283,
+      "loss_partial": 0.04638671875,
+      "step": 4670
+    },
+    {
+      "epoch": 2.4468308014667364,
+      "grad_norm": 1.6470744609832764,
+      "learning_rate": 7.867849961572232e-06,
+      "loss": 0.1037,
+      "step": 4671
+    },
+    {
+      "epoch": 2.4468308014667364,
+      "loss_partial": 0.048828125,
+      "step": 4671
+    },
+    {
+      "epoch": 2.4468308014667364,
+      "loss_partial": 0.0673828125,
+      "step": 4671
+    },
+    {
+      "epoch": 2.4473546359350444,
+      "grad_norm": 1.5106886625289917,
+      "learning_rate": 7.863383722536575e-06,
+      "loss": 0.1052,
+      "step": 4672
+    },
+    {
+      "epoch": 2.4473546359350444,
+      "loss_partial": 0.046630859375,
+      "step": 4672
+    },
+    {
+      "epoch": 2.4473546359350444,
+      "loss_partial": 0.051513671875,
+      "step": 4672
+    },
+    {
+      "epoch": 2.4478784704033525,
+      "grad_norm": 1.3436450958251953,
+      "learning_rate": 7.858917930040508e-06,
+      "loss": 0.1023,
+      "step": 4673
+    },
+    {
+      "epoch": 2.4478784704033525,
+      "loss_partial": 0.05029296875,
+      "step": 4673
+    },
+    {
+      "epoch": 2.4478784704033525,
+      "loss_partial": 0.04833984375,
+      "step": 4673
+    },
+    {
+      "epoch": 2.4484023048716606,
+      "grad_norm": 1.573409914970398,
+      "learning_rate": 7.854452585017353e-06,
+      "loss": 0.0978,
+      "step": 4674
+    },
+    {
+      "epoch": 2.4484023048716606,
+      "loss_partial": 0.047119140625,
+      "step": 4674
+    },
+    {
+      "epoch": 2.4484023048716606,
+      "loss_partial": 0.04541015625,
+      "step": 4674
+    },
+    {
+      "epoch": 2.4489261393399686,
+      "grad_norm": 1.3970975875854492,
+      "learning_rate": 7.84998768840034e-06,
+      "loss": 0.0993,
+      "step": 4675
+    },
+    {
+      "epoch": 2.4489261393399686,
+      "loss_partial": 0.055908203125,
+      "step": 4675
+    },
+    {
+      "epoch": 2.4489261393399686,
+      "loss_partial": 0.05810546875,
+      "step": 4675
+    },
+    {
+      "epoch": 2.4494499738082767,
+      "grad_norm": 1.4680520296096802,
+      "learning_rate": 7.845523241122605e-06,
+      "loss": 0.1025,
+      "step": 4676
+    },
+    {
+      "epoch": 2.4494499738082767,
+      "loss_partial": 0.05712890625,
+      "step": 4676
+    },
+    {
+      "epoch": 2.4494499738082767,
+      "loss_partial": 0.054931640625,
+      "step": 4676
+    },
+    {
+      "epoch": 2.4499738082765847,
+      "grad_norm": 1.87062406539917,
+      "learning_rate": 7.841059244117189e-06,
+      "loss": 0.1029,
+      "step": 4677
+    },
+    {
+      "epoch": 2.4499738082765847,
+      "loss_partial": 0.06640625,
+      "step": 4677
+    },
+    {
+      "epoch": 2.4499738082765847,
+      "loss_partial": 0.04833984375,
+      "step": 4677
+    },
+    {
+      "epoch": 2.450497642744893,
+      "grad_norm": 1.7464393377304077,
+      "learning_rate": 7.836595698317042e-06,
+      "loss": 0.1059,
+      "step": 4678
+    },
+    {
+      "epoch": 2.450497642744893,
+      "loss_partial": 0.05712890625,
+      "step": 4678
+    },
+    {
+      "epoch": 2.450497642744893,
+      "loss_partial": 0.05615234375,
+      "step": 4678
+    },
+    {
+      "epoch": 2.4510214772132004,
+      "grad_norm": 1.572252631187439,
+      "learning_rate": 7.832132604655017e-06,
+      "loss": 0.1041,
+      "step": 4679
+    },
+    {
+      "epoch": 2.4510214772132004,
+      "loss_partial": 0.05517578125,
+      "step": 4679
+    },
+    {
+      "epoch": 2.4510214772132004,
+      "loss_partial": 0.05029296875,
+      "step": 4679
+    },
+    {
+      "epoch": 2.4515453116815085,
+      "grad_norm": 1.3991690874099731,
+      "learning_rate": 7.827669964063866e-06,
+      "loss": 0.0995,
+      "step": 4680
+    },
+    {
+      "epoch": 2.4515453116815085,
+      "loss_partial": 0.05908203125,
+      "step": 4680
+    },
+    {
+      "epoch": 2.4515453116815085,
+      "loss_partial": 0.056640625,
+      "step": 4680
+    },
+    {
+      "epoch": 2.4520691461498165,
+      "grad_norm": 1.2283340692520142,
+      "learning_rate": 7.823207777476263e-06,
+      "loss": 0.1047,
+      "step": 4681
+    },
+    {
+      "epoch": 2.4520691461498165,
+      "loss_partial": 0.0439453125,
+      "step": 4681
+    },
+    {
+      "epoch": 2.4520691461498165,
+      "loss_partial": 0.05126953125,
+      "step": 4681
+    },
+    {
+      "epoch": 2.4525929806181246,
+      "grad_norm": 1.483657717704773,
+      "learning_rate": 7.818746045824775e-06,
+      "loss": 0.1009,
+      "step": 4682
+    },
+    {
+      "epoch": 2.4525929806181246,
+      "loss_partial": 0.05712890625,
+      "step": 4682
+    },
+    {
+      "epoch": 2.4525929806181246,
+      "loss_partial": 0.057373046875,
+      "step": 4682
+    },
+    {
+      "epoch": 2.4531168150864326,
+      "grad_norm": 1.442597508430481,
+      "learning_rate": 7.814284770041874e-06,
+      "loss": 0.101,
+      "step": 4683
+    },
+    {
+      "epoch": 2.4531168150864326,
+      "loss_partial": 0.0517578125,
+      "step": 4683
+    },
+    {
+      "epoch": 2.4531168150864326,
+      "loss_partial": 0.06201171875,
+      "step": 4683
+    },
+    {
+      "epoch": 2.4536406495547407,
+      "grad_norm": 1.3773493766784668,
+      "learning_rate": 7.809823951059939e-06,
+      "loss": 0.1046,
+      "step": 4684
+    },
+    {
+      "epoch": 2.4536406495547407,
+      "loss_partial": 0.06103515625,
+      "step": 4684
+    },
+    {
+      "epoch": 2.4536406495547407,
+      "loss_partial": 0.046875,
+      "step": 4684
+    },
+    {
+      "epoch": 2.4541644840230488,
+      "grad_norm": 1.5383949279785156,
+      "learning_rate": 7.805363589811256e-06,
+      "loss": 0.1002,
+      "step": 4685
+    },
+    {
+      "epoch": 2.4541644840230488,
+      "loss_partial": 0.046630859375,
+      "step": 4685
+    },
+    {
+      "epoch": 2.4541644840230488,
+      "loss_partial": 0.052734375,
+      "step": 4685
+    },
+    {
+      "epoch": 2.454688318491357,
+      "grad_norm": 1.6664741039276123,
+      "learning_rate": 7.800903687228012e-06,
+      "loss": 0.1029,
+      "step": 4686
+    },
+    {
+      "epoch": 2.454688318491357,
+      "loss_partial": 0.05615234375,
+      "step": 4686
+    },
+    {
+      "epoch": 2.454688318491357,
+      "loss_partial": 0.0458984375,
+      "step": 4686
+    },
+    {
+      "epoch": 2.455212152959665,
+      "grad_norm": 1.3994277715682983,
+      "learning_rate": 7.796444244242297e-06,
+      "loss": 0.1042,
+      "step": 4687
+    },
+    {
+      "epoch": 2.455212152959665,
+      "loss_partial": 0.04736328125,
+      "step": 4687
+    },
+    {
+      "epoch": 2.455212152959665,
+      "loss_partial": 0.062255859375,
+      "step": 4687
+    },
+    {
+      "epoch": 2.455735987427973,
+      "grad_norm": 1.587319016456604,
+      "learning_rate": 7.791985261786108e-06,
+      "loss": 0.1027,
+      "step": 4688
+    },
+    {
+      "epoch": 2.455735987427973,
+      "loss_partial": 0.06787109375,
+      "step": 4688
+    },
+    {
+      "epoch": 2.455735987427973,
+      "loss_partial": 0.052734375,
+      "step": 4688
+    },
+    {
+      "epoch": 2.456259821896281,
+      "grad_norm": 1.3942433595657349,
+      "learning_rate": 7.787526740791347e-06,
+      "loss": 0.0991,
+      "step": 4689
+    },
+    {
+      "epoch": 2.456259821896281,
+      "loss_partial": 0.050048828125,
+      "step": 4689
+    },
+    {
+      "epoch": 2.456259821896281,
+      "loss_partial": 0.055419921875,
+      "step": 4689
+    },
+    {
+      "epoch": 2.4567836563645886,
+      "grad_norm": 1.380881667137146,
+      "learning_rate": 7.783068682189816e-06,
+      "loss": 0.1017,
+      "step": 4690
+    },
+    {
+      "epoch": 2.4567836563645886,
+      "loss_partial": 0.05029296875,
+      "step": 4690
+    },
+    {
+      "epoch": 2.4567836563645886,
+      "loss_partial": 0.047607421875,
+      "step": 4690
+    },
+    {
+      "epoch": 2.4573074908328967,
+      "grad_norm": 1.5388023853302002,
+      "learning_rate": 7.778611086913221e-06,
+      "loss": 0.1018,
+      "step": 4691
+    },
+    {
+      "epoch": 2.4573074908328967,
+      "loss_partial": 0.051025390625,
+      "step": 4691
+    },
+    {
+      "epoch": 2.4573074908328967,
+      "loss_partial": 0.052734375,
+      "step": 4691
+    },
+    {
+      "epoch": 2.4578313253012047,
+      "grad_norm": 1.6161179542541504,
+      "learning_rate": 7.774153955893168e-06,
+      "loss": 0.1037,
+      "step": 4692
+    },
+    {
+      "epoch": 2.4578313253012047,
+      "loss_partial": 0.048095703125,
+      "step": 4692
+    },
+    {
+      "epoch": 2.4578313253012047,
+      "loss_partial": 0.0478515625,
+      "step": 4692
+    },
+    {
+      "epoch": 2.458355159769513,
+      "grad_norm": 1.7006388902664185,
+      "learning_rate": 7.769697290061177e-06,
+      "loss": 0.1028,
+      "step": 4693
+    },
+    {
+      "epoch": 2.458355159769513,
+      "loss_partial": 0.05029296875,
+      "step": 4693
+    },
+    {
+      "epoch": 2.458355159769513,
+      "loss_partial": 0.048095703125,
+      "step": 4693
+    },
+    {
+      "epoch": 2.458878994237821,
+      "grad_norm": 1.2524042129516602,
+      "learning_rate": 7.765241090348663e-06,
+      "loss": 0.1022,
+      "step": 4694
+    },
+    {
+      "epoch": 2.458878994237821,
+      "loss_partial": 0.050537109375,
+      "step": 4694
+    },
+    {
+      "epoch": 2.458878994237821,
+      "loss_partial": 0.056884765625,
+      "step": 4694
+    },
+    {
+      "epoch": 2.459402828706129,
+      "grad_norm": 1.5064036846160889,
+      "learning_rate": 7.76078535768694e-06,
+      "loss": 0.1025,
+      "step": 4695
+    },
+    {
+      "epoch": 2.459402828706129,
+      "loss_partial": 0.04736328125,
+      "step": 4695
+    },
+    {
+      "epoch": 2.459402828706129,
+      "loss_partial": 0.05810546875,
+      "step": 4695
+    },
+    {
+      "epoch": 2.459926663174437,
+      "grad_norm": 1.5361213684082031,
+      "learning_rate": 7.756330093007229e-06,
+      "loss": 0.1037,
+      "step": 4696
+    },
+    {
+      "epoch": 2.459926663174437,
+      "loss_partial": 0.05712890625,
+      "step": 4696
+    },
+    {
+      "epoch": 2.459926663174437,
+      "loss_partial": 0.05712890625,
+      "step": 4696
+    },
+    {
+      "epoch": 2.460450497642745,
+      "grad_norm": 1.1939226388931274,
+      "learning_rate": 7.751875297240656e-06,
+      "loss": 0.1042,
+      "step": 4697
+    },
+    {
+      "epoch": 2.460450497642745,
+      "loss_partial": 0.048583984375,
+      "step": 4697
+    },
+    {
+      "epoch": 2.460450497642745,
+      "loss_partial": 0.0634765625,
+      "step": 4697
+    },
+    {
+      "epoch": 2.460974332111053,
+      "grad_norm": 1.3509517908096313,
+      "learning_rate": 7.747420971318245e-06,
+      "loss": 0.1053,
+      "step": 4698
+    },
+    {
+      "epoch": 2.460974332111053,
+      "loss_partial": 0.0498046875,
+      "step": 4698
+    },
+    {
+      "epoch": 2.460974332111053,
+      "loss_partial": 0.052490234375,
+      "step": 4698
+    },
+    {
+      "epoch": 2.4614981665793607,
+      "grad_norm": 1.685531497001648,
+      "learning_rate": 7.74296711617092e-06,
+      "loss": 0.1062,
+      "step": 4699
+    },
+    {
+      "epoch": 2.4614981665793607,
+      "loss_partial": 0.04833984375,
+      "step": 4699
+    },
+    {
+      "epoch": 2.4614981665793607,
+      "loss_partial": 0.05078125,
+      "step": 4699
+    },
+    {
+      "epoch": 2.4620220010476688,
+      "grad_norm": 1.2985199689865112,
+      "learning_rate": 7.73851373272951e-06,
+      "loss": 0.1022,
+      "step": 4700
+    },
+    {
+      "epoch": 2.4620220010476688,
+      "loss_partial": 0.043701171875,
+      "step": 4700
+    },
+    {
+      "epoch": 2.4620220010476688,
+      "loss_partial": 0.0478515625,
+      "step": 4700
+    },
+    {
+      "epoch": 2.462545835515977,
+      "grad_norm": 1.832779049873352,
+      "learning_rate": 7.734060821924744e-06,
+      "loss": 0.0966,
+      "step": 4701
+    },
+    {
+      "epoch": 2.462545835515977,
+      "loss_partial": 0.0576171875,
+      "step": 4701
+    },
+    {
+      "epoch": 2.462545835515977,
+      "loss_partial": 0.056640625,
+      "step": 4701
+    },
+    {
+      "epoch": 2.463069669984285,
+      "grad_norm": 2.171231508255005,
+      "learning_rate": 7.729608384687257e-06,
+      "loss": 0.1026,
+      "step": 4702
+    },
+    {
+      "epoch": 2.463069669984285,
+      "loss_partial": 0.05078125,
+      "step": 4702
+    },
+    {
+      "epoch": 2.463069669984285,
+      "loss_partial": 0.040283203125,
+      "step": 4702
+    },
+    {
+      "epoch": 2.463593504452593,
+      "grad_norm": 1.6615930795669556,
+      "learning_rate": 7.725156421947577e-06,
+      "loss": 0.0958,
+      "step": 4703
+    },
+    {
+      "epoch": 2.463593504452593,
+      "loss_partial": 0.059326171875,
+      "step": 4703
+    },
+    {
+      "epoch": 2.463593504452593,
+      "loss_partial": 0.049560546875,
+      "step": 4703
+    },
+    {
+      "epoch": 2.464117338920901,
+      "grad_norm": 1.5097030401229858,
+      "learning_rate": 7.720704934636134e-06,
+      "loss": 0.1005,
+      "step": 4704
+    },
+    {
+      "epoch": 2.464117338920901,
+      "loss_partial": 0.054931640625,
+      "step": 4704
+    },
+    {
+      "epoch": 2.464117338920901,
+      "loss_partial": 0.048095703125,
+      "step": 4704
+    },
+    {
+      "epoch": 2.464641173389209,
+      "grad_norm": 1.2971324920654297,
+      "learning_rate": 7.716253923683268e-06,
+      "loss": 0.1024,
+      "step": 4705
+    },
+    {
+      "epoch": 2.464641173389209,
+      "loss_partial": 0.05078125,
+      "step": 4705
+    },
+    {
+      "epoch": 2.464641173389209,
+      "loss_partial": 0.052001953125,
+      "step": 4705
+    },
+    {
+      "epoch": 2.465165007857517,
+      "grad_norm": 1.5840802192687988,
+      "learning_rate": 7.711803390019212e-06,
+      "loss": 0.1022,
+      "step": 4706
+    },
+    {
+      "epoch": 2.465165007857517,
+      "loss_partial": 0.05419921875,
+      "step": 4706
+    },
+    {
+      "epoch": 2.465165007857517,
+      "loss_partial": 0.048828125,
+      "step": 4706
+    },
+    {
+      "epoch": 2.465688842325825,
+      "grad_norm": 1.389672875404358,
+      "learning_rate": 7.707353334574096e-06,
+      "loss": 0.1022,
+      "step": 4707
+    },
+    {
+      "epoch": 2.465688842325825,
+      "loss_partial": 0.048583984375,
+      "step": 4707
+    },
+    {
+      "epoch": 2.465688842325825,
+      "loss_partial": 0.06494140625,
+      "step": 4707
+    },
+    {
+      "epoch": 2.4662126767941333,
+      "grad_norm": 1.817806363105774,
+      "learning_rate": 7.702903758277954e-06,
+      "loss": 0.103,
+      "step": 4708
+    },
+    {
+      "epoch": 2.4662126767941333,
+      "loss_partial": 0.045166015625,
+      "step": 4708
+    },
+    {
+      "epoch": 2.4662126767941333,
+      "loss_partial": 0.04736328125,
+      "step": 4708
+    },
+    {
+      "epoch": 2.4667365112624413,
+      "grad_norm": 2.0603713989257812,
+      "learning_rate": 7.698454662060723e-06,
+      "loss": 0.1009,
+      "step": 4709
+    },
+    {
+      "epoch": 2.4667365112624413,
+      "loss_partial": 0.043701171875,
+      "step": 4709
+    },
+    {
+      "epoch": 2.4667365112624413,
+      "loss_partial": 0.05517578125,
+      "step": 4709
+    },
+    {
+      "epoch": 2.467260345730749,
+      "grad_norm": 1.2982561588287354,
+      "learning_rate": 7.694006046852238e-06,
+      "loss": 0.1044,
+      "step": 4710
+    },
+    {
+      "epoch": 2.467260345730749,
+      "loss_partial": 0.0517578125,
+      "step": 4710
+    },
+    {
+      "epoch": 2.467260345730749,
+      "loss_partial": 0.044189453125,
+      "step": 4710
+    },
+    {
+      "epoch": 2.467784180199057,
+      "grad_norm": 1.6999455690383911,
+      "learning_rate": 7.689557913582226e-06,
+      "loss": 0.1018,
+      "step": 4711
+    },
+    {
+      "epoch": 2.467784180199057,
+      "loss_partial": 0.04541015625,
+      "step": 4711
+    },
+    {
+      "epoch": 2.467784180199057,
+      "loss_partial": 0.056884765625,
+      "step": 4711
+    },
+    {
+      "epoch": 2.468308014667365,
+      "grad_norm": 1.7729705572128296,
+      "learning_rate": 7.68511026318032e-06,
+      "loss": 0.1021,
+      "step": 4712
+    },
+    {
+      "epoch": 2.468308014667365,
+      "loss_partial": 0.052734375,
+      "step": 4712
+    },
+    {
+      "epoch": 2.468308014667365,
+      "loss_partial": 0.053466796875,
+      "step": 4712
+    },
+    {
+      "epoch": 2.468831849135673,
+      "grad_norm": 1.5119704008102417,
+      "learning_rate": 7.68066309657606e-06,
+      "loss": 0.1043,
+      "step": 4713
+    },
+    {
+      "epoch": 2.468831849135673,
+      "loss_partial": 0.056396484375,
+      "step": 4713
+    },
+    {
+      "epoch": 2.468831849135673,
+      "loss_partial": 0.052734375,
+      "step": 4713
+    },
+    {
+      "epoch": 2.469355683603981,
+      "grad_norm": 1.3226850032806396,
+      "learning_rate": 7.676216414698869e-06,
+      "loss": 0.1046,
+      "step": 4714
+    },
+    {
+      "epoch": 2.469355683603981,
+      "loss_partial": 0.052978515625,
+      "step": 4714
+    },
+    {
+      "epoch": 2.469355683603981,
+      "loss_partial": 0.051025390625,
+      "step": 4714
+    },
+    {
+      "epoch": 2.4698795180722892,
+      "grad_norm": 1.7650749683380127,
+      "learning_rate": 7.671770218478077e-06,
+      "loss": 0.1029,
+      "step": 4715
+    },
+    {
+      "epoch": 2.4698795180722892,
+      "loss_partial": 0.051025390625,
+      "step": 4715
+    },
+    {
+      "epoch": 2.4698795180722892,
+      "loss_partial": 0.052734375,
+      "step": 4715
+    },
+    {
+      "epoch": 2.4704033525405973,
+      "grad_norm": 1.5071996450424194,
+      "learning_rate": 7.667324508842908e-06,
+      "loss": 0.0992,
+      "step": 4716
+    },
+    {
+      "epoch": 2.4704033525405973,
+      "loss_partial": 0.04833984375,
+      "step": 4716
+    },
+    {
+      "epoch": 2.4704033525405973,
+      "loss_partial": 0.05419921875,
+      "step": 4716
+    },
+    {
+      "epoch": 2.4709271870089053,
+      "grad_norm": 1.5081154108047485,
+      "learning_rate": 7.662879286722496e-06,
+      "loss": 0.1026,
+      "step": 4717
+    },
+    {
+      "epoch": 2.4709271870089053,
+      "loss_partial": 0.0498046875,
+      "step": 4717
+    },
+    {
+      "epoch": 2.4709271870089053,
+      "loss_partial": 0.042724609375,
+      "step": 4717
+    },
+    {
+      "epoch": 2.471451021477213,
+      "grad_norm": 1.6802955865859985,
+      "learning_rate": 7.658434553045862e-06,
+      "loss": 0.1007,
+      "step": 4718
+    },
+    {
+      "epoch": 2.471451021477213,
+      "loss_partial": 0.043212890625,
+      "step": 4718
+    },
+    {
+      "epoch": 2.471451021477213,
+      "loss_partial": 0.047119140625,
+      "step": 4718
+    },
+    {
+      "epoch": 2.471974855945521,
+      "grad_norm": 1.6888644695281982,
+      "learning_rate": 7.653990308741925e-06,
+      "loss": 0.1032,
+      "step": 4719
+    },
+    {
+      "epoch": 2.471974855945521,
+      "loss_partial": 0.043212890625,
+      "step": 4719
+    },
+    {
+      "epoch": 2.471974855945521,
+      "loss_partial": 0.05029296875,
+      "step": 4719
+    },
+    {
+      "epoch": 2.472498690413829,
+      "grad_norm": 1.7564140558242798,
+      "learning_rate": 7.649546554739507e-06,
+      "loss": 0.0972,
+      "step": 4720
+    },
+    {
+      "epoch": 2.472498690413829,
+      "loss_partial": 0.053955078125,
+      "step": 4720
+    },
+    {
+      "epoch": 2.472498690413829,
+      "loss_partial": 0.052734375,
+      "step": 4720
+    },
+    {
+      "epoch": 2.473022524882137,
+      "grad_norm": 1.858479619026184,
+      "learning_rate": 7.645103291967321e-06,
+      "loss": 0.1049,
+      "step": 4721
+    },
+    {
+      "epoch": 2.473022524882137,
+      "loss_partial": 0.05224609375,
+      "step": 4721
+    },
+    {
+      "epoch": 2.473022524882137,
+      "loss_partial": 0.059326171875,
+      "step": 4721
+    },
+    {
+      "epoch": 2.473546359350445,
+      "grad_norm": 1.7100528478622437,
+      "learning_rate": 7.640660521353987e-06,
+      "loss": 0.1028,
+      "step": 4722
+    },
+    {
+      "epoch": 2.473546359350445,
+      "loss_partial": 0.057373046875,
+      "step": 4722
+    },
+    {
+      "epoch": 2.473546359350445,
+      "loss_partial": 0.06298828125,
+      "step": 4722
+    },
+    {
+      "epoch": 2.4740701938187533,
+      "grad_norm": 1.682612419128418,
+      "learning_rate": 7.636218243828015e-06,
+      "loss": 0.1,
+      "step": 4723
+    },
+    {
+      "epoch": 2.4740701938187533,
+      "loss_partial": 0.0625,
+      "step": 4723
+    },
+    {
+      "epoch": 2.4740701938187533,
+      "loss_partial": 0.056640625,
+      "step": 4723
+    },
+    {
+      "epoch": 2.4745940282870613,
+      "grad_norm": 1.9128412008285522,
+      "learning_rate": 7.631776460317814e-06,
+      "loss": 0.1017,
+      "step": 4724
+    },
+    {
+      "epoch": 2.4745940282870613,
+      "loss_partial": 0.059326171875,
+      "step": 4724
+    },
+    {
+      "epoch": 2.4745940282870613,
+      "loss_partial": 0.052978515625,
+      "step": 4724
+    },
+    {
+      "epoch": 2.4751178627553694,
+      "grad_norm": 1.4338791370391846,
+      "learning_rate": 7.627335171751684e-06,
+      "loss": 0.0963,
+      "step": 4725
+    },
+    {
+      "epoch": 2.4751178627553694,
+      "loss_partial": 0.060791015625,
+      "step": 4725
+    },
+    {
+      "epoch": 2.4751178627553694,
+      "loss_partial": 0.051025390625,
+      "step": 4725
+    },
+    {
+      "epoch": 2.4756416972236774,
+      "grad_norm": 1.829272747039795,
+      "learning_rate": 7.622894379057833e-06,
+      "loss": 0.1044,
+      "step": 4726
+    },
+    {
+      "epoch": 2.4756416972236774,
+      "loss_partial": 0.048583984375,
+      "step": 4726
+    },
+    {
+      "epoch": 2.4756416972236774,
+      "loss_partial": 0.0517578125,
+      "step": 4726
+    },
+    {
+      "epoch": 2.4761655316919855,
+      "grad_norm": 1.2339503765106201,
+      "learning_rate": 7.6184540831643595e-06,
+      "loss": 0.1009,
+      "step": 4727
+    },
+    {
+      "epoch": 2.4761655316919855,
+      "loss_partial": 0.058837890625,
+      "step": 4727
+    },
+    {
+      "epoch": 2.4761655316919855,
+      "loss_partial": 0.060791015625,
+      "step": 4727
+    },
+    {
+      "epoch": 2.4766893661602936,
+      "grad_norm": 2.0843710899353027,
+      "learning_rate": 7.614014284999254e-06,
+      "loss": 0.102,
+      "step": 4728
+    },
+    {
+      "epoch": 2.4766893661602936,
+      "loss_partial": 0.0498046875,
+      "step": 4728
+    },
+    {
+      "epoch": 2.4766893661602936,
+      "loss_partial": 0.048095703125,
+      "step": 4728
+    },
+    {
+      "epoch": 2.477213200628601,
+      "grad_norm": 1.6616551876068115,
+      "learning_rate": 7.609574985490404e-06,
+      "loss": 0.1022,
+      "step": 4729
+    },
+    {
+      "epoch": 2.477213200628601,
+      "loss_partial": 0.056640625,
+      "step": 4729
+    },
+    {
+      "epoch": 2.477213200628601,
+      "loss_partial": 0.06591796875,
+      "step": 4729
+    },
+    {
+      "epoch": 2.4777370350969092,
+      "grad_norm": 1.8358222246170044,
+      "learning_rate": 7.605136185565605e-06,
+      "loss": 0.1046,
+      "step": 4730
+    },
+    {
+      "epoch": 2.4777370350969092,
+      "loss_partial": 0.052490234375,
+      "step": 4730
+    },
+    {
+      "epoch": 2.4777370350969092,
+      "loss_partial": 0.0556640625,
+      "step": 4730
+    },
+    {
+      "epoch": 2.4782608695652173,
+      "grad_norm": 1.3032833337783813,
+      "learning_rate": 7.600697886152533e-06,
+      "loss": 0.101,
+      "step": 4731
+    },
+    {
+      "epoch": 2.4782608695652173,
+      "loss_partial": 0.053466796875,
+      "step": 4731
+    },
+    {
+      "epoch": 2.4782608695652173,
+      "loss_partial": 0.0390625,
+      "step": 4731
+    },
+    {
+      "epoch": 2.4787847040335254,
+      "grad_norm": 1.8249915838241577,
+      "learning_rate": 7.596260088178766e-06,
+      "loss": 0.103,
+      "step": 4732
+    },
+    {
+      "epoch": 2.4787847040335254,
+      "loss_partial": 0.041748046875,
+      "step": 4732
+    },
+    {
+      "epoch": 2.4787847040335254,
+      "loss_partial": 0.04833984375,
+      "step": 4732
+    },
+    {
+      "epoch": 2.4793085385018334,
+      "grad_norm": 1.4255001544952393,
+      "learning_rate": 7.591822792571776e-06,
+      "loss": 0.102,
+      "step": 4733
+    },
+    {
+      "epoch": 2.4793085385018334,
+      "loss_partial": 0.046875,
+      "step": 4733
+    },
+    {
+      "epoch": 2.4793085385018334,
+      "loss_partial": 0.05078125,
+      "step": 4733
+    },
+    {
+      "epoch": 2.4798323729701415,
+      "grad_norm": 1.5250226259231567,
+      "learning_rate": 7.587386000258931e-06,
+      "loss": 0.0985,
+      "step": 4734
+    },
+    {
+      "epoch": 2.4798323729701415,
+      "loss_partial": 0.0498046875,
+      "step": 4734
+    },
+    {
+      "epoch": 2.4798323729701415,
+      "loss_partial": 0.0537109375,
+      "step": 4734
+    },
+    {
+      "epoch": 2.4803562074384495,
+      "grad_norm": 1.6871521472930908,
+      "learning_rate": 7.582949712167493e-06,
+      "loss": 0.1015,
+      "step": 4735
+    },
+    {
+      "epoch": 2.4803562074384495,
+      "loss_partial": 0.056396484375,
+      "step": 4735
+    },
+    {
+      "epoch": 2.4803562074384495,
+      "loss_partial": 0.0537109375,
+      "step": 4735
+    },
+    {
+      "epoch": 2.4808800419067576,
+      "grad_norm": 1.421401858329773,
+      "learning_rate": 7.578513929224619e-06,
+      "loss": 0.1005,
+      "step": 4736
+    },
+    {
+      "epoch": 2.4808800419067576,
+      "loss_partial": 0.05224609375,
+      "step": 4736
+    },
+    {
+      "epoch": 2.4808800419067576,
+      "loss_partial": 0.052490234375,
+      "step": 4736
+    },
+    {
+      "epoch": 2.4814038763750657,
+      "grad_norm": 1.4587335586547852,
+      "learning_rate": 7.574078652357357e-06,
+      "loss": 0.1035,
+      "step": 4737
+    },
+    {
+      "epoch": 2.4814038763750657,
+      "loss_partial": 0.05322265625,
+      "step": 4737
+    },
+    {
+      "epoch": 2.4814038763750657,
+      "loss_partial": 0.055419921875,
+      "step": 4737
+    },
+    {
+      "epoch": 2.4819277108433733,
+      "grad_norm": 1.5308952331542969,
+      "learning_rate": 7.569643882492659e-06,
+      "loss": 0.1009,
+      "step": 4738
+    },
+    {
+      "epoch": 2.4819277108433733,
+      "loss_partial": 0.051513671875,
+      "step": 4738
+    },
+    {
+      "epoch": 2.4819277108433733,
+      "loss_partial": 0.04345703125,
+      "step": 4738
+    },
+    {
+      "epoch": 2.4824515453116813,
+      "grad_norm": 1.3292019367218018,
+      "learning_rate": 7.5652096205573584e-06,
+      "loss": 0.099,
+      "step": 4739
+    },
+    {
+      "epoch": 2.4824515453116813,
+      "loss_partial": 0.05859375,
+      "step": 4739
+    },
+    {
+      "epoch": 2.4824515453116813,
+      "loss_partial": 0.047607421875,
+      "step": 4739
+    },
+    {
+      "epoch": 2.4829753797799894,
+      "grad_norm": 1.2146905660629272,
+      "learning_rate": 7.560775867478192e-06,
+      "loss": 0.1033,
+      "step": 4740
+    },
+    {
+      "epoch": 2.4829753797799894,
+      "loss_partial": 0.051513671875,
+      "step": 4740
+    },
+    {
+      "epoch": 2.4829753797799894,
+      "loss_partial": 0.0517578125,
+      "step": 4740
+    },
+    {
+      "epoch": 2.4834992142482974,
+      "grad_norm": 1.1010547876358032,
+      "learning_rate": 7.556342624181781e-06,
+      "loss": 0.1029,
+      "step": 4741
+    },
+    {
+      "epoch": 2.4834992142482974,
+      "loss_partial": 0.042724609375,
+      "step": 4741
+    },
+    {
+      "epoch": 2.4834992142482974,
+      "loss_partial": 0.042724609375,
+      "step": 4741
+    },
+    {
+      "epoch": 2.4840230487166055,
+      "grad_norm": 1.532508134841919,
+      "learning_rate": 7.551909891594655e-06,
+      "loss": 0.1025,
+      "step": 4742
+    },
+    {
+      "epoch": 2.4840230487166055,
+      "loss_partial": 0.044677734375,
+      "step": 4742
+    },
+    {
+      "epoch": 2.4840230487166055,
+      "loss_partial": 0.048828125,
+      "step": 4742
+    },
+    {
+      "epoch": 2.4845468831849136,
+      "grad_norm": 1.2278770208358765,
+      "learning_rate": 7.547477670643223e-06,
+      "loss": 0.1021,
+      "step": 4743
+    },
+    {
+      "epoch": 2.4845468831849136,
+      "loss_partial": 0.04638671875,
+      "step": 4743
+    },
+    {
+      "epoch": 2.4845468831849136,
+      "loss_partial": 0.046142578125,
+      "step": 4743
+    },
+    {
+      "epoch": 2.4850707176532216,
+      "grad_norm": 1.250858187675476,
+      "learning_rate": 7.543045962253792e-06,
+      "loss": 0.0999,
+      "step": 4744
+    },
+    {
+      "epoch": 2.4850707176532216,
+      "loss_partial": 0.052978515625,
+      "step": 4744
+    },
+    {
+      "epoch": 2.4850707176532216,
+      "loss_partial": 0.0576171875,
+      "step": 4744
+    },
+    {
+      "epoch": 2.4855945521215297,
+      "grad_norm": 1.5794453620910645,
+      "learning_rate": 7.538614767352559e-06,
+      "loss": 0.1048,
+      "step": 4745
+    },
+    {
+      "epoch": 2.4855945521215297,
+      "loss_partial": 0.044189453125,
+      "step": 4745
+    },
+    {
+      "epoch": 2.4855945521215297,
+      "loss_partial": 0.0517578125,
+      "step": 4745
+    },
+    {
+      "epoch": 2.4861183865898377,
+      "grad_norm": 1.447888731956482,
+      "learning_rate": 7.534184086865621e-06,
+      "loss": 0.1053,
+      "step": 4746
+    },
+    {
+      "epoch": 2.4861183865898377,
+      "loss_partial": 0.0546875,
+      "step": 4746
+    },
+    {
+      "epoch": 2.4861183865898377,
+      "loss_partial": 0.054443359375,
+      "step": 4746
+    },
+    {
+      "epoch": 2.486642221058146,
+      "grad_norm": 1.8879774808883667,
+      "learning_rate": 7.529753921718962e-06,
+      "loss": 0.0998,
+      "step": 4747
+    },
+    {
+      "epoch": 2.486642221058146,
+      "loss_partial": 0.052490234375,
+      "step": 4747
+    },
+    {
+      "epoch": 2.486642221058146,
+      "loss_partial": 0.052490234375,
+      "step": 4747
+    },
+    {
+      "epoch": 2.487166055526454,
+      "grad_norm": 1.4476014375686646,
+      "learning_rate": 7.5253242728384565e-06,
+      "loss": 0.1042,
+      "step": 4748
+    },
+    {
+      "epoch": 2.487166055526454,
+      "loss_partial": 0.044189453125,
+      "step": 4748
+    },
+    {
+      "epoch": 2.487166055526454,
+      "loss_partial": 0.062255859375,
+      "step": 4748
+    },
+    {
+      "epoch": 2.4876898899947615,
+      "grad_norm": 1.2706983089447021,
+      "learning_rate": 7.520895141149871e-06,
+      "loss": 0.1021,
+      "step": 4749
+    },
+    {
+      "epoch": 2.4876898899947615,
+      "loss_partial": 0.060302734375,
+      "step": 4749
+    },
+    {
+      "epoch": 2.4876898899947615,
+      "loss_partial": 0.04736328125,
+      "step": 4749
+    },
+    {
+      "epoch": 2.4882137244630695,
+      "grad_norm": 1.6863621473312378,
+      "learning_rate": 7.516466527578877e-06,
+      "loss": 0.1034,
+      "step": 4750
+    },
+    {
+      "epoch": 2.4882137244630695,
+      "loss_partial": 0.046875,
+      "step": 4750
+    },
+    {
+      "epoch": 2.4882137244630695,
+      "loss_partial": 0.0439453125,
+      "step": 4750
+    },
+    {
+      "epoch": 2.4887375589313776,
+      "grad_norm": 1.1673184633255005,
+      "learning_rate": 7.5120384330510196e-06,
+      "loss": 0.0978,
+      "step": 4751
+    },
+    {
+      "epoch": 2.4887375589313776,
+      "loss_partial": 0.04296875,
+      "step": 4751
+    },
+    {
+      "epoch": 2.4887375589313776,
+      "loss_partial": 0.04443359375,
+      "step": 4751
+    },
+    {
+      "epoch": 2.4892613933996857,
+      "grad_norm": 1.2383348941802979,
+      "learning_rate": 7.507610858491744e-06,
+      "loss": 0.0977,
+      "step": 4752
+    },
+    {
+      "epoch": 2.4892613933996857,
+      "loss_partial": 0.0537109375,
+      "step": 4752
+    },
+    {
+      "epoch": 2.4892613933996857,
+      "loss_partial": 0.057373046875,
+      "step": 4752
+    },
+    {
+      "epoch": 2.4897852278679937,
+      "grad_norm": 1.375989317893982,
+      "learning_rate": 7.503183804826382e-06,
+      "loss": 0.1033,
+      "step": 4753
+    },
+    {
+      "epoch": 2.4897852278679937,
+      "loss_partial": 0.04248046875,
+      "step": 4753
+    },
+    {
+      "epoch": 2.4897852278679937,
+      "loss_partial": 0.059326171875,
+      "step": 4753
+    },
+    {
+      "epoch": 2.490309062336302,
+      "grad_norm": 1.5244654417037964,
+      "learning_rate": 7.4987572729801705e-06,
+      "loss": 0.1001,
+      "step": 4754
+    },
+    {
+      "epoch": 2.490309062336302,
+      "loss_partial": 0.054443359375,
+      "step": 4754
+    },
+    {
+      "epoch": 2.490309062336302,
+      "loss_partial": 0.051025390625,
+      "step": 4754
+    },
+    {
+      "epoch": 2.49083289680461,
+      "grad_norm": 1.5776854753494263,
+      "learning_rate": 7.494331263878221e-06,
+      "loss": 0.1015,
+      "step": 4755
+    },
+    {
+      "epoch": 2.49083289680461,
+      "loss_partial": 0.0498046875,
+      "step": 4755
+    },
+    {
+      "epoch": 2.49083289680461,
+      "loss_partial": 0.0498046875,
+      "step": 4755
+    },
+    {
+      "epoch": 2.491356731272918,
+      "grad_norm": 1.5855138301849365,
+      "learning_rate": 7.489905778445541e-06,
+      "loss": 0.101,
+      "step": 4756
+    },
+    {
+      "epoch": 2.491356731272918,
+      "loss_partial": 0.0517578125,
+      "step": 4756
+    },
+    {
+      "epoch": 2.491356731272918,
+      "loss_partial": 0.057373046875,
+      "step": 4756
+    },
+    {
+      "epoch": 2.491880565741226,
+      "grad_norm": 1.65276300907135,
+      "learning_rate": 7.485480817607031e-06,
+      "loss": 0.1013,
+      "step": 4757
+    },
+    {
+      "epoch": 2.491880565741226,
+      "loss_partial": 0.054443359375,
+      "step": 4757
+    },
+    {
+      "epoch": 2.491880565741226,
+      "loss_partial": 0.04345703125,
+      "step": 4757
+    },
+    {
+      "epoch": 2.4924044002095336,
+      "grad_norm": 1.8881233930587769,
+      "learning_rate": 7.481056382287484e-06,
+      "loss": 0.1021,
+      "step": 4758
+    },
+    {
+      "epoch": 2.4924044002095336,
+      "loss_partial": 0.054931640625,
+      "step": 4758
+    },
+    {
+      "epoch": 2.4924044002095336,
+      "loss_partial": 0.056884765625,
+      "step": 4758
+    },
+    {
+      "epoch": 2.4929282346778416,
+      "grad_norm": 1.1665525436401367,
+      "learning_rate": 7.476632473411573e-06,
+      "loss": 0.1035,
+      "step": 4759
+    },
+    {
+      "epoch": 2.4929282346778416,
+      "loss_partial": 0.052490234375,
+      "step": 4759
+    },
+    {
+      "epoch": 2.4929282346778416,
+      "loss_partial": 0.044677734375,
+      "step": 4759
+    },
+    {
+      "epoch": 2.4934520691461497,
+      "grad_norm": 1.9726961851119995,
+      "learning_rate": 7.472209091903875e-06,
+      "loss": 0.1051,
+      "step": 4760
+    },
+    {
+      "epoch": 2.4934520691461497,
+      "loss_partial": 0.05517578125,
+      "step": 4760
+    },
+    {
+      "epoch": 2.4934520691461497,
+      "loss_partial": 0.047607421875,
+      "step": 4760
+    },
+    {
+      "epoch": 2.4939759036144578,
+      "grad_norm": 1.1835510730743408,
+      "learning_rate": 7.467786238688839e-06,
+      "loss": 0.1021,
+      "step": 4761
+    },
+    {
+      "epoch": 2.4939759036144578,
+      "loss_partial": 0.05322265625,
+      "step": 4761
+    },
+    {
+      "epoch": 2.4939759036144578,
+      "loss_partial": 0.05810546875,
+      "step": 4761
+    },
+    {
+      "epoch": 2.494499738082766,
+      "grad_norm": 1.572123646736145,
+      "learning_rate": 7.463363914690826e-06,
+      "loss": 0.1049,
+      "step": 4762
+    },
+    {
+      "epoch": 2.494499738082766,
+      "loss_partial": 0.04931640625,
+      "step": 4762
+    },
+    {
+      "epoch": 2.494499738082766,
+      "loss_partial": 0.05908203125,
+      "step": 4762
+    },
+    {
+      "epoch": 2.495023572551074,
+      "grad_norm": 1.78168523311615,
+      "learning_rate": 7.4589421208340684e-06,
+      "loss": 0.1033,
+      "step": 4763
+    },
+    {
+      "epoch": 2.495023572551074,
+      "loss_partial": 0.049560546875,
+      "step": 4763
+    },
+    {
+      "epoch": 2.495023572551074,
+      "loss_partial": 0.050048828125,
+      "step": 4763
+    },
+    {
+      "epoch": 2.495547407019382,
+      "grad_norm": 1.9074164628982544,
+      "learning_rate": 7.4545208580426934e-06,
+      "loss": 0.0995,
+      "step": 4764
+    },
+    {
+      "epoch": 2.495547407019382,
+      "loss_partial": 0.048583984375,
+      "step": 4764
+    },
+    {
+      "epoch": 2.495547407019382,
+      "loss_partial": 0.048095703125,
+      "step": 4764
+    },
+    {
+      "epoch": 2.49607124148769,
+      "grad_norm": 1.955350399017334,
+      "learning_rate": 7.450100127240713e-06,
+      "loss": 0.1028,
+      "step": 4765
+    },
+    {
+      "epoch": 2.49607124148769,
+      "loss_partial": 0.057861328125,
+      "step": 4765
+    },
+    {
+      "epoch": 2.49607124148769,
+      "loss_partial": 0.050537109375,
+      "step": 4765
+    },
+    {
+      "epoch": 2.496595075955998,
+      "grad_norm": 1.5060220956802368,
+      "learning_rate": 7.445679929352043e-06,
+      "loss": 0.1055,
+      "step": 4766
+    },
+    {
+      "epoch": 2.496595075955998,
+      "loss_partial": 0.052490234375,
+      "step": 4766
+    },
+    {
+      "epoch": 2.496595075955998,
+      "loss_partial": 0.0400390625,
+      "step": 4766
+    },
+    {
+      "epoch": 2.497118910424306,
+      "grad_norm": 1.6751863956451416,
+      "learning_rate": 7.441260265300474e-06,
+      "loss": 0.0971,
+      "step": 4767
+    },
+    {
+      "epoch": 2.497118910424306,
+      "loss_partial": 0.043701171875,
+      "step": 4767
+    },
+    {
+      "epoch": 2.497118910424306,
+      "loss_partial": 0.043701171875,
+      "step": 4767
+    },
+    {
+      "epoch": 2.497642744892614,
+      "grad_norm": 1.0349607467651367,
+      "learning_rate": 7.436841136009685e-06,
+      "loss": 0.1033,
+      "step": 4768
+    },
+    {
+      "epoch": 2.497642744892614,
+      "loss_partial": 0.04345703125,
+      "step": 4768
+    },
+    {
+      "epoch": 2.497642744892614,
+      "loss_partial": 0.04833984375,
+      "step": 4768
+    },
+    {
+      "epoch": 2.498166579360922,
+      "grad_norm": 2.1584670543670654,
+      "learning_rate": 7.432422542403246e-06,
+      "loss": 0.0999,
+      "step": 4769
+    },
+    {
+      "epoch": 2.498166579360922,
+      "loss_partial": 0.060546875,
+      "step": 4769
+    },
+    {
+      "epoch": 2.498166579360922,
+      "loss_partial": 0.047119140625,
+      "step": 4769
+    },
+    {
+      "epoch": 2.49869041382923,
+      "grad_norm": 1.3394736051559448,
+      "learning_rate": 7.428004485404621e-06,
+      "loss": 0.1017,
+      "step": 4770
+    },
+    {
+      "epoch": 2.49869041382923,
+      "loss_partial": 0.05078125,
+      "step": 4770
+    },
+    {
+      "epoch": 2.49869041382923,
+      "loss_partial": 0.04248046875,
+      "step": 4770
+    },
+    {
+      "epoch": 2.499214248297538,
+      "grad_norm": 2.0586140155792236,
+      "learning_rate": 7.423586965937154e-06,
+      "loss": 0.1017,
+      "step": 4771
+    },
+    {
+      "epoch": 2.499214248297538,
+      "loss_partial": 0.0478515625,
+      "step": 4771
+    },
+    {
+      "epoch": 2.499214248297538,
+      "loss_partial": 0.0517578125,
+      "step": 4771
+    },
+    {
+      "epoch": 2.499738082765846,
+      "grad_norm": 1.510896921157837,
+      "learning_rate": 7.4191699849240795e-06,
+      "loss": 0.0988,
+      "step": 4772
+    },
+    {
+      "epoch": 2.499738082765846,
+      "loss_partial": 0.05810546875,
+      "step": 4772
+    },
+    {
+      "epoch": 2.499738082765846,
+      "loss_partial": 0.05126953125,
+      "step": 4772
+    },
+    {
+      "epoch": 2.500261917234154,
+      "grad_norm": 1.9352073669433594,
+      "learning_rate": 7.414753543288513e-06,
+      "loss": 0.1035,
+      "step": 4773
+    },
+    {
+      "epoch": 2.500261917234154,
+      "loss_partial": 0.044189453125,
+      "step": 4773
+    },
+    {
+      "epoch": 2.500261917234154,
+      "loss_partial": 0.053466796875,
+      "step": 4773
+    },
+    {
+      "epoch": 2.500785751702462,
+      "grad_norm": 1.5037192106246948,
+      "learning_rate": 7.410337641953474e-06,
+      "loss": 0.1009,
+      "step": 4774
+    },
+    {
+      "epoch": 2.500785751702462,
+      "loss_partial": 0.049072265625,
+      "step": 4774
+    },
+    {
+      "epoch": 2.500785751702462,
+      "loss_partial": 0.0546875,
+      "step": 4774
+    },
+    {
+      "epoch": 2.50130958617077,
+      "grad_norm": 1.5510549545288086,
+      "learning_rate": 7.405922281841857e-06,
+      "loss": 0.1056,
+      "step": 4775
+    },
+    {
+      "epoch": 2.50130958617077,
+      "loss_partial": 0.055419921875,
+      "step": 4775
+    },
+    {
+      "epoch": 2.50130958617077,
+      "loss_partial": 0.050048828125,
+      "step": 4775
+    },
+    {
+      "epoch": 2.501833420639078,
+      "grad_norm": 1.8441609144210815,
+      "learning_rate": 7.401507463876438e-06,
+      "loss": 0.1022,
+      "step": 4776
+    },
+    {
+      "epoch": 2.501833420639078,
+      "loss_partial": 0.051513671875,
+      "step": 4776
+    },
+    {
+      "epoch": 2.501833420639078,
+      "loss_partial": 0.056640625,
+      "step": 4776
+    },
+    {
+      "epoch": 2.502357255107386,
+      "grad_norm": 1.5537289381027222,
+      "learning_rate": 7.397093188979893e-06,
+      "loss": 0.1017,
+      "step": 4777
+    },
+    {
+      "epoch": 2.502357255107386,
+      "loss_partial": 0.051513671875,
+      "step": 4777
+    },
+    {
+      "epoch": 2.502357255107386,
+      "loss_partial": 0.0654296875,
+      "step": 4777
+    },
+    {
+      "epoch": 2.502881089575694,
+      "grad_norm": 1.2946507930755615,
+      "learning_rate": 7.3926794580747695e-06,
+      "loss": 0.1024,
+      "step": 4778
+    },
+    {
+      "epoch": 2.502881089575694,
+      "loss_partial": 0.03759765625,
+      "step": 4778
+    },
+    {
+      "epoch": 2.502881089575694,
+      "loss_partial": 0.052734375,
+      "step": 4778
+    },
+    {
+      "epoch": 2.503404924044002,
+      "grad_norm": 1.2745630741119385,
+      "learning_rate": 7.3882662720835205e-06,
+      "loss": 0.0989,
+      "step": 4779
+    },
+    {
+      "epoch": 2.503404924044002,
+      "loss_partial": 0.0498046875,
+      "step": 4779
+    },
+    {
+      "epoch": 2.503404924044002,
+      "loss_partial": 0.05615234375,
+      "step": 4779
+    },
+    {
+      "epoch": 2.50392875851231,
+      "grad_norm": 1.6175633668899536,
+      "learning_rate": 7.383853631928471e-06,
+      "loss": 0.1014,
+      "step": 4780
+    },
+    {
+      "epoch": 2.50392875851231,
+      "loss_partial": 0.044189453125,
+      "step": 4780
+    },
+    {
+      "epoch": 2.50392875851231,
+      "loss_partial": 0.053466796875,
+      "step": 4780
+    },
+    {
+      "epoch": 2.504452592980618,
+      "grad_norm": 1.507430076599121,
+      "learning_rate": 7.379441538531833e-06,
+      "loss": 0.0999,
+      "step": 4781
+    },
+    {
+      "epoch": 2.504452592980618,
+      "loss_partial": 0.053955078125,
+      "step": 4781
+    },
+    {
+      "epoch": 2.504452592980618,
+      "loss_partial": 0.046142578125,
+      "step": 4781
+    },
+    {
+      "epoch": 2.504976427448926,
+      "grad_norm": 1.1596204042434692,
+      "learning_rate": 7.375029992815708e-06,
+      "loss": 0.1043,
+      "step": 4782
+    },
+    {
+      "epoch": 2.504976427448926,
+      "loss_partial": 0.047607421875,
+      "step": 4782
+    },
+    {
+      "epoch": 2.504976427448926,
+      "loss_partial": 0.05224609375,
+      "step": 4782
+    },
+    {
+      "epoch": 2.505500261917234,
+      "grad_norm": 1.299439787864685,
+      "learning_rate": 7.370618995702083e-06,
+      "loss": 0.102,
+      "step": 4783
+    },
+    {
+      "epoch": 2.505500261917234,
+      "loss_partial": 0.043701171875,
+      "step": 4783
+    },
+    {
+      "epoch": 2.505500261917234,
+      "loss_partial": 0.052734375,
+      "step": 4783
+    },
+    {
+      "epoch": 2.5060240963855422,
+      "grad_norm": 1.5215611457824707,
+      "learning_rate": 7.36620854811283e-06,
+      "loss": 0.1017,
+      "step": 4784
+    },
+    {
+      "epoch": 2.5060240963855422,
+      "loss_partial": 0.059814453125,
+      "step": 4784
+    },
+    {
+      "epoch": 2.5060240963855422,
+      "loss_partial": 0.05078125,
+      "step": 4784
+    },
+    {
+      "epoch": 2.5065479308538503,
+      "grad_norm": 1.4207028150558472,
+      "learning_rate": 7.3617986509697015e-06,
+      "loss": 0.1006,
+      "step": 4785
+    },
+    {
+      "epoch": 2.5065479308538503,
+      "loss_partial": 0.047119140625,
+      "step": 4785
+    },
+    {
+      "epoch": 2.5065479308538503,
+      "loss_partial": 0.046630859375,
+      "step": 4785
+    },
+    {
+      "epoch": 2.5070717653221584,
+      "grad_norm": 1.2370694875717163,
+      "learning_rate": 7.357389305194338e-06,
+      "loss": 0.0971,
+      "step": 4786
+    },
+    {
+      "epoch": 2.5070717653221584,
+      "loss_partial": 0.05908203125,
+      "step": 4786
+    },
+    {
+      "epoch": 2.5070717653221584,
+      "loss_partial": 0.05322265625,
+      "step": 4786
+    },
+    {
+      "epoch": 2.5075955997904664,
+      "grad_norm": 1.7842462062835693,
+      "learning_rate": 7.352980511708274e-06,
+      "loss": 0.101,
+      "step": 4787
+    },
+    {
+      "epoch": 2.5075955997904664,
+      "loss_partial": 0.064453125,
+      "step": 4787
+    },
+    {
+      "epoch": 2.5075955997904664,
+      "loss_partial": 0.05078125,
+      "step": 4787
+    },
+    {
+      "epoch": 2.5081194342587745,
+      "grad_norm": 1.603628158569336,
+      "learning_rate": 7.348572271432912e-06,
+      "loss": 0.1021,
+      "step": 4788
+    },
+    {
+      "epoch": 2.5081194342587745,
+      "loss_partial": 0.046875,
+      "step": 4788
+    },
+    {
+      "epoch": 2.5081194342587745,
+      "loss_partial": 0.04248046875,
+      "step": 4788
+    },
+    {
+      "epoch": 2.508643268727082,
+      "grad_norm": 1.5760470628738403,
+      "learning_rate": 7.34416458528955e-06,
+      "loss": 0.0992,
+      "step": 4789
+    },
+    {
+      "epoch": 2.508643268727082,
+      "loss_partial": 0.0498046875,
+      "step": 4789
+    },
+    {
+      "epoch": 2.508643268727082,
+      "loss_partial": 0.0517578125,
+      "step": 4789
+    },
+    {
+      "epoch": 2.50916710319539,
+      "grad_norm": 1.2950875759124756,
+      "learning_rate": 7.339757454199361e-06,
+      "loss": 0.1026,
+      "step": 4790
+    },
+    {
+      "epoch": 2.50916710319539,
+      "loss_partial": 0.04541015625,
+      "step": 4790
+    },
+    {
+      "epoch": 2.50916710319539,
+      "loss_partial": 0.05615234375,
+      "step": 4790
+    },
+    {
+      "epoch": 2.509690937663698,
+      "grad_norm": 2.008065700531006,
+      "learning_rate": 7.3353508790834195e-06,
+      "loss": 0.101,
+      "step": 4791
+    },
+    {
+      "epoch": 2.509690937663698,
+      "loss_partial": 0.04833984375,
+      "step": 4791
+    },
+    {
+      "epoch": 2.509690937663698,
+      "loss_partial": 0.044677734375,
+      "step": 4791
+    },
+    {
+      "epoch": 2.5102147721320063,
+      "grad_norm": 1.56052827835083,
+      "learning_rate": 7.330944860862665e-06,
+      "loss": 0.1006,
+      "step": 4792
+    },
+    {
+      "epoch": 2.5102147721320063,
+      "loss_partial": 0.044677734375,
+      "step": 4792
+    },
+    {
+      "epoch": 2.5102147721320063,
+      "loss_partial": 0.046875,
+      "step": 4792
+    },
+    {
+      "epoch": 2.5107386066003143,
+      "grad_norm": 1.9029736518859863,
+      "learning_rate": 7.326539400457931e-06,
+      "loss": 0.1022,
+      "step": 4793
+    },
+    {
+      "epoch": 2.5107386066003143,
+      "loss_partial": 0.046630859375,
+      "step": 4793
+    },
+    {
+      "epoch": 2.5107386066003143,
+      "loss_partial": 0.05322265625,
+      "step": 4793
+    },
+    {
+      "epoch": 2.5112624410686224,
+      "grad_norm": 1.6628915071487427,
+      "learning_rate": 7.322134498789928e-06,
+      "loss": 0.1001,
+      "step": 4794
+    },
+    {
+      "epoch": 2.5112624410686224,
+      "loss_partial": 0.048828125,
+      "step": 4794
+    },
+    {
+      "epoch": 2.5112624410686224,
+      "loss_partial": 0.051513671875,
+      "step": 4794
+    },
+    {
+      "epoch": 2.5117862755369305,
+      "grad_norm": 1.417096495628357,
+      "learning_rate": 7.317730156779257e-06,
+      "loss": 0.1016,
+      "step": 4795
+    },
+    {
+      "epoch": 2.5117862755369305,
+      "loss_partial": 0.047607421875,
+      "step": 4795
+    },
+    {
+      "epoch": 2.5117862755369305,
+      "loss_partial": 0.056884765625,
+      "step": 4795
+    },
+    {
+      "epoch": 2.512310110005238,
+      "grad_norm": 1.6210815906524658,
+      "learning_rate": 7.313326375346397e-06,
+      "loss": 0.0985,
+      "step": 4796
+    },
+    {
+      "epoch": 2.512310110005238,
+      "loss_partial": 0.045654296875,
+      "step": 4796
+    },
+    {
+      "epoch": 2.512310110005238,
+      "loss_partial": 0.04052734375,
+      "step": 4796
+    },
+    {
+      "epoch": 2.512833944473546,
+      "grad_norm": 1.3902881145477295,
+      "learning_rate": 7.30892315541171e-06,
+      "loss": 0.0992,
+      "step": 4797
+    },
+    {
+      "epoch": 2.512833944473546,
+      "loss_partial": 0.047119140625,
+      "step": 4797
+    },
+    {
+      "epoch": 2.512833944473546,
+      "loss_partial": 0.05517578125,
+      "step": 4797
+    },
+    {
+      "epoch": 2.513357778941854,
+      "grad_norm": 1.5869286060333252,
+      "learning_rate": 7.304520497895439e-06,
+      "loss": 0.1011,
+      "step": 4798
+    },
+    {
+      "epoch": 2.513357778941854,
+      "loss_partial": 0.039794921875,
+      "step": 4798
+    },
+    {
+      "epoch": 2.513357778941854,
+      "loss_partial": 0.054931640625,
+      "step": 4798
+    },
+    {
+      "epoch": 2.5138816134101623,
+      "grad_norm": 1.5163664817810059,
+      "learning_rate": 7.300118403717721e-06,
+      "loss": 0.1003,
+      "step": 4799
+    },
+    {
+      "epoch": 2.5138816134101623,
+      "loss_partial": 0.048828125,
+      "step": 4799
+    },
+    {
+      "epoch": 2.5138816134101623,
+      "loss_partial": 0.046142578125,
+      "step": 4799
+    },
+    {
+      "epoch": 2.5144054478784703,
+      "grad_norm": 1.5043227672576904,
+      "learning_rate": 7.29571687379856e-06,
+      "loss": 0.1021,
+      "step": 4800
+    },
+    {
+      "epoch": 2.5144054478784703,
+      "loss_partial": 0.051025390625,
+      "step": 4800
+    },
+    {
+      "epoch": 2.5144054478784703,
+      "loss_partial": 0.0458984375,
+      "step": 4800
+    },
+    {
+      "epoch": 2.5149292823467784,
+      "grad_norm": 1.4443175792694092,
+      "learning_rate": 7.291315909057848e-06,
+      "loss": 0.0987,
+      "step": 4801
+    },
+    {
+      "epoch": 2.5149292823467784,
+      "loss_partial": 0.047607421875,
+      "step": 4801
+    },
+    {
+      "epoch": 2.5149292823467784,
+      "loss_partial": 0.040771484375,
+      "step": 4801
+    },
+    {
+      "epoch": 2.5154531168150864,
+      "grad_norm": 1.099213719367981,
+      "learning_rate": 7.286915510415358e-06,
+      "loss": 0.1006,
+      "step": 4802
+    },
+    {
+      "epoch": 2.5154531168150864,
+      "loss_partial": 0.051025390625,
+      "step": 4802
+    },
+    {
+      "epoch": 2.5154531168150864,
+      "loss_partial": 0.058349609375,
+      "step": 4802
+    },
+    {
+      "epoch": 2.5159769512833945,
+      "grad_norm": 1.3675751686096191,
+      "learning_rate": 7.282515678790752e-06,
+      "loss": 0.1035,
+      "step": 4803
+    },
+    {
+      "epoch": 2.5159769512833945,
+      "loss_partial": 0.057373046875,
+      "step": 4803
+    },
+    {
+      "epoch": 2.5159769512833945,
+      "loss_partial": 0.050048828125,
+      "step": 4803
+    },
+    {
+      "epoch": 2.5165007857517026,
+      "grad_norm": 1.0590718984603882,
+      "learning_rate": 7.2781164151035665e-06,
+      "loss": 0.1008,
+      "step": 4804
+    },
+    {
+      "epoch": 2.5165007857517026,
+      "loss_partial": 0.05322265625,
+      "step": 4804
+    },
+    {
+      "epoch": 2.5165007857517026,
+      "loss_partial": 0.046875,
+      "step": 4804
+    },
+    {
+      "epoch": 2.5170246202200106,
+      "grad_norm": 1.6682454347610474,
+      "learning_rate": 7.2737177202732175e-06,
+      "loss": 0.1025,
+      "step": 4805
+    },
+    {
+      "epoch": 2.5170246202200106,
+      "loss_partial": 0.05859375,
+      "step": 4805
+    },
+    {
+      "epoch": 2.5170246202200106,
+      "loss_partial": 0.048828125,
+      "step": 4805
+    },
+    {
+      "epoch": 2.5175484546883187,
+      "grad_norm": 1.840259075164795,
+      "learning_rate": 7.269319595219005e-06,
+      "loss": 0.1038,
+      "step": 4806
+    },
+    {
+      "epoch": 2.5175484546883187,
+      "loss_partial": 0.05712890625,
+      "step": 4806
+    },
+    {
+      "epoch": 2.5175484546883187,
+      "loss_partial": 0.052978515625,
+      "step": 4806
+    },
+    {
+      "epoch": 2.5180722891566267,
+      "grad_norm": 1.6835196018218994,
+      "learning_rate": 7.264922040860112e-06,
+      "loss": 0.1039,
+      "step": 4807
+    },
+    {
+      "epoch": 2.5180722891566267,
+      "loss_partial": 0.041015625,
+      "step": 4807
+    },
+    {
+      "epoch": 2.5180722891566267,
+      "loss_partial": 0.04638671875,
+      "step": 4807
+    },
+    {
+      "epoch": 2.518596123624935,
+      "grad_norm": 1.409220814704895,
+      "learning_rate": 7.260525058115599e-06,
+      "loss": 0.1052,
+      "step": 4808
+    },
+    {
+      "epoch": 2.518596123624935,
+      "loss_partial": 0.059814453125,
+      "step": 4808
+    },
+    {
+      "epoch": 2.518596123624935,
+      "loss_partial": 0.050048828125,
+      "step": 4808
+    },
+    {
+      "epoch": 2.5191199580932424,
+      "grad_norm": 2.118136167526245,
+      "learning_rate": 7.25612864790441e-06,
+      "loss": 0.1043,
+      "step": 4809
+    },
+    {
+      "epoch": 2.5191199580932424,
+      "loss_partial": 0.04931640625,
+      "step": 4809
+    },
+    {
+      "epoch": 2.5191199580932424,
+      "loss_partial": 0.05908203125,
+      "step": 4809
+    },
+    {
+      "epoch": 2.5196437925615505,
+      "grad_norm": 1.3708122968673706,
+      "learning_rate": 7.251732811145362e-06,
+      "loss": 0.1002,
+      "step": 4810
+    },
+    {
+      "epoch": 2.5196437925615505,
+      "loss_partial": 0.047119140625,
+      "step": 4810
+    },
+    {
+      "epoch": 2.5196437925615505,
+      "loss_partial": 0.05810546875,
+      "step": 4810
+    },
+    {
+      "epoch": 2.5201676270298585,
+      "grad_norm": 1.3294965028762817,
+      "learning_rate": 7.247337548757165e-06,
+      "loss": 0.0992,
+      "step": 4811
+    },
+    {
+      "epoch": 2.5201676270298585,
+      "loss_partial": 0.0546875,
+      "step": 4811
+    },
+    {
+      "epoch": 2.5201676270298585,
+      "loss_partial": 0.059326171875,
+      "step": 4811
+    },
+    {
+      "epoch": 2.5206914614981666,
+      "grad_norm": 1.6535893678665161,
+      "learning_rate": 7.2429428616584e-06,
+      "loss": 0.1028,
+      "step": 4812
+    },
+    {
+      "epoch": 2.5206914614981666,
+      "loss_partial": 0.0478515625,
+      "step": 4812
+    },
+    {
+      "epoch": 2.5206914614981666,
+      "loss_partial": 0.047119140625,
+      "step": 4812
+    },
+    {
+      "epoch": 2.5212152959664746,
+      "grad_norm": 1.2870384454727173,
+      "learning_rate": 7.2385487507675265e-06,
+      "loss": 0.0968,
+      "step": 4813
+    },
+    {
+      "epoch": 2.5212152959664746,
+      "loss_partial": 0.057373046875,
+      "step": 4813
+    },
+    {
+      "epoch": 2.5212152959664746,
+      "loss_partial": 0.04248046875,
+      "step": 4813
+    },
+    {
+      "epoch": 2.5217391304347827,
+      "grad_norm": 1.522687554359436,
+      "learning_rate": 7.234155217002883e-06,
+      "loss": 0.1046,
+      "step": 4814
+    },
+    {
+      "epoch": 2.5217391304347827,
+      "loss_partial": 0.057373046875,
+      "step": 4814
+    },
+    {
+      "epoch": 2.5217391304347827,
+      "loss_partial": 0.048828125,
+      "step": 4814
+    },
+    {
+      "epoch": 2.5222629649030908,
+      "grad_norm": 1.1756314039230347,
+      "learning_rate": 7.2297622612827e-06,
+      "loss": 0.1022,
+      "step": 4815
+    },
+    {
+      "epoch": 2.5222629649030908,
+      "loss_partial": 0.052734375,
+      "step": 4815
+    },
+    {
+      "epoch": 2.5222629649030908,
+      "loss_partial": 0.0654296875,
+      "step": 4815
+    },
+    {
+      "epoch": 2.5227867993713984,
+      "grad_norm": 1.255180835723877,
+      "learning_rate": 7.225369884525073e-06,
+      "loss": 0.1014,
+      "step": 4816
+    },
+    {
+      "epoch": 2.5227867993713984,
+      "loss_partial": 0.055419921875,
+      "step": 4816
+    },
+    {
+      "epoch": 2.5227867993713984,
+      "loss_partial": 0.051025390625,
+      "step": 4816
+    },
+    {
+      "epoch": 2.5233106338397064,
+      "grad_norm": 1.445652723312378,
+      "learning_rate": 7.220978087647984e-06,
+      "loss": 0.1008,
+      "step": 4817
+    },
+    {
+      "epoch": 2.5233106338397064,
+      "loss_partial": 0.04541015625,
+      "step": 4817
+    },
+    {
+      "epoch": 2.5233106338397064,
+      "loss_partial": 0.04541015625,
+      "step": 4817
+    },
+    {
+      "epoch": 2.5238344683080145,
+      "grad_norm": 1.1381361484527588,
+      "learning_rate": 7.216586871569287e-06,
+      "loss": 0.1,
+      "step": 4818
+    },
+    {
+      "epoch": 2.5238344683080145,
+      "loss_partial": 0.0458984375,
+      "step": 4818
+    },
+    {
+      "epoch": 2.5238344683080145,
+      "loss_partial": 0.055908203125,
+      "step": 4818
+    },
+    {
+      "epoch": 2.5243583027763226,
+      "grad_norm": 1.6489479541778564,
+      "learning_rate": 7.2121962372067234e-06,
+      "loss": 0.103,
+      "step": 4819
+    },
+    {
+      "epoch": 2.5243583027763226,
+      "loss_partial": 0.050048828125,
+      "step": 4819
+    },
+    {
+      "epoch": 2.5243583027763226,
+      "loss_partial": 0.05126953125,
+      "step": 4819
+    },
+    {
+      "epoch": 2.5248821372446306,
+      "grad_norm": 1.5993529558181763,
+      "learning_rate": 7.207806185477908e-06,
+      "loss": 0.1028,
+      "step": 4820
+    },
+    {
+      "epoch": 2.5248821372446306,
+      "loss_partial": 0.041259765625,
+      "step": 4820
+    },
+    {
+      "epoch": 2.5248821372446306,
+      "loss_partial": 0.0458984375,
+      "step": 4820
+    },
+    {
+      "epoch": 2.5254059717129387,
+      "grad_norm": 1.2045632600784302,
+      "learning_rate": 7.203416717300331e-06,
+      "loss": 0.1041,
+      "step": 4821
+    },
+    {
+      "epoch": 2.5254059717129387,
+      "loss_partial": 0.05224609375,
+      "step": 4821
+    },
+    {
+      "epoch": 2.5254059717129387,
+      "loss_partial": 0.061279296875,
+      "step": 4821
+    },
+    {
+      "epoch": 2.5259298061812467,
+      "grad_norm": 1.8398629426956177,
+      "learning_rate": 7.199027833591366e-06,
+      "loss": 0.1026,
+      "step": 4822
+    },
+    {
+      "epoch": 2.5259298061812467,
+      "loss_partial": 0.060546875,
+      "step": 4822
+    },
+    {
+      "epoch": 2.5259298061812467,
+      "loss_partial": 0.050537109375,
+      "step": 4822
+    },
+    {
+      "epoch": 2.526453640649555,
+      "grad_norm": 1.8597729206085205,
+      "learning_rate": 7.194639535268264e-06,
+      "loss": 0.1006,
+      "step": 4823
+    },
+    {
+      "epoch": 2.526453640649555,
+      "loss_partial": 0.0458984375,
+      "step": 4823
+    },
+    {
+      "epoch": 2.526453640649555,
+      "loss_partial": 0.051025390625,
+      "step": 4823
+    },
+    {
+      "epoch": 2.526977475117863,
+      "grad_norm": 1.1964813470840454,
+      "learning_rate": 7.190251823248156e-06,
+      "loss": 0.0964,
+      "step": 4824
+    },
+    {
+      "epoch": 2.526977475117863,
+      "loss_partial": 0.046630859375,
+      "step": 4824
+    },
+    {
+      "epoch": 2.526977475117863,
+      "loss_partial": 0.0673828125,
+      "step": 4824
+    },
+    {
+      "epoch": 2.527501309586171,
+      "grad_norm": 1.5703444480895996,
+      "learning_rate": 7.185864698448037e-06,
+      "loss": 0.1013,
+      "step": 4825
+    },
+    {
+      "epoch": 2.527501309586171,
+      "loss_partial": 0.04150390625,
+      "step": 4825
+    },
+    {
+      "epoch": 2.527501309586171,
+      "loss_partial": 0.034912109375,
+      "step": 4825
+    },
+    {
+      "epoch": 2.528025144054479,
+      "grad_norm": 1.4212373495101929,
+      "learning_rate": 7.181478161784792e-06,
+      "loss": 0.1032,
+      "step": 4826
+    },
+    {
+      "epoch": 2.528025144054479,
+      "loss_partial": 0.052734375,
+      "step": 4826
+    },
+    {
+      "epoch": 2.528025144054479,
+      "loss_partial": 0.04638671875,
+      "step": 4826
+    },
+    {
+      "epoch": 2.528548978522787,
+      "grad_norm": 1.4379981756210327,
+      "learning_rate": 7.177092214175185e-06,
+      "loss": 0.1024,
+      "step": 4827
+    },
+    {
+      "epoch": 2.528548978522787,
+      "loss_partial": 0.05517578125,
+      "step": 4827
+    },
+    {
+      "epoch": 2.528548978522787,
+      "loss_partial": 0.051025390625,
+      "step": 4827
+    },
+    {
+      "epoch": 2.5290728129910947,
+      "grad_norm": 1.4079575538635254,
+      "learning_rate": 7.172706856535849e-06,
+      "loss": 0.1009,
+      "step": 4828
+    },
+    {
+      "epoch": 2.5290728129910947,
+      "loss_partial": 0.0517578125,
+      "step": 4828
+    },
+    {
+      "epoch": 2.5290728129910947,
+      "loss_partial": 0.047607421875,
+      "step": 4828
+    },
+    {
+      "epoch": 2.5295966474594027,
+      "grad_norm": 1.3896939754486084,
+      "learning_rate": 7.1683220897832976e-06,
+      "loss": 0.1012,
+      "step": 4829
+    },
+    {
+      "epoch": 2.5295966474594027,
+      "loss_partial": 0.0576171875,
+      "step": 4829
+    },
+    {
+      "epoch": 2.5295966474594027,
+      "loss_partial": 0.05078125,
+      "step": 4829
+    },
+    {
+      "epoch": 2.5301204819277108,
+      "grad_norm": 1.1734769344329834,
+      "learning_rate": 7.163937914833916e-06,
+      "loss": 0.1011,
+      "step": 4830
+    },
+    {
+      "epoch": 2.5301204819277108,
+      "loss_partial": 0.05419921875,
+      "step": 4830
+    },
+    {
+      "epoch": 2.5301204819277108,
+      "loss_partial": 0.048828125,
+      "step": 4830
+    },
+    {
+      "epoch": 2.530644316396019,
+      "grad_norm": 1.0345715284347534,
+      "learning_rate": 7.159554332603974e-06,
+      "loss": 0.098,
+      "step": 4831
+    },
+    {
+      "epoch": 2.530644316396019,
+      "loss_partial": 0.044189453125,
+      "step": 4831
+    },
+    {
+      "epoch": 2.530644316396019,
+      "loss_partial": 0.048828125,
+      "step": 4831
+    },
+    {
+      "epoch": 2.531168150864327,
+      "grad_norm": 1.1606229543685913,
+      "learning_rate": 7.155171344009613e-06,
+      "loss": 0.1013,
+      "step": 4832
+    },
+    {
+      "epoch": 2.531168150864327,
+      "loss_partial": 0.047607421875,
+      "step": 4832
+    },
+    {
+      "epoch": 2.531168150864327,
+      "loss_partial": 0.04833984375,
+      "step": 4832
+    },
+    {
+      "epoch": 2.531691985332635,
+      "grad_norm": 1.2761176824569702,
+      "learning_rate": 7.150788949966848e-06,
+      "loss": 0.0994,
+      "step": 4833
+    },
+    {
+      "epoch": 2.531691985332635,
+      "loss_partial": 0.05517578125,
+      "step": 4833
+    },
+    {
+      "epoch": 2.531691985332635,
+      "loss_partial": 0.0458984375,
+      "step": 4833
+    },
+    {
+      "epoch": 2.532215819800943,
+      "grad_norm": 1.25601327419281,
+      "learning_rate": 7.146407151391574e-06,
+      "loss": 0.1027,
+      "step": 4834
+    },
+    {
+      "epoch": 2.532215819800943,
+      "loss_partial": 0.05224609375,
+      "step": 4834
+    },
+    {
+      "epoch": 2.532215819800943,
+      "loss_partial": 0.056396484375,
+      "step": 4834
+    },
+    {
+      "epoch": 2.532739654269251,
+      "grad_norm": 1.2133889198303223,
+      "learning_rate": 7.1420259491995554e-06,
+      "loss": 0.1022,
+      "step": 4835
+    },
+    {
+      "epoch": 2.532739654269251,
+      "loss_partial": 0.044189453125,
+      "step": 4835
+    },
+    {
+      "epoch": 2.532739654269251,
+      "loss_partial": 0.0419921875,
+      "step": 4835
+    },
+    {
+      "epoch": 2.5332634887375587,
+      "grad_norm": 1.6216964721679688,
+      "learning_rate": 7.137645344306447e-06,
+      "loss": 0.101,
+      "step": 4836
+    },
+    {
+      "epoch": 2.5332634887375587,
+      "loss_partial": 0.05078125,
+      "step": 4836
+    },
+    {
+      "epoch": 2.5332634887375587,
+      "loss_partial": 0.052001953125,
+      "step": 4836
+    },
+    {
+      "epoch": 2.5337873232058667,
+      "grad_norm": 1.9141440391540527,
+      "learning_rate": 7.133265337627757e-06,
+      "loss": 0.0976,
+      "step": 4837
+    },
+    {
+      "epoch": 2.5337873232058667,
+      "loss_partial": 0.04541015625,
+      "step": 4837
+    },
+    {
+      "epoch": 2.5337873232058667,
+      "loss_partial": 0.05029296875,
+      "step": 4837
+    },
+    {
+      "epoch": 2.534311157674175,
+      "grad_norm": 1.3036842346191406,
+      "learning_rate": 7.128885930078886e-06,
+      "loss": 0.098,
+      "step": 4838
+    },
+    {
+      "epoch": 2.534311157674175,
+      "loss_partial": 0.044677734375,
+      "step": 4838
+    },
+    {
+      "epoch": 2.534311157674175,
+      "loss_partial": 0.052490234375,
+      "step": 4838
+    },
+    {
+      "epoch": 2.534834992142483,
+      "grad_norm": 1.9187090396881104,
+      "learning_rate": 7.124507122575098e-06,
+      "loss": 0.1028,
+      "step": 4839
+    },
+    {
+      "epoch": 2.534834992142483,
+      "loss_partial": 0.058837890625,
+      "step": 4839
+    },
+    {
+      "epoch": 2.534834992142483,
+      "loss_partial": 0.052734375,
+      "step": 4839
+    },
+    {
+      "epoch": 2.535358826610791,
+      "grad_norm": 1.2700978517532349,
+      "learning_rate": 7.120128916031541e-06,
+      "loss": 0.1062,
+      "step": 4840
+    },
+    {
+      "epoch": 2.535358826610791,
+      "loss_partial": 0.057861328125,
+      "step": 4840
+    },
+    {
+      "epoch": 2.535358826610791,
+      "loss_partial": 0.060791015625,
+      "step": 4840
+    },
+    {
+      "epoch": 2.535882661079099,
+      "grad_norm": 2.1320114135742188,
+      "learning_rate": 7.115751311363235e-06,
+      "loss": 0.103,
+      "step": 4841
+    },
+    {
+      "epoch": 2.535882661079099,
+      "loss_partial": 0.04638671875,
+      "step": 4841
+    },
+    {
+      "epoch": 2.535882661079099,
+      "loss_partial": 0.04150390625,
+      "step": 4841
+    },
+    {
+      "epoch": 2.536406495547407,
+      "grad_norm": 1.6576061248779297,
+      "learning_rate": 7.111374309485067e-06,
+      "loss": 0.1022,
+      "step": 4842
+    },
+    {
+      "epoch": 2.536406495547407,
+      "loss_partial": 0.05126953125,
+      "step": 4842
+    },
+    {
+      "epoch": 2.536406495547407,
+      "loss_partial": 0.054931640625,
+      "step": 4842
+    },
+    {
+      "epoch": 2.536930330015715,
+      "grad_norm": 1.9383962154388428,
+      "learning_rate": 7.1069979113118035e-06,
+      "loss": 0.1003,
+      "step": 4843
+    },
+    {
+      "epoch": 2.536930330015715,
+      "loss_partial": 0.04931640625,
+      "step": 4843
+    },
+    {
+      "epoch": 2.536930330015715,
+      "loss_partial": 0.0546875,
+      "step": 4843
+    },
+    {
+      "epoch": 2.537454164484023,
+      "grad_norm": 1.8256489038467407,
+      "learning_rate": 7.102622117758091e-06,
+      "loss": 0.1078,
+      "step": 4844
+    },
+    {
+      "epoch": 2.537454164484023,
+      "loss_partial": 0.04443359375,
+      "step": 4844
+    },
+    {
+      "epoch": 2.537454164484023,
+      "loss_partial": 0.04638671875,
+      "step": 4844
+    },
+    {
+      "epoch": 2.5379779989523312,
+      "grad_norm": 1.5162129402160645,
+      "learning_rate": 7.098246929738437e-06,
+      "loss": 0.1013,
+      "step": 4845
+    },
+    {
+      "epoch": 2.5379779989523312,
+      "loss_partial": 0.052734375,
+      "step": 4845
+    },
+    {
+      "epoch": 2.5379779989523312,
+      "loss_partial": 0.048583984375,
+      "step": 4845
+    },
+    {
+      "epoch": 2.5385018334206393,
+      "grad_norm": 1.1779576539993286,
+      "learning_rate": 7.093872348167232e-06,
+      "loss": 0.1,
+      "step": 4846
+    },
+    {
+      "epoch": 2.5385018334206393,
+      "loss_partial": 0.0517578125,
+      "step": 4846
+    },
+    {
+      "epoch": 2.5385018334206393,
+      "loss_partial": 0.052001953125,
+      "step": 4846
+    },
+    {
+      "epoch": 2.5390256678889473,
+      "grad_norm": 1.648642897605896,
+      "learning_rate": 7.089498373958731e-06,
+      "loss": 0.0981,
+      "step": 4847
+    },
+    {
+      "epoch": 2.5390256678889473,
+      "loss_partial": 0.054443359375,
+      "step": 4847
+    },
+    {
+      "epoch": 2.5390256678889473,
+      "loss_partial": 0.047119140625,
+      "step": 4847
+    },
+    {
+      "epoch": 2.539549502357255,
+      "grad_norm": 1.2919210195541382,
+      "learning_rate": 7.085125008027075e-06,
+      "loss": 0.1044,
+      "step": 4848
+    },
+    {
+      "epoch": 2.539549502357255,
+      "loss_partial": 0.05029296875,
+      "step": 4848
+    },
+    {
+      "epoch": 2.539549502357255,
+      "loss_partial": 0.04541015625,
+      "step": 4848
+    },
+    {
+      "epoch": 2.540073336825563,
+      "grad_norm": 1.6154170036315918,
+      "learning_rate": 7.080752251286271e-06,
+      "loss": 0.1005,
+      "step": 4849
+    },
+    {
+      "epoch": 2.540073336825563,
+      "loss_partial": 0.053955078125,
+      "step": 4849
+    },
+    {
+      "epoch": 2.540073336825563,
+      "loss_partial": 0.0498046875,
+      "step": 4849
+    },
+    {
+      "epoch": 2.540597171293871,
+      "grad_norm": 1.6119815111160278,
+      "learning_rate": 7.076380104650191e-06,
+      "loss": 0.1013,
+      "step": 4850
+    },
+    {
+      "epoch": 2.540597171293871,
+      "loss_partial": 0.0546875,
+      "step": 4850
+    },
+    {
+      "epoch": 2.540597171293871,
+      "loss_partial": 0.059326171875,
+      "step": 4850
+    },
+    {
+      "epoch": 2.541121005762179,
+      "grad_norm": 1.6151880025863647,
+      "learning_rate": 7.072008569032586e-06,
+      "loss": 0.1025,
+      "step": 4851
+    },
+    {
+      "epoch": 2.541121005762179,
+      "loss_partial": 0.052490234375,
+      "step": 4851
+    },
+    {
+      "epoch": 2.541121005762179,
+      "loss_partial": 0.0458984375,
+      "step": 4851
+    },
+    {
+      "epoch": 2.541644840230487,
+      "grad_norm": 1.7219220399856567,
+      "learning_rate": 7.067637645347088e-06,
+      "loss": 0.1023,
+      "step": 4852
+    },
+    {
+      "epoch": 2.541644840230487,
+      "loss_partial": 0.056396484375,
+      "step": 4852
+    },
+    {
+      "epoch": 2.541644840230487,
+      "loss_partial": 0.052734375,
+      "step": 4852
+    },
+    {
+      "epoch": 2.5421686746987953,
+      "grad_norm": 1.8569753170013428,
+      "learning_rate": 7.063267334507189e-06,
+      "loss": 0.1032,
+      "step": 4853
+    },
+    {
+      "epoch": 2.5421686746987953,
+      "loss_partial": 0.0556640625,
+      "step": 4853
+    },
+    {
+      "epoch": 2.5421686746987953,
+      "loss_partial": 0.04541015625,
+      "step": 4853
+    },
+    {
+      "epoch": 2.5426925091671033,
+      "grad_norm": 1.3592296838760376,
+      "learning_rate": 7.0588976374262576e-06,
+      "loss": 0.0998,
+      "step": 4854
+    },
+    {
+      "epoch": 2.5426925091671033,
+      "loss_partial": 0.058837890625,
+      "step": 4854
+    },
+    {
+      "epoch": 2.5426925091671033,
+      "loss_partial": 0.043701171875,
+      "step": 4854
+    },
+    {
+      "epoch": 2.543216343635411,
+      "grad_norm": 1.014207124710083,
+      "learning_rate": 7.0545285550175305e-06,
+      "loss": 0.0996,
+      "step": 4855
+    },
+    {
+      "epoch": 2.543216343635411,
+      "loss_partial": 0.0478515625,
+      "step": 4855
+    },
+    {
+      "epoch": 2.543216343635411,
+      "loss_partial": 0.046630859375,
+      "step": 4855
+    },
+    {
+      "epoch": 2.543740178103719,
+      "grad_norm": 2.088892698287964,
+      "learning_rate": 7.050160088194122e-06,
+      "loss": 0.1028,
+      "step": 4856
+    },
+    {
+      "epoch": 2.543740178103719,
+      "loss_partial": 0.057373046875,
+      "step": 4856
+    },
+    {
+      "epoch": 2.543740178103719,
+      "loss_partial": 0.051513671875,
+      "step": 4856
+    },
+    {
+      "epoch": 2.544264012572027,
+      "grad_norm": 1.1434175968170166,
+      "learning_rate": 7.045792237869016e-06,
+      "loss": 0.1057,
+      "step": 4857
+    },
+    {
+      "epoch": 2.544264012572027,
+      "loss_partial": 0.0419921875,
+      "step": 4857
+    },
+    {
+      "epoch": 2.544264012572027,
+      "loss_partial": 0.05126953125,
+      "step": 4857
+    },
+    {
+      "epoch": 2.544787847040335,
+      "grad_norm": 1.9481168985366821,
+      "learning_rate": 7.041425004955062e-06,
+      "loss": 0.0975,
+      "step": 4858
+    },
+    {
+      "epoch": 2.544787847040335,
+      "loss_partial": 0.045654296875,
+      "step": 4858
+    },
+    {
+      "epoch": 2.544787847040335,
+      "loss_partial": 0.048095703125,
+      "step": 4858
+    },
+    {
+      "epoch": 2.545311681508643,
+      "grad_norm": 1.7948296070098877,
+      "learning_rate": 7.037058390364984e-06,
+      "loss": 0.0994,
+      "step": 4859
+    },
+    {
+      "epoch": 2.545311681508643,
+      "loss_partial": 0.05126953125,
+      "step": 4859
+    },
+    {
+      "epoch": 2.545311681508643,
+      "loss_partial": 0.060302734375,
+      "step": 4859
+    },
+    {
+      "epoch": 2.5458355159769512,
+      "grad_norm": 1.3933415412902832,
+      "learning_rate": 7.0326923950113825e-06,
+      "loss": 0.1013,
+      "step": 4860
+    },
+    {
+      "epoch": 2.5458355159769512,
+      "loss_partial": 0.053955078125,
+      "step": 4860
+    },
+    {
+      "epoch": 2.5458355159769512,
+      "loss_partial": 0.05859375,
+      "step": 4860
+    },
+    {
+      "epoch": 2.5463593504452593,
+      "grad_norm": 1.606940507888794,
+      "learning_rate": 7.028327019806724e-06,
+      "loss": 0.1011,
+      "step": 4861
+    },
+    {
+      "epoch": 2.5463593504452593,
+      "loss_partial": 0.049560546875,
+      "step": 4861
+    },
+    {
+      "epoch": 2.5463593504452593,
+      "loss_partial": 0.052978515625,
+      "step": 4861
+    },
+    {
+      "epoch": 2.5468831849135674,
+      "grad_norm": 1.4580520391464233,
+      "learning_rate": 7.02396226566334e-06,
+      "loss": 0.0996,
+      "step": 4862
+    },
+    {
+      "epoch": 2.5468831849135674,
+      "loss_partial": 0.06005859375,
+      "step": 4862
+    },
+    {
+      "epoch": 2.5468831849135674,
+      "loss_partial": 0.050048828125,
+      "step": 4862
+    },
+    {
+      "epoch": 2.5474070193818754,
+      "grad_norm": 1.1017439365386963,
+      "learning_rate": 7.019598133493436e-06,
+      "loss": 0.1024,
+      "step": 4863
+    },
+    {
+      "epoch": 2.5474070193818754,
+      "loss_partial": 0.048583984375,
+      "step": 4863
+    },
+    {
+      "epoch": 2.5474070193818754,
+      "loss_partial": 0.042236328125,
+      "step": 4863
+    },
+    {
+      "epoch": 2.5479308538501835,
+      "grad_norm": 1.3854541778564453,
+      "learning_rate": 7.015234624209096e-06,
+      "loss": 0.1016,
+      "step": 4864
+    },
+    {
+      "epoch": 2.5479308538501835,
+      "loss_partial": 0.056884765625,
+      "step": 4864
+    },
+    {
+      "epoch": 2.5479308538501835,
+      "loss_partial": 0.04296875,
+      "step": 4864
+    },
+    {
+      "epoch": 2.5484546883184915,
+      "grad_norm": 1.2574809789657593,
+      "learning_rate": 7.010871738722261e-06,
+      "loss": 0.1017,
+      "step": 4865
+    },
+    {
+      "epoch": 2.5484546883184915,
+      "loss_partial": 0.06005859375,
+      "step": 4865
+    },
+    {
+      "epoch": 2.5484546883184915,
+      "loss_partial": 0.046875,
+      "step": 4865
+    },
+    {
+      "epoch": 2.5489785227867996,
+      "grad_norm": 1.4452450275421143,
+      "learning_rate": 7.006509477944749e-06,
+      "loss": 0.1028,
+      "step": 4866
+    },
+    {
+      "epoch": 2.5489785227867996,
+      "loss_partial": 0.0458984375,
+      "step": 4866
+    },
+    {
+      "epoch": 2.5489785227867996,
+      "loss_partial": 0.049560546875,
+      "step": 4866
+    },
+    {
+      "epoch": 2.5495023572551077,
+      "grad_norm": 1.3932995796203613,
+      "learning_rate": 7.002147842788243e-06,
+      "loss": 0.1036,
+      "step": 4867
+    },
+    {
+      "epoch": 2.5495023572551077,
+      "loss_partial": 0.05419921875,
+      "step": 4867
+    },
+    {
+      "epoch": 2.5495023572551077,
+      "loss_partial": 0.040771484375,
+      "step": 4867
+    },
+    {
+      "epoch": 2.5500261917234153,
+      "grad_norm": 1.2031445503234863,
+      "learning_rate": 6.997786834164302e-06,
+      "loss": 0.1012,
+      "step": 4868
+    },
+    {
+      "epoch": 2.5500261917234153,
+      "loss_partial": 0.052978515625,
+      "step": 4868
+    },
+    {
+      "epoch": 2.5500261917234153,
+      "loss_partial": 0.051025390625,
+      "step": 4868
+    },
+    {
+      "epoch": 2.5505500261917233,
+      "grad_norm": 1.206408977508545,
+      "learning_rate": 6.993426452984347e-06,
+      "loss": 0.0998,
+      "step": 4869
+    },
+    {
+      "epoch": 2.5505500261917233,
+      "loss_partial": 0.05126953125,
+      "step": 4869
+    },
+    {
+      "epoch": 2.5505500261917233,
+      "loss_partial": 0.056640625,
+      "step": 4869
+    },
+    {
+      "epoch": 2.5510738606600314,
+      "grad_norm": 1.8383188247680664,
+      "learning_rate": 6.989066700159671e-06,
+      "loss": 0.1006,
+      "step": 4870
+    },
+    {
+      "epoch": 2.5510738606600314,
+      "loss_partial": 0.04296875,
+      "step": 4870
+    },
+    {
+      "epoch": 2.5510738606600314,
+      "loss_partial": 0.0478515625,
+      "step": 4870
+    },
+    {
+      "epoch": 2.5515976951283394,
+      "grad_norm": 1.681834101676941,
+      "learning_rate": 6.984707576601433e-06,
+      "loss": 0.1055,
+      "step": 4871
+    },
+    {
+      "epoch": 2.5515976951283394,
+      "loss_partial": 0.041748046875,
+      "step": 4871
+    },
+    {
+      "epoch": 2.5515976951283394,
+      "loss_partial": 0.043701171875,
+      "step": 4871
+    },
+    {
+      "epoch": 2.5521215295966475,
+      "grad_norm": 1.7604483366012573,
+      "learning_rate": 6.9803490832206675e-06,
+      "loss": 0.1011,
+      "step": 4872
+    },
+    {
+      "epoch": 2.5521215295966475,
+      "loss_partial": 0.053955078125,
+      "step": 4872
+    },
+    {
+      "epoch": 2.5521215295966475,
+      "loss_partial": 0.060546875,
+      "step": 4872
+    },
+    {
+      "epoch": 2.5526453640649556,
+      "grad_norm": 1.6771079301834106,
+      "learning_rate": 6.975991220928271e-06,
+      "loss": 0.1006,
+      "step": 4873
+    },
+    {
+      "epoch": 2.5526453640649556,
+      "loss_partial": 0.050048828125,
+      "step": 4873
+    },
+    {
+      "epoch": 2.5526453640649556,
+      "loss_partial": 0.060302734375,
+      "step": 4873
+    },
+    {
+      "epoch": 2.5531691985332636,
+      "grad_norm": 1.2213844060897827,
+      "learning_rate": 6.971633990635013e-06,
+      "loss": 0.104,
+      "step": 4874
+    },
+    {
+      "epoch": 2.5531691985332636,
+      "loss_partial": 0.05029296875,
+      "step": 4874
+    },
+    {
+      "epoch": 2.5531691985332636,
+      "loss_partial": 0.045654296875,
+      "step": 4874
+    },
+    {
+      "epoch": 2.5536930330015712,
+      "grad_norm": 1.8957831859588623,
+      "learning_rate": 6.967277393251517e-06,
+      "loss": 0.1015,
+      "step": 4875
+    },
+    {
+      "epoch": 2.5536930330015712,
+      "loss_partial": 0.05029296875,
+      "step": 4875
+    },
+    {
+      "epoch": 2.5536930330015712,
+      "loss_partial": 0.049072265625,
+      "step": 4875
+    },
+    {
+      "epoch": 2.5542168674698793,
+      "grad_norm": 1.1244479417800903,
+      "learning_rate": 6.962921429688296e-06,
+      "loss": 0.1024,
+      "step": 4876
+    },
+    {
+      "epoch": 2.5542168674698793,
+      "loss_partial": 0.04638671875,
+      "step": 4876
+    },
+    {
+      "epoch": 2.5542168674698793,
+      "loss_partial": 0.050537109375,
+      "step": 4876
+    },
+    {
+      "epoch": 2.5547407019381874,
+      "grad_norm": 1.7064591646194458,
+      "learning_rate": 6.958566100855716e-06,
+      "loss": 0.1025,
+      "step": 4877
+    },
+    {
+      "epoch": 2.5547407019381874,
+      "loss_partial": 0.052001953125,
+      "step": 4877
+    },
+    {
+      "epoch": 2.5547407019381874,
+      "loss_partial": 0.052490234375,
+      "step": 4877
+    },
+    {
+      "epoch": 2.5552645364064954,
+      "grad_norm": 1.2313942909240723,
+      "learning_rate": 6.9542114076640135e-06,
+      "loss": 0.1031,
+      "step": 4878
+    },
+    {
+      "epoch": 2.5552645364064954,
+      "loss_partial": 0.046142578125,
+      "step": 4878
+    },
+    {
+      "epoch": 2.5552645364064954,
+      "loss_partial": 0.0517578125,
+      "step": 4878
+    },
+    {
+      "epoch": 2.5557883708748035,
+      "grad_norm": 1.3857237100601196,
+      "learning_rate": 6.949857351023289e-06,
+      "loss": 0.0999,
+      "step": 4879
+    },
+    {
+      "epoch": 2.5557883708748035,
+      "loss_partial": 0.0458984375,
+      "step": 4879
+    },
+    {
+      "epoch": 2.5557883708748035,
+      "loss_partial": 0.054443359375,
+      "step": 4879
+    },
+    {
+      "epoch": 2.5563122053431115,
+      "grad_norm": 1.58064603805542,
+      "learning_rate": 6.945503931843519e-06,
+      "loss": 0.1018,
+      "step": 4880
+    },
+    {
+      "epoch": 2.5563122053431115,
+      "loss_partial": 0.055908203125,
+      "step": 4880
+    },
+    {
+      "epoch": 2.5563122053431115,
+      "loss_partial": 0.058349609375,
+      "step": 4880
+    },
+    {
+      "epoch": 2.5568360398114196,
+      "grad_norm": 1.5056848526000977,
+      "learning_rate": 6.9411511510345395e-06,
+      "loss": 0.1012,
+      "step": 4881
+    },
+    {
+      "epoch": 2.5568360398114196,
+      "loss_partial": 0.05419921875,
+      "step": 4881
+    },
+    {
+      "epoch": 2.5568360398114196,
+      "loss_partial": 0.050048828125,
+      "step": 4881
+    },
+    {
+      "epoch": 2.5573598742797277,
+      "grad_norm": 1.3087526559829712,
+      "learning_rate": 6.936799009506053e-06,
+      "loss": 0.0989,
+      "step": 4882
+    },
+    {
+      "epoch": 2.5573598742797277,
+      "loss_partial": 0.040283203125,
+      "step": 4882
+    },
+    {
+      "epoch": 2.5573598742797277,
+      "loss_partial": 0.048583984375,
+      "step": 4882
+    },
+    {
+      "epoch": 2.5578837087480357,
+      "grad_norm": 1.3528800010681152,
+      "learning_rate": 6.932447508167628e-06,
+      "loss": 0.1006,
+      "step": 4883
+    },
+    {
+      "epoch": 2.5578837087480357,
+      "loss_partial": 0.0517578125,
+      "step": 4883
+    },
+    {
+      "epoch": 2.5578837087480357,
+      "loss_partial": 0.052734375,
+      "step": 4883
+    },
+    {
+      "epoch": 2.558407543216344,
+      "grad_norm": 1.6512987613677979,
+      "learning_rate": 6.928096647928708e-06,
+      "loss": 0.1021,
+      "step": 4884
+    },
+    {
+      "epoch": 2.558407543216344,
+      "loss_partial": 0.048095703125,
+      "step": 4884
+    },
+    {
+      "epoch": 2.558407543216344,
+      "loss_partial": 0.053955078125,
+      "step": 4884
+    },
+    {
+      "epoch": 2.558931377684652,
+      "grad_norm": 1.27328360080719,
+      "learning_rate": 6.923746429698595e-06,
+      "loss": 0.1003,
+      "step": 4885
+    },
+    {
+      "epoch": 2.558931377684652,
+      "loss_partial": 0.042236328125,
+      "step": 4885
+    },
+    {
+      "epoch": 2.558931377684652,
+      "loss_partial": 0.044677734375,
+      "step": 4885
+    },
+    {
+      "epoch": 2.55945521215296,
+      "grad_norm": 1.9843257665634155,
+      "learning_rate": 6.9193968543864546e-06,
+      "loss": 0.0974,
+      "step": 4886
+    },
+    {
+      "epoch": 2.55945521215296,
+      "loss_partial": 0.056640625,
+      "step": 4886
+    },
+    {
+      "epoch": 2.55945521215296,
+      "loss_partial": 0.040771484375,
+      "step": 4886
+    },
+    {
+      "epoch": 2.5599790466212675,
+      "grad_norm": 1.589637041091919,
+      "learning_rate": 6.915047922901316e-06,
+      "loss": 0.103,
+      "step": 4887
+    },
+    {
+      "epoch": 2.5599790466212675,
+      "loss_partial": 0.044189453125,
+      "step": 4887
+    },
+    {
+      "epoch": 2.5599790466212675,
+      "loss_partial": 0.044677734375,
+      "step": 4887
+    },
+    {
+      "epoch": 2.5605028810895756,
+      "grad_norm": 1.3610427379608154,
+      "learning_rate": 6.9106996361520895e-06,
+      "loss": 0.0993,
+      "step": 4888
+    },
+    {
+      "epoch": 2.5605028810895756,
+      "loss_partial": 0.051025390625,
+      "step": 4888
+    },
+    {
+      "epoch": 2.5605028810895756,
+      "loss_partial": 0.047607421875,
+      "step": 4888
+    },
+    {
+      "epoch": 2.5610267155578836,
+      "grad_norm": 1.792715311050415,
+      "learning_rate": 6.906351995047535e-06,
+      "loss": 0.0985,
+      "step": 4889
+    },
+    {
+      "epoch": 2.5610267155578836,
+      "loss_partial": 0.05517578125,
+      "step": 4889
+    },
+    {
+      "epoch": 2.5610267155578836,
+      "loss_partial": 0.052734375,
+      "step": 4889
+    },
+    {
+      "epoch": 2.5615505500261917,
+      "grad_norm": 1.0308560132980347,
+      "learning_rate": 6.902005000496282e-06,
+      "loss": 0.1011,
+      "step": 4890
+    },
+    {
+      "epoch": 2.5615505500261917,
+      "loss_partial": 0.04638671875,
+      "step": 4890
+    },
+    {
+      "epoch": 2.5615505500261917,
+      "loss_partial": 0.046630859375,
+      "step": 4890
+    },
+    {
+      "epoch": 2.5620743844944998,
+      "grad_norm": 1.5589781999588013,
+      "learning_rate": 6.897658653406828e-06,
+      "loss": 0.1018,
+      "step": 4891
+    },
+    {
+      "epoch": 2.5620743844944998,
+      "loss_partial": 0.05224609375,
+      "step": 4891
+    },
+    {
+      "epoch": 2.5620743844944998,
+      "loss_partial": 0.0439453125,
+      "step": 4891
+    },
+    {
+      "epoch": 2.562598218962808,
+      "grad_norm": 1.5287944078445435,
+      "learning_rate": 6.893312954687529e-06,
+      "loss": 0.099,
+      "step": 4892
+    },
+    {
+      "epoch": 2.562598218962808,
+      "loss_partial": 0.050537109375,
+      "step": 4892
+    },
+    {
+      "epoch": 2.562598218962808,
+      "loss_partial": 0.05517578125,
+      "step": 4892
+    },
+    {
+      "epoch": 2.563122053431116,
+      "grad_norm": 1.358941912651062,
+      "learning_rate": 6.888967905246613e-06,
+      "loss": 0.102,
+      "step": 4893
+    },
+    {
+      "epoch": 2.563122053431116,
+      "loss_partial": 0.05517578125,
+      "step": 4893
+    },
+    {
+      "epoch": 2.563122053431116,
+      "loss_partial": 0.048828125,
+      "step": 4893
+    },
+    {
+      "epoch": 2.563645887899424,
+      "grad_norm": 1.4339338541030884,
+      "learning_rate": 6.884623505992166e-06,
+      "loss": 0.1017,
+      "step": 4894
+    },
+    {
+      "epoch": 2.563645887899424,
+      "loss_partial": 0.052490234375,
+      "step": 4894
+    },
+    {
+      "epoch": 2.563645887899424,
+      "loss_partial": 0.052978515625,
+      "step": 4894
+    },
+    {
+      "epoch": 2.5641697223677316,
+      "grad_norm": 1.3585768938064575,
+      "learning_rate": 6.880279757832142e-06,
+      "loss": 0.0986,
+      "step": 4895
+    },
+    {
+      "epoch": 2.5641697223677316,
+      "loss_partial": 0.04541015625,
+      "step": 4895
+    },
+    {
+      "epoch": 2.5641697223677316,
+      "loss_partial": 0.042724609375,
+      "step": 4895
+    },
+    {
+      "epoch": 2.5646935568360396,
+      "grad_norm": 1.4659123420715332,
+      "learning_rate": 6.875936661674353e-06,
+      "loss": 0.1027,
+      "step": 4896
+    },
+    {
+      "epoch": 2.5646935568360396,
+      "loss_partial": 0.04541015625,
+      "step": 4896
+    },
+    {
+      "epoch": 2.5646935568360396,
+      "loss_partial": 0.04736328125,
+      "step": 4896
+    },
+    {
+      "epoch": 2.5652173913043477,
+      "grad_norm": 1.6495723724365234,
+      "learning_rate": 6.871594218426487e-06,
+      "loss": 0.097,
+      "step": 4897
+    },
+    {
+      "epoch": 2.5652173913043477,
+      "loss_partial": 0.050537109375,
+      "step": 4897
+    },
+    {
+      "epoch": 2.5652173913043477,
+      "loss_partial": 0.05517578125,
+      "step": 4897
+    },
+    {
+      "epoch": 2.5657412257726557,
+      "grad_norm": 1.4686951637268066,
+      "learning_rate": 6.8672524289960854e-06,
+      "loss": 0.1037,
+      "step": 4898
+    },
+    {
+      "epoch": 2.5657412257726557,
+      "loss_partial": 0.051513671875,
+      "step": 4898
+    },
+    {
+      "epoch": 2.5657412257726557,
+      "loss_partial": 0.052978515625,
+      "step": 4898
+    },
+    {
+      "epoch": 2.566265060240964,
+      "grad_norm": 1.4079796075820923,
+      "learning_rate": 6.862911294290553e-06,
+      "loss": 0.1034,
+      "step": 4899
+    },
+    {
+      "epoch": 2.566265060240964,
+      "loss_partial": 0.04931640625,
+      "step": 4899
+    },
+    {
+      "epoch": 2.566265060240964,
+      "loss_partial": 0.04052734375,
+      "step": 4899
+    },
+    {
+      "epoch": 2.566788894709272,
+      "grad_norm": 1.289152979850769,
+      "learning_rate": 6.858570815217156e-06,
+      "loss": 0.0971,
+      "step": 4900
+    },
+    {
+      "epoch": 2.566788894709272,
+      "loss_partial": 0.039794921875,
+      "step": 4900
+    },
+    {
+      "epoch": 2.566788894709272,
+      "loss_partial": 0.04345703125,
+      "step": 4900
+    },
+    {
+      "epoch": 2.56731272917758,
+      "grad_norm": 1.2464395761489868,
+      "learning_rate": 6.8542309926830365e-06,
+      "loss": 0.0984,
+      "step": 4901
+    },
+    {
+      "epoch": 2.56731272917758,
+      "loss_partial": 0.05029296875,
+      "step": 4901
+    },
+    {
+      "epoch": 2.56731272917758,
+      "loss_partial": 0.0634765625,
+      "step": 4901
+    },
+    {
+      "epoch": 2.567836563645888,
+      "grad_norm": 1.2701666355133057,
+      "learning_rate": 6.8498918275951876e-06,
+      "loss": 0.0999,
+      "step": 4902
+    },
+    {
+      "epoch": 2.567836563645888,
+      "loss_partial": 0.044921875,
+      "step": 4902
+    },
+    {
+      "epoch": 2.567836563645888,
+      "loss_partial": 0.05322265625,
+      "step": 4902
+    },
+    {
+      "epoch": 2.568360398114196,
+      "grad_norm": 1.284911036491394,
+      "learning_rate": 6.8455533208604654e-06,
+      "loss": 0.0959,
+      "step": 4903
+    },
+    {
+      "epoch": 2.568360398114196,
+      "loss_partial": 0.053466796875,
+      "step": 4903
+    },
+    {
+      "epoch": 2.568360398114196,
+      "loss_partial": 0.05126953125,
+      "step": 4903
+    },
+    {
+      "epoch": 2.568884232582504,
+      "grad_norm": 1.4952116012573242,
+      "learning_rate": 6.841215473385592e-06,
+      "loss": 0.1026,
+      "step": 4904
+    },
+    {
+      "epoch": 2.568884232582504,
+      "loss_partial": 0.04052734375,
+      "step": 4904
+    },
+    {
+      "epoch": 2.568884232582504,
+      "loss_partial": 0.044677734375,
+      "step": 4904
+    },
+    {
+      "epoch": 2.569408067050812,
+      "grad_norm": 1.4354509115219116,
+      "learning_rate": 6.836878286077154e-06,
+      "loss": 0.0996,
+      "step": 4905
+    },
+    {
+      "epoch": 2.569408067050812,
+      "loss_partial": 0.050048828125,
+      "step": 4905
+    },
+    {
+      "epoch": 2.569408067050812,
+      "loss_partial": 0.0498046875,
+      "step": 4905
+    },
+    {
+      "epoch": 2.56993190151912,
+      "grad_norm": 1.4708391427993774,
+      "learning_rate": 6.832541759841592e-06,
+      "loss": 0.1028,
+      "step": 4906
+    },
+    {
+      "epoch": 2.56993190151912,
+      "loss_partial": 0.04541015625,
+      "step": 4906
+    },
+    {
+      "epoch": 2.56993190151912,
+      "loss_partial": 0.0537109375,
+      "step": 4906
+    },
+    {
+      "epoch": 2.570455735987428,
+      "grad_norm": 1.8627933263778687,
+      "learning_rate": 6.828205895585216e-06,
+      "loss": 0.1042,
+      "step": 4907
+    },
+    {
+      "epoch": 2.570455735987428,
+      "loss_partial": 0.0537109375,
+      "step": 4907
+    },
+    {
+      "epoch": 2.570455735987428,
+      "loss_partial": 0.042724609375,
+      "step": 4907
+    },
+    {
+      "epoch": 2.570979570455736,
+      "grad_norm": 1.5881125926971436,
+      "learning_rate": 6.82387069421419e-06,
+      "loss": 0.1019,
+      "step": 4908
+    },
+    {
+      "epoch": 2.570979570455736,
+      "loss_partial": 0.0537109375,
+      "step": 4908
+    },
+    {
+      "epoch": 2.570979570455736,
+      "loss_partial": 0.04833984375,
+      "step": 4908
+    },
+    {
+      "epoch": 2.571503404924044,
+      "grad_norm": 0.9747796058654785,
+      "learning_rate": 6.819536156634553e-06,
+      "loss": 0.1012,
+      "step": 4909
+    },
+    {
+      "epoch": 2.571503404924044,
+      "loss_partial": 0.04638671875,
+      "step": 4909
+    },
+    {
+      "epoch": 2.571503404924044,
+      "loss_partial": 0.0576171875,
+      "step": 4909
+    },
+    {
+      "epoch": 2.572027239392352,
+      "grad_norm": 1.5416985750198364,
+      "learning_rate": 6.815202283752191e-06,
+      "loss": 0.1044,
+      "step": 4910
+    },
+    {
+      "epoch": 2.572027239392352,
+      "loss_partial": 0.0634765625,
+      "step": 4910
+    },
+    {
+      "epoch": 2.572027239392352,
+      "loss_partial": 0.048583984375,
+      "step": 4910
+    },
+    {
+      "epoch": 2.57255107386066,
+      "grad_norm": 1.6801990270614624,
+      "learning_rate": 6.81086907647286e-06,
+      "loss": 0.1001,
+      "step": 4911
+    },
+    {
+      "epoch": 2.57255107386066,
+      "loss_partial": 0.06640625,
+      "step": 4911
+    },
+    {
+      "epoch": 2.57255107386066,
+      "loss_partial": 0.04736328125,
+      "step": 4911
+    },
+    {
+      "epoch": 2.573074908328968,
+      "grad_norm": 1.5158450603485107,
+      "learning_rate": 6.806536535702165e-06,
+      "loss": 0.1007,
+      "step": 4912
+    },
+    {
+      "epoch": 2.573074908328968,
+      "loss_partial": 0.048095703125,
+      "step": 4912
+    },
+    {
+      "epoch": 2.573074908328968,
+      "loss_partial": 0.04931640625,
+      "step": 4912
+    },
+    {
+      "epoch": 2.573598742797276,
+      "grad_norm": 1.1214231252670288,
+      "learning_rate": 6.80220466234559e-06,
+      "loss": 0.1,
+      "step": 4913
+    },
+    {
+      "epoch": 2.573598742797276,
+      "loss_partial": 0.0537109375,
+      "step": 4913
+    },
+    {
+      "epoch": 2.573598742797276,
+      "loss_partial": 0.049072265625,
+      "step": 4913
+    },
+    {
+      "epoch": 2.574122577265584,
+      "grad_norm": 1.2649428844451904,
+      "learning_rate": 6.797873457308466e-06,
+      "loss": 0.106,
+      "step": 4914
+    },
+    {
+      "epoch": 2.574122577265584,
+      "loss_partial": 0.0400390625,
+      "step": 4914
+    },
+    {
+      "epoch": 2.574122577265584,
+      "loss_partial": 0.044189453125,
+      "step": 4914
+    },
+    {
+      "epoch": 2.574646411733892,
+      "grad_norm": 1.0452308654785156,
+      "learning_rate": 6.793542921495989e-06,
+      "loss": 0.0989,
+      "step": 4915
+    },
+    {
+      "epoch": 2.574646411733892,
+      "loss_partial": 0.051513671875,
+      "step": 4915
+    },
+    {
+      "epoch": 2.574646411733892,
+      "loss_partial": 0.0537109375,
+      "step": 4915
+    },
+    {
+      "epoch": 2.5751702462022,
+      "grad_norm": 0.9089459180831909,
+      "learning_rate": 6.789213055813211e-06,
+      "loss": 0.1031,
+      "step": 4916
+    },
+    {
+      "epoch": 2.5751702462022,
+      "loss_partial": 0.04638671875,
+      "step": 4916
+    },
+    {
+      "epoch": 2.5751702462022,
+      "loss_partial": 0.044921875,
+      "step": 4916
+    },
+    {
+      "epoch": 2.575694080670508,
+      "grad_norm": 1.435996174812317,
+      "learning_rate": 6.78488386116505e-06,
+      "loss": 0.1013,
+      "step": 4917
+    },
+    {
+      "epoch": 2.575694080670508,
+      "loss_partial": 0.04052734375,
+      "step": 4917
+    },
+    {
+      "epoch": 2.575694080670508,
+      "loss_partial": 0.041748046875,
+      "step": 4917
+    },
+    {
+      "epoch": 2.576217915138816,
+      "grad_norm": 1.3598884344100952,
+      "learning_rate": 6.780555338456282e-06,
+      "loss": 0.1026,
+      "step": 4918
+    },
+    {
+      "epoch": 2.576217915138816,
+      "loss_partial": 0.04541015625,
+      "step": 4918
+    },
+    {
+      "epoch": 2.576217915138816,
+      "loss_partial": 0.05322265625,
+      "step": 4918
+    },
+    {
+      "epoch": 2.576741749607124,
+      "grad_norm": 1.6409101486206055,
+      "learning_rate": 6.776227488591539e-06,
+      "loss": 0.1045,
+      "step": 4919
+    },
+    {
+      "epoch": 2.576741749607124,
+      "loss_partial": 0.0546875,
+      "step": 4919
+    },
+    {
+      "epoch": 2.576741749607124,
+      "loss_partial": 0.049072265625,
+      "step": 4919
+    },
+    {
+      "epoch": 2.577265584075432,
+      "grad_norm": 1.19239342212677,
+      "learning_rate": 6.7719003124753115e-06,
+      "loss": 0.1,
+      "step": 4920
+    },
+    {
+      "epoch": 2.577265584075432,
+      "loss_partial": 0.046875,
+      "step": 4920
+    },
+    {
+      "epoch": 2.577265584075432,
+      "loss_partial": 0.062255859375,
+      "step": 4920
+    },
+    {
+      "epoch": 2.57778941854374,
+      "grad_norm": 2.334714651107788,
+      "learning_rate": 6.76757381101196e-06,
+      "loss": 0.1022,
+      "step": 4921
+    },
+    {
+      "epoch": 2.57778941854374,
+      "loss_partial": 0.0458984375,
+      "step": 4921
+    },
+    {
+      "epoch": 2.57778941854374,
+      "loss_partial": 0.053466796875,
+      "step": 4921
+    },
+    {
+      "epoch": 2.5783132530120483,
+      "grad_norm": 1.4524568319320679,
+      "learning_rate": 6.763247985105695e-06,
+      "loss": 0.0999,
+      "step": 4922
+    },
+    {
+      "epoch": 2.5783132530120483,
+      "loss_partial": 0.05224609375,
+      "step": 4922
+    },
+    {
+      "epoch": 2.5783132530120483,
+      "loss_partial": 0.044677734375,
+      "step": 4922
+    },
+    {
+      "epoch": 2.5788370874803563,
+      "grad_norm": 1.9438320398330688,
+      "learning_rate": 6.758922835660587e-06,
+      "loss": 0.0996,
+      "step": 4923
+    },
+    {
+      "epoch": 2.5788370874803563,
+      "loss_partial": 0.055419921875,
+      "step": 4923
+    },
+    {
+      "epoch": 2.5788370874803563,
+      "loss_partial": 0.05712890625,
+      "step": 4923
+    },
+    {
+      "epoch": 2.5793609219486644,
+      "grad_norm": 1.3992279767990112,
+      "learning_rate": 6.754598363580558e-06,
+      "loss": 0.099,
+      "step": 4924
+    },
+    {
+      "epoch": 2.5793609219486644,
+      "loss_partial": 0.05322265625,
+      "step": 4924
+    },
+    {
+      "epoch": 2.5793609219486644,
+      "loss_partial": 0.054443359375,
+      "step": 4924
+    },
+    {
+      "epoch": 2.5798847564169725,
+      "grad_norm": 1.450713038444519,
+      "learning_rate": 6.750274569769406e-06,
+      "loss": 0.0998,
+      "step": 4925
+    },
+    {
+      "epoch": 2.5798847564169725,
+      "loss_partial": 0.05322265625,
+      "step": 4925
+    },
+    {
+      "epoch": 2.5798847564169725,
+      "loss_partial": 0.054931640625,
+      "step": 4925
+    },
+    {
+      "epoch": 2.5804085908852805,
+      "grad_norm": 1.161758303642273,
+      "learning_rate": 6.745951455130775e-06,
+      "loss": 0.1006,
+      "step": 4926
+    },
+    {
+      "epoch": 2.5804085908852805,
+      "loss_partial": 0.0546875,
+      "step": 4926
+    },
+    {
+      "epoch": 2.5804085908852805,
+      "loss_partial": 0.048828125,
+      "step": 4926
+    },
+    {
+      "epoch": 2.580932425353588,
+      "grad_norm": 1.9598435163497925,
+      "learning_rate": 6.741629020568166e-06,
+      "loss": 0.0994,
+      "step": 4927
+    },
+    {
+      "epoch": 2.580932425353588,
+      "loss_partial": 0.05322265625,
+      "step": 4927
+    },
+    {
+      "epoch": 2.580932425353588,
+      "loss_partial": 0.04931640625,
+      "step": 4927
+    },
+    {
+      "epoch": 2.581456259821896,
+      "grad_norm": 1.5077075958251953,
+      "learning_rate": 6.737307266984939e-06,
+      "loss": 0.1029,
+      "step": 4928
+    },
+    {
+      "epoch": 2.581456259821896,
+      "loss_partial": 0.047607421875,
+      "step": 4928
+    },
+    {
+      "epoch": 2.581456259821896,
+      "loss_partial": 0.04931640625,
+      "step": 4928
+    },
+    {
+      "epoch": 2.5819800942902043,
+      "grad_norm": 1.1514135599136353,
+      "learning_rate": 6.732986195284322e-06,
+      "loss": 0.0974,
+      "step": 4929
+    },
+    {
+      "epoch": 2.5819800942902043,
+      "loss_partial": 0.0498046875,
+      "step": 4929
+    },
+    {
+      "epoch": 2.5819800942902043,
+      "loss_partial": 0.053466796875,
+      "step": 4929
+    },
+    {
+      "epoch": 2.5825039287585123,
+      "grad_norm": 1.7987141609191895,
+      "learning_rate": 6.7286658063693855e-06,
+      "loss": 0.1008,
+      "step": 4930
+    },
+    {
+      "epoch": 2.5825039287585123,
+      "loss_partial": 0.046875,
+      "step": 4930
+    },
+    {
+      "epoch": 2.5825039287585123,
+      "loss_partial": 0.04931640625,
+      "step": 4930
+    },
+    {
+      "epoch": 2.5830277632268204,
+      "grad_norm": 1.6676886081695557,
+      "learning_rate": 6.724346101143067e-06,
+      "loss": 0.0983,
+      "step": 4931
+    },
+    {
+      "epoch": 2.5830277632268204,
+      "loss_partial": 0.051025390625,
+      "step": 4931
+    },
+    {
+      "epoch": 2.5830277632268204,
+      "loss_partial": 0.041748046875,
+      "step": 4931
+    },
+    {
+      "epoch": 2.5835515976951284,
+      "grad_norm": 1.3861678838729858,
+      "learning_rate": 6.720027080508153e-06,
+      "loss": 0.1035,
+      "step": 4932
+    },
+    {
+      "epoch": 2.5835515976951284,
+      "loss_partial": 0.0537109375,
+      "step": 4932
+    },
+    {
+      "epoch": 2.5835515976951284,
+      "loss_partial": 0.0537109375,
+      "step": 4932
+    },
+    {
+      "epoch": 2.5840754321634365,
+      "grad_norm": 1.6482657194137573,
+      "learning_rate": 6.7157087453672995e-06,
+      "loss": 0.0988,
+      "step": 4933
+    },
+    {
+      "epoch": 2.5840754321634365,
+      "loss_partial": 0.0537109375,
+      "step": 4933
+    },
+    {
+      "epoch": 2.5840754321634365,
+      "loss_partial": 0.044921875,
+      "step": 4933
+    },
+    {
+      "epoch": 2.584599266631744,
+      "grad_norm": 1.958420753479004,
+      "learning_rate": 6.71139109662301e-06,
+      "loss": 0.1021,
+      "step": 4934
+    },
+    {
+      "epoch": 2.584599266631744,
+      "loss_partial": 0.0439453125,
+      "step": 4934
+    },
+    {
+      "epoch": 2.584599266631744,
+      "loss_partial": 0.05126953125,
+      "step": 4934
+    },
+    {
+      "epoch": 2.585123101100052,
+      "grad_norm": 1.292141318321228,
+      "learning_rate": 6.707074135177648e-06,
+      "loss": 0.102,
+      "step": 4935
+    },
+    {
+      "epoch": 2.585123101100052,
+      "loss_partial": 0.0439453125,
+      "step": 4935
+    },
+    {
+      "epoch": 2.585123101100052,
+      "loss_partial": 0.051513671875,
+      "step": 4935
+    },
+    {
+      "epoch": 2.5856469355683602,
+      "grad_norm": 2.0035383701324463,
+      "learning_rate": 6.702757861933422e-06,
+      "loss": 0.0996,
+      "step": 4936
+    },
+    {
+      "epoch": 2.5856469355683602,
+      "loss_partial": 0.0537109375,
+      "step": 4936
+    },
+    {
+      "epoch": 2.5856469355683602,
+      "loss_partial": 0.0576171875,
+      "step": 4936
+    },
+    {
+      "epoch": 2.5861707700366683,
+      "grad_norm": 1.8627843856811523,
+      "learning_rate": 6.698442277792416e-06,
+      "loss": 0.1015,
+      "step": 4937
+    },
+    {
+      "epoch": 2.5861707700366683,
+      "loss_partial": 0.046875,
+      "step": 4937
+    },
+    {
+      "epoch": 2.5861707700366683,
+      "loss_partial": 0.046875,
+      "step": 4937
+    },
+    {
+      "epoch": 2.5866946045049763,
+      "grad_norm": 1.7511399984359741,
+      "learning_rate": 6.69412738365656e-06,
+      "loss": 0.0995,
+      "step": 4938
+    },
+    {
+      "epoch": 2.5866946045049763,
+      "loss_partial": 0.05224609375,
+      "step": 4938
+    },
+    {
+      "epoch": 2.5866946045049763,
+      "loss_partial": 0.047119140625,
+      "step": 4938
+    },
+    {
+      "epoch": 2.5872184389732844,
+      "grad_norm": 1.7211368083953857,
+      "learning_rate": 6.689813180427637e-06,
+      "loss": 0.1036,
+      "step": 4939
+    },
+    {
+      "epoch": 2.5872184389732844,
+      "loss_partial": 0.06201171875,
+      "step": 4939
+    },
+    {
+      "epoch": 2.5872184389732844,
+      "loss_partial": 0.043701171875,
+      "step": 4939
+    },
+    {
+      "epoch": 2.5877422734415925,
+      "grad_norm": 1.7608208656311035,
+      "learning_rate": 6.685499669007287e-06,
+      "loss": 0.101,
+      "step": 4940
+    },
+    {
+      "epoch": 2.5877422734415925,
+      "loss_partial": 0.0498046875,
+      "step": 4940
+    },
+    {
+      "epoch": 2.5877422734415925,
+      "loss_partial": 0.05517578125,
+      "step": 4940
+    },
+    {
+      "epoch": 2.5882661079099005,
+      "grad_norm": 1.6896686553955078,
+      "learning_rate": 6.681186850297014e-06,
+      "loss": 0.1014,
+      "step": 4941
+    },
+    {
+      "epoch": 2.5882661079099005,
+      "loss_partial": 0.0546875,
+      "step": 4941
+    },
+    {
+      "epoch": 2.5882661079099005,
+      "loss_partial": 0.039794921875,
+      "step": 4941
+    },
+    {
+      "epoch": 2.5887899423782086,
+      "grad_norm": 1.901923418045044,
+      "learning_rate": 6.6768747251981634e-06,
+      "loss": 0.0992,
+      "step": 4942
+    },
+    {
+      "epoch": 2.5887899423782086,
+      "loss_partial": 0.060546875,
+      "step": 4942
+    },
+    {
+      "epoch": 2.5887899423782086,
+      "loss_partial": 0.0546875,
+      "step": 4942
+    },
+    {
+      "epoch": 2.5893137768465166,
+      "grad_norm": 2.0337913036346436,
+      "learning_rate": 6.672563294611948e-06,
+      "loss": 0.0988,
+      "step": 4943
+    },
+    {
+      "epoch": 2.5893137768465166,
+      "loss_partial": 0.055908203125,
+      "step": 4943
+    },
+    {
+      "epoch": 2.5893137768465166,
+      "loss_partial": 0.059814453125,
+      "step": 4943
+    },
+    {
+      "epoch": 2.5898376113148247,
+      "grad_norm": 2.0266151428222656,
+      "learning_rate": 6.6682525594394235e-06,
+      "loss": 0.1004,
+      "step": 4944
+    },
+    {
+      "epoch": 2.5898376113148247,
+      "loss_partial": 0.05224609375,
+      "step": 4944
+    },
+    {
+      "epoch": 2.5898376113148247,
+      "loss_partial": 0.049560546875,
+      "step": 4944
+    },
+    {
+      "epoch": 2.5903614457831328,
+      "grad_norm": 1.5276105403900146,
+      "learning_rate": 6.663942520581514e-06,
+      "loss": 0.1024,
+      "step": 4945
+    },
+    {
+      "epoch": 2.5903614457831328,
+      "loss_partial": 0.0537109375,
+      "step": 4945
+    },
+    {
+      "epoch": 2.5903614457831328,
+      "loss_partial": 0.056884765625,
+      "step": 4945
+    },
+    {
+      "epoch": 2.5908852802514404,
+      "grad_norm": 1.6199994087219238,
+      "learning_rate": 6.659633178938989e-06,
+      "loss": 0.1042,
+      "step": 4946
+    },
+    {
+      "epoch": 2.5908852802514404,
+      "loss_partial": 0.055908203125,
+      "step": 4946
+    },
+    {
+      "epoch": 2.5908852802514404,
+      "loss_partial": 0.0556640625,
+      "step": 4946
+    },
+    {
+      "epoch": 2.5914091147197484,
+      "grad_norm": 1.8527828454971313,
+      "learning_rate": 6.655324535412473e-06,
+      "loss": 0.1071,
+      "step": 4947
+    },
+    {
+      "epoch": 2.5914091147197484,
+      "loss_partial": 0.0478515625,
+      "step": 4947
+    },
+    {
+      "epoch": 2.5914091147197484,
+      "loss_partial": 0.05712890625,
+      "step": 4947
+    },
+    {
+      "epoch": 2.5919329491880565,
+      "grad_norm": 1.3647449016571045,
+      "learning_rate": 6.651016590902448e-06,
+      "loss": 0.1025,
+      "step": 4948
+    },
+    {
+      "epoch": 2.5919329491880565,
+      "loss_partial": 0.054443359375,
+      "step": 4948
+    },
+    {
+      "epoch": 2.5919329491880565,
+      "loss_partial": 0.0546875,
+      "step": 4948
+    },
+    {
+      "epoch": 2.5924567836563646,
+      "grad_norm": 1.678478717803955,
+      "learning_rate": 6.646709346309242e-06,
+      "loss": 0.0998,
+      "step": 4949
+    },
+    {
+      "epoch": 2.5924567836563646,
+      "loss_partial": 0.055908203125,
+      "step": 4949
+    },
+    {
+      "epoch": 2.5924567836563646,
+      "loss_partial": 0.04833984375,
+      "step": 4949
+    },
+    {
+      "epoch": 2.5929806181246726,
+      "grad_norm": 1.0004894733428955,
+      "learning_rate": 6.642402802533048e-06,
+      "loss": 0.1023,
+      "step": 4950
+    },
+    {
+      "epoch": 2.5929806181246726,
+      "loss_partial": 0.054443359375,
+      "step": 4950
+    },
+    {
+      "epoch": 2.5929806181246726,
+      "loss_partial": 0.057373046875,
+      "step": 4950
+    },
+    {
+      "epoch": 2.5935044525929807,
+      "grad_norm": 1.4066777229309082,
+      "learning_rate": 6.638096960473906e-06,
+      "loss": 0.1081,
+      "step": 4951
+    },
+    {
+      "epoch": 2.5935044525929807,
+      "loss_partial": 0.04931640625,
+      "step": 4951
+    },
+    {
+      "epoch": 2.5935044525929807,
+      "loss_partial": 0.058837890625,
+      "step": 4951
+    },
+    {
+      "epoch": 2.5940282870612887,
+      "grad_norm": 1.2474950551986694,
+      "learning_rate": 6.63379182103171e-06,
+      "loss": 0.0984,
+      "step": 4952
+    },
+    {
+      "epoch": 2.5940282870612887,
+      "loss_partial": 0.062255859375,
+      "step": 4952
+    },
+    {
+      "epoch": 2.5940282870612887,
+      "loss_partial": 0.06494140625,
+      "step": 4952
+    },
+    {
+      "epoch": 2.594552121529597,
+      "grad_norm": 1.6403577327728271,
+      "learning_rate": 6.629487385106206e-06,
+      "loss": 0.1039,
+      "step": 4953
+    },
+    {
+      "epoch": 2.594552121529597,
+      "loss_partial": 0.052490234375,
+      "step": 4953
+    },
+    {
+      "epoch": 2.594552121529597,
+      "loss_partial": 0.055908203125,
+      "step": 4953
+    },
+    {
+      "epoch": 2.5950759559979044,
+      "grad_norm": 1.2427513599395752,
+      "learning_rate": 6.6251836535969986e-06,
+      "loss": 0.0991,
+      "step": 4954
+    },
+    {
+      "epoch": 2.5950759559979044,
+      "loss_partial": 0.052001953125,
+      "step": 4954
+    },
+    {
+      "epoch": 2.5950759559979044,
+      "loss_partial": 0.046875,
+      "step": 4954
+    },
+    {
+      "epoch": 2.5955997904662125,
+      "grad_norm": 1.2874634265899658,
+      "learning_rate": 6.620880627403536e-06,
+      "loss": 0.101,
+      "step": 4955
+    },
+    {
+      "epoch": 2.5955997904662125,
+      "loss_partial": 0.039306640625,
+      "step": 4955
+    },
+    {
+      "epoch": 2.5955997904662125,
+      "loss_partial": 0.048828125,
+      "step": 4955
+    },
+    {
+      "epoch": 2.5961236249345205,
+      "grad_norm": 1.4362059831619263,
+      "learning_rate": 6.616578307425128e-06,
+      "loss": 0.0966,
+      "step": 4956
+    },
+    {
+      "epoch": 2.5961236249345205,
+      "loss_partial": 0.0498046875,
+      "step": 4956
+    },
+    {
+      "epoch": 2.5961236249345205,
+      "loss_partial": 0.04833984375,
+      "step": 4956
+    },
+    {
+      "epoch": 2.5966474594028286,
+      "grad_norm": 1.3712788820266724,
+      "learning_rate": 6.612276694560927e-06,
+      "loss": 0.0965,
+      "step": 4957
+    },
+    {
+      "epoch": 2.5966474594028286,
+      "loss_partial": 0.0556640625,
+      "step": 4957
+    },
+    {
+      "epoch": 2.5966474594028286,
+      "loss_partial": 0.05126953125,
+      "step": 4957
+    },
+    {
+      "epoch": 2.5971712938711367,
+      "grad_norm": 1.7709980010986328,
+      "learning_rate": 6.607975789709953e-06,
+      "loss": 0.1017,
+      "step": 4958
+    },
+    {
+      "epoch": 2.5971712938711367,
+      "loss_partial": 0.04443359375,
+      "step": 4958
+    },
+    {
+      "epoch": 2.5971712938711367,
+      "loss_partial": 0.052490234375,
+      "step": 4958
+    },
+    {
+      "epoch": 2.5976951283394447,
+      "grad_norm": 1.4813032150268555,
+      "learning_rate": 6.6036755937710616e-06,
+      "loss": 0.1001,
+      "step": 4959
+    },
+    {
+      "epoch": 2.5976951283394447,
+      "loss_partial": 0.051513671875,
+      "step": 4959
+    },
+    {
+      "epoch": 2.5976951283394447,
+      "loss_partial": 0.056640625,
+      "step": 4959
+    },
+    {
+      "epoch": 2.5982189628077528,
+      "grad_norm": 1.2548625469207764,
+      "learning_rate": 6.599376107642972e-06,
+      "loss": 0.1012,
+      "step": 4960
+    },
+    {
+      "epoch": 2.5982189628077528,
+      "loss_partial": 0.05029296875,
+      "step": 4960
+    },
+    {
+      "epoch": 2.5982189628077528,
+      "loss_partial": 0.053955078125,
+      "step": 4960
+    },
+    {
+      "epoch": 2.598742797276061,
+      "grad_norm": 1.6192189455032349,
+      "learning_rate": 6.59507733222424e-06,
+      "loss": 0.1027,
+      "step": 4961
+    },
+    {
+      "epoch": 2.598742797276061,
+      "loss_partial": 0.05126953125,
+      "step": 4961
+    },
+    {
+      "epoch": 2.598742797276061,
+      "loss_partial": 0.053466796875,
+      "step": 4961
+    },
+    {
+      "epoch": 2.599266631744369,
+      "grad_norm": 1.502748966217041,
+      "learning_rate": 6.590779268413294e-06,
+      "loss": 0.1035,
+      "step": 4962
+    },
+    {
+      "epoch": 2.599266631744369,
+      "loss_partial": 0.05615234375,
+      "step": 4962
+    },
+    {
+      "epoch": 2.599266631744369,
+      "loss_partial": 0.048828125,
+      "step": 4962
+    },
+    {
+      "epoch": 2.599790466212677,
+      "grad_norm": 1.611935019493103,
+      "learning_rate": 6.586481917108401e-06,
+      "loss": 0.1003,
+      "step": 4963
+    },
+    {
+      "epoch": 2.599790466212677,
+      "loss_partial": 0.05078125,
+      "step": 4963
+    },
+    {
+      "epoch": 2.599790466212677,
+      "loss_partial": 0.053955078125,
+      "step": 4963
+    },
+    {
+      "epoch": 2.600314300680985,
+      "grad_norm": 1.7115532159805298,
+      "learning_rate": 6.582185279207677e-06,
+      "loss": 0.1005,
+      "step": 4964
+    },
+    {
+      "epoch": 2.600314300680985,
+      "loss_partial": 0.048828125,
+      "step": 4964
+    },
+    {
+      "epoch": 2.600314300680985,
+      "loss_partial": 0.048583984375,
+      "step": 4964
+    },
+    {
+      "epoch": 2.600838135149293,
+      "grad_norm": 1.465645432472229,
+      "learning_rate": 6.577889355609093e-06,
+      "loss": 0.1053,
+      "step": 4965
+    },
+    {
+      "epoch": 2.600838135149293,
+      "loss_partial": 0.05224609375,
+      "step": 4965
+    },
+    {
+      "epoch": 2.600838135149293,
+      "loss_partial": 0.046630859375,
+      "step": 4965
+    },
+    {
+      "epoch": 2.6013619696176007,
+      "grad_norm": 1.9375065565109253,
+      "learning_rate": 6.573594147210473e-06,
+      "loss": 0.1005,
+      "step": 4966
+    },
+    {
+      "epoch": 2.6013619696176007,
+      "loss_partial": 0.054931640625,
+      "step": 4966
+    },
+    {
+      "epoch": 2.6013619696176007,
+      "loss_partial": 0.0478515625,
+      "step": 4966
+    },
+    {
+      "epoch": 2.6018858040859087,
+      "grad_norm": 1.6786726713180542,
+      "learning_rate": 6.569299654909489e-06,
+      "loss": 0.0981,
+      "step": 4967
+    },
+    {
+      "epoch": 2.6018858040859087,
+      "loss_partial": 0.04443359375,
+      "step": 4967
+    },
+    {
+      "epoch": 2.6018858040859087,
+      "loss_partial": 0.04052734375,
+      "step": 4967
+    },
+    {
+      "epoch": 2.602409638554217,
+      "grad_norm": 1.066249132156372,
+      "learning_rate": 6.565005879603663e-06,
+      "loss": 0.0991,
+      "step": 4968
+    },
+    {
+      "epoch": 2.602409638554217,
+      "loss_partial": 0.06396484375,
+      "step": 4968
+    },
+    {
+      "epoch": 2.602409638554217,
+      "loss_partial": 0.051513671875,
+      "step": 4968
+    },
+    {
+      "epoch": 2.602933473022525,
+      "grad_norm": 1.9171593189239502,
+      "learning_rate": 6.56071282219036e-06,
+      "loss": 0.1001,
+      "step": 4969
+    },
+    {
+      "epoch": 2.602933473022525,
+      "loss_partial": 0.041748046875,
+      "step": 4969
+    },
+    {
+      "epoch": 2.602933473022525,
+      "loss_partial": 0.046875,
+      "step": 4969
+    },
+    {
+      "epoch": 2.603457307490833,
+      "grad_norm": 1.2695817947387695,
+      "learning_rate": 6.556420483566817e-06,
+      "loss": 0.1,
+      "step": 4970
+    },
+    {
+      "epoch": 2.603457307490833,
+      "loss_partial": 0.04541015625,
+      "step": 4970
+    },
+    {
+      "epoch": 2.603457307490833,
+      "loss_partial": 0.048828125,
+      "step": 4970
+    },
+    {
+      "epoch": 2.603981141959141,
+      "grad_norm": 1.5288066864013672,
+      "learning_rate": 6.552128864630096e-06,
+      "loss": 0.1015,
+      "step": 4971
+    },
+    {
+      "epoch": 2.603981141959141,
+      "loss_partial": 0.04736328125,
+      "step": 4971
+    },
+    {
+      "epoch": 2.603981141959141,
+      "loss_partial": 0.054443359375,
+      "step": 4971
+    },
+    {
+      "epoch": 2.604504976427449,
+      "grad_norm": 1.677095651626587,
+      "learning_rate": 6.547837966277122e-06,
+      "loss": 0.1027,
+      "step": 4972
+    },
+    {
+      "epoch": 2.604504976427449,
+      "loss_partial": 0.0517578125,
+      "step": 4972
+    },
+    {
+      "epoch": 2.604504976427449,
+      "loss_partial": 0.044921875,
+      "step": 4972
+    },
+    {
+      "epoch": 2.6050288108957567,
+      "grad_norm": 1.7671376466751099,
+      "learning_rate": 6.543547789404663e-06,
+      "loss": 0.1017,
+      "step": 4973
+    },
+    {
+      "epoch": 2.6050288108957567,
+      "loss_partial": 0.04833984375,
+      "step": 4973
+    },
+    {
+      "epoch": 2.6050288108957567,
+      "loss_partial": 0.049560546875,
+      "step": 4973
+    },
+    {
+      "epoch": 2.6055526453640647,
+      "grad_norm": 1.4981157779693604,
+      "learning_rate": 6.539258334909344e-06,
+      "loss": 0.0994,
+      "step": 4974
+    },
+    {
+      "epoch": 2.6055526453640647,
+      "loss_partial": 0.04541015625,
+      "step": 4974
+    },
+    {
+      "epoch": 2.6055526453640647,
+      "loss_partial": 0.06005859375,
+      "step": 4974
+    },
+    {
+      "epoch": 2.606076479832373,
+      "grad_norm": 1.5952342748641968,
+      "learning_rate": 6.534969603687633e-06,
+      "loss": 0.1036,
+      "step": 4975
+    },
+    {
+      "epoch": 2.606076479832373,
+      "loss_partial": 0.0625,
+      "step": 4975
+    },
+    {
+      "epoch": 2.606076479832373,
+      "loss_partial": 0.0517578125,
+      "step": 4975
+    },
+    {
+      "epoch": 2.606600314300681,
+      "grad_norm": 1.9029772281646729,
+      "learning_rate": 6.530681596635849e-06,
+      "loss": 0.1018,
+      "step": 4976
+    },
+    {
+      "epoch": 2.606600314300681,
+      "loss_partial": 0.054931640625,
+      "step": 4976
+    },
+    {
+      "epoch": 2.606600314300681,
+      "loss_partial": 0.05712890625,
+      "step": 4976
+    },
+    {
+      "epoch": 2.607124148768989,
+      "grad_norm": 1.2941737174987793,
+      "learning_rate": 6.526394314650153e-06,
+      "loss": 0.099,
+      "step": 4977
+    },
+    {
+      "epoch": 2.607124148768989,
+      "loss_partial": 0.050048828125,
+      "step": 4977
+    },
+    {
+      "epoch": 2.607124148768989,
+      "loss_partial": 0.052001953125,
+      "step": 4977
+    },
+    {
+      "epoch": 2.607647983237297,
+      "grad_norm": 1.4756627082824707,
+      "learning_rate": 6.522107758626569e-06,
+      "loss": 0.096,
+      "step": 4978
+    },
+    {
+      "epoch": 2.607647983237297,
+      "loss_partial": 0.0546875,
+      "step": 4978
+    },
+    {
+      "epoch": 2.607647983237297,
+      "loss_partial": 0.04638671875,
+      "step": 4978
+    },
+    {
+      "epoch": 2.608171817705605,
+      "grad_norm": 1.2057651281356812,
+      "learning_rate": 6.517821929460957e-06,
+      "loss": 0.1025,
+      "step": 4979
+    },
+    {
+      "epoch": 2.608171817705605,
+      "loss_partial": 0.05126953125,
+      "step": 4979
+    },
+    {
+      "epoch": 2.608171817705605,
+      "loss_partial": 0.04931640625,
+      "step": 4979
+    },
+    {
+      "epoch": 2.608695652173913,
+      "grad_norm": 1.2002522945404053,
+      "learning_rate": 6.513536828049028e-06,
+      "loss": 0.0991,
+      "step": 4980
+    },
+    {
+      "epoch": 2.608695652173913,
+      "loss_partial": 0.053955078125,
+      "step": 4980
+    },
+    {
+      "epoch": 2.608695652173913,
+      "loss_partial": 0.04931640625,
+      "step": 4980
+    },
+    {
+      "epoch": 2.609219486642221,
+      "grad_norm": 1.6792365312576294,
+      "learning_rate": 6.509252455286337e-06,
+      "loss": 0.1013,
+      "step": 4981
+    },
+    {
+      "epoch": 2.609219486642221,
+      "loss_partial": 0.0517578125,
+      "step": 4981
+    },
+    {
+      "epoch": 2.609219486642221,
+      "loss_partial": 0.054443359375,
+      "step": 4981
+    },
+    {
+      "epoch": 2.609743321110529,
+      "grad_norm": 1.6510201692581177,
+      "learning_rate": 6.504968812068302e-06,
+      "loss": 0.1017,
+      "step": 4982
+    },
+    {
+      "epoch": 2.609743321110529,
+      "loss_partial": 0.064453125,
+      "step": 4982
+    },
+    {
+      "epoch": 2.609743321110529,
+      "loss_partial": 0.055419921875,
+      "step": 4982
+    },
+    {
+      "epoch": 2.6102671555788373,
+      "grad_norm": 1.7226994037628174,
+      "learning_rate": 6.500685899290172e-06,
+      "loss": 0.105,
+      "step": 4983
+    },
+    {
+      "epoch": 2.6102671555788373,
+      "loss_partial": 0.03759765625,
+      "step": 4983
+    },
+    {
+      "epoch": 2.6102671555788373,
+      "loss_partial": 0.047119140625,
+      "step": 4983
+    },
+    {
+      "epoch": 2.6107909900471453,
+      "grad_norm": 2.2776780128479004,
+      "learning_rate": 6.496403717847051e-06,
+      "loss": 0.099,
+      "step": 4984
+    },
+    {
+      "epoch": 2.6107909900471453,
+      "loss_partial": 0.057373046875,
+      "step": 4984
+    },
+    {
+      "epoch": 2.6107909900471453,
+      "loss_partial": 0.050537109375,
+      "step": 4984
+    },
+    {
+      "epoch": 2.6113148245154534,
+      "grad_norm": 1.5841161012649536,
+      "learning_rate": 6.492122268633885e-06,
+      "loss": 0.1034,
+      "step": 4985
+    },
+    {
+      "epoch": 2.6113148245154534,
+      "loss_partial": 0.048828125,
+      "step": 4985
+    },
+    {
+      "epoch": 2.6113148245154534,
+      "loss_partial": 0.04541015625,
+      "step": 4985
+    },
+    {
+      "epoch": 2.611838658983761,
+      "grad_norm": 1.6230340003967285,
+      "learning_rate": 6.487841552545474e-06,
+      "loss": 0.1014,
+      "step": 4986
+    },
+    {
+      "epoch": 2.611838658983761,
+      "loss_partial": 0.0537109375,
+      "step": 4986
+    },
+    {
+      "epoch": 2.611838658983761,
+      "loss_partial": 0.050048828125,
+      "step": 4986
+    },
+    {
+      "epoch": 2.612362493452069,
+      "grad_norm": 2.1374878883361816,
+      "learning_rate": 6.483561570476461e-06,
+      "loss": 0.1041,
+      "step": 4987
+    },
+    {
+      "epoch": 2.612362493452069,
+      "loss_partial": 0.047607421875,
+      "step": 4987
+    },
+    {
+      "epoch": 2.612362493452069,
+      "loss_partial": 0.05712890625,
+      "step": 4987
+    },
+    {
+      "epoch": 2.612886327920377,
+      "grad_norm": 1.4946489334106445,
+      "learning_rate": 6.479282323321334e-06,
+      "loss": 0.1065,
+      "step": 4988
+    },
+    {
+      "epoch": 2.612886327920377,
+      "loss_partial": 0.039306640625,
+      "step": 4988
+    },
+    {
+      "epoch": 2.612886327920377,
+      "loss_partial": 0.048095703125,
+      "step": 4988
+    },
+    {
+      "epoch": 2.613410162388685,
+      "grad_norm": 1.3665833473205566,
+      "learning_rate": 6.475003811974429e-06,
+      "loss": 0.0984,
+      "step": 4989
+    },
+    {
+      "epoch": 2.613410162388685,
+      "loss_partial": 0.048828125,
+      "step": 4989
+    },
+    {
+      "epoch": 2.613410162388685,
+      "loss_partial": 0.051025390625,
+      "step": 4989
+    },
+    {
+      "epoch": 2.6139339968569932,
+      "grad_norm": 1.234934687614441,
+      "learning_rate": 6.47072603732993e-06,
+      "loss": 0.1001,
+      "step": 4990
+    },
+    {
+      "epoch": 2.6139339968569932,
+      "loss_partial": 0.06494140625,
+      "step": 4990
+    },
+    {
+      "epoch": 2.6139339968569932,
+      "loss_partial": 0.044677734375,
+      "step": 4990
+    },
+    {
+      "epoch": 2.6144578313253013,
+      "grad_norm": 1.3547977209091187,
+      "learning_rate": 6.466449000281866e-06,
+      "loss": 0.0993,
+      "step": 4991
+    },
+    {
+      "epoch": 2.6144578313253013,
+      "loss_partial": 0.06494140625,
+      "step": 4991
+    },
+    {
+      "epoch": 2.6144578313253013,
+      "loss_partial": 0.056884765625,
+      "step": 4991
+    },
+    {
+      "epoch": 2.6149816657936094,
+      "grad_norm": 1.3021423816680908,
+      "learning_rate": 6.4621727017241095e-06,
+      "loss": 0.1031,
+      "step": 4992
+    },
+    {
+      "epoch": 2.6149816657936094,
+      "loss_partial": 0.050048828125,
+      "step": 4992
+    },
+    {
+      "epoch": 2.6149816657936094,
+      "loss_partial": 0.047607421875,
+      "step": 4992
+    },
+    {
+      "epoch": 2.615505500261917,
+      "grad_norm": 1.7061759233474731,
+      "learning_rate": 6.457897142550377e-06,
+      "loss": 0.0963,
+      "step": 4993
+    },
+    {
+      "epoch": 2.615505500261917,
+      "loss_partial": 0.059326171875,
+      "step": 4993
+    },
+    {
+      "epoch": 2.615505500261917,
+      "loss_partial": 0.04833984375,
+      "step": 4993
+    },
+    {
+      "epoch": 2.616029334730225,
+      "grad_norm": 1.7478082180023193,
+      "learning_rate": 6.4536223236542424e-06,
+      "loss": 0.1013,
+      "step": 4994
+    },
+    {
+      "epoch": 2.616029334730225,
+      "loss_partial": 0.056396484375,
+      "step": 4994
+    },
+    {
+      "epoch": 2.616029334730225,
+      "loss_partial": 0.053955078125,
+      "step": 4994
+    },
+    {
+      "epoch": 2.616553169198533,
+      "grad_norm": 1.1667578220367432,
+      "learning_rate": 6.449348245929112e-06,
+      "loss": 0.103,
+      "step": 4995
+    },
+    {
+      "epoch": 2.616553169198533,
+      "loss_partial": 0.04248046875,
+      "step": 4995
+    },
+    {
+      "epoch": 2.616553169198533,
+      "loss_partial": 0.047119140625,
+      "step": 4995
+    },
+    {
+      "epoch": 2.617077003666841,
+      "grad_norm": 1.4552414417266846,
+      "learning_rate": 6.445074910268241e-06,
+      "loss": 0.1007,
+      "step": 4996
+    },
+    {
+      "epoch": 2.617077003666841,
+      "loss_partial": 0.052978515625,
+      "step": 4996
+    },
+    {
+      "epoch": 2.617077003666841,
+      "loss_partial": 0.055419921875,
+      "step": 4996
+    },
+    {
+      "epoch": 2.617600838135149,
+      "grad_norm": 1.2722892761230469,
+      "learning_rate": 6.44080231756473e-06,
+      "loss": 0.1019,
+      "step": 4997
+    },
+    {
+      "epoch": 2.617600838135149,
+      "loss_partial": 0.052001953125,
+      "step": 4997
+    },
+    {
+      "epoch": 2.617600838135149,
+      "loss_partial": 0.048095703125,
+      "step": 4997
+    },
+    {
+      "epoch": 2.6181246726034573,
+      "grad_norm": 1.4199885129928589,
+      "learning_rate": 6.4365304687115256e-06,
+      "loss": 0.1043,
+      "step": 4998
+    },
+    {
+      "epoch": 2.6181246726034573,
+      "loss_partial": 0.05517578125,
+      "step": 4998
+    },
+    {
+      "epoch": 2.6181246726034573,
+      "loss_partial": 0.06591796875,
+      "step": 4998
+    },
+    {
+      "epoch": 2.6186485070717653,
+      "grad_norm": 1.1861895322799683,
+      "learning_rate": 6.432259364601419e-06,
+      "loss": 0.1004,
+      "step": 4999
+    },
+    {
+      "epoch": 2.6186485070717653,
+      "loss_partial": 0.057373046875,
+      "step": 4999
+    },
+    {
+      "epoch": 2.6186485070717653,
+      "loss_partial": 0.054443359375,
+      "step": 4999
+    },
+    {
+      "epoch": 2.6191723415400734,
+      "grad_norm": 1.5973321199417114,
+      "learning_rate": 6.4279890061270425e-06,
+      "loss": 0.0994,
+      "step": 5000
+    },
+    {
+      "epoch": 2.6191723415400734,
+      "loss_partial": 0.0390625,
+      "step": 5000
+    },
+    {
+      "epoch": 2.6191723415400734,
+      "loss_partial": 0.050048828125,
+      "step": 5000
+    },
+    {
+      "epoch": 2.6196961760083814,
+      "grad_norm": 1.3694350719451904,
+      "learning_rate": 6.423719394180872e-06,
+      "loss": 0.1039,
+      "step": 5001
+    },
+    {
+      "epoch": 2.6196961760083814,
+      "loss_partial": 0.04443359375,
+      "step": 5001
+    },
+    {
+      "epoch": 2.6196961760083814,
+      "loss_partial": 0.05078125,
+      "step": 5001
+    },
+    {
+      "epoch": 2.6202200104766895,
+      "grad_norm": 1.2993440628051758,
+      "learning_rate": 6.419450529655236e-06,
+      "loss": 0.1007,
+      "step": 5002
+    },
+    {
+      "epoch": 2.6202200104766895,
+      "loss_partial": 0.04931640625,
+      "step": 5002
+    },
+    {
+      "epoch": 2.6202200104766895,
+      "loss_partial": 0.050537109375,
+      "step": 5002
+    },
+    {
+      "epoch": 2.6207438449449976,
+      "grad_norm": 1.0663962364196777,
+      "learning_rate": 6.415182413442298e-06,
+      "loss": 0.0995,
+      "step": 5003
+    },
+    {
+      "epoch": 2.6207438449449976,
+      "loss_partial": 0.051513671875,
+      "step": 5003
+    },
+    {
+      "epoch": 2.6207438449449976,
+      "loss_partial": 0.05517578125,
+      "step": 5003
+    },
+    {
+      "epoch": 2.6212676794133056,
+      "grad_norm": 1.5554026365280151,
+      "learning_rate": 6.410915046434068e-06,
+      "loss": 0.1003,
+      "step": 5004
+    },
+    {
+      "epoch": 2.6212676794133056,
+      "loss_partial": 0.04443359375,
+      "step": 5004
+    },
+    {
+      "epoch": 2.6212676794133056,
+      "loss_partial": 0.052734375,
+      "step": 5004
+    },
+    {
+      "epoch": 2.6217915138816132,
+      "grad_norm": 1.4697322845458984,
+      "learning_rate": 6.406648429522397e-06,
+      "loss": 0.1004,
+      "step": 5005
+    },
+    {
+      "epoch": 2.6217915138816132,
+      "loss_partial": 0.046875,
+      "step": 5005
+    },
+    {
+      "epoch": 2.6217915138816132,
+      "loss_partial": 0.053466796875,
+      "step": 5005
+    },
+    {
+      "epoch": 2.6223153483499213,
+      "grad_norm": 1.3017001152038574,
+      "learning_rate": 6.402382563598981e-06,
+      "loss": 0.1034,
+      "step": 5006
+    },
+    {
+      "epoch": 2.6223153483499213,
+      "loss_partial": 0.05322265625,
+      "step": 5006
+    },
+    {
+      "epoch": 2.6223153483499213,
+      "loss_partial": 0.04833984375,
+      "step": 5006
+    },
+    {
+      "epoch": 2.6228391828182294,
+      "grad_norm": 1.3957303762435913,
+      "learning_rate": 6.398117449555366e-06,
+      "loss": 0.102,
+      "step": 5007
+    },
+    {
+      "epoch": 2.6228391828182294,
+      "loss_partial": 0.058837890625,
+      "step": 5007
+    },
+    {
+      "epoch": 2.6228391828182294,
+      "loss_partial": 0.051513671875,
+      "step": 5007
+    },
+    {
+      "epoch": 2.6233630172865374,
+      "grad_norm": 1.6179311275482178,
+      "learning_rate": 6.393853088282931e-06,
+      "loss": 0.1025,
+      "step": 5008
+    },
+    {
+      "epoch": 2.6233630172865374,
+      "loss_partial": 0.05078125,
+      "step": 5008
+    },
+    {
+      "epoch": 2.6233630172865374,
+      "loss_partial": 0.0478515625,
+      "step": 5008
+    },
+    {
+      "epoch": 2.6238868517548455,
+      "grad_norm": 1.2145930528640747,
+      "learning_rate": 6.389589480672902e-06,
+      "loss": 0.1013,
+      "step": 5009
+    },
+    {
+      "epoch": 2.6238868517548455,
+      "loss_partial": 0.055419921875,
+      "step": 5009
+    },
+    {
+      "epoch": 2.6238868517548455,
+      "loss_partial": 0.054443359375,
+      "step": 5009
+    },
+    {
+      "epoch": 2.6244106862231535,
+      "grad_norm": 1.2930710315704346,
+      "learning_rate": 6.385326627616342e-06,
+      "loss": 0.1011,
+      "step": 5010
+    },
+    {
+      "epoch": 2.6244106862231535,
+      "loss_partial": 0.052490234375,
+      "step": 5010
+    },
+    {
+      "epoch": 2.6244106862231535,
+      "loss_partial": 0.05029296875,
+      "step": 5010
+    },
+    {
+      "epoch": 2.6249345206914616,
+      "grad_norm": 1.4911309480667114,
+      "learning_rate": 6.381064530004166e-06,
+      "loss": 0.1045,
+      "step": 5011
+    },
+    {
+      "epoch": 2.6249345206914616,
+      "loss_partial": 0.044189453125,
+      "step": 5011
+    },
+    {
+      "epoch": 2.6249345206914616,
+      "loss_partial": 0.048828125,
+      "step": 5011
+    },
+    {
+      "epoch": 2.6254583551597697,
+      "grad_norm": 1.363584280014038,
+      "learning_rate": 6.376803188727125e-06,
+      "loss": 0.1026,
+      "step": 5012
+    },
+    {
+      "epoch": 2.6254583551597697,
+      "loss_partial": 0.053955078125,
+      "step": 5012
+    },
+    {
+      "epoch": 2.6254583551597697,
+      "loss_partial": 0.068359375,
+      "step": 5012
+    },
+    {
+      "epoch": 2.6259821896280773,
+      "grad_norm": 1.2457737922668457,
+      "learning_rate": 6.372542604675812e-06,
+      "loss": 0.1013,
+      "step": 5013
+    },
+    {
+      "epoch": 2.6259821896280773,
+      "loss_partial": 0.05029296875,
+      "step": 5013
+    },
+    {
+      "epoch": 2.6259821896280773,
+      "loss_partial": 0.044189453125,
+      "step": 5013
+    },
+    {
+      "epoch": 2.6265060240963853,
+      "grad_norm": 1.5312104225158691,
+      "learning_rate": 6.368282778740662e-06,
+      "loss": 0.102,
+      "step": 5014
+    },
+    {
+      "epoch": 2.6265060240963853,
+      "loss_partial": 0.048828125,
+      "step": 5014
+    },
+    {
+      "epoch": 2.6265060240963853,
+      "loss_partial": 0.047607421875,
+      "step": 5014
+    },
+    {
+      "epoch": 2.6270298585646934,
+      "grad_norm": 1.197471022605896,
+      "learning_rate": 6.364023711811955e-06,
+      "loss": 0.0988,
+      "step": 5015
+    },
+    {
+      "epoch": 2.6270298585646934,
+      "loss_partial": 0.04833984375,
+      "step": 5015
+    },
+    {
+      "epoch": 2.6270298585646934,
+      "loss_partial": 0.053955078125,
+      "step": 5015
+    },
+    {
+      "epoch": 2.6275536930330015,
+      "grad_norm": 1.4917184114456177,
+      "learning_rate": 6.359765404779807e-06,
+      "loss": 0.106,
+      "step": 5016
+    },
+    {
+      "epoch": 2.6275536930330015,
+      "loss_partial": 0.0517578125,
+      "step": 5016
+    },
+    {
+      "epoch": 2.6275536930330015,
+      "loss_partial": 0.044189453125,
+      "step": 5016
+    },
+    {
+      "epoch": 2.6280775275013095,
+      "grad_norm": 1.7528951168060303,
+      "learning_rate": 6.35550785853418e-06,
+      "loss": 0.1023,
+      "step": 5017
+    },
+    {
+      "epoch": 2.6280775275013095,
+      "loss_partial": 0.06005859375,
+      "step": 5017
+    },
+    {
+      "epoch": 2.6280775275013095,
+      "loss_partial": 0.041015625,
+      "step": 5017
+    },
+    {
+      "epoch": 2.6286013619696176,
+      "grad_norm": 1.7427102327346802,
+      "learning_rate": 6.35125107396487e-06,
+      "loss": 0.104,
+      "step": 5018
+    },
+    {
+      "epoch": 2.6286013619696176,
+      "loss_partial": 0.051025390625,
+      "step": 5018
+    },
+    {
+      "epoch": 2.6286013619696176,
+      "loss_partial": 0.052734375,
+      "step": 5018
+    },
+    {
+      "epoch": 2.6291251964379256,
+      "grad_norm": 1.2774176597595215,
+      "learning_rate": 6.346995051961525e-06,
+      "loss": 0.099,
+      "step": 5019
+    },
+    {
+      "epoch": 2.6291251964379256,
+      "loss_partial": 0.0419921875,
+      "step": 5019
+    },
+    {
+      "epoch": 2.6291251964379256,
+      "loss_partial": 0.05126953125,
+      "step": 5019
+    },
+    {
+      "epoch": 2.6296490309062337,
+      "grad_norm": 1.311876654624939,
+      "learning_rate": 6.342739793413626e-06,
+      "loss": 0.1056,
+      "step": 5020
+    },
+    {
+      "epoch": 2.6296490309062337,
+      "loss_partial": 0.052490234375,
+      "step": 5020
+    },
+    {
+      "epoch": 2.6296490309062337,
+      "loss_partial": 0.0439453125,
+      "step": 5020
+    },
+    {
+      "epoch": 2.6301728653745418,
+      "grad_norm": 1.947481393814087,
+      "learning_rate": 6.3384852992104924e-06,
+      "loss": 0.1031,
+      "step": 5021
+    },
+    {
+      "epoch": 2.6301728653745418,
+      "loss_partial": 0.04931640625,
+      "step": 5021
+    },
+    {
+      "epoch": 2.6301728653745418,
+      "loss_partial": 0.05419921875,
+      "step": 5021
+    },
+    {
+      "epoch": 2.63069669984285,
+      "grad_norm": 1.3890984058380127,
+      "learning_rate": 6.3342315702412894e-06,
+      "loss": 0.1006,
+      "step": 5022
+    },
+    {
+      "epoch": 2.63069669984285,
+      "loss_partial": 0.048095703125,
+      "step": 5022
+    },
+    {
+      "epoch": 2.63069669984285,
+      "loss_partial": 0.0537109375,
+      "step": 5022
+    },
+    {
+      "epoch": 2.631220534311158,
+      "grad_norm": 1.9813169240951538,
+      "learning_rate": 6.329978607395022e-06,
+      "loss": 0.1012,
+      "step": 5023
+    },
+    {
+      "epoch": 2.631220534311158,
+      "loss_partial": 0.050537109375,
+      "step": 5023
+    },
+    {
+      "epoch": 2.631220534311158,
+      "loss_partial": 0.0478515625,
+      "step": 5023
+    },
+    {
+      "epoch": 2.631744368779466,
+      "grad_norm": 1.9009305238723755,
+      "learning_rate": 6.325726411560529e-06,
+      "loss": 0.1029,
+      "step": 5024
+    },
+    {
+      "epoch": 2.631744368779466,
+      "loss_partial": 0.053466796875,
+      "step": 5024
+    },
+    {
+      "epoch": 2.631744368779466,
+      "loss_partial": 0.054931640625,
+      "step": 5024
+    },
+    {
+      "epoch": 2.6322682032477736,
+      "grad_norm": 1.6250214576721191,
+      "learning_rate": 6.321474983626497e-06,
+      "loss": 0.1016,
+      "step": 5025
+    },
+    {
+      "epoch": 2.6322682032477736,
+      "loss_partial": 0.049072265625,
+      "step": 5025
+    },
+    {
+      "epoch": 2.6322682032477736,
+      "loss_partial": 0.054931640625,
+      "step": 5025
+    },
+    {
+      "epoch": 2.6327920377160816,
+      "grad_norm": 1.4481147527694702,
+      "learning_rate": 6.317224324481442e-06,
+      "loss": 0.1022,
+      "step": 5026
+    },
+    {
+      "epoch": 2.6327920377160816,
+      "loss_partial": 0.056884765625,
+      "step": 5026
+    },
+    {
+      "epoch": 2.6327920377160816,
+      "loss_partial": 0.049072265625,
+      "step": 5026
+    },
+    {
+      "epoch": 2.6333158721843897,
+      "grad_norm": 1.8216729164123535,
+      "learning_rate": 6.312974435013732e-06,
+      "loss": 0.0974,
+      "step": 5027
+    },
+    {
+      "epoch": 2.6333158721843897,
+      "loss_partial": 0.0478515625,
+      "step": 5027
+    },
+    {
+      "epoch": 2.6333158721843897,
+      "loss_partial": 0.052978515625,
+      "step": 5027
+    },
+    {
+      "epoch": 2.6338397066526977,
+      "grad_norm": 1.1301440000534058,
+      "learning_rate": 6.3087253161115654e-06,
+      "loss": 0.1007,
+      "step": 5028
+    },
+    {
+      "epoch": 2.6338397066526977,
+      "loss_partial": 0.046142578125,
+      "step": 5028
+    },
+    {
+      "epoch": 2.6338397066526977,
+      "loss_partial": 0.064453125,
+      "step": 5028
+    },
+    {
+      "epoch": 2.634363541121006,
+      "grad_norm": 1.131736159324646,
+      "learning_rate": 6.304476968662983e-06,
+      "loss": 0.099,
+      "step": 5029
+    },
+    {
+      "epoch": 2.634363541121006,
+      "loss_partial": 0.046630859375,
+      "step": 5029
+    },
+    {
+      "epoch": 2.634363541121006,
+      "loss_partial": 0.047119140625,
+      "step": 5029
+    },
+    {
+      "epoch": 2.634887375589314,
+      "grad_norm": 1.28972327709198,
+      "learning_rate": 6.3002293935558565e-06,
+      "loss": 0.0985,
+      "step": 5030
+    },
+    {
+      "epoch": 2.634887375589314,
+      "loss_partial": 0.04638671875,
+      "step": 5030
+    },
+    {
+      "epoch": 2.634887375589314,
+      "loss_partial": 0.040771484375,
+      "step": 5030
+    },
+    {
+      "epoch": 2.635411210057622,
+      "grad_norm": 1.6172621250152588,
+      "learning_rate": 6.295982591677911e-06,
+      "loss": 0.1001,
+      "step": 5031
+    },
+    {
+      "epoch": 2.635411210057622,
+      "loss_partial": 0.046875,
+      "step": 5031
+    },
+    {
+      "epoch": 2.635411210057622,
+      "loss_partial": 0.046630859375,
+      "step": 5031
+    },
+    {
+      "epoch": 2.63593504452593,
+      "grad_norm": 1.8282787799835205,
+      "learning_rate": 6.291736563916701e-06,
+      "loss": 0.1001,
+      "step": 5032
+    },
+    {
+      "epoch": 2.63593504452593,
+      "loss_partial": 0.049072265625,
+      "step": 5032
+    },
+    {
+      "epoch": 2.63593504452593,
+      "loss_partial": 0.047607421875,
+      "step": 5032
+    },
+    {
+      "epoch": 2.6364588789942376,
+      "grad_norm": 1.512916922569275,
+      "learning_rate": 6.287491311159617e-06,
+      "loss": 0.1021,
+      "step": 5033
+    },
+    {
+      "epoch": 2.6364588789942376,
+      "loss_partial": 0.049072265625,
+      "step": 5033
+    },
+    {
+      "epoch": 2.6364588789942376,
+      "loss_partial": 0.052978515625,
+      "step": 5033
+    },
+    {
+      "epoch": 2.6369827134625456,
+      "grad_norm": 1.8654098510742188,
+      "learning_rate": 6.283246834293889e-06,
+      "loss": 0.0999,
+      "step": 5034
+    },
+    {
+      "epoch": 2.6369827134625456,
+      "loss_partial": 0.05517578125,
+      "step": 5034
+    },
+    {
+      "epoch": 2.6369827134625456,
+      "loss_partial": 0.051513671875,
+      "step": 5034
+    },
+    {
+      "epoch": 2.6375065479308537,
+      "grad_norm": 1.4436650276184082,
+      "learning_rate": 6.279003134206591e-06,
+      "loss": 0.0996,
+      "step": 5035
+    },
+    {
+      "epoch": 2.6375065479308537,
+      "loss_partial": 0.04931640625,
+      "step": 5035
+    },
+    {
+      "epoch": 2.6375065479308537,
+      "loss_partial": 0.053466796875,
+      "step": 5035
+    },
+    {
+      "epoch": 2.6380303823991618,
+      "grad_norm": 1.9637337923049927,
+      "learning_rate": 6.27476021178463e-06,
+      "loss": 0.0999,
+      "step": 5036
+    },
+    {
+      "epoch": 2.6380303823991618,
+      "loss_partial": 0.050048828125,
+      "step": 5036
+    },
+    {
+      "epoch": 2.6380303823991618,
+      "loss_partial": 0.046630859375,
+      "step": 5036
+    },
+    {
+      "epoch": 2.63855421686747,
+      "grad_norm": 1.1558465957641602,
+      "learning_rate": 6.2705180679147455e-06,
+      "loss": 0.0993,
+      "step": 5037
+    },
+    {
+      "epoch": 2.63855421686747,
+      "loss_partial": 0.049072265625,
+      "step": 5037
+    },
+    {
+      "epoch": 2.63855421686747,
+      "loss_partial": 0.0419921875,
+      "step": 5037
+    },
+    {
+      "epoch": 2.639078051335778,
+      "grad_norm": 1.0941872596740723,
+      "learning_rate": 6.266276703483522e-06,
+      "loss": 0.1008,
+      "step": 5038
+    },
+    {
+      "epoch": 2.639078051335778,
+      "loss_partial": 0.052490234375,
+      "step": 5038
+    },
+    {
+      "epoch": 2.639078051335778,
+      "loss_partial": 0.05078125,
+      "step": 5038
+    },
+    {
+      "epoch": 2.639601885804086,
+      "grad_norm": 1.7095681428909302,
+      "learning_rate": 6.262036119377382e-06,
+      "loss": 0.102,
+      "step": 5039
+    },
+    {
+      "epoch": 2.639601885804086,
+      "loss_partial": 0.05419921875,
+      "step": 5039
+    },
+    {
+      "epoch": 2.639601885804086,
+      "loss_partial": 0.04931640625,
+      "step": 5039
+    },
+    {
+      "epoch": 2.640125720272394,
+      "grad_norm": 1.3580856323242188,
+      "learning_rate": 6.257796316482576e-06,
+      "loss": 0.102,
+      "step": 5040
+    },
+    {
+      "epoch": 2.640125720272394,
+      "loss_partial": 0.042236328125,
+      "step": 5040
+    },
+    {
+      "epoch": 2.640125720272394,
+      "loss_partial": 0.047607421875,
+      "step": 5040
+    },
+    {
+      "epoch": 2.640649554740702,
+      "grad_norm": 1.3150520324707031,
+      "learning_rate": 6.2535572956852e-06,
+      "loss": 0.099,
+      "step": 5041
+    },
+    {
+      "epoch": 2.640649554740702,
+      "loss_partial": 0.046630859375,
+      "step": 5041
+    },
+    {
+      "epoch": 2.640649554740702,
+      "loss_partial": 0.048583984375,
+      "step": 5041
+    },
+    {
+      "epoch": 2.64117338920901,
+      "grad_norm": 1.2075419425964355,
+      "learning_rate": 6.249319057871177e-06,
+      "loss": 0.1048,
+      "step": 5042
+    },
+    {
+      "epoch": 2.64117338920901,
+      "loss_partial": 0.05078125,
+      "step": 5042
+    },
+    {
+      "epoch": 2.64117338920901,
+      "loss_partial": 0.0478515625,
+      "step": 5042
+    },
+    {
+      "epoch": 2.641697223677318,
+      "grad_norm": 1.2993756532669067,
+      "learning_rate": 6.245081603926282e-06,
+      "loss": 0.0981,
+      "step": 5043
+    },
+    {
+      "epoch": 2.641697223677318,
+      "loss_partial": 0.054443359375,
+      "step": 5043
+    },
+    {
+      "epoch": 2.641697223677318,
+      "loss_partial": 0.054443359375,
+      "step": 5043
+    },
+    {
+      "epoch": 2.6422210581456262,
+      "grad_norm": 1.2954399585723877,
+      "learning_rate": 6.2408449347361125e-06,
+      "loss": 0.1003,
+      "step": 5044
+    },
+    {
+      "epoch": 2.6422210581456262,
+      "loss_partial": 0.05126953125,
+      "step": 5044
+    },
+    {
+      "epoch": 2.6422210581456262,
+      "loss_partial": 0.047607421875,
+      "step": 5044
+    },
+    {
+      "epoch": 2.642744892613934,
+      "grad_norm": 1.5458462238311768,
+      "learning_rate": 6.236609051186105e-06,
+      "loss": 0.1012,
+      "step": 5045
+    },
+    {
+      "epoch": 2.642744892613934,
+      "loss_partial": 0.043212890625,
+      "step": 5045
+    },
+    {
+      "epoch": 2.642744892613934,
+      "loss_partial": 0.0634765625,
+      "step": 5045
+    },
+    {
+      "epoch": 2.643268727082242,
+      "grad_norm": 1.8064478635787964,
+      "learning_rate": 6.232373954161532e-06,
+      "loss": 0.1053,
+      "step": 5046
+    },
+    {
+      "epoch": 2.643268727082242,
+      "loss_partial": 0.04736328125,
+      "step": 5046
+    },
+    {
+      "epoch": 2.643268727082242,
+      "loss_partial": 0.0634765625,
+      "step": 5046
+    },
+    {
+      "epoch": 2.64379256155055,
+      "grad_norm": 1.6127519607543945,
+      "learning_rate": 6.228139644547505e-06,
+      "loss": 0.1065,
+      "step": 5047
+    },
+    {
+      "epoch": 2.64379256155055,
+      "loss_partial": 0.053955078125,
+      "step": 5047
+    },
+    {
+      "epoch": 2.64379256155055,
+      "loss_partial": 0.059326171875,
+      "step": 5047
+    },
+    {
+      "epoch": 2.644316396018858,
+      "grad_norm": 1.4075621366500854,
+      "learning_rate": 6.223906123228969e-06,
+      "loss": 0.0995,
+      "step": 5048
+    },
+    {
+      "epoch": 2.644316396018858,
+      "loss_partial": 0.048095703125,
+      "step": 5048
+    },
+    {
+      "epoch": 2.644316396018858,
+      "loss_partial": 0.057861328125,
+      "step": 5048
+    },
+    {
+      "epoch": 2.644840230487166,
+      "grad_norm": 1.6154760122299194,
+      "learning_rate": 6.219673391090703e-06,
+      "loss": 0.1015,
+      "step": 5049
+    },
+    {
+      "epoch": 2.644840230487166,
+      "loss_partial": 0.043701171875,
+      "step": 5049
+    },
+    {
+      "epoch": 2.644840230487166,
+      "loss_partial": 0.0390625,
+      "step": 5049
+    },
+    {
+      "epoch": 2.645364064955474,
+      "grad_norm": 1.424925446510315,
+      "learning_rate": 6.215441449017319e-06,
+      "loss": 0.098,
+      "step": 5050
+    },
+    {
+      "epoch": 2.645364064955474,
+      "loss_partial": 0.0478515625,
+      "step": 5050
+    },
+    {
+      "epoch": 2.645364064955474,
+      "loss_partial": 0.049072265625,
+      "step": 5050
+    },
+    {
+      "epoch": 2.645887899423782,
+      "grad_norm": 1.6284409761428833,
+      "learning_rate": 6.211210297893271e-06,
+      "loss": 0.0989,
+      "step": 5051
+    },
+    {
+      "epoch": 2.645887899423782,
+      "loss_partial": 0.048583984375,
+      "step": 5051
+    },
+    {
+      "epoch": 2.645887899423782,
+      "loss_partial": 0.057373046875,
+      "step": 5051
+    },
+    {
+      "epoch": 2.64641173389209,
+      "grad_norm": 1.6954902410507202,
+      "learning_rate": 6.206979938602842e-06,
+      "loss": 0.0999,
+      "step": 5052
+    },
+    {
+      "epoch": 2.64641173389209,
+      "loss_partial": 0.0439453125,
+      "step": 5052
+    },
+    {
+      "epoch": 2.64641173389209,
+      "loss_partial": 0.056640625,
+      "step": 5052
+    },
+    {
+      "epoch": 2.646935568360398,
+      "grad_norm": 1.3802552223205566,
+      "learning_rate": 6.202750372030149e-06,
+      "loss": 0.1036,
+      "step": 5053
+    },
+    {
+      "epoch": 2.646935568360398,
+      "loss_partial": 0.056640625,
+      "step": 5053
+    },
+    {
+      "epoch": 2.646935568360398,
+      "loss_partial": 0.058837890625,
+      "step": 5053
+    },
+    {
+      "epoch": 2.647459402828706,
+      "grad_norm": 1.3033915758132935,
+      "learning_rate": 6.198521599059144e-06,
+      "loss": 0.1036,
+      "step": 5054
+    },
+    {
+      "epoch": 2.647459402828706,
+      "loss_partial": 0.0498046875,
+      "step": 5054
+    },
+    {
+      "epoch": 2.647459402828706,
+      "loss_partial": 0.047119140625,
+      "step": 5054
+    },
+    {
+      "epoch": 2.647983237297014,
+      "grad_norm": 1.7066242694854736,
+      "learning_rate": 6.194293620573621e-06,
+      "loss": 0.1007,
+      "step": 5055
+    },
+    {
+      "epoch": 2.647983237297014,
+      "loss_partial": 0.053955078125,
+      "step": 5055
+    },
+    {
+      "epoch": 2.647983237297014,
+      "loss_partial": 0.047119140625,
+      "step": 5055
+    },
+    {
+      "epoch": 2.648507071765322,
+      "grad_norm": 1.8222078084945679,
+      "learning_rate": 6.190066437457198e-06,
+      "loss": 0.1038,
+      "step": 5056
+    },
+    {
+      "epoch": 2.648507071765322,
+      "loss_partial": 0.0498046875,
+      "step": 5056
+    },
+    {
+      "epoch": 2.648507071765322,
+      "loss_partial": 0.050537109375,
+      "step": 5056
+    },
+    {
+      "epoch": 2.64903090623363,
+      "grad_norm": 1.557166337966919,
+      "learning_rate": 6.18584005059333e-06,
+      "loss": 0.0973,
+      "step": 5057
+    },
+    {
+      "epoch": 2.64903090623363,
+      "loss_partial": 0.049560546875,
+      "step": 5057
+    },
+    {
+      "epoch": 2.64903090623363,
+      "loss_partial": 0.0693359375,
+      "step": 5057
+    },
+    {
+      "epoch": 2.649554740701938,
+      "grad_norm": 1.899184226989746,
+      "learning_rate": 6.1816144608653015e-06,
+      "loss": 0.1036,
+      "step": 5058
+    },
+    {
+      "epoch": 2.649554740701938,
+      "loss_partial": 0.051513671875,
+      "step": 5058
+    },
+    {
+      "epoch": 2.649554740701938,
+      "loss_partial": 0.047119140625,
+      "step": 5058
+    },
+    {
+      "epoch": 2.6500785751702463,
+      "grad_norm": 1.2403132915496826,
+      "learning_rate": 6.177389669156242e-06,
+      "loss": 0.0966,
+      "step": 5059
+    },
+    {
+      "epoch": 2.6500785751702463,
+      "loss_partial": 0.03466796875,
+      "step": 5059
+    },
+    {
+      "epoch": 2.6500785751702463,
+      "loss_partial": 0.047119140625,
+      "step": 5059
+    },
+    {
+      "epoch": 2.6506024096385543,
+      "grad_norm": 1.7211748361587524,
+      "learning_rate": 6.173165676349103e-06,
+      "loss": 0.1007,
+      "step": 5060
+    },
+    {
+      "epoch": 2.6506024096385543,
+      "loss_partial": 0.052001953125,
+      "step": 5060
+    },
+    {
+      "epoch": 2.6506024096385543,
+      "loss_partial": 0.046142578125,
+      "step": 5060
+    },
+    {
+      "epoch": 2.6511262441068624,
+      "grad_norm": 1.5578477382659912,
+      "learning_rate": 6.168942483326674e-06,
+      "loss": 0.0999,
+      "step": 5061
+    },
+    {
+      "epoch": 2.6511262441068624,
+      "loss_partial": 0.05322265625,
+      "step": 5061
+    },
+    {
+      "epoch": 2.6511262441068624,
+      "loss_partial": 0.047607421875,
+      "step": 5061
+    },
+    {
+      "epoch": 2.6516500785751704,
+      "grad_norm": 1.5657672882080078,
+      "learning_rate": 6.164720090971574e-06,
+      "loss": 0.1043,
+      "step": 5062
+    },
+    {
+      "epoch": 2.6516500785751704,
+      "loss_partial": 0.0517578125,
+      "step": 5062
+    },
+    {
+      "epoch": 2.6516500785751704,
+      "loss_partial": 0.04345703125,
+      "step": 5062
+    },
+    {
+      "epoch": 2.6521739130434785,
+      "grad_norm": 1.0518759489059448,
+      "learning_rate": 6.160498500166256e-06,
+      "loss": 0.0988,
+      "step": 5063
+    },
+    {
+      "epoch": 2.6521739130434785,
+      "loss_partial": 0.060546875,
+      "step": 5063
+    },
+    {
+      "epoch": 2.6521739130434785,
+      "loss_partial": 0.0498046875,
+      "step": 5063
+    },
+    {
+      "epoch": 2.6526977475117866,
+      "grad_norm": 2.2153282165527344,
+      "learning_rate": 6.1562777117930125e-06,
+      "loss": 0.1039,
+      "step": 5064
+    },
+    {
+      "epoch": 2.6526977475117866,
+      "loss_partial": 0.0576171875,
+      "step": 5064
+    },
+    {
+      "epoch": 2.6526977475117866,
+      "loss_partial": 0.055908203125,
+      "step": 5064
+    },
+    {
+      "epoch": 2.653221581980094,
+      "grad_norm": 1.7212187051773071,
+      "learning_rate": 6.1520577267339575e-06,
+      "loss": 0.101,
+      "step": 5065
+    },
+    {
+      "epoch": 2.653221581980094,
+      "loss_partial": 0.0478515625,
+      "step": 5065
+    },
+    {
+      "epoch": 2.653221581980094,
+      "loss_partial": 0.041015625,
+      "step": 5065
+    },
+    {
+      "epoch": 2.6537454164484022,
+      "grad_norm": 1.7665674686431885,
+      "learning_rate": 6.147838545871043e-06,
+      "loss": 0.0995,
+      "step": 5066
+    },
+    {
+      "epoch": 2.6537454164484022,
+      "loss_partial": 0.0556640625,
+      "step": 5066
+    },
+    {
+      "epoch": 2.6537454164484022,
+      "loss_partial": 0.0478515625,
+      "step": 5066
+    },
+    {
+      "epoch": 2.6542692509167103,
+      "grad_norm": 1.7856864929199219,
+      "learning_rate": 6.143620170086047e-06,
+      "loss": 0.1024,
+      "step": 5067
+    },
+    {
+      "epoch": 2.6542692509167103,
+      "loss_partial": 0.048828125,
+      "step": 5067
+    },
+    {
+      "epoch": 2.6542692509167103,
+      "loss_partial": 0.044189453125,
+      "step": 5067
+    },
+    {
+      "epoch": 2.6547930853850183,
+      "grad_norm": 1.8974976539611816,
+      "learning_rate": 6.139402600260593e-06,
+      "loss": 0.1021,
+      "step": 5068
+    },
+    {
+      "epoch": 2.6547930853850183,
+      "loss_partial": 0.04638671875,
+      "step": 5068
+    },
+    {
+      "epoch": 2.6547930853850183,
+      "loss_partial": 0.0634765625,
+      "step": 5068
+    },
+    {
+      "epoch": 2.6553169198533264,
+      "grad_norm": 1.375257968902588,
+      "learning_rate": 6.135185837276121e-06,
+      "loss": 0.1026,
+      "step": 5069
+    },
+    {
+      "epoch": 2.6553169198533264,
+      "loss_partial": 0.049560546875,
+      "step": 5069
+    },
+    {
+      "epoch": 2.6553169198533264,
+      "loss_partial": 0.044921875,
+      "step": 5069
+    },
+    {
+      "epoch": 2.6558407543216345,
+      "grad_norm": 1.6427364349365234,
+      "learning_rate": 6.130969882013911e-06,
+      "loss": 0.0993,
+      "step": 5070
+    },
+    {
+      "epoch": 2.6558407543216345,
+      "loss_partial": 0.052490234375,
+      "step": 5070
+    },
+    {
+      "epoch": 2.6558407543216345,
+      "loss_partial": 0.05078125,
+      "step": 5070
+    },
+    {
+      "epoch": 2.6563645887899425,
+      "grad_norm": 1.9332833290100098,
+      "learning_rate": 6.126754735355067e-06,
+      "loss": 0.1,
+      "step": 5071
+    },
+    {
+      "epoch": 2.6563645887899425,
+      "loss_partial": 0.05078125,
+      "step": 5071
+    },
+    {
+      "epoch": 2.6563645887899425,
+      "loss_partial": 0.045654296875,
+      "step": 5071
+    },
+    {
+      "epoch": 2.65688842325825,
+      "grad_norm": 1.2862414121627808,
+      "learning_rate": 6.1225403981805345e-06,
+      "loss": 0.1022,
+      "step": 5072
+    },
+    {
+      "epoch": 2.65688842325825,
+      "loss_partial": 0.05126953125,
+      "step": 5072
+    },
+    {
+      "epoch": 2.65688842325825,
+      "loss_partial": 0.04931640625,
+      "step": 5072
+    },
+    {
+      "epoch": 2.657412257726558,
+      "grad_norm": 1.6194247007369995,
+      "learning_rate": 6.118326871371082e-06,
+      "loss": 0.0992,
+      "step": 5073
+    },
+    {
+      "epoch": 2.657412257726558,
+      "loss_partial": 0.055419921875,
+      "step": 5073
+    },
+    {
+      "epoch": 2.657412257726558,
+      "loss_partial": 0.04736328125,
+      "step": 5073
+    },
+    {
+      "epoch": 2.6579360921948663,
+      "grad_norm": 1.3125807046890259,
+      "learning_rate": 6.114114155807309e-06,
+      "loss": 0.0997,
+      "step": 5074
+    },
+    {
+      "epoch": 2.6579360921948663,
+      "loss_partial": 0.05029296875,
+      "step": 5074
+    },
+    {
+      "epoch": 2.6579360921948663,
+      "loss_partial": 0.049560546875,
+      "step": 5074
+    },
+    {
+      "epoch": 2.6584599266631743,
+      "grad_norm": 1.7911317348480225,
+      "learning_rate": 6.109902252369646e-06,
+      "loss": 0.104,
+      "step": 5075
+    },
+    {
+      "epoch": 2.6584599266631743,
+      "loss_partial": 0.047607421875,
+      "step": 5075
+    },
+    {
+      "epoch": 2.6584599266631743,
+      "loss_partial": 0.054443359375,
+      "step": 5075
+    },
+    {
+      "epoch": 2.6589837611314824,
+      "grad_norm": 1.7776662111282349,
+      "learning_rate": 6.105691161938359e-06,
+      "loss": 0.1002,
+      "step": 5076
+    },
+    {
+      "epoch": 2.6589837611314824,
+      "loss_partial": 0.054931640625,
+      "step": 5076
+    },
+    {
+      "epoch": 2.6589837611314824,
+      "loss_partial": 0.052490234375,
+      "step": 5076
+    },
+    {
+      "epoch": 2.6595075955997904,
+      "grad_norm": 1.4651401042938232,
+      "learning_rate": 6.101480885393537e-06,
+      "loss": 0.1028,
+      "step": 5077
+    },
+    {
+      "epoch": 2.6595075955997904,
+      "loss_partial": 0.047607421875,
+      "step": 5077
+    },
+    {
+      "epoch": 2.6595075955997904,
+      "loss_partial": 0.042724609375,
+      "step": 5077
+    },
+    {
+      "epoch": 2.6600314300680985,
+      "grad_norm": 1.617023229598999,
+      "learning_rate": 6.097271423615103e-06,
+      "loss": 0.0997,
+      "step": 5078
+    },
+    {
+      "epoch": 2.6600314300680985,
+      "loss_partial": 0.054931640625,
+      "step": 5078
+    },
+    {
+      "epoch": 2.6600314300680985,
+      "loss_partial": 0.055908203125,
+      "step": 5078
+    },
+    {
+      "epoch": 2.6605552645364066,
+      "grad_norm": 1.4349576234817505,
+      "learning_rate": 6.093062777482802e-06,
+      "loss": 0.1024,
+      "step": 5079
+    },
+    {
+      "epoch": 2.6605552645364066,
+      "loss_partial": 0.0498046875,
+      "step": 5079
+    },
+    {
+      "epoch": 2.6605552645364066,
+      "loss_partial": 0.044677734375,
+      "step": 5079
+    },
+    {
+      "epoch": 2.6610790990047146,
+      "grad_norm": 1.5526080131530762,
+      "learning_rate": 6.088854947876226e-06,
+      "loss": 0.1002,
+      "step": 5080
+    },
+    {
+      "epoch": 2.6610790990047146,
+      "loss_partial": 0.05029296875,
+      "step": 5080
+    },
+    {
+      "epoch": 2.6610790990047146,
+      "loss_partial": 0.05078125,
+      "step": 5080
+    },
+    {
+      "epoch": 2.6616029334730227,
+      "grad_norm": 1.2975051403045654,
+      "learning_rate": 6.08464793567478e-06,
+      "loss": 0.1018,
+      "step": 5081
+    },
+    {
+      "epoch": 2.6616029334730227,
+      "loss_partial": 0.057861328125,
+      "step": 5081
+    },
+    {
+      "epoch": 2.6616029334730227,
+      "loss_partial": 0.05419921875,
+      "step": 5081
+    },
+    {
+      "epoch": 2.6621267679413307,
+      "grad_norm": 1.6602596044540405,
+      "learning_rate": 6.080441741757703e-06,
+      "loss": 0.1012,
+      "step": 5082
+    },
+    {
+      "epoch": 2.6621267679413307,
+      "loss_partial": 0.056640625,
+      "step": 5082
+    },
+    {
+      "epoch": 2.6621267679413307,
+      "loss_partial": 0.047119140625,
+      "step": 5082
+    },
+    {
+      "epoch": 2.662650602409639,
+      "grad_norm": 1.7220956087112427,
+      "learning_rate": 6.076236367004064e-06,
+      "loss": 0.1014,
+      "step": 5083
+    },
+    {
+      "epoch": 2.662650602409639,
+      "loss_partial": 0.0556640625,
+      "step": 5083
+    },
+    {
+      "epoch": 2.662650602409639,
+      "loss_partial": 0.057373046875,
+      "step": 5083
+    },
+    {
+      "epoch": 2.6631744368779464,
+      "grad_norm": 1.338773250579834,
+      "learning_rate": 6.07203181229276e-06,
+      "loss": 0.1021,
+      "step": 5084
+    },
+    {
+      "epoch": 2.6631744368779464,
+      "loss_partial": 0.0498046875,
+      "step": 5084
+    },
+    {
+      "epoch": 2.6631744368779464,
+      "loss_partial": 0.0498046875,
+      "step": 5084
+    },
+    {
+      "epoch": 2.6636982713462545,
+      "grad_norm": 1.563969373703003,
+      "learning_rate": 6.06782807850252e-06,
+      "loss": 0.099,
+      "step": 5085
+    },
+    {
+      "epoch": 2.6636982713462545,
+      "loss_partial": 0.051513671875,
+      "step": 5085
+    },
+    {
+      "epoch": 2.6636982713462545,
+      "loss_partial": 0.048828125,
+      "step": 5085
+    },
+    {
+      "epoch": 2.6642221058145625,
+      "grad_norm": 1.6719434261322021,
+      "learning_rate": 6.063625166511894e-06,
+      "loss": 0.1049,
+      "step": 5086
+    },
+    {
+      "epoch": 2.6642221058145625,
+      "loss_partial": 0.052490234375,
+      "step": 5086
+    },
+    {
+      "epoch": 2.6642221058145625,
+      "loss_partial": 0.043701171875,
+      "step": 5086
+    },
+    {
+      "epoch": 2.6647459402828706,
+      "grad_norm": 1.1366245746612549,
+      "learning_rate": 6.059423077199264e-06,
+      "loss": 0.1024,
+      "step": 5087
+    },
+    {
+      "epoch": 2.6647459402828706,
+      "loss_partial": 0.0537109375,
+      "step": 5087
+    },
+    {
+      "epoch": 2.6647459402828706,
+      "loss_partial": 0.046630859375,
+      "step": 5087
+    },
+    {
+      "epoch": 2.6652697747511787,
+      "grad_norm": 2.055469036102295,
+      "learning_rate": 6.055221811442846e-06,
+      "loss": 0.1024,
+      "step": 5088
+    },
+    {
+      "epoch": 2.6652697747511787,
+      "loss_partial": 0.05419921875,
+      "step": 5088
+    },
+    {
+      "epoch": 2.6652697747511787,
+      "loss_partial": 0.04736328125,
+      "step": 5088
+    },
+    {
+      "epoch": 2.6657936092194867,
+      "grad_norm": 2.4946672916412354,
+      "learning_rate": 6.0510213701206755e-06,
+      "loss": 0.1033,
+      "step": 5089
+    },
+    {
+      "epoch": 2.6657936092194867,
+      "loss_partial": 0.05322265625,
+      "step": 5089
+    },
+    {
+      "epoch": 2.6657936092194867,
+      "loss_partial": 0.05859375,
+      "step": 5089
+    },
+    {
+      "epoch": 2.6663174436877948,
+      "grad_norm": 1.4497753381729126,
+      "learning_rate": 6.0468217541106185e-06,
+      "loss": 0.0993,
+      "step": 5090
+    },
+    {
+      "epoch": 2.6663174436877948,
+      "loss_partial": 0.0517578125,
+      "step": 5090
+    },
+    {
+      "epoch": 2.6663174436877948,
+      "loss_partial": 0.049560546875,
+      "step": 5090
+    },
+    {
+      "epoch": 2.666841278156103,
+      "grad_norm": 1.920600414276123,
+      "learning_rate": 6.042622964290366e-06,
+      "loss": 0.1043,
+      "step": 5091
+    },
+    {
+      "epoch": 2.666841278156103,
+      "loss_partial": 0.058349609375,
+      "step": 5091
+    },
+    {
+      "epoch": 2.666841278156103,
+      "loss_partial": 0.0517578125,
+      "step": 5091
+    },
+    {
+      "epoch": 2.6673651126244105,
+      "grad_norm": 1.3738726377487183,
+      "learning_rate": 6.038425001537447e-06,
+      "loss": 0.1059,
+      "step": 5092
+    },
+    {
+      "epoch": 2.6673651126244105,
+      "loss_partial": 0.05029296875,
+      "step": 5092
+    },
+    {
+      "epoch": 2.6673651126244105,
+      "loss_partial": 0.0478515625,
+      "step": 5092
+    },
+    {
+      "epoch": 2.6678889470927185,
+      "grad_norm": 1.8037298917770386,
+      "learning_rate": 6.034227866729203e-06,
+      "loss": 0.1038,
+      "step": 5093
+    },
+    {
+      "epoch": 2.6678889470927185,
+      "loss_partial": 0.048583984375,
+      "step": 5093
+    },
+    {
+      "epoch": 2.6678889470927185,
+      "loss_partial": 0.047607421875,
+      "step": 5093
+    },
+    {
+      "epoch": 2.6684127815610266,
+      "grad_norm": 1.748593807220459,
+      "learning_rate": 6.030031560742813e-06,
+      "loss": 0.0977,
+      "step": 5094
+    },
+    {
+      "epoch": 2.6684127815610266,
+      "loss_partial": 0.04736328125,
+      "step": 5094
+    },
+    {
+      "epoch": 2.6684127815610266,
+      "loss_partial": 0.05859375,
+      "step": 5094
+    },
+    {
+      "epoch": 2.6689366160293346,
+      "grad_norm": 1.334985613822937,
+      "learning_rate": 6.0258360844552734e-06,
+      "loss": 0.1015,
+      "step": 5095
+    },
+    {
+      "epoch": 2.6689366160293346,
+      "loss_partial": 0.052001953125,
+      "step": 5095
+    },
+    {
+      "epoch": 2.6689366160293346,
+      "loss_partial": 0.048583984375,
+      "step": 5095
+    },
+    {
+      "epoch": 2.6694604504976427,
+      "grad_norm": 1.6394455432891846,
+      "learning_rate": 6.0216414387434195e-06,
+      "loss": 0.1007,
+      "step": 5096
+    },
+    {
+      "epoch": 2.6694604504976427,
+      "loss_partial": 0.044677734375,
+      "step": 5096
+    },
+    {
+      "epoch": 2.6694604504976427,
+      "loss_partial": 0.054931640625,
+      "step": 5096
+    },
+    {
+      "epoch": 2.6699842849659507,
+      "grad_norm": 1.6563001871109009,
+      "learning_rate": 6.017447624483903e-06,
+      "loss": 0.1013,
+      "step": 5097
+    },
+    {
+      "epoch": 2.6699842849659507,
+      "loss_partial": 0.04833984375,
+      "step": 5097
+    },
+    {
+      "epoch": 2.6699842849659507,
+      "loss_partial": 0.047607421875,
+      "step": 5097
+    },
+    {
+      "epoch": 2.670508119434259,
+      "grad_norm": 1.80418062210083,
+      "learning_rate": 6.013254642553207e-06,
+      "loss": 0.0993,
+      "step": 5098
+    },
+    {
+      "epoch": 2.670508119434259,
+      "loss_partial": 0.061767578125,
+      "step": 5098
+    },
+    {
+      "epoch": 2.670508119434259,
+      "loss_partial": 0.04931640625,
+      "step": 5098
+    },
+    {
+      "epoch": 2.671031953902567,
+      "grad_norm": 1.417734980583191,
+      "learning_rate": 6.009062493827635e-06,
+      "loss": 0.1005,
+      "step": 5099
+    },
+    {
+      "epoch": 2.671031953902567,
+      "loss_partial": 0.047119140625,
+      "step": 5099
+    },
+    {
+      "epoch": 2.671031953902567,
+      "loss_partial": 0.04443359375,
+      "step": 5099
+    },
+    {
+      "epoch": 2.671555788370875,
+      "grad_norm": 1.2240098714828491,
+      "learning_rate": 6.004871179183324e-06,
+      "loss": 0.1039,
+      "step": 5100
+    },
+    {
+      "epoch": 2.671555788370875,
+      "loss_partial": 0.05322265625,
+      "step": 5100
+    },
+    {
+      "epoch": 2.671555788370875,
+      "loss_partial": 0.0576171875,
+      "step": 5100
+    },
+    {
+      "epoch": 2.672079622839183,
+      "grad_norm": 1.4564402103424072,
+      "learning_rate": 6.00068069949623e-06,
+      "loss": 0.0995,
+      "step": 5101
+    },
+    {
+      "epoch": 2.672079622839183,
+      "loss_partial": 0.06298828125,
+      "step": 5101
+    },
+    {
+      "epoch": 2.672079622839183,
+      "loss_partial": 0.05419921875,
+      "step": 5101
+    },
+    {
+      "epoch": 2.672603457307491,
+      "grad_norm": 1.2260953187942505,
+      "learning_rate": 5.996491055642142e-06,
+      "loss": 0.1031,
+      "step": 5102
+    },
+    {
+      "epoch": 2.672603457307491,
+      "loss_partial": 0.054443359375,
+      "step": 5102
+    },
+    {
+      "epoch": 2.672603457307491,
+      "loss_partial": 0.05078125,
+      "step": 5102
+    },
+    {
+      "epoch": 2.673127291775799,
+      "grad_norm": 1.3831454515457153,
+      "learning_rate": 5.992302248496661e-06,
+      "loss": 0.103,
+      "step": 5103
+    },
+    {
+      "epoch": 2.673127291775799,
+      "loss_partial": 0.04443359375,
+      "step": 5103
+    },
+    {
+      "epoch": 2.673127291775799,
+      "loss_partial": 0.048828125,
+      "step": 5103
+    },
+    {
+      "epoch": 2.6736511262441067,
+      "grad_norm": 1.2858411073684692,
+      "learning_rate": 5.988114278935231e-06,
+      "loss": 0.1023,
+      "step": 5104
+    },
+    {
+      "epoch": 2.6736511262441067,
+      "loss_partial": 0.04345703125,
+      "step": 5104
+    },
+    {
+      "epoch": 2.6736511262441067,
+      "loss_partial": 0.0419921875,
+      "step": 5104
+    },
+    {
+      "epoch": 2.674174960712415,
+      "grad_norm": 1.289152979850769,
+      "learning_rate": 5.9839271478331085e-06,
+      "loss": 0.0996,
+      "step": 5105
+    },
+    {
+      "epoch": 2.674174960712415,
+      "loss_partial": 0.053466796875,
+      "step": 5105
+    },
+    {
+      "epoch": 2.674174960712415,
+      "loss_partial": 0.048828125,
+      "step": 5105
+    },
+    {
+      "epoch": 2.674698795180723,
+      "grad_norm": 1.3512592315673828,
+      "learning_rate": 5.979740856065377e-06,
+      "loss": 0.0999,
+      "step": 5106
+    },
+    {
+      "epoch": 2.674698795180723,
+      "loss_partial": 0.050048828125,
+      "step": 5106
+    },
+    {
+      "epoch": 2.674698795180723,
+      "loss_partial": 0.060546875,
+      "step": 5106
+    },
+    {
+      "epoch": 2.675222629649031,
+      "grad_norm": 1.3696345090866089,
+      "learning_rate": 5.975555404506943e-06,
+      "loss": 0.1011,
+      "step": 5107
+    },
+    {
+      "epoch": 2.675222629649031,
+      "loss_partial": 0.052978515625,
+      "step": 5107
+    },
+    {
+      "epoch": 2.675222629649031,
+      "loss_partial": 0.044921875,
+      "step": 5107
+    },
+    {
+      "epoch": 2.675746464117339,
+      "grad_norm": 1.5498214960098267,
+      "learning_rate": 5.971370794032546e-06,
+      "loss": 0.0984,
+      "step": 5108
+    },
+    {
+      "epoch": 2.675746464117339,
+      "loss_partial": 0.048583984375,
+      "step": 5108
+    },
+    {
+      "epoch": 2.675746464117339,
+      "loss_partial": 0.050048828125,
+      "step": 5108
+    },
+    {
+      "epoch": 2.676270298585647,
+      "grad_norm": 1.336927056312561,
+      "learning_rate": 5.967187025516741e-06,
+      "loss": 0.0987,
+      "step": 5109
+    },
+    {
+      "epoch": 2.676270298585647,
+      "loss_partial": 0.047119140625,
+      "step": 5109
+    },
+    {
+      "epoch": 2.676270298585647,
+      "loss_partial": 0.04833984375,
+      "step": 5109
+    },
+    {
+      "epoch": 2.676794133053955,
+      "grad_norm": 1.7871243953704834,
+      "learning_rate": 5.963004099833909e-06,
+      "loss": 0.1038,
+      "step": 5110
+    },
+    {
+      "epoch": 2.676794133053955,
+      "loss_partial": 0.046875,
+      "step": 5110
+    },
+    {
+      "epoch": 2.676794133053955,
+      "loss_partial": 0.06103515625,
+      "step": 5110
+    },
+    {
+      "epoch": 2.6773179675222627,
+      "grad_norm": 1.7710208892822266,
+      "learning_rate": 5.958822017858255e-06,
+      "loss": 0.1014,
+      "step": 5111
+    },
+    {
+      "epoch": 2.6773179675222627,
+      "loss_partial": 0.05322265625,
+      "step": 5111
+    },
+    {
+      "epoch": 2.6773179675222627,
+      "loss_partial": 0.057373046875,
+      "step": 5111
+    },
+    {
+      "epoch": 2.6778418019905708,
+      "grad_norm": 2.0106887817382812,
+      "learning_rate": 5.95464078046381e-06,
+      "loss": 0.1009,
+      "step": 5112
+    },
+    {
+      "epoch": 2.6778418019905708,
+      "loss_partial": 0.050048828125,
+      "step": 5112
+    },
+    {
+      "epoch": 2.6778418019905708,
+      "loss_partial": 0.054443359375,
+      "step": 5112
+    },
+    {
+      "epoch": 2.678365636458879,
+      "grad_norm": 1.9040828943252563,
+      "learning_rate": 5.950460388524427e-06,
+      "loss": 0.101,
+      "step": 5113
+    },
+    {
+      "epoch": 2.678365636458879,
+      "loss_partial": 0.05810546875,
+      "step": 5113
+    },
+    {
+      "epoch": 2.678365636458879,
+      "loss_partial": 0.051025390625,
+      "step": 5113
+    },
+    {
+      "epoch": 2.678889470927187,
+      "grad_norm": 2.4949705600738525,
+      "learning_rate": 5.946280842913781e-06,
+      "loss": 0.1025,
+      "step": 5114
+    },
+    {
+      "epoch": 2.678889470927187,
+      "loss_partial": 0.04541015625,
+      "step": 5114
+    },
+    {
+      "epoch": 2.678889470927187,
+      "loss_partial": 0.052978515625,
+      "step": 5114
+    },
+    {
+      "epoch": 2.679413305395495,
+      "grad_norm": 1.3710099458694458,
+      "learning_rate": 5.942102144505368e-06,
+      "loss": 0.1006,
+      "step": 5115
+    },
+    {
+      "epoch": 2.679413305395495,
+      "loss_partial": 0.046142578125,
+      "step": 5115
+    },
+    {
+      "epoch": 2.679413305395495,
+      "loss_partial": 0.052734375,
+      "step": 5115
+    },
+    {
+      "epoch": 2.679937139863803,
+      "grad_norm": 1.7955117225646973,
+      "learning_rate": 5.937924294172517e-06,
+      "loss": 0.1001,
+      "step": 5116
+    },
+    {
+      "epoch": 2.679937139863803,
+      "loss_partial": 0.0537109375,
+      "step": 5116
+    },
+    {
+      "epoch": 2.679937139863803,
+      "loss_partial": 0.052001953125,
+      "step": 5116
+    },
+    {
+      "epoch": 2.680460974332111,
+      "grad_norm": 2.217419147491455,
+      "learning_rate": 5.933747292788369e-06,
+      "loss": 0.1054,
+      "step": 5117
+    },
+    {
+      "epoch": 2.680460974332111,
+      "loss_partial": 0.053466796875,
+      "step": 5117
+    },
+    {
+      "epoch": 2.680460974332111,
+      "loss_partial": 0.05126953125,
+      "step": 5117
+    },
+    {
+      "epoch": 2.680984808800419,
+      "grad_norm": 1.8307700157165527,
+      "learning_rate": 5.929571141225892e-06,
+      "loss": 0.0962,
+      "step": 5118
+    },
+    {
+      "epoch": 2.680984808800419,
+      "loss_partial": 0.06005859375,
+      "step": 5118
+    },
+    {
+      "epoch": 2.680984808800419,
+      "loss_partial": 0.068359375,
+      "step": 5118
+    },
+    {
+      "epoch": 2.681508643268727,
+      "grad_norm": 1.6064574718475342,
+      "learning_rate": 5.925395840357877e-06,
+      "loss": 0.1032,
+      "step": 5119
+    },
+    {
+      "epoch": 2.681508643268727,
+      "loss_partial": 0.051513671875,
+      "step": 5119
+    },
+    {
+      "epoch": 2.681508643268727,
+      "loss_partial": 0.0634765625,
+      "step": 5119
+    },
+    {
+      "epoch": 2.6820324777370352,
+      "grad_norm": 1.565971851348877,
+      "learning_rate": 5.92122139105693e-06,
+      "loss": 0.0999,
+      "step": 5120
+    },
+    {
+      "epoch": 2.6820324777370352,
+      "loss_partial": 0.053955078125,
+      "step": 5120
+    },
+    {
+      "epoch": 2.6820324777370352,
+      "loss_partial": 0.04296875,
+      "step": 5120
+    },
+    {
+      "epoch": 2.6825563122053433,
+      "grad_norm": 1.3539642095565796,
+      "learning_rate": 5.917047794195494e-06,
+      "loss": 0.102,
+      "step": 5121
+    },
+    {
+      "epoch": 2.6825563122053433,
+      "loss_partial": 0.05078125,
+      "step": 5121
+    },
+    {
+      "epoch": 2.6825563122053433,
+      "loss_partial": 0.044677734375,
+      "step": 5121
+    },
+    {
+      "epoch": 2.6830801466736514,
+      "grad_norm": 1.4603965282440186,
+      "learning_rate": 5.912875050645821e-06,
+      "loss": 0.1012,
+      "step": 5122
+    },
+    {
+      "epoch": 2.6830801466736514,
+      "loss_partial": 0.04638671875,
+      "step": 5122
+    },
+    {
+      "epoch": 2.6830801466736514,
+      "loss_partial": 0.04541015625,
+      "step": 5122
+    },
+    {
+      "epoch": 2.6836039811419594,
+      "grad_norm": 2.0013132095336914,
+      "learning_rate": 5.908703161279987e-06,
+      "loss": 0.1026,
+      "step": 5123
+    },
+    {
+      "epoch": 2.6836039811419594,
+      "loss_partial": 0.0478515625,
+      "step": 5123
+    },
+    {
+      "epoch": 2.6836039811419594,
+      "loss_partial": 0.05810546875,
+      "step": 5123
+    },
+    {
+      "epoch": 2.684127815610267,
+      "grad_norm": 1.1139878034591675,
+      "learning_rate": 5.9045321269698915e-06,
+      "loss": 0.1043,
+      "step": 5124
+    },
+    {
+      "epoch": 2.684127815610267,
+      "loss_partial": 0.047119140625,
+      "step": 5124
+    },
+    {
+      "epoch": 2.684127815610267,
+      "loss_partial": 0.0458984375,
+      "step": 5124
+    },
+    {
+      "epoch": 2.684651650078575,
+      "grad_norm": 1.5006074905395508,
+      "learning_rate": 5.900361948587259e-06,
+      "loss": 0.0991,
+      "step": 5125
+    },
+    {
+      "epoch": 2.684651650078575,
+      "loss_partial": 0.05859375,
+      "step": 5125
+    },
+    {
+      "epoch": 2.684651650078575,
+      "loss_partial": 0.049560546875,
+      "step": 5125
+    },
+    {
+      "epoch": 2.685175484546883,
+      "grad_norm": 1.4102413654327393,
+      "learning_rate": 5.896192627003625e-06,
+      "loss": 0.1033,
+      "step": 5126
+    },
+    {
+      "epoch": 2.685175484546883,
+      "loss_partial": 0.05810546875,
+      "step": 5126
+    },
+    {
+      "epoch": 2.685175484546883,
+      "loss_partial": 0.044677734375,
+      "step": 5126
+    },
+    {
+      "epoch": 2.685699319015191,
+      "grad_norm": 1.8015676736831665,
+      "learning_rate": 5.892024163090357e-06,
+      "loss": 0.104,
+      "step": 5127
+    },
+    {
+      "epoch": 2.685699319015191,
+      "loss_partial": 0.053955078125,
+      "step": 5127
+    },
+    {
+      "epoch": 2.685699319015191,
+      "loss_partial": 0.05322265625,
+      "step": 5127
+    },
+    {
+      "epoch": 2.6862231534834993,
+      "grad_norm": 1.5072813034057617,
+      "learning_rate": 5.887856557718631e-06,
+      "loss": 0.1006,
+      "step": 5128
+    },
+    {
+      "epoch": 2.6862231534834993,
+      "loss_partial": 0.055419921875,
+      "step": 5128
+    },
+    {
+      "epoch": 2.6862231534834993,
+      "loss_partial": 0.052734375,
+      "step": 5128
+    },
+    {
+      "epoch": 2.6867469879518073,
+      "grad_norm": 1.4994456768035889,
+      "learning_rate": 5.8836898117594596e-06,
+      "loss": 0.1041,
+      "step": 5129
+    },
+    {
+      "epoch": 2.6867469879518073,
+      "loss_partial": 0.05712890625,
+      "step": 5129
+    },
+    {
+      "epoch": 2.6867469879518073,
+      "loss_partial": 0.04345703125,
+      "step": 5129
+    },
+    {
+      "epoch": 2.6872708224201154,
+      "grad_norm": 1.897438406944275,
+      "learning_rate": 5.879523926083663e-06,
+      "loss": 0.0992,
+      "step": 5130
+    },
+    {
+      "epoch": 2.6872708224201154,
+      "loss_partial": 0.0556640625,
+      "step": 5130
+    },
+    {
+      "epoch": 2.6872708224201154,
+      "loss_partial": 0.04052734375,
+      "step": 5130
+    },
+    {
+      "epoch": 2.687794656888423,
+      "grad_norm": 1.1339386701583862,
+      "learning_rate": 5.8753589015618865e-06,
+      "loss": 0.1049,
+      "step": 5131
+    },
+    {
+      "epoch": 2.687794656888423,
+      "loss_partial": 0.0517578125,
+      "step": 5131
+    },
+    {
+      "epoch": 2.687794656888423,
+      "loss_partial": 0.05908203125,
+      "step": 5131
+    },
+    {
+      "epoch": 2.688318491356731,
+      "grad_norm": 1.192391037940979,
+      "learning_rate": 5.871194739064591e-06,
+      "loss": 0.105,
+      "step": 5132
+    },
+    {
+      "epoch": 2.688318491356731,
+      "loss_partial": 0.051025390625,
+      "step": 5132
+    },
+    {
+      "epoch": 2.688318491356731,
+      "loss_partial": 0.054443359375,
+      "step": 5132
+    },
+    {
+      "epoch": 2.688842325825039,
+      "grad_norm": 1.4004995822906494,
+      "learning_rate": 5.867031439462065e-06,
+      "loss": 0.1022,
+      "step": 5133
+    },
+    {
+      "epoch": 2.688842325825039,
+      "loss_partial": 0.054931640625,
+      "step": 5133
+    },
+    {
+      "epoch": 2.688842325825039,
+      "loss_partial": 0.039306640625,
+      "step": 5133
+    },
+    {
+      "epoch": 2.689366160293347,
+      "grad_norm": 1.2510367631912231,
+      "learning_rate": 5.862869003624412e-06,
+      "loss": 0.0973,
+      "step": 5134
+    },
+    {
+      "epoch": 2.689366160293347,
+      "loss_partial": 0.04248046875,
+      "step": 5134
+    },
+    {
+      "epoch": 2.689366160293347,
+      "loss_partial": 0.068359375,
+      "step": 5134
+    },
+    {
+      "epoch": 2.6898899947616552,
+      "grad_norm": 0.9450027942657471,
+      "learning_rate": 5.858707432421552e-06,
+      "loss": 0.1002,
+      "step": 5135
+    },
+    {
+      "epoch": 2.6898899947616552,
+      "loss_partial": 0.054931640625,
+      "step": 5135
+    },
+    {
+      "epoch": 2.6898899947616552,
+      "loss_partial": 0.05322265625,
+      "step": 5135
+    },
+    {
+      "epoch": 2.6904138292299633,
+      "grad_norm": 1.626201868057251,
+      "learning_rate": 5.854546726723229e-06,
+      "loss": 0.1012,
+      "step": 5136
+    },
+    {
+      "epoch": 2.6904138292299633,
+      "loss_partial": 0.059814453125,
+      "step": 5136
+    },
+    {
+      "epoch": 2.6904138292299633,
+      "loss_partial": 0.0634765625,
+      "step": 5136
+    },
+    {
+      "epoch": 2.6909376636982714,
+      "grad_norm": 1.2735594511032104,
+      "learning_rate": 5.850386887399006e-06,
+      "loss": 0.103,
+      "step": 5137
+    },
+    {
+      "epoch": 2.6909376636982714,
+      "loss_partial": 0.04541015625,
+      "step": 5137
+    },
+    {
+      "epoch": 2.6909376636982714,
+      "loss_partial": 0.053466796875,
+      "step": 5137
+    },
+    {
+      "epoch": 2.6914614981665794,
+      "grad_norm": 1.1745012998580933,
+      "learning_rate": 5.846227915318265e-06,
+      "loss": 0.1034,
+      "step": 5138
+    },
+    {
+      "epoch": 2.6914614981665794,
+      "loss_partial": 0.0478515625,
+      "step": 5138
+    },
+    {
+      "epoch": 2.6914614981665794,
+      "loss_partial": 0.048095703125,
+      "step": 5138
+    },
+    {
+      "epoch": 2.6919853326348875,
+      "grad_norm": 0.9361257553100586,
+      "learning_rate": 5.842069811350202e-06,
+      "loss": 0.0992,
+      "step": 5139
+    },
+    {
+      "epoch": 2.6919853326348875,
+      "loss_partial": 0.04541015625,
+      "step": 5139
+    },
+    {
+      "epoch": 2.6919853326348875,
+      "loss_partial": 0.05517578125,
+      "step": 5139
+    },
+    {
+      "epoch": 2.6925091671031955,
+      "grad_norm": 1.263627529144287,
+      "learning_rate": 5.837912576363836e-06,
+      "loss": 0.1013,
+      "step": 5140
+    },
+    {
+      "epoch": 2.6925091671031955,
+      "loss_partial": 0.043701171875,
+      "step": 5140
+    },
+    {
+      "epoch": 2.6925091671031955,
+      "loss_partial": 0.047119140625,
+      "step": 5140
+    },
+    {
+      "epoch": 2.6930330015715036,
+      "grad_norm": 1.3924027681350708,
+      "learning_rate": 5.833756211228005e-06,
+      "loss": 0.0986,
+      "step": 5141
+    },
+    {
+      "epoch": 2.6930330015715036,
+      "loss_partial": 0.059814453125,
+      "step": 5141
+    },
+    {
+      "epoch": 2.6930330015715036,
+      "loss_partial": 0.0439453125,
+      "step": 5141
+    },
+    {
+      "epoch": 2.6935568360398117,
+      "grad_norm": 1.0745112895965576,
+      "learning_rate": 5.829600716811362e-06,
+      "loss": 0.1021,
+      "step": 5142
+    },
+    {
+      "epoch": 2.6935568360398117,
+      "loss_partial": 0.046142578125,
+      "step": 5142
+    },
+    {
+      "epoch": 2.6935568360398117,
+      "loss_partial": 0.04931640625,
+      "step": 5142
+    },
+    {
+      "epoch": 2.6940806705081193,
+      "grad_norm": 1.4358851909637451,
+      "learning_rate": 5.825446093982381e-06,
+      "loss": 0.0986,
+      "step": 5143
+    },
+    {
+      "epoch": 2.6940806705081193,
+      "loss_partial": 0.05712890625,
+      "step": 5143
+    },
+    {
+      "epoch": 2.6940806705081193,
+      "loss_partial": 0.048095703125,
+      "step": 5143
+    },
+    {
+      "epoch": 2.6946045049764273,
+      "grad_norm": 1.5506690740585327,
+      "learning_rate": 5.821292343609348e-06,
+      "loss": 0.1006,
+      "step": 5144
+    },
+    {
+      "epoch": 2.6946045049764273,
+      "loss_partial": 0.04541015625,
+      "step": 5144
+    },
+    {
+      "epoch": 2.6946045049764273,
+      "loss_partial": 0.0478515625,
+      "step": 5144
+    },
+    {
+      "epoch": 2.6951283394447354,
+      "grad_norm": 0.9391441941261292,
+      "learning_rate": 5.817139466560378e-06,
+      "loss": 0.1005,
+      "step": 5145
+    },
+    {
+      "epoch": 2.6951283394447354,
+      "loss_partial": 0.04296875,
+      "step": 5145
+    },
+    {
+      "epoch": 2.6951283394447354,
+      "loss_partial": 0.049560546875,
+      "step": 5145
+    },
+    {
+      "epoch": 2.6956521739130435,
+      "grad_norm": 1.2818080186843872,
+      "learning_rate": 5.812987463703395e-06,
+      "loss": 0.0979,
+      "step": 5146
+    },
+    {
+      "epoch": 2.6956521739130435,
+      "loss_partial": 0.042236328125,
+      "step": 5146
+    },
+    {
+      "epoch": 2.6956521739130435,
+      "loss_partial": 0.043212890625,
+      "step": 5146
+    },
+    {
+      "epoch": 2.6961760083813515,
+      "grad_norm": 1.3007534742355347,
+      "learning_rate": 5.8088363359061384e-06,
+      "loss": 0.1004,
+      "step": 5147
+    },
+    {
+      "epoch": 2.6961760083813515,
+      "loss_partial": 0.05517578125,
+      "step": 5147
+    },
+    {
+      "epoch": 2.6961760083813515,
+      "loss_partial": 0.05126953125,
+      "step": 5147
+    },
+    {
+      "epoch": 2.6966998428496596,
+      "grad_norm": 1.3004359006881714,
+      "learning_rate": 5.804686084036166e-06,
+      "loss": 0.0977,
+      "step": 5148
+    },
+    {
+      "epoch": 2.6966998428496596,
+      "loss_partial": 0.04638671875,
+      "step": 5148
+    },
+    {
+      "epoch": 2.6966998428496596,
+      "loss_partial": 0.048095703125,
+      "step": 5148
+    },
+    {
+      "epoch": 2.6972236773179676,
+      "grad_norm": 1.4523046016693115,
+      "learning_rate": 5.800536708960864e-06,
+      "loss": 0.1003,
+      "step": 5149
+    },
+    {
+      "epoch": 2.6972236773179676,
+      "loss_partial": 0.053466796875,
+      "step": 5149
+    },
+    {
+      "epoch": 2.6972236773179676,
+      "loss_partial": 0.045166015625,
+      "step": 5149
+    },
+    {
+      "epoch": 2.6977475117862757,
+      "grad_norm": 1.4850858449935913,
+      "learning_rate": 5.796388211547421e-06,
+      "loss": 0.1008,
+      "step": 5150
+    },
+    {
+      "epoch": 2.6977475117862757,
+      "loss_partial": 0.037841796875,
+      "step": 5150
+    },
+    {
+      "epoch": 2.6977475117862757,
+      "loss_partial": 0.052734375,
+      "step": 5150
+    },
+    {
+      "epoch": 2.6982713462545833,
+      "grad_norm": 1.4024499654769897,
+      "learning_rate": 5.792240592662849e-06,
+      "loss": 0.0967,
+      "step": 5151
+    },
+    {
+      "epoch": 2.6982713462545833,
+      "loss_partial": 0.05078125,
+      "step": 5151
+    },
+    {
+      "epoch": 2.6982713462545833,
+      "loss_partial": 0.049072265625,
+      "step": 5151
+    },
+    {
+      "epoch": 2.6987951807228914,
+      "grad_norm": 1.759426474571228,
+      "learning_rate": 5.788093853173971e-06,
+      "loss": 0.1014,
+      "step": 5152
+    },
+    {
+      "epoch": 2.6987951807228914,
+      "loss_partial": 0.047607421875,
+      "step": 5152
+    },
+    {
+      "epoch": 2.6987951807228914,
+      "loss_partial": 0.052490234375,
+      "step": 5152
+    },
+    {
+      "epoch": 2.6993190151911994,
+      "grad_norm": 1.196576714515686,
+      "learning_rate": 5.783947993947433e-06,
+      "loss": 0.1018,
+      "step": 5153
+    },
+    {
+      "epoch": 2.6993190151911994,
+      "loss_partial": 0.054931640625,
+      "step": 5153
+    },
+    {
+      "epoch": 2.6993190151911994,
+      "loss_partial": 0.052734375,
+      "step": 5153
+    },
+    {
+      "epoch": 2.6998428496595075,
+      "grad_norm": 1.3397084474563599,
+      "learning_rate": 5.779803015849693e-06,
+      "loss": 0.1027,
+      "step": 5154
+    },
+    {
+      "epoch": 2.6998428496595075,
+      "loss_partial": 0.050537109375,
+      "step": 5154
+    },
+    {
+      "epoch": 2.6998428496595075,
+      "loss_partial": 0.049072265625,
+      "step": 5154
+    },
+    {
+      "epoch": 2.7003666841278156,
+      "grad_norm": 1.6989003419876099,
+      "learning_rate": 5.775658919747026e-06,
+      "loss": 0.1006,
+      "step": 5155
+    },
+    {
+      "epoch": 2.7003666841278156,
+      "loss_partial": 0.06640625,
+      "step": 5155
+    },
+    {
+      "epoch": 2.7003666841278156,
+      "loss_partial": 0.044677734375,
+      "step": 5155
+    },
+    {
+      "epoch": 2.7008905185961236,
+      "grad_norm": 1.4577434062957764,
+      "learning_rate": 5.7715157065055175e-06,
+      "loss": 0.1045,
+      "step": 5156
+    },
+    {
+      "epoch": 2.7008905185961236,
+      "loss_partial": 0.044189453125,
+      "step": 5156
+    },
+    {
+      "epoch": 2.7008905185961236,
+      "loss_partial": 0.053466796875,
+      "step": 5156
+    },
+    {
+      "epoch": 2.7014143530644317,
+      "grad_norm": 1.814559817314148,
+      "learning_rate": 5.767373376991082e-06,
+      "loss": 0.0991,
+      "step": 5157
+    },
+    {
+      "epoch": 2.7014143530644317,
+      "loss_partial": 0.0634765625,
+      "step": 5157
+    },
+    {
+      "epoch": 2.7014143530644317,
+      "loss_partial": 0.0517578125,
+      "step": 5157
+    },
+    {
+      "epoch": 2.7019381875327397,
+      "grad_norm": 1.3557658195495605,
+      "learning_rate": 5.763231932069438e-06,
+      "loss": 0.1036,
+      "step": 5158
+    },
+    {
+      "epoch": 2.7019381875327397,
+      "loss_partial": 0.06298828125,
+      "step": 5158
+    },
+    {
+      "epoch": 2.7019381875327397,
+      "loss_partial": 0.049560546875,
+      "step": 5158
+    },
+    {
+      "epoch": 2.702462022001048,
+      "grad_norm": 2.250032424926758,
+      "learning_rate": 5.759091372606121e-06,
+      "loss": 0.103,
+      "step": 5159
+    },
+    {
+      "epoch": 2.702462022001048,
+      "loss_partial": 0.045166015625,
+      "step": 5159
+    },
+    {
+      "epoch": 2.702462022001048,
+      "loss_partial": 0.04541015625,
+      "step": 5159
+    },
+    {
+      "epoch": 2.702985856469356,
+      "grad_norm": 1.4534988403320312,
+      "learning_rate": 5.7549516994664765e-06,
+      "loss": 0.101,
+      "step": 5160
+    },
+    {
+      "epoch": 2.702985856469356,
+      "loss_partial": 0.046875,
+      "step": 5160
+    },
+    {
+      "epoch": 2.702985856469356,
+      "loss_partial": 0.044677734375,
+      "step": 5160
+    },
+    {
+      "epoch": 2.703509690937664,
+      "grad_norm": 1.996535301208496,
+      "learning_rate": 5.750812913515681e-06,
+      "loss": 0.1004,
+      "step": 5161
+    },
+    {
+      "epoch": 2.703509690937664,
+      "loss_partial": 0.046875,
+      "step": 5161
+    },
+    {
+      "epoch": 2.703509690937664,
+      "loss_partial": 0.052978515625,
+      "step": 5161
+    },
+    {
+      "epoch": 2.704033525405972,
+      "grad_norm": 1.8181157112121582,
+      "learning_rate": 5.74667501561871e-06,
+      "loss": 0.1003,
+      "step": 5162
+    },
+    {
+      "epoch": 2.704033525405972,
+      "loss_partial": 0.048828125,
+      "step": 5162
+    },
+    {
+      "epoch": 2.704033525405972,
+      "loss_partial": 0.051025390625,
+      "step": 5162
+    },
+    {
+      "epoch": 2.7045573598742796,
+      "grad_norm": 2.065812587738037,
+      "learning_rate": 5.74253800664036e-06,
+      "loss": 0.1011,
+      "step": 5163
+    },
+    {
+      "epoch": 2.7045573598742796,
+      "loss_partial": 0.052734375,
+      "step": 5163
+    },
+    {
+      "epoch": 2.7045573598742796,
+      "loss_partial": 0.049560546875,
+      "step": 5163
+    },
+    {
+      "epoch": 2.7050811943425876,
+      "grad_norm": 1.8653485774993896,
+      "learning_rate": 5.738401887445238e-06,
+      "loss": 0.1008,
+      "step": 5164
+    },
+    {
+      "epoch": 2.7050811943425876,
+      "loss_partial": 0.057861328125,
+      "step": 5164
+    },
+    {
+      "epoch": 2.7050811943425876,
+      "loss_partial": 0.0546875,
+      "step": 5164
+    },
+    {
+      "epoch": 2.7056050288108957,
+      "grad_norm": 2.1192543506622314,
+      "learning_rate": 5.7342666588977705e-06,
+      "loss": 0.102,
+      "step": 5165
+    },
+    {
+      "epoch": 2.7056050288108957,
+      "loss_partial": 0.044677734375,
+      "step": 5165
+    },
+    {
+      "epoch": 2.7056050288108957,
+      "loss_partial": 0.057861328125,
+      "step": 5165
+    },
+    {
+      "epoch": 2.7061288632792038,
+      "grad_norm": 1.42652428150177,
+      "learning_rate": 5.730132321862192e-06,
+      "loss": 0.1001,
+      "step": 5166
+    },
+    {
+      "epoch": 2.7061288632792038,
+      "loss_partial": 0.052001953125,
+      "step": 5166
+    },
+    {
+      "epoch": 2.7061288632792038,
+      "loss_partial": 0.056640625,
+      "step": 5166
+    },
+    {
+      "epoch": 2.706652697747512,
+      "grad_norm": 1.8222856521606445,
+      "learning_rate": 5.725998877202552e-06,
+      "loss": 0.1031,
+      "step": 5167
+    },
+    {
+      "epoch": 2.706652697747512,
+      "loss_partial": 0.061279296875,
+      "step": 5167
+    },
+    {
+      "epoch": 2.706652697747512,
+      "loss_partial": 0.047607421875,
+      "step": 5167
+    },
+    {
+      "epoch": 2.70717653221582,
+      "grad_norm": 2.3057966232299805,
+      "learning_rate": 5.721866325782716e-06,
+      "loss": 0.1016,
+      "step": 5168
+    },
+    {
+      "epoch": 2.70717653221582,
+      "loss_partial": 0.055908203125,
+      "step": 5168
+    },
+    {
+      "epoch": 2.70717653221582,
+      "loss_partial": 0.049072265625,
+      "step": 5168
+    },
+    {
+      "epoch": 2.707700366684128,
+      "grad_norm": 1.5558295249938965,
+      "learning_rate": 5.717734668466366e-06,
+      "loss": 0.0999,
+      "step": 5169
+    },
+    {
+      "epoch": 2.707700366684128,
+      "loss_partial": 0.0546875,
+      "step": 5169
+    },
+    {
+      "epoch": 2.707700366684128,
+      "loss_partial": 0.045654296875,
+      "step": 5169
+    },
+    {
+      "epoch": 2.7082242011524356,
+      "grad_norm": 2.137779474258423,
+      "learning_rate": 5.713603906116989e-06,
+      "loss": 0.0997,
+      "step": 5170
+    },
+    {
+      "epoch": 2.7082242011524356,
+      "loss_partial": 0.058837890625,
+      "step": 5170
+    },
+    {
+      "epoch": 2.7082242011524356,
+      "loss_partial": 0.053955078125,
+      "step": 5170
+    },
+    {
+      "epoch": 2.7087480356207436,
+      "grad_norm": 1.7185823917388916,
+      "learning_rate": 5.709474039597889e-06,
+      "loss": 0.1032,
+      "step": 5171
+    },
+    {
+      "epoch": 2.7087480356207436,
+      "loss_partial": 0.048583984375,
+      "step": 5171
+    },
+    {
+      "epoch": 2.7087480356207436,
+      "loss_partial": 0.041748046875,
+      "step": 5171
+    },
+    {
+      "epoch": 2.7092718700890517,
+      "grad_norm": 1.992856502532959,
+      "learning_rate": 5.705345069772179e-06,
+      "loss": 0.0983,
+      "step": 5172
+    },
+    {
+      "epoch": 2.7092718700890517,
+      "loss_partial": 0.0498046875,
+      "step": 5172
+    },
+    {
+      "epoch": 2.7092718700890517,
+      "loss_partial": 0.05126953125,
+      "step": 5172
+    },
+    {
+      "epoch": 2.7097957045573597,
+      "grad_norm": 1.4740389585494995,
+      "learning_rate": 5.7012169975027965e-06,
+      "loss": 0.1013,
+      "step": 5173
+    },
+    {
+      "epoch": 2.7097957045573597,
+      "loss_partial": 0.045166015625,
+      "step": 5173
+    },
+    {
+      "epoch": 2.7097957045573597,
+      "loss_partial": 0.048095703125,
+      "step": 5173
+    },
+    {
+      "epoch": 2.710319539025668,
+      "grad_norm": 1.9717679023742676,
+      "learning_rate": 5.697089823652479e-06,
+      "loss": 0.1013,
+      "step": 5174
+    },
+    {
+      "epoch": 2.710319539025668,
+      "loss_partial": 0.0625,
+      "step": 5174
+    },
+    {
+      "epoch": 2.710319539025668,
+      "loss_partial": 0.04736328125,
+      "step": 5174
+    },
+    {
+      "epoch": 2.710843373493976,
+      "grad_norm": 1.576818585395813,
+      "learning_rate": 5.6929635490837785e-06,
+      "loss": 0.1013,
+      "step": 5175
+    },
+    {
+      "epoch": 2.710843373493976,
+      "loss_partial": 0.05224609375,
+      "step": 5175
+    },
+    {
+      "epoch": 2.710843373493976,
+      "loss_partial": 0.056396484375,
+      "step": 5175
+    },
+    {
+      "epoch": 2.711367207962284,
+      "grad_norm": 1.9385721683502197,
+      "learning_rate": 5.688838174659064e-06,
+      "loss": 0.1011,
+      "step": 5176
+    },
+    {
+      "epoch": 2.711367207962284,
+      "loss_partial": 0.051513671875,
+      "step": 5176
+    },
+    {
+      "epoch": 2.711367207962284,
+      "loss_partial": 0.05322265625,
+      "step": 5176
+    },
+    {
+      "epoch": 2.711891042430592,
+      "grad_norm": 1.7649697065353394,
+      "learning_rate": 5.68471370124051e-06,
+      "loss": 0.1002,
+      "step": 5177
+    },
+    {
+      "epoch": 2.711891042430592,
+      "loss_partial": 0.046875,
+      "step": 5177
+    },
+    {
+      "epoch": 2.711891042430592,
+      "loss_partial": 0.048095703125,
+      "step": 5177
+    },
+    {
+      "epoch": 2.7124148768989,
+      "grad_norm": 1.924977421760559,
+      "learning_rate": 5.680590129690109e-06,
+      "loss": 0.0965,
+      "step": 5178
+    },
+    {
+      "epoch": 2.7124148768989,
+      "loss_partial": 0.05615234375,
+      "step": 5178
+    },
+    {
+      "epoch": 2.7124148768989,
+      "loss_partial": 0.0537109375,
+      "step": 5178
+    },
+    {
+      "epoch": 2.712938711367208,
+      "grad_norm": 1.5096074342727661,
+      "learning_rate": 5.6764674608696605e-06,
+      "loss": 0.0994,
+      "step": 5179
+    },
+    {
+      "epoch": 2.712938711367208,
+      "loss_partial": 0.05126953125,
+      "step": 5179
+    },
+    {
+      "epoch": 2.712938711367208,
+      "loss_partial": 0.05615234375,
+      "step": 5179
+    },
+    {
+      "epoch": 2.713462545835516,
+      "grad_norm": 1.398897409439087,
+      "learning_rate": 5.6723456956407755e-06,
+      "loss": 0.0996,
+      "step": 5180
+    },
+    {
+      "epoch": 2.713462545835516,
+      "loss_partial": 0.046142578125,
+      "step": 5180
+    },
+    {
+      "epoch": 2.713462545835516,
+      "loss_partial": 0.054443359375,
+      "step": 5180
+    },
+    {
+      "epoch": 2.713986380303824,
+      "grad_norm": 2.279258966445923,
+      "learning_rate": 5.668224834864875e-06,
+      "loss": 0.1016,
+      "step": 5181
+    },
+    {
+      "epoch": 2.713986380303824,
+      "loss_partial": 0.0537109375,
+      "step": 5181
+    },
+    {
+      "epoch": 2.713986380303824,
+      "loss_partial": 0.056884765625,
+      "step": 5181
+    },
+    {
+      "epoch": 2.7145102147721323,
+      "grad_norm": 1.360904335975647,
+      "learning_rate": 5.664104879403203e-06,
+      "loss": 0.0993,
+      "step": 5182
+    },
+    {
+      "epoch": 2.7145102147721323,
+      "loss_partial": 0.0439453125,
+      "step": 5182
+    },
+    {
+      "epoch": 2.7145102147721323,
+      "loss_partial": 0.05224609375,
+      "step": 5182
+    },
+    {
+      "epoch": 2.71503404924044,
+      "grad_norm": 1.5069891214370728,
+      "learning_rate": 5.659985830116796e-06,
+      "loss": 0.101,
+      "step": 5183
+    },
+    {
+      "epoch": 2.71503404924044,
+      "loss_partial": 0.0556640625,
+      "step": 5183
+    },
+    {
+      "epoch": 2.71503404924044,
+      "loss_partial": 0.06689453125,
+      "step": 5183
+    },
+    {
+      "epoch": 2.715557883708748,
+      "grad_norm": 1.4704906940460205,
+      "learning_rate": 5.655867687866514e-06,
+      "loss": 0.1039,
+      "step": 5184
+    },
+    {
+      "epoch": 2.715557883708748,
+      "loss_partial": 0.04638671875,
+      "step": 5184
+    },
+    {
+      "epoch": 2.715557883708748,
+      "loss_partial": 0.049072265625,
+      "step": 5184
+    },
+    {
+      "epoch": 2.716081718177056,
+      "grad_norm": 1.434053897857666,
+      "learning_rate": 5.651750453513019e-06,
+      "loss": 0.1027,
+      "step": 5185
+    },
+    {
+      "epoch": 2.716081718177056,
+      "loss_partial": 0.04541015625,
+      "step": 5185
+    },
+    {
+      "epoch": 2.716081718177056,
+      "loss_partial": 0.04541015625,
+      "step": 5185
+    },
+    {
+      "epoch": 2.716605552645364,
+      "grad_norm": 1.608617901802063,
+      "learning_rate": 5.647634127916793e-06,
+      "loss": 0.0992,
+      "step": 5186
+    },
+    {
+      "epoch": 2.716605552645364,
+      "loss_partial": 0.0400390625,
+      "step": 5186
+    },
+    {
+      "epoch": 2.716605552645364,
+      "loss_partial": 0.041748046875,
+      "step": 5186
+    },
+    {
+      "epoch": 2.717129387113672,
+      "grad_norm": 1.3126925230026245,
+      "learning_rate": 5.64351871193812e-06,
+      "loss": 0.1005,
+      "step": 5187
+    },
+    {
+      "epoch": 2.717129387113672,
+      "loss_partial": 0.0498046875,
+      "step": 5187
+    },
+    {
+      "epoch": 2.717129387113672,
+      "loss_partial": 0.052978515625,
+      "step": 5187
+    },
+    {
+      "epoch": 2.71765322158198,
+      "grad_norm": 1.7756940126419067,
+      "learning_rate": 5.639404206437098e-06,
+      "loss": 0.1026,
+      "step": 5188
+    },
+    {
+      "epoch": 2.71765322158198,
+      "loss_partial": 0.04345703125,
+      "step": 5188
+    },
+    {
+      "epoch": 2.71765322158198,
+      "loss_partial": 0.05029296875,
+      "step": 5188
+    },
+    {
+      "epoch": 2.7181770560502883,
+      "grad_norm": 1.2675230503082275,
+      "learning_rate": 5.635290612273629e-06,
+      "loss": 0.1022,
+      "step": 5189
+    },
+    {
+      "epoch": 2.7181770560502883,
+      "loss_partial": 0.0595703125,
+      "step": 5189
+    },
+    {
+      "epoch": 2.7181770560502883,
+      "loss_partial": 0.04443359375,
+      "step": 5189
+    },
+    {
+      "epoch": 2.718700890518596,
+      "grad_norm": 1.5546133518218994,
+      "learning_rate": 5.631177930307435e-06,
+      "loss": 0.0997,
+      "step": 5190
+    },
+    {
+      "epoch": 2.718700890518596,
+      "loss_partial": 0.052001953125,
+      "step": 5190
+    },
+    {
+      "epoch": 2.718700890518596,
+      "loss_partial": 0.048095703125,
+      "step": 5190
+    },
+    {
+      "epoch": 2.719224724986904,
+      "grad_norm": 1.3255176544189453,
+      "learning_rate": 5.627066161398035e-06,
+      "loss": 0.0997,
+      "step": 5191
+    },
+    {
+      "epoch": 2.719224724986904,
+      "loss_partial": 0.05322265625,
+      "step": 5191
+    },
+    {
+      "epoch": 2.719224724986904,
+      "loss_partial": 0.052001953125,
+      "step": 5191
+    },
+    {
+      "epoch": 2.719748559455212,
+      "grad_norm": 1.4458589553833008,
+      "learning_rate": 5.622955306404767e-06,
+      "loss": 0.1037,
+      "step": 5192
+    },
+    {
+      "epoch": 2.719748559455212,
+      "loss_partial": 0.057373046875,
+      "step": 5192
+    },
+    {
+      "epoch": 2.719748559455212,
+      "loss_partial": 0.0546875,
+      "step": 5192
+    },
+    {
+      "epoch": 2.72027239392352,
+      "grad_norm": 1.8821038007736206,
+      "learning_rate": 5.618845366186768e-06,
+      "loss": 0.104,
+      "step": 5193
+    },
+    {
+      "epoch": 2.72027239392352,
+      "loss_partial": 0.05078125,
+      "step": 5193
+    },
+    {
+      "epoch": 2.72027239392352,
+      "loss_partial": 0.05322265625,
+      "step": 5193
+    },
+    {
+      "epoch": 2.720796228391828,
+      "grad_norm": 1.3328200578689575,
+      "learning_rate": 5.614736341603e-06,
+      "loss": 0.1003,
+      "step": 5194
+    },
+    {
+      "epoch": 2.720796228391828,
+      "loss_partial": 0.049072265625,
+      "step": 5194
+    },
+    {
+      "epoch": 2.720796228391828,
+      "loss_partial": 0.051513671875,
+      "step": 5194
+    },
+    {
+      "epoch": 2.721320062860136,
+      "grad_norm": 1.4582477807998657,
+      "learning_rate": 5.610628233512219e-06,
+      "loss": 0.0982,
+      "step": 5195
+    },
+    {
+      "epoch": 2.721320062860136,
+      "loss_partial": 0.04736328125,
+      "step": 5195
+    },
+    {
+      "epoch": 2.721320062860136,
+      "loss_partial": 0.04296875,
+      "step": 5195
+    },
+    {
+      "epoch": 2.7218438973284442,
+      "grad_norm": 1.5928739309310913,
+      "learning_rate": 5.606521042772994e-06,
+      "loss": 0.098,
+      "step": 5196
+    },
+    {
+      "epoch": 2.7218438973284442,
+      "loss_partial": 0.051025390625,
+      "step": 5196
+    },
+    {
+      "epoch": 2.7218438973284442,
+      "loss_partial": 0.04833984375,
+      "step": 5196
+    },
+    {
+      "epoch": 2.7223677317967523,
+      "grad_norm": 1.0890388488769531,
+      "learning_rate": 5.602414770243698e-06,
+      "loss": 0.1013,
+      "step": 5197
+    },
+    {
+      "epoch": 2.7223677317967523,
+      "loss_partial": 0.05126953125,
+      "step": 5197
+    },
+    {
+      "epoch": 2.7223677317967523,
+      "loss_partial": 0.04638671875,
+      "step": 5197
+    },
+    {
+      "epoch": 2.7228915662650603,
+      "grad_norm": 1.4428671598434448,
+      "learning_rate": 5.598309416782526e-06,
+      "loss": 0.1024,
+      "step": 5198
+    },
+    {
+      "epoch": 2.7228915662650603,
+      "loss_partial": 0.064453125,
+      "step": 5198
+    },
+    {
+      "epoch": 2.7228915662650603,
+      "loss_partial": 0.052978515625,
+      "step": 5198
+    },
+    {
+      "epoch": 2.7234154007333684,
+      "grad_norm": 1.1735591888427734,
+      "learning_rate": 5.594204983247466e-06,
+      "loss": 0.1033,
+      "step": 5199
+    },
+    {
+      "epoch": 2.7234154007333684,
+      "loss_partial": 0.042724609375,
+      "step": 5199
+    },
+    {
+      "epoch": 2.7234154007333684,
+      "loss_partial": 0.049560546875,
+      "step": 5199
+    },
+    {
+      "epoch": 2.7239392352016765,
+      "grad_norm": 1.8219960927963257,
+      "learning_rate": 5.590101470496321e-06,
+      "loss": 0.1003,
+      "step": 5200
+    },
+    {
+      "epoch": 2.7239392352016765,
+      "loss_partial": 0.054931640625,
+      "step": 5200
+    },
+    {
+      "epoch": 2.7239392352016765,
+      "loss_partial": 0.0458984375,
+      "step": 5200
+    },
+    {
+      "epoch": 2.7244630696699845,
+      "grad_norm": 1.3409608602523804,
+      "learning_rate": 5.585998879386698e-06,
+      "loss": 0.1006,
+      "step": 5201
+    },
+    {
+      "epoch": 2.7244630696699845,
+      "loss_partial": 0.046142578125,
+      "step": 5201
+    },
+    {
+      "epoch": 2.7244630696699845,
+      "loss_partial": 0.052490234375,
+      "step": 5201
+    },
+    {
+      "epoch": 2.724986904138292,
+      "grad_norm": 1.6202212572097778,
+      "learning_rate": 5.581897210776015e-06,
+      "loss": 0.1021,
+      "step": 5202
+    },
+    {
+      "epoch": 2.724986904138292,
+      "loss_partial": 0.04296875,
+      "step": 5202
+    },
+    {
+      "epoch": 2.724986904138292,
+      "loss_partial": 0.046630859375,
+      "step": 5202
+    },
+    {
+      "epoch": 2.7255107386066,
+      "grad_norm": 1.4444124698638916,
+      "learning_rate": 5.577796465521492e-06,
+      "loss": 0.0975,
+      "step": 5203
+    },
+    {
+      "epoch": 2.7255107386066,
+      "loss_partial": 0.052978515625,
+      "step": 5203
+    },
+    {
+      "epoch": 2.7255107386066,
+      "loss_partial": 0.058349609375,
+      "step": 5203
+    },
+    {
+      "epoch": 2.7260345730749083,
+      "grad_norm": 1.5480196475982666,
+      "learning_rate": 5.573696644480163e-06,
+      "loss": 0.1021,
+      "step": 5204
+    },
+    {
+      "epoch": 2.7260345730749083,
+      "loss_partial": 0.038818359375,
+      "step": 5204
+    },
+    {
+      "epoch": 2.7260345730749083,
+      "loss_partial": 0.04931640625,
+      "step": 5204
+    },
+    {
+      "epoch": 2.7265584075432163,
+      "grad_norm": 1.3006184101104736,
+      "learning_rate": 5.569597748508859e-06,
+      "loss": 0.0975,
+      "step": 5205
+    },
+    {
+      "epoch": 2.7265584075432163,
+      "loss_partial": 0.05322265625,
+      "step": 5205
+    },
+    {
+      "epoch": 2.7265584075432163,
+      "loss_partial": 0.055419921875,
+      "step": 5205
+    },
+    {
+      "epoch": 2.7270822420115244,
+      "grad_norm": 1.2737233638763428,
+      "learning_rate": 5.565499778464234e-06,
+      "loss": 0.0971,
+      "step": 5206
+    },
+    {
+      "epoch": 2.7270822420115244,
+      "loss_partial": 0.046142578125,
+      "step": 5206
+    },
+    {
+      "epoch": 2.7270822420115244,
+      "loss_partial": 0.05126953125,
+      "step": 5206
+    },
+    {
+      "epoch": 2.7276060764798324,
+      "grad_norm": 1.9489216804504395,
+      "learning_rate": 5.561402735202732e-06,
+      "loss": 0.1008,
+      "step": 5207
+    },
+    {
+      "epoch": 2.7276060764798324,
+      "loss_partial": 0.051513671875,
+      "step": 5207
+    },
+    {
+      "epoch": 2.7276060764798324,
+      "loss_partial": 0.04248046875,
+      "step": 5207
+    },
+    {
+      "epoch": 2.7281299109481405,
+      "grad_norm": 1.335628628730774,
+      "learning_rate": 5.557306619580613e-06,
+      "loss": 0.0986,
+      "step": 5208
+    },
+    {
+      "epoch": 2.7281299109481405,
+      "loss_partial": 0.0595703125,
+      "step": 5208
+    },
+    {
+      "epoch": 2.7281299109481405,
+      "loss_partial": 0.05126953125,
+      "step": 5208
+    },
+    {
+      "epoch": 2.7286537454164486,
+      "grad_norm": 1.7476102113723755,
+      "learning_rate": 5.553211432453931e-06,
+      "loss": 0.0999,
+      "step": 5209
+    },
+    {
+      "epoch": 2.7286537454164486,
+      "loss_partial": 0.052978515625,
+      "step": 5209
+    },
+    {
+      "epoch": 2.7286537454164486,
+      "loss_partial": 0.0498046875,
+      "step": 5209
+    },
+    {
+      "epoch": 2.729177579884756,
+      "grad_norm": 1.738525152206421,
+      "learning_rate": 5.549117174678566e-06,
+      "loss": 0.1025,
+      "step": 5210
+    },
+    {
+      "epoch": 2.729177579884756,
+      "loss_partial": 0.05224609375,
+      "step": 5210
+    },
+    {
+      "epoch": 2.729177579884756,
+      "loss_partial": 0.044921875,
+      "step": 5210
+    },
+    {
+      "epoch": 2.7297014143530642,
+      "grad_norm": 1.7375999689102173,
+      "learning_rate": 5.545023847110189e-06,
+      "loss": 0.1027,
+      "step": 5211
+    },
+    {
+      "epoch": 2.7297014143530642,
+      "loss_partial": 0.04931640625,
+      "step": 5211
+    },
+    {
+      "epoch": 2.7297014143530642,
+      "loss_partial": 0.042724609375,
+      "step": 5211
+    },
+    {
+      "epoch": 2.7302252488213723,
+      "grad_norm": 1.721333622932434,
+      "learning_rate": 5.54093145060428e-06,
+      "loss": 0.1008,
+      "step": 5212
+    },
+    {
+      "epoch": 2.7302252488213723,
+      "loss_partial": 0.04931640625,
+      "step": 5212
+    },
+    {
+      "epoch": 2.7302252488213723,
+      "loss_partial": 0.05078125,
+      "step": 5212
+    },
+    {
+      "epoch": 2.7307490832896804,
+      "grad_norm": 1.91456937789917,
+      "learning_rate": 5.5368399860161225e-06,
+      "loss": 0.1016,
+      "step": 5213
+    },
+    {
+      "epoch": 2.7307490832896804,
+      "loss_partial": 0.05322265625,
+      "step": 5213
+    },
+    {
+      "epoch": 2.7307490832896804,
+      "loss_partial": 0.06591796875,
+      "step": 5213
+    },
+    {
+      "epoch": 2.7312729177579884,
+      "grad_norm": 1.605788230895996,
+      "learning_rate": 5.532749454200809e-06,
+      "loss": 0.0997,
+      "step": 5214
+    },
+    {
+      "epoch": 2.7312729177579884,
+      "loss_partial": 0.049072265625,
+      "step": 5214
+    },
+    {
+      "epoch": 2.7312729177579884,
+      "loss_partial": 0.05029296875,
+      "step": 5214
+    },
+    {
+      "epoch": 2.7317967522262965,
+      "grad_norm": 1.415568232536316,
+      "learning_rate": 5.528659856013234e-06,
+      "loss": 0.0999,
+      "step": 5215
+    },
+    {
+      "epoch": 2.7317967522262965,
+      "loss_partial": 0.060302734375,
+      "step": 5215
+    },
+    {
+      "epoch": 2.7317967522262965,
+      "loss_partial": 0.048828125,
+      "step": 5215
+    },
+    {
+      "epoch": 2.7323205866946045,
+      "grad_norm": 1.377583384513855,
+      "learning_rate": 5.5245711923081e-06,
+      "loss": 0.1019,
+      "step": 5216
+    },
+    {
+      "epoch": 2.7323205866946045,
+      "loss_partial": 0.0478515625,
+      "step": 5216
+    },
+    {
+      "epoch": 2.7323205866946045,
+      "loss_partial": 0.057373046875,
+      "step": 5216
+    },
+    {
+      "epoch": 2.7328444211629126,
+      "grad_norm": 1.482014536857605,
+      "learning_rate": 5.520483463939907e-06,
+      "loss": 0.1012,
+      "step": 5217
+    },
+    {
+      "epoch": 2.7328444211629126,
+      "loss_partial": 0.05517578125,
+      "step": 5217
+    },
+    {
+      "epoch": 2.7328444211629126,
+      "loss_partial": 0.0419921875,
+      "step": 5217
+    },
+    {
+      "epoch": 2.7333682556312207,
+      "grad_norm": 1.3575574159622192,
+      "learning_rate": 5.516396671762975e-06,
+      "loss": 0.1022,
+      "step": 5218
+    },
+    {
+      "epoch": 2.7333682556312207,
+      "loss_partial": 0.04736328125,
+      "step": 5218
+    },
+    {
+      "epoch": 2.7333682556312207,
+      "loss_partial": 0.064453125,
+      "step": 5218
+    },
+    {
+      "epoch": 2.7338920900995287,
+      "grad_norm": 1.3860487937927246,
+      "learning_rate": 5.512310816631412e-06,
+      "loss": 0.0983,
+      "step": 5219
+    },
+    {
+      "epoch": 2.7338920900995287,
+      "loss_partial": 0.05029296875,
+      "step": 5219
+    },
+    {
+      "epoch": 2.7338920900995287,
+      "loss_partial": 0.04638671875,
+      "step": 5219
+    },
+    {
+      "epoch": 2.7344159245678368,
+      "grad_norm": 1.4833011627197266,
+      "learning_rate": 5.508225899399136e-06,
+      "loss": 0.0981,
+      "step": 5220
+    },
+    {
+      "epoch": 2.7344159245678368,
+      "loss_partial": 0.048095703125,
+      "step": 5220
+    },
+    {
+      "epoch": 2.7344159245678368,
+      "loss_partial": 0.046630859375,
+      "step": 5220
+    },
+    {
+      "epoch": 2.734939759036145,
+      "grad_norm": 2.4208755493164062,
+      "learning_rate": 5.5041419209198684e-06,
+      "loss": 0.1024,
+      "step": 5221
+    },
+    {
+      "epoch": 2.734939759036145,
+      "loss_partial": 0.054931640625,
+      "step": 5221
+    },
+    {
+      "epoch": 2.734939759036145,
+      "loss_partial": 0.0439453125,
+      "step": 5221
+    },
+    {
+      "epoch": 2.7354635935044525,
+      "grad_norm": 1.5347118377685547,
+      "learning_rate": 5.500058882047141e-06,
+      "loss": 0.1015,
+      "step": 5222
+    },
+    {
+      "epoch": 2.7354635935044525,
+      "loss_partial": 0.0458984375,
+      "step": 5222
+    },
+    {
+      "epoch": 2.7354635935044525,
+      "loss_partial": 0.053955078125,
+      "step": 5222
+    },
+    {
+      "epoch": 2.7359874279727605,
+      "grad_norm": 1.778318166732788,
+      "learning_rate": 5.495976783634282e-06,
+      "loss": 0.1059,
+      "step": 5223
+    },
+    {
+      "epoch": 2.7359874279727605,
+      "loss_partial": 0.046142578125,
+      "step": 5223
+    },
+    {
+      "epoch": 2.7359874279727605,
+      "loss_partial": 0.046630859375,
+      "step": 5223
+    },
+    {
+      "epoch": 2.7365112624410686,
+      "grad_norm": 1.8373955488204956,
+      "learning_rate": 5.49189562653442e-06,
+      "loss": 0.1036,
+      "step": 5224
+    },
+    {
+      "epoch": 2.7365112624410686,
+      "loss_partial": 0.058837890625,
+      "step": 5224
+    },
+    {
+      "epoch": 2.7365112624410686,
+      "loss_partial": 0.045654296875,
+      "step": 5224
+    },
+    {
+      "epoch": 2.7370350969093766,
+      "grad_norm": 1.8450877666473389,
+      "learning_rate": 5.4878154116004975e-06,
+      "loss": 0.1007,
+      "step": 5225
+    },
+    {
+      "epoch": 2.7370350969093766,
+      "loss_partial": 0.041015625,
+      "step": 5225
+    },
+    {
+      "epoch": 2.7370350969093766,
+      "loss_partial": 0.044189453125,
+      "step": 5225
+    },
+    {
+      "epoch": 2.7375589313776847,
+      "grad_norm": 1.7491931915283203,
+      "learning_rate": 5.483736139685251e-06,
+      "loss": 0.0994,
+      "step": 5226
+    },
+    {
+      "epoch": 2.7375589313776847,
+      "loss_partial": 0.049072265625,
+      "step": 5226
+    },
+    {
+      "epoch": 2.7375589313776847,
+      "loss_partial": 0.057373046875,
+      "step": 5226
+    },
+    {
+      "epoch": 2.7380827658459927,
+      "grad_norm": 2.077334403991699,
+      "learning_rate": 5.479657811641223e-06,
+      "loss": 0.1005,
+      "step": 5227
+    },
+    {
+      "epoch": 2.7380827658459927,
+      "loss_partial": 0.05712890625,
+      "step": 5227
+    },
+    {
+      "epoch": 2.7380827658459927,
+      "loss_partial": 0.0537109375,
+      "step": 5227
+    },
+    {
+      "epoch": 2.738606600314301,
+      "grad_norm": 1.5168898105621338,
+      "learning_rate": 5.475580428320759e-06,
+      "loss": 0.1021,
+      "step": 5228
+    },
+    {
+      "epoch": 2.738606600314301,
+      "loss_partial": 0.05419921875,
+      "step": 5228
+    },
+    {
+      "epoch": 2.738606600314301,
+      "loss_partial": 0.0439453125,
+      "step": 5228
+    },
+    {
+      "epoch": 2.7391304347826084,
+      "grad_norm": 2.463688850402832,
+      "learning_rate": 5.471503990576005e-06,
+      "loss": 0.1001,
+      "step": 5229
+    },
+    {
+      "epoch": 2.7391304347826084,
+      "loss_partial": 0.046875,
+      "step": 5229
+    },
+    {
+      "epoch": 2.7391304347826084,
+      "loss_partial": 0.051025390625,
+      "step": 5229
+    },
+    {
+      "epoch": 2.7396542692509165,
+      "grad_norm": 1.3839027881622314,
+      "learning_rate": 5.467428499258916e-06,
+      "loss": 0.0991,
+      "step": 5230
+    },
+    {
+      "epoch": 2.7396542692509165,
+      "loss_partial": 0.04638671875,
+      "step": 5230
+    },
+    {
+      "epoch": 2.7396542692509165,
+      "loss_partial": 0.052734375,
+      "step": 5230
+    },
+    {
+      "epoch": 2.7401781037192245,
+      "grad_norm": 1.2342360019683838,
+      "learning_rate": 5.4633539552212425e-06,
+      "loss": 0.0984,
+      "step": 5231
+    },
+    {
+      "epoch": 2.7401781037192245,
+      "loss_partial": 0.052490234375,
+      "step": 5231
+    },
+    {
+      "epoch": 2.7401781037192245,
+      "loss_partial": 0.045166015625,
+      "step": 5231
+    },
+    {
+      "epoch": 2.7407019381875326,
+      "grad_norm": 1.4910300970077515,
+      "learning_rate": 5.459280359314538e-06,
+      "loss": 0.1038,
+      "step": 5232
+    },
+    {
+      "epoch": 2.7407019381875326,
+      "loss_partial": 0.0556640625,
+      "step": 5232
+    },
+    {
+      "epoch": 2.7407019381875326,
+      "loss_partial": 0.05859375,
+      "step": 5232
+    },
+    {
+      "epoch": 2.7412257726558407,
+      "grad_norm": 1.0703483819961548,
+      "learning_rate": 5.45520771239016e-06,
+      "loss": 0.1012,
+      "step": 5233
+    },
+    {
+      "epoch": 2.7412257726558407,
+      "loss_partial": 0.060302734375,
+      "step": 5233
+    },
+    {
+      "epoch": 2.7412257726558407,
+      "loss_partial": 0.055908203125,
+      "step": 5233
+    },
+    {
+      "epoch": 2.7417496071241487,
+      "grad_norm": 1.288394808769226,
+      "learning_rate": 5.451136015299263e-06,
+      "loss": 0.1052,
+      "step": 5234
+    },
+    {
+      "epoch": 2.7417496071241487,
+      "loss_partial": 0.052001953125,
+      "step": 5234
+    },
+    {
+      "epoch": 2.7417496071241487,
+      "loss_partial": 0.0498046875,
+      "step": 5234
+    },
+    {
+      "epoch": 2.742273441592457,
+      "grad_norm": 1.3463680744171143,
+      "learning_rate": 5.447065268892813e-06,
+      "loss": 0.0984,
+      "step": 5235
+    },
+    {
+      "epoch": 2.742273441592457,
+      "loss_partial": 0.04931640625,
+      "step": 5235
+    },
+    {
+      "epoch": 2.742273441592457,
+      "loss_partial": 0.04345703125,
+      "step": 5235
+    },
+    {
+      "epoch": 2.742797276060765,
+      "grad_norm": 1.2031760215759277,
+      "learning_rate": 5.442995474021566e-06,
+      "loss": 0.1016,
+      "step": 5236
+    },
+    {
+      "epoch": 2.742797276060765,
+      "loss_partial": 0.04833984375,
+      "step": 5236
+    },
+    {
+      "epoch": 2.742797276060765,
+      "loss_partial": 0.044921875,
+      "step": 5236
+    },
+    {
+      "epoch": 2.743321110529073,
+      "grad_norm": 1.1336536407470703,
+      "learning_rate": 5.438926631536087e-06,
+      "loss": 0.0973,
+      "step": 5237
+    },
+    {
+      "epoch": 2.743321110529073,
+      "loss_partial": 0.05615234375,
+      "step": 5237
+    },
+    {
+      "epoch": 2.743321110529073,
+      "loss_partial": 0.048828125,
+      "step": 5237
+    },
+    {
+      "epoch": 2.743844944997381,
+      "grad_norm": 1.2924784421920776,
+      "learning_rate": 5.434858742286735e-06,
+      "loss": 0.1043,
+      "step": 5238
+    },
+    {
+      "epoch": 2.743844944997381,
+      "loss_partial": 0.04931640625,
+      "step": 5238
+    },
+    {
+      "epoch": 2.743844944997381,
+      "loss_partial": 0.049072265625,
+      "step": 5238
+    },
+    {
+      "epoch": 2.744368779465689,
+      "grad_norm": 1.7023720741271973,
+      "learning_rate": 5.430791807123678e-06,
+      "loss": 0.1006,
+      "step": 5239
+    },
+    {
+      "epoch": 2.744368779465689,
+      "loss_partial": 0.05078125,
+      "step": 5239
+    },
+    {
+      "epoch": 2.744368779465689,
+      "loss_partial": 0.038330078125,
+      "step": 5239
+    },
+    {
+      "epoch": 2.744892613933997,
+      "grad_norm": 1.633498191833496,
+      "learning_rate": 5.426725826896877e-06,
+      "loss": 0.1024,
+      "step": 5240
+    },
+    {
+      "epoch": 2.744892613933997,
+      "loss_partial": 0.0458984375,
+      "step": 5240
+    },
+    {
+      "epoch": 2.744892613933997,
+      "loss_partial": 0.043212890625,
+      "step": 5240
+    },
+    {
+      "epoch": 2.745416448402305,
+      "grad_norm": 1.4652763605117798,
+      "learning_rate": 5.422660802456099e-06,
+      "loss": 0.0953,
+      "step": 5241
+    },
+    {
+      "epoch": 2.745416448402305,
+      "loss_partial": 0.052001953125,
+      "step": 5241
+    },
+    {
+      "epoch": 2.745416448402305,
+      "loss_partial": 0.050537109375,
+      "step": 5241
+    },
+    {
+      "epoch": 2.7459402828706128,
+      "grad_norm": 1.610593557357788,
+      "learning_rate": 5.418596734650903e-06,
+      "loss": 0.0982,
+      "step": 5242
+    },
+    {
+      "epoch": 2.7459402828706128,
+      "loss_partial": 0.05029296875,
+      "step": 5242
+    },
+    {
+      "epoch": 2.7459402828706128,
+      "loss_partial": 0.056396484375,
+      "step": 5242
+    },
+    {
+      "epoch": 2.746464117338921,
+      "grad_norm": 1.1498560905456543,
+      "learning_rate": 5.414533624330663e-06,
+      "loss": 0.1008,
+      "step": 5243
+    },
+    {
+      "epoch": 2.746464117338921,
+      "loss_partial": 0.05419921875,
+      "step": 5243
+    },
+    {
+      "epoch": 2.746464117338921,
+      "loss_partial": 0.0498046875,
+      "step": 5243
+    },
+    {
+      "epoch": 2.746987951807229,
+      "grad_norm": 1.9127659797668457,
+      "learning_rate": 5.410471472344539e-06,
+      "loss": 0.1043,
+      "step": 5244
+    },
+    {
+      "epoch": 2.746987951807229,
+      "loss_partial": 0.041259765625,
+      "step": 5244
+    },
+    {
+      "epoch": 2.746987951807229,
+      "loss_partial": 0.056640625,
+      "step": 5244
+    },
+    {
+      "epoch": 2.747511786275537,
+      "grad_norm": 1.3487673997879028,
+      "learning_rate": 5.406410279541498e-06,
+      "loss": 0.0985,
+      "step": 5245
+    },
+    {
+      "epoch": 2.747511786275537,
+      "loss_partial": 0.048828125,
+      "step": 5245
+    },
+    {
+      "epoch": 2.747511786275537,
+      "loss_partial": 0.0498046875,
+      "step": 5245
+    },
+    {
+      "epoch": 2.748035620743845,
+      "grad_norm": 1.5018492937088013,
+      "learning_rate": 5.402350046770297e-06,
+      "loss": 0.1026,
+      "step": 5246
+    },
+    {
+      "epoch": 2.748035620743845,
+      "loss_partial": 0.054443359375,
+      "step": 5246
+    },
+    {
+      "epoch": 2.748035620743845,
+      "loss_partial": 0.050048828125,
+      "step": 5246
+    },
+    {
+      "epoch": 2.748559455212153,
+      "grad_norm": 1.460072636604309,
+      "learning_rate": 5.398290774879511e-06,
+      "loss": 0.0995,
+      "step": 5247
+    },
+    {
+      "epoch": 2.748559455212153,
+      "loss_partial": 0.0625,
+      "step": 5247
+    },
+    {
+      "epoch": 2.748559455212153,
+      "loss_partial": 0.04931640625,
+      "step": 5247
+    },
+    {
+      "epoch": 2.749083289680461,
+      "grad_norm": 1.4277451038360596,
+      "learning_rate": 5.394232464717496e-06,
+      "loss": 0.1,
+      "step": 5248
+    },
+    {
+      "epoch": 2.749083289680461,
+      "loss_partial": 0.05078125,
+      "step": 5248
+    },
+    {
+      "epoch": 2.749083289680461,
+      "loss_partial": 0.04541015625,
+      "step": 5248
+    },
+    {
+      "epoch": 2.7496071241487687,
+      "grad_norm": 1.3446547985076904,
+      "learning_rate": 5.390175117132415e-06,
+      "loss": 0.1013,
+      "step": 5249
+    },
+    {
+      "epoch": 2.7496071241487687,
+      "loss_partial": 0.05078125,
+      "step": 5249
+    },
+    {
+      "epoch": 2.7496071241487687,
+      "loss_partial": 0.042724609375,
+      "step": 5249
+    },
+    {
+      "epoch": 2.750130958617077,
+      "grad_norm": 1.3506314754486084,
+      "learning_rate": 5.3861187329722276e-06,
+      "loss": 0.0973,
+      "step": 5250
+    },
+    {
+      "epoch": 2.750130958617077,
+      "loss_partial": 0.04736328125,
+      "step": 5250
+    },
+    {
+      "epoch": 2.750130958617077,
+      "loss_partial": 0.05517578125,
+      "step": 5250
+    },
+    {
+      "epoch": 2.750654793085385,
+      "grad_norm": 1.6099951267242432,
+      "learning_rate": 5.382063313084695e-06,
+      "loss": 0.1006,
+      "step": 5251
+    },
+    {
+      "epoch": 2.750654793085385,
+      "loss_partial": 0.042724609375,
+      "step": 5251
+    },
+    {
+      "epoch": 2.750654793085385,
+      "loss_partial": 0.052734375,
+      "step": 5251
+    },
+    {
+      "epoch": 2.751178627553693,
+      "grad_norm": 1.392758846282959,
+      "learning_rate": 5.378008858317372e-06,
+      "loss": 0.1008,
+      "step": 5252
+    },
+    {
+      "epoch": 2.751178627553693,
+      "loss_partial": 0.04833984375,
+      "step": 5252
+    },
+    {
+      "epoch": 2.751178627553693,
+      "loss_partial": 0.05322265625,
+      "step": 5252
+    },
+    {
+      "epoch": 2.751702462022001,
+      "grad_norm": 1.3279714584350586,
+      "learning_rate": 5.373955369517616e-06,
+      "loss": 0.0999,
+      "step": 5253
+    },
+    {
+      "epoch": 2.751702462022001,
+      "loss_partial": 0.05615234375,
+      "step": 5253
+    },
+    {
+      "epoch": 2.751702462022001,
+      "loss_partial": 0.056640625,
+      "step": 5253
+    },
+    {
+      "epoch": 2.752226296490309,
+      "grad_norm": 1.6323341131210327,
+      "learning_rate": 5.369902847532579e-06,
+      "loss": 0.1022,
+      "step": 5254
+    },
+    {
+      "epoch": 2.752226296490309,
+      "loss_partial": 0.041259765625,
+      "step": 5254
+    },
+    {
+      "epoch": 2.752226296490309,
+      "loss_partial": 0.055908203125,
+      "step": 5254
+    },
+    {
+      "epoch": 2.752750130958617,
+      "grad_norm": 1.4680441617965698,
+      "learning_rate": 5.365851293209218e-06,
+      "loss": 0.1024,
+      "step": 5255
+    },
+    {
+      "epoch": 2.752750130958617,
+      "loss_partial": 0.056884765625,
+      "step": 5255
+    },
+    {
+      "epoch": 2.752750130958617,
+      "loss_partial": 0.04736328125,
+      "step": 5255
+    },
+    {
+      "epoch": 2.753273965426925,
+      "grad_norm": 1.5115967988967896,
+      "learning_rate": 5.3618007073942805e-06,
+      "loss": 0.1008,
+      "step": 5256
+    },
+    {
+      "epoch": 2.753273965426925,
+      "loss_partial": 0.04736328125,
+      "step": 5256
+    },
+    {
+      "epoch": 2.753273965426925,
+      "loss_partial": 0.05517578125,
+      "step": 5256
+    },
+    {
+      "epoch": 2.753797799895233,
+      "grad_norm": 1.2955667972564697,
+      "learning_rate": 5.357751090934312e-06,
+      "loss": 0.1013,
+      "step": 5257
+    },
+    {
+      "epoch": 2.753797799895233,
+      "loss_partial": 0.048583984375,
+      "step": 5257
+    },
+    {
+      "epoch": 2.753797799895233,
+      "loss_partial": 0.053466796875,
+      "step": 5257
+    },
+    {
+      "epoch": 2.7543216343635413,
+      "grad_norm": 1.8946170806884766,
+      "learning_rate": 5.353702444675656e-06,
+      "loss": 0.0967,
+      "step": 5258
+    },
+    {
+      "epoch": 2.7543216343635413,
+      "loss_partial": 0.050537109375,
+      "step": 5258
+    },
+    {
+      "epoch": 2.7543216343635413,
+      "loss_partial": 0.05029296875,
+      "step": 5258
+    },
+    {
+      "epoch": 2.7548454688318493,
+      "grad_norm": 1.6877611875534058,
+      "learning_rate": 5.349654769464461e-06,
+      "loss": 0.1015,
+      "step": 5259
+    },
+    {
+      "epoch": 2.7548454688318493,
+      "loss_partial": 0.04150390625,
+      "step": 5259
+    },
+    {
+      "epoch": 2.7548454688318493,
+      "loss_partial": 0.0546875,
+      "step": 5259
+    },
+    {
+      "epoch": 2.7553693033001574,
+      "grad_norm": 1.3589242696762085,
+      "learning_rate": 5.345608066146662e-06,
+      "loss": 0.102,
+      "step": 5260
+    },
+    {
+      "epoch": 2.7553693033001574,
+      "loss_partial": 0.058349609375,
+      "step": 5260
+    },
+    {
+      "epoch": 2.7553693033001574,
+      "loss_partial": 0.051513671875,
+      "step": 5260
+    },
+    {
+      "epoch": 2.755893137768465,
+      "grad_norm": 2.245070695877075,
+      "learning_rate": 5.341562335567996e-06,
+      "loss": 0.1025,
+      "step": 5261
+    },
+    {
+      "epoch": 2.755893137768465,
+      "loss_partial": 0.050537109375,
+      "step": 5261
+    },
+    {
+      "epoch": 2.755893137768465,
+      "loss_partial": 0.056640625,
+      "step": 5261
+    },
+    {
+      "epoch": 2.756416972236773,
+      "grad_norm": 1.8057489395141602,
+      "learning_rate": 5.337517578573994e-06,
+      "loss": 0.1024,
+      "step": 5262
+    },
+    {
+      "epoch": 2.756416972236773,
+      "loss_partial": 0.06298828125,
+      "step": 5262
+    },
+    {
+      "epoch": 2.756416972236773,
+      "loss_partial": 0.0537109375,
+      "step": 5262
+    },
+    {
+      "epoch": 2.756940806705081,
+      "grad_norm": 1.6319595575332642,
+      "learning_rate": 5.333473796009987e-06,
+      "loss": 0.1025,
+      "step": 5263
+    },
+    {
+      "epoch": 2.756940806705081,
+      "loss_partial": 0.06640625,
+      "step": 5263
+    },
+    {
+      "epoch": 2.756940806705081,
+      "loss_partial": 0.053466796875,
+      "step": 5263
+    },
+    {
+      "epoch": 2.757464641173389,
+      "grad_norm": 1.3168994188308716,
+      "learning_rate": 5.329430988721098e-06,
+      "loss": 0.1046,
+      "step": 5264
+    },
+    {
+      "epoch": 2.757464641173389,
+      "loss_partial": 0.054443359375,
+      "step": 5264
+    },
+    {
+      "epoch": 2.757464641173389,
+      "loss_partial": 0.05517578125,
+      "step": 5264
+    },
+    {
+      "epoch": 2.7579884756416972,
+      "grad_norm": 2.4339802265167236,
+      "learning_rate": 5.325389157552252e-06,
+      "loss": 0.0988,
+      "step": 5265
+    },
+    {
+      "epoch": 2.7579884756416972,
+      "loss_partial": 0.05517578125,
+      "step": 5265
+    },
+    {
+      "epoch": 2.7579884756416972,
+      "loss_partial": 0.04931640625,
+      "step": 5265
+    },
+    {
+      "epoch": 2.7585123101100053,
+      "grad_norm": 1.827087163925171,
+      "learning_rate": 5.321348303348162e-06,
+      "loss": 0.0977,
+      "step": 5266
+    },
+    {
+      "epoch": 2.7585123101100053,
+      "loss_partial": 0.048828125,
+      "step": 5266
+    },
+    {
+      "epoch": 2.7585123101100053,
+      "loss_partial": 0.05078125,
+      "step": 5266
+    },
+    {
+      "epoch": 2.7590361445783134,
+      "grad_norm": 2.2737276554107666,
+      "learning_rate": 5.317308426953347e-06,
+      "loss": 0.1028,
+      "step": 5267
+    },
+    {
+      "epoch": 2.7590361445783134,
+      "loss_partial": 0.045654296875,
+      "step": 5267
+    },
+    {
+      "epoch": 2.7590361445783134,
+      "loss_partial": 0.044677734375,
+      "step": 5267
+    },
+    {
+      "epoch": 2.7595599790466214,
+      "grad_norm": 1.4603850841522217,
+      "learning_rate": 5.313269529212114e-06,
+      "loss": 0.1007,
+      "step": 5268
+    },
+    {
+      "epoch": 2.7595599790466214,
+      "loss_partial": 0.06298828125,
+      "step": 5268
+    },
+    {
+      "epoch": 2.7595599790466214,
+      "loss_partial": 0.046142578125,
+      "step": 5268
+    },
+    {
+      "epoch": 2.760083813514929,
+      "grad_norm": 1.698393702507019,
+      "learning_rate": 5.309231610968568e-06,
+      "loss": 0.104,
+      "step": 5269
+    },
+    {
+      "epoch": 2.760083813514929,
+      "loss_partial": 0.05029296875,
+      "step": 5269
+    },
+    {
+      "epoch": 2.760083813514929,
+      "loss_partial": 0.048583984375,
+      "step": 5269
+    },
+    {
+      "epoch": 2.760607647983237,
+      "grad_norm": 1.6044859886169434,
+      "learning_rate": 5.305194673066605e-06,
+      "loss": 0.0995,
+      "step": 5270
+    },
+    {
+      "epoch": 2.760607647983237,
+      "loss_partial": 0.04541015625,
+      "step": 5270
+    },
+    {
+      "epoch": 2.760607647983237,
+      "loss_partial": 0.049560546875,
+      "step": 5270
+    },
+    {
+      "epoch": 2.761131482451545,
+      "grad_norm": 1.7483792304992676,
+      "learning_rate": 5.301158716349925e-06,
+      "loss": 0.1027,
+      "step": 5271
+    },
+    {
+      "epoch": 2.761131482451545,
+      "loss_partial": 0.043212890625,
+      "step": 5271
+    },
+    {
+      "epoch": 2.761131482451545,
+      "loss_partial": 0.04638671875,
+      "step": 5271
+    },
+    {
+      "epoch": 2.761655316919853,
+      "grad_norm": 1.5551588535308838,
+      "learning_rate": 5.29712374166202e-06,
+      "loss": 0.0988,
+      "step": 5272
+    },
+    {
+      "epoch": 2.761655316919853,
+      "loss_partial": 0.052490234375,
+      "step": 5272
+    },
+    {
+      "epoch": 2.761655316919853,
+      "loss_partial": 0.054443359375,
+      "step": 5272
+    },
+    {
+      "epoch": 2.7621791513881613,
+      "grad_norm": 1.2641140222549438,
+      "learning_rate": 5.29308974984617e-06,
+      "loss": 0.0997,
+      "step": 5273
+    },
+    {
+      "epoch": 2.7621791513881613,
+      "loss_partial": 0.051513671875,
+      "step": 5273
+    },
+    {
+      "epoch": 2.7621791513881613,
+      "loss_partial": 0.046142578125,
+      "step": 5273
+    },
+    {
+      "epoch": 2.7627029858564693,
+      "grad_norm": 1.4172834157943726,
+      "learning_rate": 5.2890567417454555e-06,
+      "loss": 0.1008,
+      "step": 5274
+    },
+    {
+      "epoch": 2.7627029858564693,
+      "loss_partial": 0.05126953125,
+      "step": 5274
+    },
+    {
+      "epoch": 2.7627029858564693,
+      "loss_partial": 0.052490234375,
+      "step": 5274
+    },
+    {
+      "epoch": 2.7632268203247774,
+      "grad_norm": 1.0961134433746338,
+      "learning_rate": 5.2850247182027495e-06,
+      "loss": 0.1011,
+      "step": 5275
+    },
+    {
+      "epoch": 2.7632268203247774,
+      "loss_partial": 0.0546875,
+      "step": 5275
+    },
+    {
+      "epoch": 2.7632268203247774,
+      "loss_partial": 0.0478515625,
+      "step": 5275
+    },
+    {
+      "epoch": 2.7637506547930855,
+      "grad_norm": 1.6615808010101318,
+      "learning_rate": 5.280993680060722e-06,
+      "loss": 0.1015,
+      "step": 5276
+    },
+    {
+      "epoch": 2.7637506547930855,
+      "loss_partial": 0.049560546875,
+      "step": 5276
+    },
+    {
+      "epoch": 2.7637506547930855,
+      "loss_partial": 0.053955078125,
+      "step": 5276
+    },
+    {
+      "epoch": 2.7642744892613935,
+      "grad_norm": 1.4936459064483643,
+      "learning_rate": 5.276963628161833e-06,
+      "loss": 0.104,
+      "step": 5277
+    },
+    {
+      "epoch": 2.7642744892613935,
+      "loss_partial": 0.05322265625,
+      "step": 5277
+    },
+    {
+      "epoch": 2.7642744892613935,
+      "loss_partial": 0.047119140625,
+      "step": 5277
+    },
+    {
+      "epoch": 2.7647983237297016,
+      "grad_norm": 1.5794295072555542,
+      "learning_rate": 5.272934563348336e-06,
+      "loss": 0.0991,
+      "step": 5278
+    },
+    {
+      "epoch": 2.7647983237297016,
+      "loss_partial": 0.049072265625,
+      "step": 5278
+    },
+    {
+      "epoch": 2.7647983237297016,
+      "loss_partial": 0.05224609375,
+      "step": 5278
+    },
+    {
+      "epoch": 2.7653221581980096,
+      "grad_norm": 1.5694738626480103,
+      "learning_rate": 5.268906486462288e-06,
+      "loss": 0.1008,
+      "step": 5279
+    },
+    {
+      "epoch": 2.7653221581980096,
+      "loss_partial": 0.04736328125,
+      "step": 5279
+    },
+    {
+      "epoch": 2.7653221581980096,
+      "loss_partial": 0.048828125,
+      "step": 5279
+    },
+    {
+      "epoch": 2.7658459926663177,
+      "grad_norm": 1.528174877166748,
+      "learning_rate": 5.264879398345528e-06,
+      "loss": 0.1052,
+      "step": 5280
+    },
+    {
+      "epoch": 2.7658459926663177,
+      "loss_partial": 0.044189453125,
+      "step": 5280
+    },
+    {
+      "epoch": 2.7658459926663177,
+      "loss_partial": 0.05322265625,
+      "step": 5280
+    },
+    {
+      "epoch": 2.7663698271346253,
+      "grad_norm": 1.849381685256958,
+      "learning_rate": 5.2608532998396926e-06,
+      "loss": 0.1034,
+      "step": 5281
+    },
+    {
+      "epoch": 2.7663698271346253,
+      "loss_partial": 0.05419921875,
+      "step": 5281
+    },
+    {
+      "epoch": 2.7663698271346253,
+      "loss_partial": 0.0419921875,
+      "step": 5281
+    },
+    {
+      "epoch": 2.7668936616029334,
+      "grad_norm": 1.615626335144043,
+      "learning_rate": 5.256828191786207e-06,
+      "loss": 0.102,
+      "step": 5282
+    },
+    {
+      "epoch": 2.7668936616029334,
+      "loss_partial": 0.06396484375,
+      "step": 5282
+    },
+    {
+      "epoch": 2.7668936616029334,
+      "loss_partial": 0.044677734375,
+      "step": 5282
+    },
+    {
+      "epoch": 2.7674174960712414,
+      "grad_norm": 1.18269681930542,
+      "learning_rate": 5.252804075026303e-06,
+      "loss": 0.1007,
+      "step": 5283
+    },
+    {
+      "epoch": 2.7674174960712414,
+      "loss_partial": 0.043701171875,
+      "step": 5283
+    },
+    {
+      "epoch": 2.7674174960712414,
+      "loss_partial": 0.057373046875,
+      "step": 5283
+    },
+    {
+      "epoch": 2.7679413305395495,
+      "grad_norm": 1.7552708387374878,
+      "learning_rate": 5.2487809504009915e-06,
+      "loss": 0.0996,
+      "step": 5284
+    },
+    {
+      "epoch": 2.7679413305395495,
+      "loss_partial": 0.0458984375,
+      "step": 5284
+    },
+    {
+      "epoch": 2.7679413305395495,
+      "loss_partial": 0.04736328125,
+      "step": 5284
+    },
+    {
+      "epoch": 2.7684651650078576,
+      "grad_norm": 1.1907657384872437,
+      "learning_rate": 5.2447588187510815e-06,
+      "loss": 0.1015,
+      "step": 5285
+    },
+    {
+      "epoch": 2.7684651650078576,
+      "loss_partial": 0.0458984375,
+      "step": 5285
+    },
+    {
+      "epoch": 2.7684651650078576,
+      "loss_partial": 0.04638671875,
+      "step": 5285
+    },
+    {
+      "epoch": 2.7689889994761656,
+      "grad_norm": 1.483500361442566,
+      "learning_rate": 5.2407376809171715e-06,
+      "loss": 0.0981,
+      "step": 5286
+    },
+    {
+      "epoch": 2.7689889994761656,
+      "loss_partial": 0.044921875,
+      "step": 5286
+    },
+    {
+      "epoch": 2.7689889994761656,
+      "loss_partial": 0.052734375,
+      "step": 5286
+    },
+    {
+      "epoch": 2.7695128339444737,
+      "grad_norm": 1.2657173871994019,
+      "learning_rate": 5.236717537739656e-06,
+      "loss": 0.099,
+      "step": 5287
+    },
+    {
+      "epoch": 2.7695128339444737,
+      "loss_partial": 0.043701171875,
+      "step": 5287
+    },
+    {
+      "epoch": 2.7695128339444737,
+      "loss_partial": 0.052001953125,
+      "step": 5287
+    },
+    {
+      "epoch": 2.7700366684127813,
+      "grad_norm": 2.119218349456787,
+      "learning_rate": 5.232698390058722e-06,
+      "loss": 0.1031,
+      "step": 5288
+    },
+    {
+      "epoch": 2.7700366684127813,
+      "loss_partial": 0.04150390625,
+      "step": 5288
+    },
+    {
+      "epoch": 2.7700366684127813,
+      "loss_partial": 0.043212890625,
+      "step": 5288
+    },
+    {
+      "epoch": 2.7705605028810893,
+      "grad_norm": 1.5547899007797241,
+      "learning_rate": 5.228680238714341e-06,
+      "loss": 0.0975,
+      "step": 5289
+    },
+    {
+      "epoch": 2.7705605028810893,
+      "loss_partial": 0.0458984375,
+      "step": 5289
+    },
+    {
+      "epoch": 2.7705605028810893,
+      "loss_partial": 0.0654296875,
+      "step": 5289
+    },
+    {
+      "epoch": 2.7710843373493974,
+      "grad_norm": 1.7033331394195557,
+      "learning_rate": 5.224663084546287e-06,
+      "loss": 0.1012,
+      "step": 5290
+    },
+    {
+      "epoch": 2.7710843373493974,
+      "loss_partial": 0.052490234375,
+      "step": 5290
+    },
+    {
+      "epoch": 2.7710843373493974,
+      "loss_partial": 0.051025390625,
+      "step": 5290
+    },
+    {
+      "epoch": 2.7716081718177055,
+      "grad_norm": 1.1742374897003174,
+      "learning_rate": 5.220646928394114e-06,
+      "loss": 0.1003,
+      "step": 5291
+    },
+    {
+      "epoch": 2.7716081718177055,
+      "loss_partial": 0.052490234375,
+      "step": 5291
+    },
+    {
+      "epoch": 2.7716081718177055,
+      "loss_partial": 0.04638671875,
+      "step": 5291
+    },
+    {
+      "epoch": 2.7721320062860135,
+      "grad_norm": 1.4360926151275635,
+      "learning_rate": 5.216631771097183e-06,
+      "loss": 0.0958,
+      "step": 5292
+    },
+    {
+      "epoch": 2.7721320062860135,
+      "loss_partial": 0.042236328125,
+      "step": 5292
+    },
+    {
+      "epoch": 2.7721320062860135,
+      "loss_partial": 0.05615234375,
+      "step": 5292
+    },
+    {
+      "epoch": 2.7726558407543216,
+      "grad_norm": 2.0137553215026855,
+      "learning_rate": 5.212617613494632e-06,
+      "loss": 0.1004,
+      "step": 5293
+    },
+    {
+      "epoch": 2.7726558407543216,
+      "loss_partial": 0.054443359375,
+      "step": 5293
+    },
+    {
+      "epoch": 2.7726558407543216,
+      "loss_partial": 0.05029296875,
+      "step": 5293
+    },
+    {
+      "epoch": 2.7731796752226296,
+      "grad_norm": 1.5464189052581787,
+      "learning_rate": 5.208604456425395e-06,
+      "loss": 0.0997,
+      "step": 5294
+    },
+    {
+      "epoch": 2.7731796752226296,
+      "loss_partial": 0.0498046875,
+      "step": 5294
+    },
+    {
+      "epoch": 2.7731796752226296,
+      "loss_partial": 0.0478515625,
+      "step": 5294
+    },
+    {
+      "epoch": 2.7737035096909377,
+      "grad_norm": 1.5058790445327759,
+      "learning_rate": 5.204592300728195e-06,
+      "loss": 0.1053,
+      "step": 5295
+    },
+    {
+      "epoch": 2.7737035096909377,
+      "loss_partial": 0.048583984375,
+      "step": 5295
+    },
+    {
+      "epoch": 2.7737035096909377,
+      "loss_partial": 0.046875,
+      "step": 5295
+    },
+    {
+      "epoch": 2.7742273441592458,
+      "grad_norm": 1.4651217460632324,
+      "learning_rate": 5.200581147241553e-06,
+      "loss": 0.0997,
+      "step": 5296
+    },
+    {
+      "epoch": 2.7742273441592458,
+      "loss_partial": 0.0498046875,
+      "step": 5296
+    },
+    {
+      "epoch": 2.7742273441592458,
+      "loss_partial": 0.048828125,
+      "step": 5296
+    },
+    {
+      "epoch": 2.774751178627554,
+      "grad_norm": 1.5931185483932495,
+      "learning_rate": 5.196570996803772e-06,
+      "loss": 0.1015,
+      "step": 5297
+    },
+    {
+      "epoch": 2.774751178627554,
+      "loss_partial": 0.049560546875,
+      "step": 5297
+    },
+    {
+      "epoch": 2.774751178627554,
+      "loss_partial": 0.057373046875,
+      "step": 5297
+    },
+    {
+      "epoch": 2.775275013095862,
+      "grad_norm": 1.147142767906189,
+      "learning_rate": 5.192561850252951e-06,
+      "loss": 0.1028,
+      "step": 5298
+    },
+    {
+      "epoch": 2.775275013095862,
+      "loss_partial": 0.054931640625,
+      "step": 5298
+    },
+    {
+      "epoch": 2.775275013095862,
+      "loss_partial": 0.05078125,
+      "step": 5298
+    },
+    {
+      "epoch": 2.77579884756417,
+      "grad_norm": 1.2362492084503174,
+      "learning_rate": 5.188553708426973e-06,
+      "loss": 0.1008,
+      "step": 5299
+    },
+    {
+      "epoch": 2.77579884756417,
+      "loss_partial": 0.05810546875,
+      "step": 5299
+    },
+    {
+      "epoch": 2.77579884756417,
+      "loss_partial": 0.044921875,
+      "step": 5299
+    },
+    {
+      "epoch": 2.776322682032478,
+      "grad_norm": 1.5786256790161133,
+      "learning_rate": 5.184546572163518e-06,
+      "loss": 0.0992,
+      "step": 5300
+    },
+    {
+      "epoch": 2.776322682032478,
+      "loss_partial": 0.0576171875,
+      "step": 5300
+    },
+    {
+      "epoch": 2.776322682032478,
+      "loss_partial": 0.046142578125,
+      "step": 5300
+    },
+    {
+      "epoch": 2.7768465165007856,
+      "grad_norm": 1.5312973260879517,
+      "learning_rate": 5.18054044230005e-06,
+      "loss": 0.098,
+      "step": 5301
+    },
+    {
+      "epoch": 2.7768465165007856,
+      "loss_partial": 0.05029296875,
+      "step": 5301
+    },
+    {
+      "epoch": 2.7768465165007856,
+      "loss_partial": 0.051513671875,
+      "step": 5301
+    },
+    {
+      "epoch": 2.7773703509690937,
+      "grad_norm": 1.4681435823440552,
+      "learning_rate": 5.17653531967383e-06,
+      "loss": 0.1023,
+      "step": 5302
+    },
+    {
+      "epoch": 2.7773703509690937,
+      "loss_partial": 0.046875,
+      "step": 5302
+    },
+    {
+      "epoch": 2.7773703509690937,
+      "loss_partial": 0.055419921875,
+      "step": 5302
+    },
+    {
+      "epoch": 2.7778941854374017,
+      "grad_norm": 1.238735556602478,
+      "learning_rate": 5.1725312051218944e-06,
+      "loss": 0.1002,
+      "step": 5303
+    },
+    {
+      "epoch": 2.7778941854374017,
+      "loss_partial": 0.044677734375,
+      "step": 5303
+    },
+    {
+      "epoch": 2.7778941854374017,
+      "loss_partial": 0.056884765625,
+      "step": 5303
+    },
+    {
+      "epoch": 2.77841801990571,
+      "grad_norm": 1.2857427597045898,
+      "learning_rate": 5.168528099481089e-06,
+      "loss": 0.1002,
+      "step": 5304
+    },
+    {
+      "epoch": 2.77841801990571,
+      "loss_partial": 0.056884765625,
+      "step": 5304
+    },
+    {
+      "epoch": 2.77841801990571,
+      "loss_partial": 0.047607421875,
+      "step": 5304
+    },
+    {
+      "epoch": 2.778941854374018,
+      "grad_norm": 1.428964614868164,
+      "learning_rate": 5.1645260035880355e-06,
+      "loss": 0.1011,
+      "step": 5305
+    },
+    {
+      "epoch": 2.778941854374018,
+      "loss_partial": 0.04638671875,
+      "step": 5305
+    },
+    {
+      "epoch": 2.778941854374018,
+      "loss_partial": 0.0458984375,
+      "step": 5305
+    },
+    {
+      "epoch": 2.779465688842326,
+      "grad_norm": 1.4707444906234741,
+      "learning_rate": 5.160524918279144e-06,
+      "loss": 0.1003,
+      "step": 5306
+    },
+    {
+      "epoch": 2.779465688842326,
+      "loss_partial": 0.042236328125,
+      "step": 5306
+    },
+    {
+      "epoch": 2.779465688842326,
+      "loss_partial": 0.050048828125,
+      "step": 5306
+    },
+    {
+      "epoch": 2.779989523310634,
+      "grad_norm": 1.2272069454193115,
+      "learning_rate": 5.156524844390614e-06,
+      "loss": 0.0997,
+      "step": 5307
+    },
+    {
+      "epoch": 2.779989523310634,
+      "loss_partial": 0.042724609375,
+      "step": 5307
+    },
+    {
+      "epoch": 2.779989523310634,
+      "loss_partial": 0.045654296875,
+      "step": 5307
+    },
+    {
+      "epoch": 2.7805133577789416,
+      "grad_norm": 1.5166089534759521,
+      "learning_rate": 5.152525782758444e-06,
+      "loss": 0.1012,
+      "step": 5308
+    },
+    {
+      "epoch": 2.7805133577789416,
+      "loss_partial": 0.055908203125,
+      "step": 5308
+    },
+    {
+      "epoch": 2.7805133577789416,
+      "loss_partial": 0.049072265625,
+      "step": 5308
+    },
+    {
+      "epoch": 2.7810371922472497,
+      "grad_norm": 1.3063297271728516,
+      "learning_rate": 5.148527734218412e-06,
+      "loss": 0.101,
+      "step": 5309
+    },
+    {
+      "epoch": 2.7810371922472497,
+      "loss_partial": 0.04638671875,
+      "step": 5309
+    },
+    {
+      "epoch": 2.7810371922472497,
+      "loss_partial": 0.050048828125,
+      "step": 5309
+    },
+    {
+      "epoch": 2.7815610267155577,
+      "grad_norm": 1.4069355726242065,
+      "learning_rate": 5.144530699606081e-06,
+      "loss": 0.1001,
+      "step": 5310
+    },
+    {
+      "epoch": 2.7815610267155577,
+      "loss_partial": 0.046875,
+      "step": 5310
+    },
+    {
+      "epoch": 2.7815610267155577,
+      "loss_partial": 0.054931640625,
+      "step": 5310
+    },
+    {
+      "epoch": 2.7820848611838658,
+      "grad_norm": 1.5383219718933105,
+      "learning_rate": 5.140534679756808e-06,
+      "loss": 0.1028,
+      "step": 5311
+    },
+    {
+      "epoch": 2.7820848611838658,
+      "loss_partial": 0.04052734375,
+      "step": 5311
+    },
+    {
+      "epoch": 2.7820848611838658,
+      "loss_partial": 0.04443359375,
+      "step": 5311
+    },
+    {
+      "epoch": 2.782608695652174,
+      "grad_norm": 1.4236176013946533,
+      "learning_rate": 5.136539675505739e-06,
+      "loss": 0.0938,
+      "step": 5312
+    },
+    {
+      "epoch": 2.782608695652174,
+      "loss_partial": 0.053955078125,
+      "step": 5312
+    },
+    {
+      "epoch": 2.782608695652174,
+      "loss_partial": 0.04931640625,
+      "step": 5312
+    },
+    {
+      "epoch": 2.783132530120482,
+      "grad_norm": 1.6571015119552612,
+      "learning_rate": 5.132545687687801e-06,
+      "loss": 0.1012,
+      "step": 5313
+    },
+    {
+      "epoch": 2.783132530120482,
+      "loss_partial": 0.04443359375,
+      "step": 5313
+    },
+    {
+      "epoch": 2.783132530120482,
+      "loss_partial": 0.052001953125,
+      "step": 5313
+    },
+    {
+      "epoch": 2.78365636458879,
+      "grad_norm": 1.5427159070968628,
+      "learning_rate": 5.1285527171377155e-06,
+      "loss": 0.0993,
+      "step": 5314
+    },
+    {
+      "epoch": 2.78365636458879,
+      "loss_partial": 0.0546875,
+      "step": 5314
+    },
+    {
+      "epoch": 2.78365636458879,
+      "loss_partial": 0.055419921875,
+      "step": 5314
+    },
+    {
+      "epoch": 2.784180199057098,
+      "grad_norm": 1.5176913738250732,
+      "learning_rate": 5.124560764689984e-06,
+      "loss": 0.1023,
+      "step": 5315
+    },
+    {
+      "epoch": 2.784180199057098,
+      "loss_partial": 0.04736328125,
+      "step": 5315
+    },
+    {
+      "epoch": 2.784180199057098,
+      "loss_partial": 0.046875,
+      "step": 5315
+    },
+    {
+      "epoch": 2.784704033525406,
+      "grad_norm": 1.5103129148483276,
+      "learning_rate": 5.120569831178908e-06,
+      "loss": 0.1008,
+      "step": 5316
+    },
+    {
+      "epoch": 2.784704033525406,
+      "loss_partial": 0.056640625,
+      "step": 5316
+    },
+    {
+      "epoch": 2.784704033525406,
+      "loss_partial": 0.0576171875,
+      "step": 5316
+    },
+    {
+      "epoch": 2.785227867993714,
+      "grad_norm": 1.546228051185608,
+      "learning_rate": 5.116579917438564e-06,
+      "loss": 0.1021,
+      "step": 5317
+    },
+    {
+      "epoch": 2.785227867993714,
+      "loss_partial": 0.0439453125,
+      "step": 5317
+    },
+    {
+      "epoch": 2.785227867993714,
+      "loss_partial": 0.046875,
+      "step": 5317
+    },
+    {
+      "epoch": 2.785751702462022,
+      "grad_norm": 1.1891005039215088,
+      "learning_rate": 5.112591024302819e-06,
+      "loss": 0.1014,
+      "step": 5318
+    },
+    {
+      "epoch": 2.785751702462022,
+      "loss_partial": 0.044921875,
+      "step": 5318
+    },
+    {
+      "epoch": 2.785751702462022,
+      "loss_partial": 0.04931640625,
+      "step": 5318
+    },
+    {
+      "epoch": 2.7862755369303303,
+      "grad_norm": 2.0039169788360596,
+      "learning_rate": 5.108603152605323e-06,
+      "loss": 0.1024,
+      "step": 5319
+    },
+    {
+      "epoch": 2.7862755369303303,
+      "loss_partial": 0.0556640625,
+      "step": 5319
+    },
+    {
+      "epoch": 2.7862755369303303,
+      "loss_partial": 0.046630859375,
+      "step": 5319
+    },
+    {
+      "epoch": 2.786799371398638,
+      "grad_norm": 1.1175647974014282,
+      "learning_rate": 5.104616303179524e-06,
+      "loss": 0.1015,
+      "step": 5320
+    },
+    {
+      "epoch": 2.786799371398638,
+      "loss_partial": 0.044189453125,
+      "step": 5320
+    },
+    {
+      "epoch": 2.786799371398638,
+      "loss_partial": 0.04052734375,
+      "step": 5320
+    },
+    {
+      "epoch": 2.787323205866946,
+      "grad_norm": 1.7723972797393799,
+      "learning_rate": 5.100630476858646e-06,
+      "loss": 0.1006,
+      "step": 5321
+    },
+    {
+      "epoch": 2.787323205866946,
+      "loss_partial": 0.04833984375,
+      "step": 5321
+    },
+    {
+      "epoch": 2.787323205866946,
+      "loss_partial": 0.04541015625,
+      "step": 5321
+    },
+    {
+      "epoch": 2.787847040335254,
+      "grad_norm": 1.370697259902954,
+      "learning_rate": 5.096645674475702e-06,
+      "loss": 0.1028,
+      "step": 5322
+    },
+    {
+      "epoch": 2.787847040335254,
+      "loss_partial": 0.053955078125,
+      "step": 5322
+    },
+    {
+      "epoch": 2.787847040335254,
+      "loss_partial": 0.048828125,
+      "step": 5322
+    },
+    {
+      "epoch": 2.788370874803562,
+      "grad_norm": 1.9830586910247803,
+      "learning_rate": 5.092661896863491e-06,
+      "loss": 0.1035,
+      "step": 5323
+    },
+    {
+      "epoch": 2.788370874803562,
+      "loss_partial": 0.04052734375,
+      "step": 5323
+    },
+    {
+      "epoch": 2.788370874803562,
+      "loss_partial": 0.050048828125,
+      "step": 5323
+    },
+    {
+      "epoch": 2.78889470927187,
+      "grad_norm": 1.27532160282135,
+      "learning_rate": 5.088679144854596e-06,
+      "loss": 0.1023,
+      "step": 5324
+    },
+    {
+      "epoch": 2.78889470927187,
+      "loss_partial": 0.048583984375,
+      "step": 5324
+    },
+    {
+      "epoch": 2.78889470927187,
+      "loss_partial": 0.0517578125,
+      "step": 5324
+    },
+    {
+      "epoch": 2.789418543740178,
+      "grad_norm": 2.0780982971191406,
+      "learning_rate": 5.08469741928139e-06,
+      "loss": 0.0995,
+      "step": 5325
+    },
+    {
+      "epoch": 2.789418543740178,
+      "loss_partial": 0.051025390625,
+      "step": 5325
+    },
+    {
+      "epoch": 2.789418543740178,
+      "loss_partial": 0.054443359375,
+      "step": 5325
+    },
+    {
+      "epoch": 2.7899423782084862,
+      "grad_norm": 1.8176355361938477,
+      "learning_rate": 5.080716720976028e-06,
+      "loss": 0.1022,
+      "step": 5326
+    },
+    {
+      "epoch": 2.7899423782084862,
+      "loss_partial": 0.05419921875,
+      "step": 5326
+    },
+    {
+      "epoch": 2.7899423782084862,
+      "loss_partial": 0.0537109375,
+      "step": 5326
+    },
+    {
+      "epoch": 2.7904662126767943,
+      "grad_norm": 1.4168765544891357,
+      "learning_rate": 5.076737050770446e-06,
+      "loss": 0.1001,
+      "step": 5327
+    },
+    {
+      "epoch": 2.7904662126767943,
+      "loss_partial": 0.044677734375,
+      "step": 5327
+    },
+    {
+      "epoch": 2.7904662126767943,
+      "loss_partial": 0.046142578125,
+      "step": 5327
+    },
+    {
+      "epoch": 2.790990047145102,
+      "grad_norm": 1.1886051893234253,
+      "learning_rate": 5.072758409496381e-06,
+      "loss": 0.0985,
+      "step": 5328
+    },
+    {
+      "epoch": 2.790990047145102,
+      "loss_partial": 0.047607421875,
+      "step": 5328
+    },
+    {
+      "epoch": 2.790990047145102,
+      "loss_partial": 0.052978515625,
+      "step": 5328
+    },
+    {
+      "epoch": 2.79151388161341,
+      "grad_norm": 1.3162994384765625,
+      "learning_rate": 5.06878079798534e-06,
+      "loss": 0.0999,
+      "step": 5329
+    },
+    {
+      "epoch": 2.79151388161341,
+      "loss_partial": 0.04931640625,
+      "step": 5329
+    },
+    {
+      "epoch": 2.79151388161341,
+      "loss_partial": 0.04052734375,
+      "step": 5329
+    },
+    {
+      "epoch": 2.792037716081718,
+      "grad_norm": 1.2026506662368774,
+      "learning_rate": 5.064804217068617e-06,
+      "loss": 0.0996,
+      "step": 5330
+    },
+    {
+      "epoch": 2.792037716081718,
+      "loss_partial": 0.053955078125,
+      "step": 5330
+    },
+    {
+      "epoch": 2.792037716081718,
+      "loss_partial": 0.056396484375,
+      "step": 5330
+    },
+    {
+      "epoch": 2.792561550550026,
+      "grad_norm": 1.6319572925567627,
+      "learning_rate": 5.0608286675772926e-06,
+      "loss": 0.1017,
+      "step": 5331
+    },
+    {
+      "epoch": 2.792561550550026,
+      "loss_partial": 0.045654296875,
+      "step": 5331
+    },
+    {
+      "epoch": 2.792561550550026,
+      "loss_partial": 0.049072265625,
+      "step": 5331
+    },
+    {
+      "epoch": 2.793085385018334,
+      "grad_norm": 1.7556886672973633,
+      "learning_rate": 5.056854150342237e-06,
+      "loss": 0.0948,
+      "step": 5332
+    },
+    {
+      "epoch": 2.793085385018334,
+      "loss_partial": 0.0546875,
+      "step": 5332
+    },
+    {
+      "epoch": 2.793085385018334,
+      "loss_partial": 0.07373046875,
+      "step": 5332
+    },
+    {
+      "epoch": 2.793609219486642,
+      "grad_norm": 1.7603321075439453,
+      "learning_rate": 5.0528806661940945e-06,
+      "loss": 0.1061,
+      "step": 5333
+    },
+    {
+      "epoch": 2.793609219486642,
+      "loss_partial": 0.054443359375,
+      "step": 5333
+    },
+    {
+      "epoch": 2.793609219486642,
+      "loss_partial": 0.044921875,
+      "step": 5333
+    },
+    {
+      "epoch": 2.7941330539549503,
+      "grad_norm": 1.5084682703018188,
+      "learning_rate": 5.048908215963302e-06,
+      "loss": 0.1021,
+      "step": 5334
+    },
+    {
+      "epoch": 2.7941330539549503,
+      "loss_partial": 0.0546875,
+      "step": 5334
+    },
+    {
+      "epoch": 2.7941330539549503,
+      "loss_partial": 0.057373046875,
+      "step": 5334
+    },
+    {
+      "epoch": 2.7946568884232583,
+      "grad_norm": 2.0784223079681396,
+      "learning_rate": 5.044936800480076e-06,
+      "loss": 0.101,
+      "step": 5335
+    },
+    {
+      "epoch": 2.7946568884232583,
+      "loss_partial": 0.047119140625,
+      "step": 5335
+    },
+    {
+      "epoch": 2.7946568884232583,
+      "loss_partial": 0.059326171875,
+      "step": 5335
+    },
+    {
+      "epoch": 2.7951807228915664,
+      "grad_norm": 1.8948593139648438,
+      "learning_rate": 5.0409664205744165e-06,
+      "loss": 0.0983,
+      "step": 5336
+    },
+    {
+      "epoch": 2.7951807228915664,
+      "loss_partial": 0.05224609375,
+      "step": 5336
+    },
+    {
+      "epoch": 2.7951807228915664,
+      "loss_partial": 0.049072265625,
+      "step": 5336
+    },
+    {
+      "epoch": 2.7957045573598744,
+      "grad_norm": 1.5081195831298828,
+      "learning_rate": 5.03699707707611e-06,
+      "loss": 0.1036,
+      "step": 5337
+    },
+    {
+      "epoch": 2.7957045573598744,
+      "loss_partial": 0.0517578125,
+      "step": 5337
+    },
+    {
+      "epoch": 2.7957045573598744,
+      "loss_partial": 0.05126953125,
+      "step": 5337
+    },
+    {
+      "epoch": 2.7962283918281825,
+      "grad_norm": 2.1385462284088135,
+      "learning_rate": 5.033028770814725e-06,
+      "loss": 0.0981,
+      "step": 5338
+    },
+    {
+      "epoch": 2.7962283918281825,
+      "loss_partial": 0.049560546875,
+      "step": 5338
+    },
+    {
+      "epoch": 2.7962283918281825,
+      "loss_partial": 0.052490234375,
+      "step": 5338
+    },
+    {
+      "epoch": 2.7967522262964906,
+      "grad_norm": 2.114044666290283,
+      "learning_rate": 5.0290615026196065e-06,
+      "loss": 0.1024,
+      "step": 5339
+    },
+    {
+      "epoch": 2.7967522262964906,
+      "loss_partial": 0.0517578125,
+      "step": 5339
+    },
+    {
+      "epoch": 2.7967522262964906,
+      "loss_partial": 0.0537109375,
+      "step": 5339
+    },
+    {
+      "epoch": 2.797276060764798,
+      "grad_norm": 1.281339406967163,
+      "learning_rate": 5.0250952733199e-06,
+      "loss": 0.1024,
+      "step": 5340
+    },
+    {
+      "epoch": 2.797276060764798,
+      "loss_partial": 0.05029296875,
+      "step": 5340
+    },
+    {
+      "epoch": 2.797276060764798,
+      "loss_partial": 0.052734375,
+      "step": 5340
+    },
+    {
+      "epoch": 2.7977998952331062,
+      "grad_norm": 1.743058204650879,
+      "learning_rate": 5.021130083744517e-06,
+      "loss": 0.1021,
+      "step": 5341
+    },
+    {
+      "epoch": 2.7977998952331062,
+      "loss_partial": 0.046630859375,
+      "step": 5341
+    },
+    {
+      "epoch": 2.7977998952331062,
+      "loss_partial": 0.04736328125,
+      "step": 5341
+    },
+    {
+      "epoch": 2.7983237297014143,
+      "grad_norm": 1.6318920850753784,
+      "learning_rate": 5.017165934722157e-06,
+      "loss": 0.1009,
+      "step": 5342
+    },
+    {
+      "epoch": 2.7983237297014143,
+      "loss_partial": 0.04345703125,
+      "step": 5342
+    },
+    {
+      "epoch": 2.7983237297014143,
+      "loss_partial": 0.0546875,
+      "step": 5342
+    },
+    {
+      "epoch": 2.7988475641697224,
+      "grad_norm": 1.460496187210083,
+      "learning_rate": 5.013202827081301e-06,
+      "loss": 0.0999,
+      "step": 5343
+    },
+    {
+      "epoch": 2.7988475641697224,
+      "loss_partial": 0.052001953125,
+      "step": 5343
+    },
+    {
+      "epoch": 2.7988475641697224,
+      "loss_partial": 0.049560546875,
+      "step": 5343
+    },
+    {
+      "epoch": 2.7993713986380304,
+      "grad_norm": 1.6793696880340576,
+      "learning_rate": 5.00924076165022e-06,
+      "loss": 0.101,
+      "step": 5344
+    },
+    {
+      "epoch": 2.7993713986380304,
+      "loss_partial": 0.049072265625,
+      "step": 5344
+    },
+    {
+      "epoch": 2.7993713986380304,
+      "loss_partial": 0.053955078125,
+      "step": 5344
+    },
+    {
+      "epoch": 2.7998952331063385,
+      "grad_norm": 1.7009825706481934,
+      "learning_rate": 5.005279739256959e-06,
+      "loss": 0.1042,
+      "step": 5345
+    },
+    {
+      "epoch": 2.7998952331063385,
+      "loss_partial": 0.0537109375,
+      "step": 5345
+    },
+    {
+      "epoch": 2.7998952331063385,
+      "loss_partial": 0.05126953125,
+      "step": 5345
+    },
+    {
+      "epoch": 2.8004190675746465,
+      "grad_norm": 1.7055604457855225,
+      "learning_rate": 5.001319760729345e-06,
+      "loss": 0.1032,
+      "step": 5346
+    },
+    {
+      "epoch": 2.8004190675746465,
+      "loss_partial": 0.048095703125,
+      "step": 5346
+    },
+    {
+      "epoch": 2.8004190675746465,
+      "loss_partial": 0.053955078125,
+      "step": 5346
+    },
+    {
+      "epoch": 2.800942902042954,
+      "grad_norm": 1.7448620796203613,
+      "learning_rate": 4.99736082689499e-06,
+      "loss": 0.0988,
+      "step": 5347
+    },
+    {
+      "epoch": 2.800942902042954,
+      "loss_partial": 0.052978515625,
+      "step": 5347
+    },
+    {
+      "epoch": 2.800942902042954,
+      "loss_partial": 0.050537109375,
+      "step": 5347
+    },
+    {
+      "epoch": 2.801466736511262,
+      "grad_norm": 1.664597988128662,
+      "learning_rate": 4.993402938581288e-06,
+      "loss": 0.0986,
+      "step": 5348
+    },
+    {
+      "epoch": 2.801466736511262,
+      "loss_partial": 0.054931640625,
+      "step": 5348
+    },
+    {
+      "epoch": 2.801466736511262,
+      "loss_partial": 0.051025390625,
+      "step": 5348
+    },
+    {
+      "epoch": 2.8019905709795703,
+      "grad_norm": 1.9507865905761719,
+      "learning_rate": 4.9894460966154105e-06,
+      "loss": 0.102,
+      "step": 5349
+    },
+    {
+      "epoch": 2.8019905709795703,
+      "loss_partial": 0.0615234375,
+      "step": 5349
+    },
+    {
+      "epoch": 2.8019905709795703,
+      "loss_partial": 0.052734375,
+      "step": 5349
+    },
+    {
+      "epoch": 2.8025144054478783,
+      "grad_norm": 1.4546144008636475,
+      "learning_rate": 4.985490301824315e-06,
+      "loss": 0.1,
+      "step": 5350
+    },
+    {
+      "epoch": 2.8025144054478783,
+      "loss_partial": 0.048583984375,
+      "step": 5350
+    },
+    {
+      "epoch": 2.8025144054478783,
+      "loss_partial": 0.04736328125,
+      "step": 5350
+    },
+    {
+      "epoch": 2.8030382399161864,
+      "grad_norm": 1.2363289594650269,
+      "learning_rate": 4.981535555034738e-06,
+      "loss": 0.0999,
+      "step": 5351
+    },
+    {
+      "epoch": 2.8030382399161864,
+      "loss_partial": 0.044921875,
+      "step": 5351
+    },
+    {
+      "epoch": 2.8030382399161864,
+      "loss_partial": 0.0419921875,
+      "step": 5351
+    },
+    {
+      "epoch": 2.8035620743844945,
+      "grad_norm": 1.5558793544769287,
+      "learning_rate": 4.977581857073192e-06,
+      "loss": 0.1004,
+      "step": 5352
+    },
+    {
+      "epoch": 2.8035620743844945,
+      "loss_partial": 0.049072265625,
+      "step": 5352
+    },
+    {
+      "epoch": 2.8035620743844945,
+      "loss_partial": 0.04931640625,
+      "step": 5352
+    },
+    {
+      "epoch": 2.8040859088528025,
+      "grad_norm": 1.4557323455810547,
+      "learning_rate": 4.973629208765986e-06,
+      "loss": 0.0998,
+      "step": 5353
+    },
+    {
+      "epoch": 2.8040859088528025,
+      "loss_partial": 0.050537109375,
+      "step": 5353
+    },
+    {
+      "epoch": 2.8040859088528025,
+      "loss_partial": 0.0419921875,
+      "step": 5353
+    },
+    {
+      "epoch": 2.8046097433211106,
+      "grad_norm": 1.690584421157837,
+      "learning_rate": 4.969677610939191e-06,
+      "loss": 0.1017,
+      "step": 5354
+    },
+    {
+      "epoch": 2.8046097433211106,
+      "loss_partial": 0.046142578125,
+      "step": 5354
+    },
+    {
+      "epoch": 2.8046097433211106,
+      "loss_partial": 0.047607421875,
+      "step": 5354
+    },
+    {
+      "epoch": 2.8051335777894186,
+      "grad_norm": 1.4365246295928955,
+      "learning_rate": 4.965727064418669e-06,
+      "loss": 0.1015,
+      "step": 5355
+    },
+    {
+      "epoch": 2.8051335777894186,
+      "loss_partial": 0.054931640625,
+      "step": 5355
+    },
+    {
+      "epoch": 2.8051335777894186,
+      "loss_partial": 0.048828125,
+      "step": 5355
+    },
+    {
+      "epoch": 2.8056574122577267,
+      "grad_norm": 1.077073335647583,
+      "learning_rate": 4.961777570030057e-06,
+      "loss": 0.0981,
+      "step": 5356
+    },
+    {
+      "epoch": 2.8056574122577267,
+      "loss_partial": 0.04638671875,
+      "step": 5356
+    },
+    {
+      "epoch": 2.8056574122577267,
+      "loss_partial": 0.044677734375,
+      "step": 5356
+    },
+    {
+      "epoch": 2.8061812467260348,
+      "grad_norm": 2.2596216201782227,
+      "learning_rate": 4.957829128598781e-06,
+      "loss": 0.0989,
+      "step": 5357
+    },
+    {
+      "epoch": 2.8061812467260348,
+      "loss_partial": 0.0400390625,
+      "step": 5357
+    },
+    {
+      "epoch": 2.8061812467260348,
+      "loss_partial": 0.050537109375,
+      "step": 5357
+    },
+    {
+      "epoch": 2.806705081194343,
+      "grad_norm": 1.6047881841659546,
+      "learning_rate": 4.953881740950036e-06,
+      "loss": 0.1066,
+      "step": 5358
+    },
+    {
+      "epoch": 2.806705081194343,
+      "loss_partial": 0.055908203125,
+      "step": 5358
+    },
+    {
+      "epoch": 2.806705081194343,
+      "loss_partial": 0.045166015625,
+      "step": 5358
+    },
+    {
+      "epoch": 2.807228915662651,
+      "grad_norm": 2.1709163188934326,
+      "learning_rate": 4.9499354079088046e-06,
+      "loss": 0.1026,
+      "step": 5359
+    },
+    {
+      "epoch": 2.807228915662651,
+      "loss_partial": 0.0458984375,
+      "step": 5359
+    },
+    {
+      "epoch": 2.807228915662651,
+      "loss_partial": 0.050537109375,
+      "step": 5359
+    },
+    {
+      "epoch": 2.8077527501309585,
+      "grad_norm": 1.5570186376571655,
+      "learning_rate": 4.945990130299842e-06,
+      "loss": 0.1023,
+      "step": 5360
+    },
+    {
+      "epoch": 2.8077527501309585,
+      "loss_partial": 0.046142578125,
+      "step": 5360
+    },
+    {
+      "epoch": 2.8077527501309585,
+      "loss_partial": 0.04931640625,
+      "step": 5360
+    },
+    {
+      "epoch": 2.8082765845992665,
+      "grad_norm": 2.86972713470459,
+      "learning_rate": 4.942045908947691e-06,
+      "loss": 0.0977,
+      "step": 5361
+    },
+    {
+      "epoch": 2.8082765845992665,
+      "loss_partial": 0.039306640625,
+      "step": 5361
+    },
+    {
+      "epoch": 2.8082765845992665,
+      "loss_partial": 0.0419921875,
+      "step": 5361
+    },
+    {
+      "epoch": 2.8088004190675746,
+      "grad_norm": 1.3593136072158813,
+      "learning_rate": 4.938102744676668e-06,
+      "loss": 0.1017,
+      "step": 5362
+    },
+    {
+      "epoch": 2.8088004190675746,
+      "loss_partial": 0.050048828125,
+      "step": 5362
+    },
+    {
+      "epoch": 2.8088004190675746,
+      "loss_partial": 0.051025390625,
+      "step": 5362
+    },
+    {
+      "epoch": 2.8093242535358827,
+      "grad_norm": 1.9236317873001099,
+      "learning_rate": 4.9341606383108686e-06,
+      "loss": 0.0992,
+      "step": 5363
+    },
+    {
+      "epoch": 2.8093242535358827,
+      "loss_partial": 0.04931640625,
+      "step": 5363
+    },
+    {
+      "epoch": 2.8093242535358827,
+      "loss_partial": 0.0576171875,
+      "step": 5363
+    },
+    {
+      "epoch": 2.8098480880041907,
+      "grad_norm": 1.4510643482208252,
+      "learning_rate": 4.930219590674165e-06,
+      "loss": 0.1002,
+      "step": 5364
+    },
+    {
+      "epoch": 2.8098480880041907,
+      "loss_partial": 0.041748046875,
+      "step": 5364
+    },
+    {
+      "epoch": 2.8098480880041907,
+      "loss_partial": 0.04541015625,
+      "step": 5364
+    },
+    {
+      "epoch": 2.810371922472499,
+      "grad_norm": 1.610817313194275,
+      "learning_rate": 4.926279602590222e-06,
+      "loss": 0.1007,
+      "step": 5365
+    },
+    {
+      "epoch": 2.810371922472499,
+      "loss_partial": 0.052978515625,
+      "step": 5365
+    },
+    {
+      "epoch": 2.810371922472499,
+      "loss_partial": 0.0498046875,
+      "step": 5365
+    },
+    {
+      "epoch": 2.810895756940807,
+      "grad_norm": 1.179850459098816,
+      "learning_rate": 4.922340674882465e-06,
+      "loss": 0.1008,
+      "step": 5366
+    },
+    {
+      "epoch": 2.810895756940807,
+      "loss_partial": 0.0439453125,
+      "step": 5366
+    },
+    {
+      "epoch": 2.810895756940807,
+      "loss_partial": 0.05615234375,
+      "step": 5366
+    },
+    {
+      "epoch": 2.8114195914091145,
+      "grad_norm": 1.2442047595977783,
+      "learning_rate": 4.91840280837411e-06,
+      "loss": 0.101,
+      "step": 5367
+    },
+    {
+      "epoch": 2.8114195914091145,
+      "loss_partial": 0.0419921875,
+      "step": 5367
+    },
+    {
+      "epoch": 2.8114195914091145,
+      "loss_partial": 0.047119140625,
+      "step": 5367
+    },
+    {
+      "epoch": 2.8119434258774225,
+      "grad_norm": 1.8202027082443237,
+      "learning_rate": 4.914466003888138e-06,
+      "loss": 0.098,
+      "step": 5368
+    },
+    {
+      "epoch": 2.8119434258774225,
+      "loss_partial": 0.04541015625,
+      "step": 5368
+    },
+    {
+      "epoch": 2.8119434258774225,
+      "loss_partial": 0.0517578125,
+      "step": 5368
+    },
+    {
+      "epoch": 2.8124672603457306,
+      "grad_norm": 1.2907758951187134,
+      "learning_rate": 4.910530262247329e-06,
+      "loss": 0.1039,
+      "step": 5369
+    },
+    {
+      "epoch": 2.8124672603457306,
+      "loss_partial": 0.04931640625,
+      "step": 5369
+    },
+    {
+      "epoch": 2.8124672603457306,
+      "loss_partial": 0.050048828125,
+      "step": 5369
+    },
+    {
+      "epoch": 2.8129910948140386,
+      "grad_norm": 1.6268577575683594,
+      "learning_rate": 4.906595584274223e-06,
+      "loss": 0.0975,
+      "step": 5370
+    },
+    {
+      "epoch": 2.8129910948140386,
+      "loss_partial": 0.041259765625,
+      "step": 5370
+    },
+    {
+      "epoch": 2.8129910948140386,
+      "loss_partial": 0.04248046875,
+      "step": 5370
+    },
+    {
+      "epoch": 2.8135149292823467,
+      "grad_norm": 1.6821156740188599,
+      "learning_rate": 4.902661970791142e-06,
+      "loss": 0.1007,
+      "step": 5371
+    },
+    {
+      "epoch": 2.8135149292823467,
+      "loss_partial": 0.05908203125,
+      "step": 5371
+    },
+    {
+      "epoch": 2.8135149292823467,
+      "loss_partial": 0.04638671875,
+      "step": 5371
+    },
+    {
+      "epoch": 2.8140387637506548,
+      "grad_norm": 1.8829576969146729,
+      "learning_rate": 4.898729422620189e-06,
+      "loss": 0.1,
+      "step": 5372
+    },
+    {
+      "epoch": 2.8140387637506548,
+      "loss_partial": 0.05224609375,
+      "step": 5372
+    },
+    {
+      "epoch": 2.8140387637506548,
+      "loss_partial": 0.051025390625,
+      "step": 5372
+    },
+    {
+      "epoch": 2.814562598218963,
+      "grad_norm": 2.185447931289673,
+      "learning_rate": 4.8947979405832405e-06,
+      "loss": 0.1067,
+      "step": 5373
+    },
+    {
+      "epoch": 2.814562598218963,
+      "loss_partial": 0.050537109375,
+      "step": 5373
+    },
+    {
+      "epoch": 2.814562598218963,
+      "loss_partial": 0.0419921875,
+      "step": 5373
+    },
+    {
+      "epoch": 2.815086432687271,
+      "grad_norm": 1.492300271987915,
+      "learning_rate": 4.890867525501953e-06,
+      "loss": 0.1003,
+      "step": 5374
+    },
+    {
+      "epoch": 2.815086432687271,
+      "loss_partial": 0.059814453125,
+      "step": 5374
+    },
+    {
+      "epoch": 2.815086432687271,
+      "loss_partial": 0.05126953125,
+      "step": 5374
+    },
+    {
+      "epoch": 2.815610267155579,
+      "grad_norm": 1.9531466960906982,
+      "learning_rate": 4.8869381781977584e-06,
+      "loss": 0.1001,
+      "step": 5375
+    },
+    {
+      "epoch": 2.815610267155579,
+      "loss_partial": 0.041748046875,
+      "step": 5375
+    },
+    {
+      "epoch": 2.815610267155579,
+      "loss_partial": 0.057861328125,
+      "step": 5375
+    },
+    {
+      "epoch": 2.816134101623887,
+      "grad_norm": 1.5954135656356812,
+      "learning_rate": 4.883009899491864e-06,
+      "loss": 0.1011,
+      "step": 5376
+    },
+    {
+      "epoch": 2.816134101623887,
+      "loss_partial": 0.04541015625,
+      "step": 5376
+    },
+    {
+      "epoch": 2.816134101623887,
+      "loss_partial": 0.0498046875,
+      "step": 5376
+    },
+    {
+      "epoch": 2.816657936092195,
+      "grad_norm": 1.5021684169769287,
+      "learning_rate": 4.879082690205262e-06,
+      "loss": 0.1032,
+      "step": 5377
+    },
+    {
+      "epoch": 2.816657936092195,
+      "loss_partial": 0.051025390625,
+      "step": 5377
+    },
+    {
+      "epoch": 2.816657936092195,
+      "loss_partial": 0.057373046875,
+      "step": 5377
+    },
+    {
+      "epoch": 2.817181770560503,
+      "grad_norm": 1.4806231260299683,
+      "learning_rate": 4.8751565511587085e-06,
+      "loss": 0.102,
+      "step": 5378
+    },
+    {
+      "epoch": 2.817181770560503,
+      "loss_partial": 0.045654296875,
+      "step": 5378
+    },
+    {
+      "epoch": 2.817181770560503,
+      "loss_partial": 0.0458984375,
+      "step": 5378
+    },
+    {
+      "epoch": 2.8177056050288107,
+      "grad_norm": 1.2912237644195557,
+      "learning_rate": 4.871231483172747e-06,
+      "loss": 0.0999,
+      "step": 5379
+    },
+    {
+      "epoch": 2.8177056050288107,
+      "loss_partial": 0.057373046875,
+      "step": 5379
+    },
+    {
+      "epoch": 2.8177056050288107,
+      "loss_partial": 0.06298828125,
+      "step": 5379
+    },
+    {
+      "epoch": 2.818229439497119,
+      "grad_norm": 1.8221521377563477,
+      "learning_rate": 4.867307487067685e-06,
+      "loss": 0.1023,
+      "step": 5380
+    },
+    {
+      "epoch": 2.818229439497119,
+      "loss_partial": 0.0595703125,
+      "step": 5380
+    },
+    {
+      "epoch": 2.818229439497119,
+      "loss_partial": 0.042236328125,
+      "step": 5380
+    },
+    {
+      "epoch": 2.818753273965427,
+      "grad_norm": 1.6540957689285278,
+      "learning_rate": 4.863384563663623e-06,
+      "loss": 0.1064,
+      "step": 5381
+    },
+    {
+      "epoch": 2.818753273965427,
+      "loss_partial": 0.044189453125,
+      "step": 5381
+    },
+    {
+      "epoch": 2.818753273965427,
+      "loss_partial": 0.055419921875,
+      "step": 5381
+    },
+    {
+      "epoch": 2.819277108433735,
+      "grad_norm": 1.4961998462677002,
+      "learning_rate": 4.859462713780423e-06,
+      "loss": 0.1028,
+      "step": 5382
+    },
+    {
+      "epoch": 2.819277108433735,
+      "loss_partial": 0.04931640625,
+      "step": 5382
+    },
+    {
+      "epoch": 2.819277108433735,
+      "loss_partial": 0.039306640625,
+      "step": 5382
+    },
+    {
+      "epoch": 2.819800942902043,
+      "grad_norm": 1.9721201658248901,
+      "learning_rate": 4.855541938237728e-06,
+      "loss": 0.1051,
+      "step": 5383
+    },
+    {
+      "epoch": 2.819800942902043,
+      "loss_partial": 0.049560546875,
+      "step": 5383
+    },
+    {
+      "epoch": 2.819800942902043,
+      "loss_partial": 0.056396484375,
+      "step": 5383
+    },
+    {
+      "epoch": 2.820324777370351,
+      "grad_norm": 1.248085618019104,
+      "learning_rate": 4.851622237854951e-06,
+      "loss": 0.1013,
+      "step": 5384
+    },
+    {
+      "epoch": 2.820324777370351,
+      "loss_partial": 0.045166015625,
+      "step": 5384
+    },
+    {
+      "epoch": 2.820324777370351,
+      "loss_partial": 0.0498046875,
+      "step": 5384
+    },
+    {
+      "epoch": 2.820848611838659,
+      "grad_norm": 1.545913815498352,
+      "learning_rate": 4.847703613451299e-06,
+      "loss": 0.1001,
+      "step": 5385
+    },
+    {
+      "epoch": 2.820848611838659,
+      "loss_partial": 0.054931640625,
+      "step": 5385
+    },
+    {
+      "epoch": 2.820848611838659,
+      "loss_partial": 0.04736328125,
+      "step": 5385
+    },
+    {
+      "epoch": 2.821372446306967,
+      "grad_norm": 1.430363416671753,
+      "learning_rate": 4.843786065845726e-06,
+      "loss": 0.0992,
+      "step": 5386
+    },
+    {
+      "epoch": 2.821372446306967,
+      "loss_partial": 0.042724609375,
+      "step": 5386
+    },
+    {
+      "epoch": 2.821372446306967,
+      "loss_partial": 0.054931640625,
+      "step": 5386
+    },
+    {
+      "epoch": 2.8218962807752748,
+      "grad_norm": 1.2672821283340454,
+      "learning_rate": 4.83986959585698e-06,
+      "loss": 0.1009,
+      "step": 5387
+    },
+    {
+      "epoch": 2.8218962807752748,
+      "loss_partial": 0.041015625,
+      "step": 5387
+    },
+    {
+      "epoch": 2.8218962807752748,
+      "loss_partial": 0.042724609375,
+      "step": 5387
+    },
+    {
+      "epoch": 2.822420115243583,
+      "grad_norm": 1.172764539718628,
+      "learning_rate": 4.835954204303576e-06,
+      "loss": 0.0964,
+      "step": 5388
+    },
+    {
+      "epoch": 2.822420115243583,
+      "loss_partial": 0.046630859375,
+      "step": 5388
+    },
+    {
+      "epoch": 2.822420115243583,
+      "loss_partial": 0.050537109375,
+      "step": 5388
+    },
+    {
+      "epoch": 2.822943949711891,
+      "grad_norm": 1.5640385150909424,
+      "learning_rate": 4.832039892003813e-06,
+      "loss": 0.1036,
+      "step": 5389
+    },
+    {
+      "epoch": 2.822943949711891,
+      "loss_partial": 0.05419921875,
+      "step": 5389
+    },
+    {
+      "epoch": 2.822943949711891,
+      "loss_partial": 0.04638671875,
+      "step": 5389
+    },
+    {
+      "epoch": 2.823467784180199,
+      "grad_norm": 1.338155746459961,
+      "learning_rate": 4.828126659775755e-06,
+      "loss": 0.1008,
+      "step": 5390
+    },
+    {
+      "epoch": 2.823467784180199,
+      "loss_partial": 0.048095703125,
+      "step": 5390
+    },
+    {
+      "epoch": 2.823467784180199,
+      "loss_partial": 0.05078125,
+      "step": 5390
+    },
+    {
+      "epoch": 2.823991618648507,
+      "grad_norm": 1.1829806566238403,
+      "learning_rate": 4.8242145084372435e-06,
+      "loss": 0.0997,
+      "step": 5391
+    },
+    {
+      "epoch": 2.823991618648507,
+      "loss_partial": 0.059326171875,
+      "step": 5391
+    },
+    {
+      "epoch": 2.823991618648507,
+      "loss_partial": 0.043701171875,
+      "step": 5391
+    },
+    {
+      "epoch": 2.824515453116815,
+      "grad_norm": 1.4092864990234375,
+      "learning_rate": 4.82030343880589e-06,
+      "loss": 0.1016,
+      "step": 5392
+    },
+    {
+      "epoch": 2.824515453116815,
+      "loss_partial": 0.064453125,
+      "step": 5392
+    },
+    {
+      "epoch": 2.824515453116815,
+      "loss_partial": 0.043701171875,
+      "step": 5392
+    },
+    {
+      "epoch": 2.825039287585123,
+      "grad_norm": 1.6069544553756714,
+      "learning_rate": 4.81639345169909e-06,
+      "loss": 0.0993,
+      "step": 5393
+    },
+    {
+      "epoch": 2.825039287585123,
+      "loss_partial": 0.046630859375,
+      "step": 5393
+    },
+    {
+      "epoch": 2.825039287585123,
+      "loss_partial": 0.052734375,
+      "step": 5393
+    },
+    {
+      "epoch": 2.825563122053431,
+      "grad_norm": 1.5624653100967407,
+      "learning_rate": 4.8124845479340055e-06,
+      "loss": 0.0996,
+      "step": 5394
+    },
+    {
+      "epoch": 2.825563122053431,
+      "loss_partial": 0.050048828125,
+      "step": 5394
+    },
+    {
+      "epoch": 2.825563122053431,
+      "loss_partial": 0.04345703125,
+      "step": 5394
+    },
+    {
+      "epoch": 2.8260869565217392,
+      "grad_norm": 1.3825781345367432,
+      "learning_rate": 4.808576728327572e-06,
+      "loss": 0.0997,
+      "step": 5395
+    },
+    {
+      "epoch": 2.8260869565217392,
+      "loss_partial": 0.048828125,
+      "step": 5395
+    },
+    {
+      "epoch": 2.8260869565217392,
+      "loss_partial": 0.05078125,
+      "step": 5395
+    },
+    {
+      "epoch": 2.8266107909900473,
+      "grad_norm": 1.2663674354553223,
+      "learning_rate": 4.804669993696495e-06,
+      "loss": 0.1051,
+      "step": 5396
+    },
+    {
+      "epoch": 2.8266107909900473,
+      "loss_partial": 0.054443359375,
+      "step": 5396
+    },
+    {
+      "epoch": 2.8266107909900473,
+      "loss_partial": 0.048828125,
+      "step": 5396
+    },
+    {
+      "epoch": 2.8271346254583554,
+      "grad_norm": 1.2236310243606567,
+      "learning_rate": 4.80076434485727e-06,
+      "loss": 0.1018,
+      "step": 5397
+    },
+    {
+      "epoch": 2.8271346254583554,
+      "loss_partial": 0.053955078125,
+      "step": 5397
+    },
+    {
+      "epoch": 2.8271346254583554,
+      "loss_partial": 0.052734375,
+      "step": 5397
+    },
+    {
+      "epoch": 2.8276584599266634,
+      "grad_norm": 1.4147934913635254,
+      "learning_rate": 4.796859782626142e-06,
+      "loss": 0.1056,
+      "step": 5398
+    },
+    {
+      "epoch": 2.8276584599266634,
+      "loss_partial": 0.048828125,
+      "step": 5398
+    },
+    {
+      "epoch": 2.8276584599266634,
+      "loss_partial": 0.046630859375,
+      "step": 5398
+    },
+    {
+      "epoch": 2.828182294394971,
+      "grad_norm": 0.9987791776657104,
+      "learning_rate": 4.792956307819144e-06,
+      "loss": 0.0995,
+      "step": 5399
+    },
+    {
+      "epoch": 2.828182294394971,
+      "loss_partial": 0.050048828125,
+      "step": 5399
+    },
+    {
+      "epoch": 2.828182294394971,
+      "loss_partial": 0.054443359375,
+      "step": 5399
+    },
+    {
+      "epoch": 2.828706128863279,
+      "grad_norm": 1.2611472606658936,
+      "learning_rate": 4.789053921252076e-06,
+      "loss": 0.1024,
+      "step": 5400
+    },
+    {
+      "epoch": 2.828706128863279,
+      "loss_partial": 0.055419921875,
+      "step": 5400
+    },
+    {
+      "epoch": 2.828706128863279,
+      "loss_partial": 0.044921875,
+      "step": 5400
+    },
+    {
+      "epoch": 2.829229963331587,
+      "grad_norm": 1.7867472171783447,
+      "learning_rate": 4.785152623740517e-06,
+      "loss": 0.1032,
+      "step": 5401
+    },
+    {
+      "epoch": 2.829229963331587,
+      "loss_partial": 0.0654296875,
+      "step": 5401
+    },
+    {
+      "epoch": 2.829229963331587,
+      "loss_partial": 0.043701171875,
+      "step": 5401
+    },
+    {
+      "epoch": 2.829753797799895,
+      "grad_norm": 1.0379019975662231,
+      "learning_rate": 4.781252416099812e-06,
+      "loss": 0.0982,
+      "step": 5402
+    },
+    {
+      "epoch": 2.829753797799895,
+      "loss_partial": 0.0478515625,
+      "step": 5402
+    },
+    {
+      "epoch": 2.829753797799895,
+      "loss_partial": 0.0439453125,
+      "step": 5402
+    },
+    {
+      "epoch": 2.8302776322682033,
+      "grad_norm": 1.6040197610855103,
+      "learning_rate": 4.777353299145081e-06,
+      "loss": 0.0985,
+      "step": 5403
+    },
+    {
+      "epoch": 2.8302776322682033,
+      "loss_partial": 0.04296875,
+      "step": 5403
+    },
+    {
+      "epoch": 2.8302776322682033,
+      "loss_partial": 0.05126953125,
+      "step": 5403
+    },
+    {
+      "epoch": 2.8308014667365113,
+      "grad_norm": 1.115466594696045,
+      "learning_rate": 4.7734552736912144e-06,
+      "loss": 0.1035,
+      "step": 5404
+    },
+    {
+      "epoch": 2.8308014667365113,
+      "loss_partial": 0.045654296875,
+      "step": 5404
+    },
+    {
+      "epoch": 2.8308014667365113,
+      "loss_partial": 0.05029296875,
+      "step": 5404
+    },
+    {
+      "epoch": 2.8313253012048194,
+      "grad_norm": 1.2231404781341553,
+      "learning_rate": 4.769558340552872e-06,
+      "loss": 0.0969,
+      "step": 5405
+    },
+    {
+      "epoch": 2.8313253012048194,
+      "loss_partial": 0.04931640625,
+      "step": 5405
+    },
+    {
+      "epoch": 2.8313253012048194,
+      "loss_partial": 0.05419921875,
+      "step": 5405
+    },
+    {
+      "epoch": 2.831849135673127,
+      "grad_norm": 1.3342407941818237,
+      "learning_rate": 4.7656625005444955e-06,
+      "loss": 0.1053,
+      "step": 5406
+    },
+    {
+      "epoch": 2.831849135673127,
+      "loss_partial": 0.05419921875,
+      "step": 5406
+    },
+    {
+      "epoch": 2.831849135673127,
+      "loss_partial": 0.05322265625,
+      "step": 5406
+    },
+    {
+      "epoch": 2.832372970141435,
+      "grad_norm": 1.4874136447906494,
+      "learning_rate": 4.7617677544802884e-06,
+      "loss": 0.1032,
+      "step": 5407
+    },
+    {
+      "epoch": 2.832372970141435,
+      "loss_partial": 0.055908203125,
+      "step": 5407
+    },
+    {
+      "epoch": 2.832372970141435,
+      "loss_partial": 0.056396484375,
+      "step": 5407
+    },
+    {
+      "epoch": 2.832896804609743,
+      "grad_norm": 1.168366551399231,
+      "learning_rate": 4.757874103174229e-06,
+      "loss": 0.1009,
+      "step": 5408
+    },
+    {
+      "epoch": 2.832896804609743,
+      "loss_partial": 0.053466796875,
+      "step": 5408
+    },
+    {
+      "epoch": 2.832896804609743,
+      "loss_partial": 0.046875,
+      "step": 5408
+    },
+    {
+      "epoch": 2.833420639078051,
+      "grad_norm": 1.5906517505645752,
+      "learning_rate": 4.753981547440062e-06,
+      "loss": 0.1031,
+      "step": 5409
+    },
+    {
+      "epoch": 2.833420639078051,
+      "loss_partial": 0.050048828125,
+      "step": 5409
+    },
+    {
+      "epoch": 2.833420639078051,
+      "loss_partial": 0.048583984375,
+      "step": 5409
+    },
+    {
+      "epoch": 2.8339444735463593,
+      "grad_norm": 1.500758171081543,
+      "learning_rate": 4.750090088091318e-06,
+      "loss": 0.0995,
+      "step": 5410
+    },
+    {
+      "epoch": 2.8339444735463593,
+      "loss_partial": 0.048095703125,
+      "step": 5410
+    },
+    {
+      "epoch": 2.8339444735463593,
+      "loss_partial": 0.043212890625,
+      "step": 5410
+    },
+    {
+      "epoch": 2.8344683080146673,
+      "grad_norm": 1.013719081878662,
+      "learning_rate": 4.746199725941278e-06,
+      "loss": 0.1037,
+      "step": 5411
+    },
+    {
+      "epoch": 2.8344683080146673,
+      "loss_partial": 0.044677734375,
+      "step": 5411
+    },
+    {
+      "epoch": 2.8344683080146673,
+      "loss_partial": 0.049072265625,
+      "step": 5411
+    },
+    {
+      "epoch": 2.8349921424829754,
+      "grad_norm": 1.9665182828903198,
+      "learning_rate": 4.742310461803007e-06,
+      "loss": 0.1015,
+      "step": 5412
+    },
+    {
+      "epoch": 2.8349921424829754,
+      "loss_partial": 0.04931640625,
+      "step": 5412
+    },
+    {
+      "epoch": 2.8349921424829754,
+      "loss_partial": 0.05322265625,
+      "step": 5412
+    },
+    {
+      "epoch": 2.8355159769512834,
+      "grad_norm": 1.6394968032836914,
+      "learning_rate": 4.7384222964893324e-06,
+      "loss": 0.0995,
+      "step": 5413
+    },
+    {
+      "epoch": 2.8355159769512834,
+      "loss_partial": 0.05224609375,
+      "step": 5413
+    },
+    {
+      "epoch": 2.8355159769512834,
+      "loss_partial": 0.04443359375,
+      "step": 5413
+    },
+    {
+      "epoch": 2.8360398114195915,
+      "grad_norm": 1.2752339839935303,
+      "learning_rate": 4.734535230812865e-06,
+      "loss": 0.1006,
+      "step": 5414
+    },
+    {
+      "epoch": 2.8360398114195915,
+      "loss_partial": 0.04443359375,
+      "step": 5414
+    },
+    {
+      "epoch": 2.8360398114195915,
+      "loss_partial": 0.051025390625,
+      "step": 5414
+    },
+    {
+      "epoch": 2.8365636458878996,
+      "grad_norm": 1.4906612634658813,
+      "learning_rate": 4.7306492655859735e-06,
+      "loss": 0.1011,
+      "step": 5415
+    },
+    {
+      "epoch": 2.8365636458878996,
+      "loss_partial": 0.04638671875,
+      "step": 5415
+    },
+    {
+      "epoch": 2.8365636458878996,
+      "loss_partial": 0.052001953125,
+      "step": 5415
+    },
+    {
+      "epoch": 2.8370874803562076,
+      "grad_norm": 1.613943099975586,
+      "learning_rate": 4.726764401620802e-06,
+      "loss": 0.1022,
+      "step": 5416
+    },
+    {
+      "epoch": 2.8370874803562076,
+      "loss_partial": 0.05126953125,
+      "step": 5416
+    },
+    {
+      "epoch": 2.8370874803562076,
+      "loss_partial": 0.0458984375,
+      "step": 5416
+    },
+    {
+      "epoch": 2.8376113148245157,
+      "grad_norm": 1.1367088556289673,
+      "learning_rate": 4.722880639729256e-06,
+      "loss": 0.0996,
+      "step": 5417
+    },
+    {
+      "epoch": 2.8376113148245157,
+      "loss_partial": 0.059326171875,
+      "step": 5417
+    },
+    {
+      "epoch": 2.8376113148245157,
+      "loss_partial": 0.07763671875,
+      "step": 5417
+    },
+    {
+      "epoch": 2.8381351492928237,
+      "grad_norm": 1.4404901266098022,
+      "learning_rate": 4.718997980723029e-06,
+      "loss": 0.1041,
+      "step": 5418
+    },
+    {
+      "epoch": 2.8381351492928237,
+      "loss_partial": 0.057861328125,
+      "step": 5418
+    },
+    {
+      "epoch": 2.8381351492928237,
+      "loss_partial": 0.044189453125,
+      "step": 5418
+    },
+    {
+      "epoch": 2.8386589837611313,
+      "grad_norm": 1.463174819946289,
+      "learning_rate": 4.715116425413567e-06,
+      "loss": 0.0982,
+      "step": 5419
+    },
+    {
+      "epoch": 2.8386589837611313,
+      "loss_partial": 0.04833984375,
+      "step": 5419
+    },
+    {
+      "epoch": 2.8386589837611313,
+      "loss_partial": 0.04833984375,
+      "step": 5419
+    },
+    {
+      "epoch": 2.8391828182294394,
+      "grad_norm": 1.3753182888031006,
+      "learning_rate": 4.711235974612091e-06,
+      "loss": 0.1049,
+      "step": 5420
+    },
+    {
+      "epoch": 2.8391828182294394,
+      "loss_partial": 0.044921875,
+      "step": 5420
+    },
+    {
+      "epoch": 2.8391828182294394,
+      "loss_partial": 0.041015625,
+      "step": 5420
+    },
+    {
+      "epoch": 2.8397066526977475,
+      "grad_norm": 1.5846871137619019,
+      "learning_rate": 4.707356629129588e-06,
+      "loss": 0.1028,
+      "step": 5421
+    },
+    {
+      "epoch": 2.8397066526977475,
+      "loss_partial": 0.0517578125,
+      "step": 5421
+    },
+    {
+      "epoch": 2.8397066526977475,
+      "loss_partial": 0.05419921875,
+      "step": 5421
+    },
+    {
+      "epoch": 2.8402304871660555,
+      "grad_norm": 1.3094528913497925,
+      "learning_rate": 4.7034783897768285e-06,
+      "loss": 0.104,
+      "step": 5422
+    },
+    {
+      "epoch": 2.8402304871660555,
+      "loss_partial": 0.044189453125,
+      "step": 5422
+    },
+    {
+      "epoch": 2.8402304871660555,
+      "loss_partial": 0.05029296875,
+      "step": 5422
+    },
+    {
+      "epoch": 2.8407543216343636,
+      "grad_norm": 1.4209988117218018,
+      "learning_rate": 4.699601257364329e-06,
+      "loss": 0.1011,
+      "step": 5423
+    },
+    {
+      "epoch": 2.8407543216343636,
+      "loss_partial": 0.05859375,
+      "step": 5423
+    },
+    {
+      "epoch": 2.8407543216343636,
+      "loss_partial": 0.06103515625,
+      "step": 5423
+    },
+    {
+      "epoch": 2.8412781561026716,
+      "grad_norm": 1.7943212985992432,
+      "learning_rate": 4.695725232702389e-06,
+      "loss": 0.1007,
+      "step": 5424
+    },
+    {
+      "epoch": 2.8412781561026716,
+      "loss_partial": 0.04931640625,
+      "step": 5424
+    },
+    {
+      "epoch": 2.8412781561026716,
+      "loss_partial": 0.04345703125,
+      "step": 5424
+    },
+    {
+      "epoch": 2.8418019905709797,
+      "grad_norm": 1.5447946786880493,
+      "learning_rate": 4.691850316601073e-06,
+      "loss": 0.1012,
+      "step": 5425
+    },
+    {
+      "epoch": 2.8418019905709797,
+      "loss_partial": 0.06005859375,
+      "step": 5425
+    },
+    {
+      "epoch": 2.8418019905709797,
+      "loss_partial": 0.048095703125,
+      "step": 5425
+    },
+    {
+      "epoch": 2.8423258250392873,
+      "grad_norm": 1.3304163217544556,
+      "learning_rate": 4.687976509870219e-06,
+      "loss": 0.1017,
+      "step": 5426
+    },
+    {
+      "epoch": 2.8423258250392873,
+      "loss_partial": 0.05078125,
+      "step": 5426
+    },
+    {
+      "epoch": 2.8423258250392873,
+      "loss_partial": 0.052978515625,
+      "step": 5426
+    },
+    {
+      "epoch": 2.8428496595075954,
+      "grad_norm": 1.3123916387557983,
+      "learning_rate": 4.684103813319425e-06,
+      "loss": 0.1011,
+      "step": 5427
+    },
+    {
+      "epoch": 2.8428496595075954,
+      "loss_partial": 0.04443359375,
+      "step": 5427
+    },
+    {
+      "epoch": 2.8428496595075954,
+      "loss_partial": 0.052490234375,
+      "step": 5427
+    },
+    {
+      "epoch": 2.8433734939759034,
+      "grad_norm": 1.857090950012207,
+      "learning_rate": 4.680232227758062e-06,
+      "loss": 0.1018,
+      "step": 5428
+    },
+    {
+      "epoch": 2.8433734939759034,
+      "loss_partial": 0.055908203125,
+      "step": 5428
+    },
+    {
+      "epoch": 2.8433734939759034,
+      "loss_partial": 0.04736328125,
+      "step": 5428
+    },
+    {
+      "epoch": 2.8438973284442115,
+      "grad_norm": 1.6706814765930176,
+      "learning_rate": 4.676361753995264e-06,
+      "loss": 0.098,
+      "step": 5429
+    },
+    {
+      "epoch": 2.8438973284442115,
+      "loss_partial": 0.06640625,
+      "step": 5429
+    },
+    {
+      "epoch": 2.8438973284442115,
+      "loss_partial": 0.048583984375,
+      "step": 5429
+    },
+    {
+      "epoch": 2.8444211629125196,
+      "grad_norm": 2.136056900024414,
+      "learning_rate": 4.6724923928399425e-06,
+      "loss": 0.105,
+      "step": 5430
+    },
+    {
+      "epoch": 2.8444211629125196,
+      "loss_partial": 0.039794921875,
+      "step": 5430
+    },
+    {
+      "epoch": 2.8444211629125196,
+      "loss_partial": 0.052490234375,
+      "step": 5430
+    },
+    {
+      "epoch": 2.8449449973808276,
+      "grad_norm": 1.6418062448501587,
+      "learning_rate": 4.668624145100767e-06,
+      "loss": 0.1025,
+      "step": 5431
+    },
+    {
+      "epoch": 2.8449449973808276,
+      "loss_partial": 0.057373046875,
+      "step": 5431
+    },
+    {
+      "epoch": 2.8449449973808276,
+      "loss_partial": 0.048828125,
+      "step": 5431
+    },
+    {
+      "epoch": 2.8454688318491357,
+      "grad_norm": 1.2658404111862183,
+      "learning_rate": 4.664757011586175e-06,
+      "loss": 0.099,
+      "step": 5432
+    },
+    {
+      "epoch": 2.8454688318491357,
+      "loss_partial": 0.0478515625,
+      "step": 5432
+    },
+    {
+      "epoch": 2.8454688318491357,
+      "loss_partial": 0.054443359375,
+      "step": 5432
+    },
+    {
+      "epoch": 2.8459926663174437,
+      "grad_norm": 1.2831213474273682,
+      "learning_rate": 4.6608909931043735e-06,
+      "loss": 0.0992,
+      "step": 5433
+    },
+    {
+      "epoch": 2.8459926663174437,
+      "loss_partial": 0.0419921875,
+      "step": 5433
+    },
+    {
+      "epoch": 2.8459926663174437,
+      "loss_partial": 0.046142578125,
+      "step": 5433
+    },
+    {
+      "epoch": 2.846516500785752,
+      "grad_norm": 1.344188928604126,
+      "learning_rate": 4.657026090463345e-06,
+      "loss": 0.0992,
+      "step": 5434
+    },
+    {
+      "epoch": 2.846516500785752,
+      "loss_partial": 0.048828125,
+      "step": 5434
+    },
+    {
+      "epoch": 2.846516500785752,
+      "loss_partial": 0.0478515625,
+      "step": 5434
+    },
+    {
+      "epoch": 2.84704033525406,
+      "grad_norm": 1.46793532371521,
+      "learning_rate": 4.653162304470821e-06,
+      "loss": 0.1042,
+      "step": 5435
+    },
+    {
+      "epoch": 2.84704033525406,
+      "loss_partial": 0.052001953125,
+      "step": 5435
+    },
+    {
+      "epoch": 2.84704033525406,
+      "loss_partial": 0.042724609375,
+      "step": 5435
+    },
+    {
+      "epoch": 2.847564169722368,
+      "grad_norm": 1.2752553224563599,
+      "learning_rate": 4.64929963593431e-06,
+      "loss": 0.0985,
+      "step": 5436
+    },
+    {
+      "epoch": 2.847564169722368,
+      "loss_partial": 0.041748046875,
+      "step": 5436
+    },
+    {
+      "epoch": 2.847564169722368,
+      "loss_partial": 0.047119140625,
+      "step": 5436
+    },
+    {
+      "epoch": 2.848088004190676,
+      "grad_norm": 1.0850138664245605,
+      "learning_rate": 4.645438085661085e-06,
+      "loss": 0.0968,
+      "step": 5437
+    },
+    {
+      "epoch": 2.848088004190676,
+      "loss_partial": 0.05615234375,
+      "step": 5437
+    },
+    {
+      "epoch": 2.848088004190676,
+      "loss_partial": 0.046875,
+      "step": 5437
+    },
+    {
+      "epoch": 2.8486118386589836,
+      "grad_norm": 1.7159733772277832,
+      "learning_rate": 4.641577654458192e-06,
+      "loss": 0.1013,
+      "step": 5438
+    },
+    {
+      "epoch": 2.8486118386589836,
+      "loss_partial": 0.057373046875,
+      "step": 5438
+    },
+    {
+      "epoch": 2.8486118386589836,
+      "loss_partial": 0.05126953125,
+      "step": 5438
+    },
+    {
+      "epoch": 2.8491356731272917,
+      "grad_norm": 1.325321912765503,
+      "learning_rate": 4.6377183431324326e-06,
+      "loss": 0.1013,
+      "step": 5439
+    },
+    {
+      "epoch": 2.8491356731272917,
+      "loss_partial": 0.055419921875,
+      "step": 5439
+    },
+    {
+      "epoch": 2.8491356731272917,
+      "loss_partial": 0.052734375,
+      "step": 5439
+    },
+    {
+      "epoch": 2.8496595075955997,
+      "grad_norm": 1.1306085586547852,
+      "learning_rate": 4.6338601524903805e-06,
+      "loss": 0.0989,
+      "step": 5440
+    },
+    {
+      "epoch": 2.8496595075955997,
+      "loss_partial": 0.039306640625,
+      "step": 5440
+    },
+    {
+      "epoch": 2.8496595075955997,
+      "loss_partial": 0.058837890625,
+      "step": 5440
+    },
+    {
+      "epoch": 2.8501833420639078,
+      "grad_norm": 1.5805615186691284,
+      "learning_rate": 4.63000308333837e-06,
+      "loss": 0.102,
+      "step": 5441
+    },
+    {
+      "epoch": 2.8501833420639078,
+      "loss_partial": 0.052001953125,
+      "step": 5441
+    },
+    {
+      "epoch": 2.8501833420639078,
+      "loss_partial": 0.049072265625,
+      "step": 5441
+    },
+    {
+      "epoch": 2.850707176532216,
+      "grad_norm": 1.047638177871704,
+      "learning_rate": 4.626147136482512e-06,
+      "loss": 0.0995,
+      "step": 5442
+    },
+    {
+      "epoch": 2.850707176532216,
+      "loss_partial": 0.043701171875,
+      "step": 5442
+    },
+    {
+      "epoch": 2.850707176532216,
+      "loss_partial": 0.050537109375,
+      "step": 5442
+    },
+    {
+      "epoch": 2.851231011000524,
+      "grad_norm": 1.5697622299194336,
+      "learning_rate": 4.622292312728671e-06,
+      "loss": 0.0979,
+      "step": 5443
+    },
+    {
+      "epoch": 2.851231011000524,
+      "loss_partial": 0.059814453125,
+      "step": 5443
+    },
+    {
+      "epoch": 2.851231011000524,
+      "loss_partial": 0.04150390625,
+      "step": 5443
+    },
+    {
+      "epoch": 2.851754845468832,
+      "grad_norm": 1.1164042949676514,
+      "learning_rate": 4.618438612882482e-06,
+      "loss": 0.0993,
+      "step": 5444
+    },
+    {
+      "epoch": 2.851754845468832,
+      "loss_partial": 0.044921875,
+      "step": 5444
+    },
+    {
+      "epoch": 2.851754845468832,
+      "loss_partial": 0.049072265625,
+      "step": 5444
+    },
+    {
+      "epoch": 2.85227867993714,
+      "grad_norm": 1.3666397333145142,
+      "learning_rate": 4.61458603774934e-06,
+      "loss": 0.1017,
+      "step": 5445
+    },
+    {
+      "epoch": 2.85227867993714,
+      "loss_partial": 0.04833984375,
+      "step": 5445
+    },
+    {
+      "epoch": 2.85227867993714,
+      "loss_partial": 0.061767578125,
+      "step": 5445
+    },
+    {
+      "epoch": 2.8528025144054476,
+      "grad_norm": 1.7180607318878174,
+      "learning_rate": 4.610734588134419e-06,
+      "loss": 0.1011,
+      "step": 5446
+    },
+    {
+      "epoch": 2.8528025144054476,
+      "loss_partial": 0.050537109375,
+      "step": 5446
+    },
+    {
+      "epoch": 2.8528025144054476,
+      "loss_partial": 0.06787109375,
+      "step": 5446
+    },
+    {
+      "epoch": 2.8533263488737557,
+      "grad_norm": 1.2099425792694092,
+      "learning_rate": 4.606884264842646e-06,
+      "loss": 0.1053,
+      "step": 5447
+    },
+    {
+      "epoch": 2.8533263488737557,
+      "loss_partial": 0.049072265625,
+      "step": 5447
+    },
+    {
+      "epoch": 2.8533263488737557,
+      "loss_partial": 0.048828125,
+      "step": 5447
+    },
+    {
+      "epoch": 2.8538501833420638,
+      "grad_norm": 1.2740826606750488,
+      "learning_rate": 4.6030350686787076e-06,
+      "loss": 0.0992,
+      "step": 5448
+    },
+    {
+      "epoch": 2.8538501833420638,
+      "loss_partial": 0.052001953125,
+      "step": 5448
+    },
+    {
+      "epoch": 2.8538501833420638,
+      "loss_partial": 0.051513671875,
+      "step": 5448
+    },
+    {
+      "epoch": 2.854374017810372,
+      "grad_norm": 1.7402061223983765,
+      "learning_rate": 4.599187000447062e-06,
+      "loss": 0.1024,
+      "step": 5449
+    },
+    {
+      "epoch": 2.854374017810372,
+      "loss_partial": 0.0439453125,
+      "step": 5449
+    },
+    {
+      "epoch": 2.854374017810372,
+      "loss_partial": 0.047119140625,
+      "step": 5449
+    },
+    {
+      "epoch": 2.85489785227868,
+      "grad_norm": 1.6378302574157715,
+      "learning_rate": 4.59534006095194e-06,
+      "loss": 0.0983,
+      "step": 5450
+    },
+    {
+      "epoch": 2.85489785227868,
+      "loss_partial": 0.0498046875,
+      "step": 5450
+    },
+    {
+      "epoch": 2.85489785227868,
+      "loss_partial": 0.04638671875,
+      "step": 5450
+    },
+    {
+      "epoch": 2.855421686746988,
+      "grad_norm": 1.4625452756881714,
+      "learning_rate": 4.591494250997324e-06,
+      "loss": 0.098,
+      "step": 5451
+    },
+    {
+      "epoch": 2.855421686746988,
+      "loss_partial": 0.045654296875,
+      "step": 5451
+    },
+    {
+      "epoch": 2.855421686746988,
+      "loss_partial": 0.048583984375,
+      "step": 5451
+    },
+    {
+      "epoch": 2.855945521215296,
+      "grad_norm": 1.9998650550842285,
+      "learning_rate": 4.587649571386963e-06,
+      "loss": 0.1003,
+      "step": 5452
+    },
+    {
+      "epoch": 2.855945521215296,
+      "loss_partial": 0.044677734375,
+      "step": 5452
+    },
+    {
+      "epoch": 2.855945521215296,
+      "loss_partial": 0.0537109375,
+      "step": 5452
+    },
+    {
+      "epoch": 2.856469355683604,
+      "grad_norm": 1.3501108884811401,
+      "learning_rate": 4.58380602292437e-06,
+      "loss": 0.1002,
+      "step": 5453
+    },
+    {
+      "epoch": 2.856469355683604,
+      "loss_partial": 0.04833984375,
+      "step": 5453
+    },
+    {
+      "epoch": 2.856469355683604,
+      "loss_partial": 0.06689453125,
+      "step": 5453
+    },
+    {
+      "epoch": 2.856993190151912,
+      "grad_norm": 1.1250618696212769,
+      "learning_rate": 4.579963606412829e-06,
+      "loss": 0.1047,
+      "step": 5454
+    },
+    {
+      "epoch": 2.856993190151912,
+      "loss_partial": 0.0537109375,
+      "step": 5454
+    },
+    {
+      "epoch": 2.856993190151912,
+      "loss_partial": 0.043212890625,
+      "step": 5454
+    },
+    {
+      "epoch": 2.85751702462022,
+      "grad_norm": 1.5116596221923828,
+      "learning_rate": 4.576122322655378e-06,
+      "loss": 0.1008,
+      "step": 5455
+    },
+    {
+      "epoch": 2.85751702462022,
+      "loss_partial": 0.052734375,
+      "step": 5455
+    },
+    {
+      "epoch": 2.85751702462022,
+      "loss_partial": 0.048828125,
+      "step": 5455
+    },
+    {
+      "epoch": 2.8580408590885282,
+      "grad_norm": 1.319466233253479,
+      "learning_rate": 4.5722821724548215e-06,
+      "loss": 0.1058,
+      "step": 5456
+    },
+    {
+      "epoch": 2.8580408590885282,
+      "loss_partial": 0.046630859375,
+      "step": 5456
+    },
+    {
+      "epoch": 2.8580408590885282,
+      "loss_partial": 0.0419921875,
+      "step": 5456
+    },
+    {
+      "epoch": 2.8585646935568363,
+      "grad_norm": 1.4158093929290771,
+      "learning_rate": 4.568443156613722e-06,
+      "loss": 0.0967,
+      "step": 5457
+    },
+    {
+      "epoch": 2.8585646935568363,
+      "loss_partial": 0.04736328125,
+      "step": 5457
+    },
+    {
+      "epoch": 2.8585646935568363,
+      "loss_partial": 0.05126953125,
+      "step": 5457
+    },
+    {
+      "epoch": 2.859088528025144,
+      "grad_norm": 1.276081919670105,
+      "learning_rate": 4.5646052759344206e-06,
+      "loss": 0.1021,
+      "step": 5458
+    },
+    {
+      "epoch": 2.859088528025144,
+      "loss_partial": 0.048095703125,
+      "step": 5458
+    },
+    {
+      "epoch": 2.859088528025144,
+      "loss_partial": 0.054443359375,
+      "step": 5458
+    },
+    {
+      "epoch": 2.859612362493452,
+      "grad_norm": 1.1503660678863525,
+      "learning_rate": 4.560768531219008e-06,
+      "loss": 0.1009,
+      "step": 5459
+    },
+    {
+      "epoch": 2.859612362493452,
+      "loss_partial": 0.052978515625,
+      "step": 5459
+    },
+    {
+      "epoch": 2.859612362493452,
+      "loss_partial": 0.046630859375,
+      "step": 5459
+    },
+    {
+      "epoch": 2.86013619696176,
+      "grad_norm": 1.3775506019592285,
+      "learning_rate": 4.556932923269332e-06,
+      "loss": 0.0978,
+      "step": 5460
+    },
+    {
+      "epoch": 2.86013619696176,
+      "loss_partial": 0.048583984375,
+      "step": 5460
+    },
+    {
+      "epoch": 2.86013619696176,
+      "loss_partial": 0.041259765625,
+      "step": 5460
+    },
+    {
+      "epoch": 2.860660031430068,
+      "grad_norm": 1.3392674922943115,
+      "learning_rate": 4.553098452887018e-06,
+      "loss": 0.0997,
+      "step": 5461
+    },
+    {
+      "epoch": 2.860660031430068,
+      "loss_partial": 0.048095703125,
+      "step": 5461
+    },
+    {
+      "epoch": 2.860660031430068,
+      "loss_partial": 0.06298828125,
+      "step": 5461
+    },
+    {
+      "epoch": 2.861183865898376,
+      "grad_norm": 1.3828221559524536,
+      "learning_rate": 4.54926512087344e-06,
+      "loss": 0.0996,
+      "step": 5462
+    },
+    {
+      "epoch": 2.861183865898376,
+      "loss_partial": 0.055908203125,
+      "step": 5462
+    },
+    {
+      "epoch": 2.861183865898376,
+      "loss_partial": 0.050537109375,
+      "step": 5462
+    },
+    {
+      "epoch": 2.861707700366684,
+      "grad_norm": 1.5087831020355225,
+      "learning_rate": 4.545432928029749e-06,
+      "loss": 0.1028,
+      "step": 5463
+    },
+    {
+      "epoch": 2.861707700366684,
+      "loss_partial": 0.05126953125,
+      "step": 5463
+    },
+    {
+      "epoch": 2.861707700366684,
+      "loss_partial": 0.039794921875,
+      "step": 5463
+    },
+    {
+      "epoch": 2.8622315348349923,
+      "grad_norm": 1.7980390787124634,
+      "learning_rate": 4.541601875156847e-06,
+      "loss": 0.1023,
+      "step": 5464
+    },
+    {
+      "epoch": 2.8622315348349923,
+      "loss_partial": 0.0556640625,
+      "step": 5464
+    },
+    {
+      "epoch": 2.8622315348349923,
+      "loss_partial": 0.046875,
+      "step": 5464
+    },
+    {
+      "epoch": 2.8627553693033003,
+      "grad_norm": 1.3347560167312622,
+      "learning_rate": 4.537771963055399e-06,
+      "loss": 0.0985,
+      "step": 5465
+    },
+    {
+      "epoch": 2.8627553693033003,
+      "loss_partial": 0.061279296875,
+      "step": 5465
+    },
+    {
+      "epoch": 2.8627553693033003,
+      "loss_partial": 0.04931640625,
+      "step": 5465
+    },
+    {
+      "epoch": 2.863279203771608,
+      "grad_norm": 1.2932184934616089,
+      "learning_rate": 4.5339431925258295e-06,
+      "loss": 0.1006,
+      "step": 5466
+    },
+    {
+      "epoch": 2.863279203771608,
+      "loss_partial": 0.051513671875,
+      "step": 5466
+    },
+    {
+      "epoch": 2.863279203771608,
+      "loss_partial": 0.055419921875,
+      "step": 5466
+    },
+    {
+      "epoch": 2.863803038239916,
+      "grad_norm": 1.4662680625915527,
+      "learning_rate": 4.530115564368336e-06,
+      "loss": 0.1022,
+      "step": 5467
+    },
+    {
+      "epoch": 2.863803038239916,
+      "loss_partial": 0.042724609375,
+      "step": 5467
+    },
+    {
+      "epoch": 2.863803038239916,
+      "loss_partial": 0.0615234375,
+      "step": 5467
+    },
+    {
+      "epoch": 2.864326872708224,
+      "grad_norm": 1.2169549465179443,
+      "learning_rate": 4.526289079382864e-06,
+      "loss": 0.1057,
+      "step": 5468
+    },
+    {
+      "epoch": 2.864326872708224,
+      "loss_partial": 0.04736328125,
+      "step": 5468
+    },
+    {
+      "epoch": 2.864326872708224,
+      "loss_partial": 0.052734375,
+      "step": 5468
+    },
+    {
+      "epoch": 2.864850707176532,
+      "grad_norm": 1.347649335861206,
+      "learning_rate": 4.522463738369126e-06,
+      "loss": 0.1,
+      "step": 5469
+    },
+    {
+      "epoch": 2.864850707176532,
+      "loss_partial": 0.056884765625,
+      "step": 5469
+    },
+    {
+      "epoch": 2.864850707176532,
+      "loss_partial": 0.055419921875,
+      "step": 5469
+    },
+    {
+      "epoch": 2.86537454164484,
+      "grad_norm": 1.394509196281433,
+      "learning_rate": 4.5186395421265915e-06,
+      "loss": 0.1011,
+      "step": 5470
+    },
+    {
+      "epoch": 2.86537454164484,
+      "loss_partial": 0.0458984375,
+      "step": 5470
+    },
+    {
+      "epoch": 2.86537454164484,
+      "loss_partial": 0.05859375,
+      "step": 5470
+    },
+    {
+      "epoch": 2.8658983761131482,
+      "grad_norm": 1.1062631607055664,
+      "learning_rate": 4.514816491454504e-06,
+      "loss": 0.1004,
+      "step": 5471
+    },
+    {
+      "epoch": 2.8658983761131482,
+      "loss_partial": 0.04931640625,
+      "step": 5471
+    },
+    {
+      "epoch": 2.8658983761131482,
+      "loss_partial": 0.046875,
+      "step": 5471
+    },
+    {
+      "epoch": 2.8664222105814563,
+      "grad_norm": 1.550408124923706,
+      "learning_rate": 4.510994587151846e-06,
+      "loss": 0.1024,
+      "step": 5472
+    },
+    {
+      "epoch": 2.8664222105814563,
+      "loss_partial": 0.049560546875,
+      "step": 5472
+    },
+    {
+      "epoch": 2.8664222105814563,
+      "loss_partial": 0.046630859375,
+      "step": 5472
+    },
+    {
+      "epoch": 2.8669460450497644,
+      "grad_norm": 1.4526468515396118,
+      "learning_rate": 4.507173830017378e-06,
+      "loss": 0.1036,
+      "step": 5473
+    },
+    {
+      "epoch": 2.8669460450497644,
+      "loss_partial": 0.0537109375,
+      "step": 5473
+    },
+    {
+      "epoch": 2.8669460450497644,
+      "loss_partial": 0.0498046875,
+      "step": 5473
+    },
+    {
+      "epoch": 2.8674698795180724,
+      "grad_norm": 1.5874261856079102,
+      "learning_rate": 4.503354220849607e-06,
+      "loss": 0.1018,
+      "step": 5474
+    },
+    {
+      "epoch": 2.8674698795180724,
+      "loss_partial": 0.04736328125,
+      "step": 5474
+    },
+    {
+      "epoch": 2.8674698795180724,
+      "loss_partial": 0.05224609375,
+      "step": 5474
+    },
+    {
+      "epoch": 2.8679937139863805,
+      "grad_norm": 1.7556689977645874,
+      "learning_rate": 4.499535760446819e-06,
+      "loss": 0.0988,
+      "step": 5475
+    },
+    {
+      "epoch": 2.8679937139863805,
+      "loss_partial": 0.052490234375,
+      "step": 5475
+    },
+    {
+      "epoch": 2.8679937139863805,
+      "loss_partial": 0.04443359375,
+      "step": 5475
+    },
+    {
+      "epoch": 2.8685175484546885,
+      "grad_norm": 1.571257472038269,
+      "learning_rate": 4.495718449607043e-06,
+      "loss": 0.1011,
+      "step": 5476
+    },
+    {
+      "epoch": 2.8685175484546885,
+      "loss_partial": 0.0439453125,
+      "step": 5476
+    },
+    {
+      "epoch": 2.8685175484546885,
+      "loss_partial": 0.047119140625,
+      "step": 5476
+    },
+    {
+      "epoch": 2.8690413829229966,
+      "grad_norm": 1.1131646633148193,
+      "learning_rate": 4.4919022891280725e-06,
+      "loss": 0.1037,
+      "step": 5477
+    },
+    {
+      "epoch": 2.8690413829229966,
+      "loss_partial": 0.05419921875,
+      "step": 5477
+    },
+    {
+      "epoch": 2.8690413829229966,
+      "loss_partial": 0.04736328125,
+      "step": 5477
+    },
+    {
+      "epoch": 2.869565217391304,
+      "grad_norm": 1.4558385610580444,
+      "learning_rate": 4.488087279807461e-06,
+      "loss": 0.101,
+      "step": 5478
+    },
+    {
+      "epoch": 2.869565217391304,
+      "loss_partial": 0.05712890625,
+      "step": 5478
+    },
+    {
+      "epoch": 2.869565217391304,
+      "loss_partial": 0.047119140625,
+      "step": 5478
+    },
+    {
+      "epoch": 2.8700890518596123,
+      "grad_norm": 1.6554723978042603,
+      "learning_rate": 4.4842734224425255e-06,
+      "loss": 0.1022,
+      "step": 5479
+    },
+    {
+      "epoch": 2.8700890518596123,
+      "loss_partial": 0.044189453125,
+      "step": 5479
+    },
+    {
+      "epoch": 2.8700890518596123,
+      "loss_partial": 0.04931640625,
+      "step": 5479
+    },
+    {
+      "epoch": 2.8706128863279203,
+      "grad_norm": 1.3773289918899536,
+      "learning_rate": 4.4804607178303365e-06,
+      "loss": 0.1001,
+      "step": 5480
+    },
+    {
+      "epoch": 2.8706128863279203,
+      "loss_partial": 0.05615234375,
+      "step": 5480
+    },
+    {
+      "epoch": 2.8706128863279203,
+      "loss_partial": 0.048095703125,
+      "step": 5480
+    },
+    {
+      "epoch": 2.8711367207962284,
+      "grad_norm": 1.4563506841659546,
+      "learning_rate": 4.4766491667677256e-06,
+      "loss": 0.1003,
+      "step": 5481
+    },
+    {
+      "epoch": 2.8711367207962284,
+      "loss_partial": 0.05615234375,
+      "step": 5481
+    },
+    {
+      "epoch": 2.8711367207962284,
+      "loss_partial": 0.03759765625,
+      "step": 5481
+    },
+    {
+      "epoch": 2.8716605552645365,
+      "grad_norm": 1.251489520072937,
+      "learning_rate": 4.47283877005128e-06,
+      "loss": 0.0998,
+      "step": 5482
+    },
+    {
+      "epoch": 2.8716605552645365,
+      "loss_partial": 0.04833984375,
+      "step": 5482
+    },
+    {
+      "epoch": 2.8716605552645365,
+      "loss_partial": 0.048095703125,
+      "step": 5482
+    },
+    {
+      "epoch": 2.8721843897328445,
+      "grad_norm": 1.6317731142044067,
+      "learning_rate": 4.469029528477357e-06,
+      "loss": 0.1013,
+      "step": 5483
+    },
+    {
+      "epoch": 2.8721843897328445,
+      "loss_partial": 0.0498046875,
+      "step": 5483
+    },
+    {
+      "epoch": 2.8721843897328445,
+      "loss_partial": 0.045654296875,
+      "step": 5483
+    },
+    {
+      "epoch": 2.8727082242011526,
+      "grad_norm": 1.5037152767181396,
+      "learning_rate": 4.465221442842062e-06,
+      "loss": 0.1,
+      "step": 5484
+    },
+    {
+      "epoch": 2.8727082242011526,
+      "loss_partial": 0.048095703125,
+      "step": 5484
+    },
+    {
+      "epoch": 2.8727082242011526,
+      "loss_partial": 0.05126953125,
+      "step": 5484
+    },
+    {
+      "epoch": 2.87323205866946,
+      "grad_norm": 1.4735671281814575,
+      "learning_rate": 4.4614145139412575e-06,
+      "loss": 0.1002,
+      "step": 5485
+    },
+    {
+      "epoch": 2.87323205866946,
+      "loss_partial": 0.056640625,
+      "step": 5485
+    },
+    {
+      "epoch": 2.87323205866946,
+      "loss_partial": 0.060302734375,
+      "step": 5485
+    },
+    {
+      "epoch": 2.8737558931377682,
+      "grad_norm": 1.4026641845703125,
+      "learning_rate": 4.457608742570565e-06,
+      "loss": 0.1044,
+      "step": 5486
+    },
+    {
+      "epoch": 2.8737558931377682,
+      "loss_partial": 0.05126953125,
+      "step": 5486
+    },
+    {
+      "epoch": 2.8737558931377682,
+      "loss_partial": 0.045166015625,
+      "step": 5486
+    },
+    {
+      "epoch": 2.8742797276060763,
+      "grad_norm": 1.3645344972610474,
+      "learning_rate": 4.453804129525376e-06,
+      "loss": 0.1002,
+      "step": 5487
+    },
+    {
+      "epoch": 2.8742797276060763,
+      "loss_partial": 0.05078125,
+      "step": 5487
+    },
+    {
+      "epoch": 2.8742797276060763,
+      "loss_partial": 0.045166015625,
+      "step": 5487
+    },
+    {
+      "epoch": 2.8748035620743844,
+      "grad_norm": 1.2895325422286987,
+      "learning_rate": 4.450000675600828e-06,
+      "loss": 0.0988,
+      "step": 5488
+    },
+    {
+      "epoch": 2.8748035620743844,
+      "loss_partial": 0.05224609375,
+      "step": 5488
+    },
+    {
+      "epoch": 2.8748035620743844,
+      "loss_partial": 0.0517578125,
+      "step": 5488
+    },
+    {
+      "epoch": 2.8753273965426924,
+      "grad_norm": 1.2301220893859863,
+      "learning_rate": 4.446198381591818e-06,
+      "loss": 0.0997,
+      "step": 5489
+    },
+    {
+      "epoch": 2.8753273965426924,
+      "loss_partial": 0.05224609375,
+      "step": 5489
+    },
+    {
+      "epoch": 2.8753273965426924,
+      "loss_partial": 0.05712890625,
+      "step": 5489
+    },
+    {
+      "epoch": 2.8758512310110005,
+      "grad_norm": 1.4061498641967773,
+      "learning_rate": 4.442397248293e-06,
+      "loss": 0.101,
+      "step": 5490
+    },
+    {
+      "epoch": 2.8758512310110005,
+      "loss_partial": 0.044189453125,
+      "step": 5490
+    },
+    {
+      "epoch": 2.8758512310110005,
+      "loss_partial": 0.048583984375,
+      "step": 5490
+    },
+    {
+      "epoch": 2.8763750654793085,
+      "grad_norm": 1.4624900817871094,
+      "learning_rate": 4.438597276498793e-06,
+      "loss": 0.1019,
+      "step": 5491
+    },
+    {
+      "epoch": 2.8763750654793085,
+      "loss_partial": 0.046630859375,
+      "step": 5491
+    },
+    {
+      "epoch": 2.8763750654793085,
+      "loss_partial": 0.06103515625,
+      "step": 5491
+    },
+    {
+      "epoch": 2.8768988999476166,
+      "grad_norm": 1.570765495300293,
+      "learning_rate": 4.434798467003365e-06,
+      "loss": 0.1028,
+      "step": 5492
+    },
+    {
+      "epoch": 2.8768988999476166,
+      "loss_partial": 0.056396484375,
+      "step": 5492
+    },
+    {
+      "epoch": 2.8768988999476166,
+      "loss_partial": 0.050537109375,
+      "step": 5492
+    },
+    {
+      "epoch": 2.8774227344159247,
+      "grad_norm": 1.512205719947815,
+      "learning_rate": 4.431000820600642e-06,
+      "loss": 0.1012,
+      "step": 5493
+    },
+    {
+      "epoch": 2.8774227344159247,
+      "loss_partial": 0.044921875,
+      "step": 5493
+    },
+    {
+      "epoch": 2.8774227344159247,
+      "loss_partial": 0.048828125,
+      "step": 5493
+    },
+    {
+      "epoch": 2.8779465688842327,
+      "grad_norm": 1.2429802417755127,
+      "learning_rate": 4.427204338084307e-06,
+      "loss": 0.0975,
+      "step": 5494
+    },
+    {
+      "epoch": 2.8779465688842327,
+      "loss_partial": 0.057861328125,
+      "step": 5494
+    },
+    {
+      "epoch": 2.8779465688842327,
+      "loss_partial": 0.06298828125,
+      "step": 5494
+    },
+    {
+      "epoch": 2.878470403352541,
+      "grad_norm": 2.4802727699279785,
+      "learning_rate": 4.423409020247807e-06,
+      "loss": 0.1041,
+      "step": 5495
+    },
+    {
+      "epoch": 2.878470403352541,
+      "loss_partial": 0.0498046875,
+      "step": 5495
+    },
+    {
+      "epoch": 2.878470403352541,
+      "loss_partial": 0.037109375,
+      "step": 5495
+    },
+    {
+      "epoch": 2.878994237820849,
+      "grad_norm": 1.4455126523971558,
+      "learning_rate": 4.419614867884342e-06,
+      "loss": 0.0999,
+      "step": 5496
+    },
+    {
+      "epoch": 2.878994237820849,
+      "loss_partial": 0.041748046875,
+      "step": 5496
+    },
+    {
+      "epoch": 2.878994237820849,
+      "loss_partial": 0.048095703125,
+      "step": 5496
+    },
+    {
+      "epoch": 2.8795180722891565,
+      "grad_norm": 1.1754450798034668,
+      "learning_rate": 4.415821881786857e-06,
+      "loss": 0.1029,
+      "step": 5497
+    },
+    {
+      "epoch": 2.8795180722891565,
+      "loss_partial": 0.050048828125,
+      "step": 5497
+    },
+    {
+      "epoch": 2.8795180722891565,
+      "loss_partial": 0.051513671875,
+      "step": 5497
+    },
+    {
+      "epoch": 2.8800419067574645,
+      "grad_norm": 1.3231782913208008,
+      "learning_rate": 4.412030062748064e-06,
+      "loss": 0.0975,
+      "step": 5498
+    },
+    {
+      "epoch": 2.8800419067574645,
+      "loss_partial": 0.054443359375,
+      "step": 5498
+    },
+    {
+      "epoch": 2.8800419067574645,
+      "loss_partial": 0.04736328125,
+      "step": 5498
+    },
+    {
+      "epoch": 2.8805657412257726,
+      "grad_norm": 1.8461140394210815,
+      "learning_rate": 4.408239411560436e-06,
+      "loss": 0.1017,
+      "step": 5499
+    },
+    {
+      "epoch": 2.8805657412257726,
+      "loss_partial": 0.052734375,
+      "step": 5499
+    },
+    {
+      "epoch": 2.8805657412257726,
+      "loss_partial": 0.043701171875,
+      "step": 5499
+    },
+    {
+      "epoch": 2.8810895756940806,
+      "grad_norm": 1.1586189270019531,
+      "learning_rate": 4.4044499290161935e-06,
+      "loss": 0.1007,
+      "step": 5500
+    },
+    {
+      "epoch": 2.8810895756940806,
+      "loss_partial": 0.0458984375,
+      "step": 5500
+    },
+    {
+      "epoch": 2.8810895756940806,
+      "loss_partial": 0.05126953125,
+      "step": 5500
+    },
+    {
+      "epoch": 2.8816134101623887,
+      "grad_norm": 1.3779757022857666,
+      "learning_rate": 4.400661615907315e-06,
+      "loss": 0.1034,
+      "step": 5501
+    },
+    {
+      "epoch": 2.8816134101623887,
+      "loss_partial": 0.04248046875,
+      "step": 5501
+    },
+    {
+      "epoch": 2.8816134101623887,
+      "loss_partial": 0.04736328125,
+      "step": 5501
+    },
+    {
+      "epoch": 2.8821372446306968,
+      "grad_norm": 1.48667311668396,
+      "learning_rate": 4.396874473025527e-06,
+      "loss": 0.0974,
+      "step": 5502
+    },
+    {
+      "epoch": 2.8821372446306968,
+      "loss_partial": 0.048095703125,
+      "step": 5502
+    },
+    {
+      "epoch": 2.8821372446306968,
+      "loss_partial": 0.043701171875,
+      "step": 5502
+    },
+    {
+      "epoch": 2.882661079099005,
+      "grad_norm": 1.2867305278778076,
+      "learning_rate": 4.393088501162331e-06,
+      "loss": 0.099,
+      "step": 5503
+    },
+    {
+      "epoch": 2.882661079099005,
+      "loss_partial": 0.04736328125,
+      "step": 5503
+    },
+    {
+      "epoch": 2.882661079099005,
+      "loss_partial": 0.04150390625,
+      "step": 5503
+    },
+    {
+      "epoch": 2.883184913567313,
+      "grad_norm": 1.7338377237319946,
+      "learning_rate": 4.389303701108966e-06,
+      "loss": 0.0946,
+      "step": 5504
+    },
+    {
+      "epoch": 2.883184913567313,
+      "loss_partial": 0.0517578125,
+      "step": 5504
+    },
+    {
+      "epoch": 2.883184913567313,
+      "loss_partial": 0.051025390625,
+      "step": 5504
+    },
+    {
+      "epoch": 2.8837087480356205,
+      "grad_norm": 1.2719056606292725,
+      "learning_rate": 4.385520073656432e-06,
+      "loss": 0.0994,
+      "step": 5505
+    },
+    {
+      "epoch": 2.8837087480356205,
+      "loss_partial": 0.043212890625,
+      "step": 5505
+    },
+    {
+      "epoch": 2.8837087480356205,
+      "loss_partial": 0.0517578125,
+      "step": 5505
+    },
+    {
+      "epoch": 2.8842325825039286,
+      "grad_norm": 1.8273680210113525,
+      "learning_rate": 4.381737619595479e-06,
+      "loss": 0.0973,
+      "step": 5506
+    },
+    {
+      "epoch": 2.8842325825039286,
+      "loss_partial": 0.050048828125,
+      "step": 5506
+    },
+    {
+      "epoch": 2.8842325825039286,
+      "loss_partial": 0.052734375,
+      "step": 5506
+    },
+    {
+      "epoch": 2.8847564169722366,
+      "grad_norm": 1.230118751525879,
+      "learning_rate": 4.377956339716626e-06,
+      "loss": 0.101,
+      "step": 5507
+    },
+    {
+      "epoch": 2.8847564169722366,
+      "loss_partial": 0.06005859375,
+      "step": 5507
+    },
+    {
+      "epoch": 2.8847564169722366,
+      "loss_partial": 0.043212890625,
+      "step": 5507
+    },
+    {
+      "epoch": 2.8852802514405447,
+      "grad_norm": 1.0586553812026978,
+      "learning_rate": 4.3741762348101315e-06,
+      "loss": 0.0979,
+      "step": 5508
+    },
+    {
+      "epoch": 2.8852802514405447,
+      "loss_partial": 0.061279296875,
+      "step": 5508
+    },
+    {
+      "epoch": 2.8852802514405447,
+      "loss_partial": 0.042724609375,
+      "step": 5508
+    },
+    {
+      "epoch": 2.8858040859088527,
+      "grad_norm": 1.427607536315918,
+      "learning_rate": 4.370397305666019e-06,
+      "loss": 0.1008,
+      "step": 5509
+    },
+    {
+      "epoch": 2.8858040859088527,
+      "loss_partial": 0.046630859375,
+      "step": 5509
+    },
+    {
+      "epoch": 2.8858040859088527,
+      "loss_partial": 0.054443359375,
+      "step": 5509
+    },
+    {
+      "epoch": 2.886327920377161,
+      "grad_norm": 1.0848182439804077,
+      "learning_rate": 4.366619553074049e-06,
+      "loss": 0.102,
+      "step": 5510
+    },
+    {
+      "epoch": 2.886327920377161,
+      "loss_partial": 0.044189453125,
+      "step": 5510
+    },
+    {
+      "epoch": 2.886327920377161,
+      "loss_partial": 0.048095703125,
+      "step": 5510
+    },
+    {
+      "epoch": 2.886851754845469,
+      "grad_norm": 1.4268625974655151,
+      "learning_rate": 4.362842977823761e-06,
+      "loss": 0.0964,
+      "step": 5511
+    },
+    {
+      "epoch": 2.886851754845469,
+      "loss_partial": 0.044189453125,
+      "step": 5511
+    },
+    {
+      "epoch": 2.886851754845469,
+      "loss_partial": 0.05810546875,
+      "step": 5511
+    },
+    {
+      "epoch": 2.887375589313777,
+      "grad_norm": 1.481766939163208,
+      "learning_rate": 4.3590675807044324e-06,
+      "loss": 0.1032,
+      "step": 5512
+    },
+    {
+      "epoch": 2.887375589313777,
+      "loss_partial": 0.052001953125,
+      "step": 5512
+    },
+    {
+      "epoch": 2.887375589313777,
+      "loss_partial": 0.046630859375,
+      "step": 5512
+    },
+    {
+      "epoch": 2.887899423782085,
+      "grad_norm": 1.8139877319335938,
+      "learning_rate": 4.355293362505097e-06,
+      "loss": 0.1005,
+      "step": 5513
+    },
+    {
+      "epoch": 2.887899423782085,
+      "loss_partial": 0.054931640625,
+      "step": 5513
+    },
+    {
+      "epoch": 2.887899423782085,
+      "loss_partial": 0.043212890625,
+      "step": 5513
+    },
+    {
+      "epoch": 2.888423258250393,
+      "grad_norm": 1.289674162864685,
+      "learning_rate": 4.351520324014538e-06,
+      "loss": 0.1016,
+      "step": 5514
+    },
+    {
+      "epoch": 2.888423258250393,
+      "loss_partial": 0.0439453125,
+      "step": 5514
+    },
+    {
+      "epoch": 2.888423258250393,
+      "loss_partial": 0.045166015625,
+      "step": 5514
+    },
+    {
+      "epoch": 2.888947092718701,
+      "grad_norm": 1.654218316078186,
+      "learning_rate": 4.3477484660213064e-06,
+      "loss": 0.0992,
+      "step": 5515
+    },
+    {
+      "epoch": 2.888947092718701,
+      "loss_partial": 0.048583984375,
+      "step": 5515
+    },
+    {
+      "epoch": 2.888947092718701,
+      "loss_partial": 0.047119140625,
+      "step": 5515
+    },
+    {
+      "epoch": 2.889470927187009,
+      "grad_norm": 2.0273473262786865,
+      "learning_rate": 4.343977789313693e-06,
+      "loss": 0.0999,
+      "step": 5516
+    },
+    {
+      "epoch": 2.889470927187009,
+      "loss_partial": 0.05224609375,
+      "step": 5516
+    },
+    {
+      "epoch": 2.889470927187009,
+      "loss_partial": 0.05126953125,
+      "step": 5516
+    },
+    {
+      "epoch": 2.8899947616553168,
+      "grad_norm": 1.6661409139633179,
+      "learning_rate": 4.340208294679745e-06,
+      "loss": 0.0997,
+      "step": 5517
+    },
+    {
+      "epoch": 2.8899947616553168,
+      "loss_partial": 0.044921875,
+      "step": 5517
+    },
+    {
+      "epoch": 2.8899947616553168,
+      "loss_partial": 0.06298828125,
+      "step": 5517
+    },
+    {
+      "epoch": 2.890518596123625,
+      "grad_norm": 1.6898887157440186,
+      "learning_rate": 4.336439982907261e-06,
+      "loss": 0.103,
+      "step": 5518
+    },
+    {
+      "epoch": 2.890518596123625,
+      "loss_partial": 0.04296875,
+      "step": 5518
+    },
+    {
+      "epoch": 2.890518596123625,
+      "loss_partial": 0.052978515625,
+      "step": 5518
+    },
+    {
+      "epoch": 2.891042430591933,
+      "grad_norm": 1.3262641429901123,
+      "learning_rate": 4.332672854783803e-06,
+      "loss": 0.0997,
+      "step": 5519
+    },
+    {
+      "epoch": 2.891042430591933,
+      "loss_partial": 0.04443359375,
+      "step": 5519
+    },
+    {
+      "epoch": 2.891042430591933,
+      "loss_partial": 0.046630859375,
+      "step": 5519
+    },
+    {
+      "epoch": 2.891566265060241,
+      "grad_norm": 1.6994922161102295,
+      "learning_rate": 4.328906911096672e-06,
+      "loss": 0.0995,
+      "step": 5520
+    },
+    {
+      "epoch": 2.891566265060241,
+      "loss_partial": 0.047607421875,
+      "step": 5520
+    },
+    {
+      "epoch": 2.891566265060241,
+      "loss_partial": 0.040771484375,
+      "step": 5520
+    },
+    {
+      "epoch": 2.892090099528549,
+      "grad_norm": 1.212968349456787,
+      "learning_rate": 4.325142152632931e-06,
+      "loss": 0.1037,
+      "step": 5521
+    },
+    {
+      "epoch": 2.892090099528549,
+      "loss_partial": 0.051513671875,
+      "step": 5521
+    },
+    {
+      "epoch": 2.892090099528549,
+      "loss_partial": 0.050537109375,
+      "step": 5521
+    },
+    {
+      "epoch": 2.892613933996857,
+      "grad_norm": 1.4223328828811646,
+      "learning_rate": 4.321378580179384e-06,
+      "loss": 0.1003,
+      "step": 5522
+    },
+    {
+      "epoch": 2.892613933996857,
+      "loss_partial": 0.056396484375,
+      "step": 5522
+    },
+    {
+      "epoch": 2.892613933996857,
+      "loss_partial": 0.049072265625,
+      "step": 5522
+    },
+    {
+      "epoch": 2.893137768465165,
+      "grad_norm": 1.227493166923523,
+      "learning_rate": 4.317616194522595e-06,
+      "loss": 0.1029,
+      "step": 5523
+    },
+    {
+      "epoch": 2.893137768465165,
+      "loss_partial": 0.043701171875,
+      "step": 5523
+    },
+    {
+      "epoch": 2.893137768465165,
+      "loss_partial": 0.045166015625,
+      "step": 5523
+    },
+    {
+      "epoch": 2.893661602933473,
+      "grad_norm": 1.0549821853637695,
+      "learning_rate": 4.313854996448886e-06,
+      "loss": 0.0983,
+      "step": 5524
+    },
+    {
+      "epoch": 2.893661602933473,
+      "loss_partial": 0.051025390625,
+      "step": 5524
+    },
+    {
+      "epoch": 2.893661602933473,
+      "loss_partial": 0.054443359375,
+      "step": 5524
+    },
+    {
+      "epoch": 2.894185437401781,
+      "grad_norm": 1.3005344867706299,
+      "learning_rate": 4.310094986744321e-06,
+      "loss": 0.1016,
+      "step": 5525
+    },
+    {
+      "epoch": 2.894185437401781,
+      "loss_partial": 0.05517578125,
+      "step": 5525
+    },
+    {
+      "epoch": 2.894185437401781,
+      "loss_partial": 0.0595703125,
+      "step": 5525
+    },
+    {
+      "epoch": 2.894709271870089,
+      "grad_norm": 1.20004141330719,
+      "learning_rate": 4.3063361661947175e-06,
+      "loss": 0.1045,
+      "step": 5526
+    },
+    {
+      "epoch": 2.894709271870089,
+      "loss_partial": 0.049072265625,
+      "step": 5526
+    },
+    {
+      "epoch": 2.894709271870089,
+      "loss_partial": 0.048828125,
+      "step": 5526
+    },
+    {
+      "epoch": 2.895233106338397,
+      "grad_norm": 1.5218167304992676,
+      "learning_rate": 4.302578535585644e-06,
+      "loss": 0.0975,
+      "step": 5527
+    },
+    {
+      "epoch": 2.895233106338397,
+      "loss_partial": 0.053466796875,
+      "step": 5527
+    },
+    {
+      "epoch": 2.895233106338397,
+      "loss_partial": 0.042236328125,
+      "step": 5527
+    },
+    {
+      "epoch": 2.895756940806705,
+      "grad_norm": 1.1413313150405884,
+      "learning_rate": 4.298822095702429e-06,
+      "loss": 0.0988,
+      "step": 5528
+    },
+    {
+      "epoch": 2.895756940806705,
+      "loss_partial": 0.050537109375,
+      "step": 5528
+    },
+    {
+      "epoch": 2.895756940806705,
+      "loss_partial": 0.050048828125,
+      "step": 5528
+    },
+    {
+      "epoch": 2.896280775275013,
+      "grad_norm": 1.6328805685043335,
+      "learning_rate": 4.295066847330139e-06,
+      "loss": 0.1025,
+      "step": 5529
+    },
+    {
+      "epoch": 2.896280775275013,
+      "loss_partial": 0.047607421875,
+      "step": 5529
+    },
+    {
+      "epoch": 2.896280775275013,
+      "loss_partial": 0.045166015625,
+      "step": 5529
+    },
+    {
+      "epoch": 2.896804609743321,
+      "grad_norm": 1.684382438659668,
+      "learning_rate": 4.291312791253601e-06,
+      "loss": 0.0995,
+      "step": 5530
+    },
+    {
+      "epoch": 2.896804609743321,
+      "loss_partial": 0.045654296875,
+      "step": 5530
+    },
+    {
+      "epoch": 2.896804609743321,
+      "loss_partial": 0.050537109375,
+      "step": 5530
+    },
+    {
+      "epoch": 2.897328444211629,
+      "grad_norm": 1.4358477592468262,
+      "learning_rate": 4.287559928257384e-06,
+      "loss": 0.0972,
+      "step": 5531
+    },
+    {
+      "epoch": 2.897328444211629,
+      "loss_partial": 0.056640625,
+      "step": 5531
+    },
+    {
+      "epoch": 2.897328444211629,
+      "loss_partial": 0.052490234375,
+      "step": 5531
+    },
+    {
+      "epoch": 2.897852278679937,
+      "grad_norm": 1.185627818107605,
+      "learning_rate": 4.283808259125822e-06,
+      "loss": 0.1019,
+      "step": 5532
+    },
+    {
+      "epoch": 2.897852278679937,
+      "loss_partial": 0.044921875,
+      "step": 5532
+    },
+    {
+      "epoch": 2.897852278679937,
+      "loss_partial": 0.050537109375,
+      "step": 5532
+    },
+    {
+      "epoch": 2.8983761131482453,
+      "grad_norm": 2.0588109493255615,
+      "learning_rate": 4.280057784642988e-06,
+      "loss": 0.0979,
+      "step": 5533
+    },
+    {
+      "epoch": 2.8983761131482453,
+      "loss_partial": 0.041259765625,
+      "step": 5533
+    },
+    {
+      "epoch": 2.8983761131482453,
+      "loss_partial": 0.05419921875,
+      "step": 5533
+    },
+    {
+      "epoch": 2.8988999476165533,
+      "grad_norm": 1.365464210510254,
+      "learning_rate": 4.2763085055927035e-06,
+      "loss": 0.1003,
+      "step": 5534
+    },
+    {
+      "epoch": 2.8988999476165533,
+      "loss_partial": 0.050048828125,
+      "step": 5534
+    },
+    {
+      "epoch": 2.8988999476165533,
+      "loss_partial": 0.04296875,
+      "step": 5534
+    },
+    {
+      "epoch": 2.8994237820848614,
+      "grad_norm": 1.5084351301193237,
+      "learning_rate": 4.272560422758543e-06,
+      "loss": 0.1007,
+      "step": 5535
+    },
+    {
+      "epoch": 2.8994237820848614,
+      "loss_partial": 0.043701171875,
+      "step": 5535
+    },
+    {
+      "epoch": 2.8994237820848614,
+      "loss_partial": 0.060302734375,
+      "step": 5535
+    },
+    {
+      "epoch": 2.8999476165531695,
+      "grad_norm": 1.73582124710083,
+      "learning_rate": 4.268813536923841e-06,
+      "loss": 0.0952,
+      "step": 5536
+    },
+    {
+      "epoch": 2.8999476165531695,
+      "loss_partial": 0.05029296875,
+      "step": 5536
+    },
+    {
+      "epoch": 2.8999476165531695,
+      "loss_partial": 0.048583984375,
+      "step": 5536
+    },
+    {
+      "epoch": 2.900471451021477,
+      "grad_norm": 1.6447336673736572,
+      "learning_rate": 4.2650678488716675e-06,
+      "loss": 0.0994,
+      "step": 5537
+    },
+    {
+      "epoch": 2.900471451021477,
+      "loss_partial": 0.04638671875,
+      "step": 5537
+    },
+    {
+      "epoch": 2.900471451021477,
+      "loss_partial": 0.05224609375,
+      "step": 5537
+    },
+    {
+      "epoch": 2.900995285489785,
+      "grad_norm": 1.6800694465637207,
+      "learning_rate": 4.261323359384853e-06,
+      "loss": 0.1024,
+      "step": 5538
+    },
+    {
+      "epoch": 2.900995285489785,
+      "loss_partial": 0.040771484375,
+      "step": 5538
+    },
+    {
+      "epoch": 2.900995285489785,
+      "loss_partial": 0.049560546875,
+      "step": 5538
+    },
+    {
+      "epoch": 2.901519119958093,
+      "grad_norm": 1.56015145778656,
+      "learning_rate": 4.257580069245962e-06,
+      "loss": 0.0987,
+      "step": 5539
+    },
+    {
+      "epoch": 2.901519119958093,
+      "loss_partial": 0.046875,
+      "step": 5539
+    },
+    {
+      "epoch": 2.901519119958093,
+      "loss_partial": 0.05078125,
+      "step": 5539
+    },
+    {
+      "epoch": 2.9020429544264013,
+      "grad_norm": 1.5441652536392212,
+      "learning_rate": 4.253837979237333e-06,
+      "loss": 0.1003,
+      "step": 5540
+    },
+    {
+      "epoch": 2.9020429544264013,
+      "loss_partial": 0.044189453125,
+      "step": 5540
+    },
+    {
+      "epoch": 2.9020429544264013,
+      "loss_partial": 0.044677734375,
+      "step": 5540
+    },
+    {
+      "epoch": 2.9025667888947093,
+      "grad_norm": 1.3819023370742798,
+      "learning_rate": 4.250097090141032e-06,
+      "loss": 0.1022,
+      "step": 5541
+    },
+    {
+      "epoch": 2.9025667888947093,
+      "loss_partial": 0.0439453125,
+      "step": 5541
+    },
+    {
+      "epoch": 2.9025667888947093,
+      "loss_partial": 0.052001953125,
+      "step": 5541
+    },
+    {
+      "epoch": 2.9030906233630174,
+      "grad_norm": 1.261030673980713,
+      "learning_rate": 4.246357402738883e-06,
+      "loss": 0.102,
+      "step": 5542
+    },
+    {
+      "epoch": 2.9030906233630174,
+      "loss_partial": 0.056396484375,
+      "step": 5542
+    },
+    {
+      "epoch": 2.9030906233630174,
+      "loss_partial": 0.047607421875,
+      "step": 5542
+    },
+    {
+      "epoch": 2.9036144578313254,
+      "grad_norm": 1.569046139717102,
+      "learning_rate": 4.242618917812453e-06,
+      "loss": 0.1054,
+      "step": 5543
+    },
+    {
+      "epoch": 2.9036144578313254,
+      "loss_partial": 0.05029296875,
+      "step": 5543
+    },
+    {
+      "epoch": 2.9036144578313254,
+      "loss_partial": 0.046142578125,
+      "step": 5543
+    },
+    {
+      "epoch": 2.904138292299633,
+      "grad_norm": 1.4551100730895996,
+      "learning_rate": 4.2388816361430716e-06,
+      "loss": 0.1012,
+      "step": 5544
+    },
+    {
+      "epoch": 2.904138292299633,
+      "loss_partial": 0.044189453125,
+      "step": 5544
+    },
+    {
+      "epoch": 2.904138292299633,
+      "loss_partial": 0.057373046875,
+      "step": 5544
+    },
+    {
+      "epoch": 2.904662126767941,
+      "grad_norm": 1.5543817281723022,
+      "learning_rate": 4.235145558511802e-06,
+      "loss": 0.104,
+      "step": 5545
+    },
+    {
+      "epoch": 2.904662126767941,
+      "loss_partial": 0.0419921875,
+      "step": 5545
+    },
+    {
+      "epoch": 2.904662126767941,
+      "loss_partial": 0.049072265625,
+      "step": 5545
+    },
+    {
+      "epoch": 2.905185961236249,
+      "grad_norm": 1.3994300365447998,
+      "learning_rate": 4.231410685699465e-06,
+      "loss": 0.0989,
+      "step": 5546
+    },
+    {
+      "epoch": 2.905185961236249,
+      "loss_partial": 0.041259765625,
+      "step": 5546
+    },
+    {
+      "epoch": 2.905185961236249,
+      "loss_partial": 0.04931640625,
+      "step": 5546
+    },
+    {
+      "epoch": 2.9057097957045572,
+      "grad_norm": 1.4606037139892578,
+      "learning_rate": 4.227677018486616e-06,
+      "loss": 0.0964,
+      "step": 5547
+    },
+    {
+      "epoch": 2.9057097957045572,
+      "loss_partial": 0.051513671875,
+      "step": 5547
+    },
+    {
+      "epoch": 2.9057097957045572,
+      "loss_partial": 0.06201171875,
+      "step": 5547
+    },
+    {
+      "epoch": 2.9062336301728653,
+      "grad_norm": 1.1556960344314575,
+      "learning_rate": 4.223944557653578e-06,
+      "loss": 0.0992,
+      "step": 5548
+    },
+    {
+      "epoch": 2.9062336301728653,
+      "loss_partial": 0.052978515625,
+      "step": 5548
+    },
+    {
+      "epoch": 2.9062336301728653,
+      "loss_partial": 0.05029296875,
+      "step": 5548
+    },
+    {
+      "epoch": 2.9067574646411733,
+      "grad_norm": 1.4716501235961914,
+      "learning_rate": 4.2202133039804105e-06,
+      "loss": 0.1004,
+      "step": 5549
+    },
+    {
+      "epoch": 2.9067574646411733,
+      "loss_partial": 0.046875,
+      "step": 5549
+    },
+    {
+      "epoch": 2.9067574646411733,
+      "loss_partial": 0.048583984375,
+      "step": 5549
+    },
+    {
+      "epoch": 2.9072812991094814,
+      "grad_norm": 1.4113564491271973,
+      "learning_rate": 4.216483258246921e-06,
+      "loss": 0.1019,
+      "step": 5550
+    },
+    {
+      "epoch": 2.9072812991094814,
+      "loss_partial": 0.052978515625,
+      "step": 5550
+    },
+    {
+      "epoch": 2.9072812991094814,
+      "loss_partial": 0.04931640625,
+      "step": 5550
+    },
+    {
+      "epoch": 2.9078051335777895,
+      "grad_norm": 1.1083109378814697,
+      "learning_rate": 4.212754421232663e-06,
+      "loss": 0.1013,
+      "step": 5551
+    },
+    {
+      "epoch": 2.9078051335777895,
+      "loss_partial": 0.037841796875,
+      "step": 5551
+    },
+    {
+      "epoch": 2.9078051335777895,
+      "loss_partial": 0.04931640625,
+      "step": 5551
+    },
+    {
+      "epoch": 2.9083289680460975,
+      "grad_norm": 1.2244949340820312,
+      "learning_rate": 4.209026793716947e-06,
+      "loss": 0.1001,
+      "step": 5552
+    },
+    {
+      "epoch": 2.9083289680460975,
+      "loss_partial": 0.056884765625,
+      "step": 5552
+    },
+    {
+      "epoch": 2.9083289680460975,
+      "loss_partial": 0.058837890625,
+      "step": 5552
+    },
+    {
+      "epoch": 2.9088528025144056,
+      "grad_norm": 1.284732460975647,
+      "learning_rate": 4.205300376478822e-06,
+      "loss": 0.1045,
+      "step": 5553
+    },
+    {
+      "epoch": 2.9088528025144056,
+      "loss_partial": 0.0439453125,
+      "step": 5553
+    },
+    {
+      "epoch": 2.9088528025144056,
+      "loss_partial": 0.051025390625,
+      "step": 5553
+    },
+    {
+      "epoch": 2.9093766369827136,
+      "grad_norm": 1.3758716583251953,
+      "learning_rate": 4.201575170297083e-06,
+      "loss": 0.0984,
+      "step": 5554
+    },
+    {
+      "epoch": 2.9093766369827136,
+      "loss_partial": 0.0498046875,
+      "step": 5554
+    },
+    {
+      "epoch": 2.9093766369827136,
+      "loss_partial": 0.05908203125,
+      "step": 5554
+    },
+    {
+      "epoch": 2.9099004714510217,
+      "grad_norm": 1.187752604484558,
+      "learning_rate": 4.197851175950274e-06,
+      "loss": 0.0989,
+      "step": 5555
+    },
+    {
+      "epoch": 2.9099004714510217,
+      "loss_partial": 0.045166015625,
+      "step": 5555
+    },
+    {
+      "epoch": 2.9099004714510217,
+      "loss_partial": 0.042724609375,
+      "step": 5555
+    },
+    {
+      "epoch": 2.9104243059193298,
+      "grad_norm": 1.5491394996643066,
+      "learning_rate": 4.1941283942166936e-06,
+      "loss": 0.1035,
+      "step": 5556
+    },
+    {
+      "epoch": 2.9104243059193298,
+      "loss_partial": 0.0537109375,
+      "step": 5556
+    },
+    {
+      "epoch": 2.9104243059193298,
+      "loss_partial": 0.046875,
+      "step": 5556
+    },
+    {
+      "epoch": 2.9109481403876374,
+      "grad_norm": 1.317083477973938,
+      "learning_rate": 4.190406825874377e-06,
+      "loss": 0.0983,
+      "step": 5557
+    },
+    {
+      "epoch": 2.9109481403876374,
+      "loss_partial": 0.048095703125,
+      "step": 5557
+    },
+    {
+      "epoch": 2.9109481403876374,
+      "loss_partial": 0.0498046875,
+      "step": 5557
+    },
+    {
+      "epoch": 2.9114719748559454,
+      "grad_norm": 0.9755622744560242,
+      "learning_rate": 4.186686471701111e-06,
+      "loss": 0.1004,
+      "step": 5558
+    },
+    {
+      "epoch": 2.9114719748559454,
+      "loss_partial": 0.052490234375,
+      "step": 5558
+    },
+    {
+      "epoch": 2.9114719748559454,
+      "loss_partial": 0.05029296875,
+      "step": 5558
+    },
+    {
+      "epoch": 2.9119958093242535,
+      "grad_norm": 1.1656349897384644,
+      "learning_rate": 4.1829673324744165e-06,
+      "loss": 0.1012,
+      "step": 5559
+    },
+    {
+      "epoch": 2.9119958093242535,
+      "loss_partial": 0.050537109375,
+      "step": 5559
+    },
+    {
+      "epoch": 2.9119958093242535,
+      "loss_partial": 0.053466796875,
+      "step": 5559
+    },
+    {
+      "epoch": 2.9125196437925616,
+      "grad_norm": 1.4981470108032227,
+      "learning_rate": 4.179249408971583e-06,
+      "loss": 0.102,
+      "step": 5560
+    },
+    {
+      "epoch": 2.9125196437925616,
+      "loss_partial": 0.05078125,
+      "step": 5560
+    },
+    {
+      "epoch": 2.9125196437925616,
+      "loss_partial": 0.052978515625,
+      "step": 5560
+    },
+    {
+      "epoch": 2.9130434782608696,
+      "grad_norm": 1.1917973756790161,
+      "learning_rate": 4.175532701969628e-06,
+      "loss": 0.1009,
+      "step": 5561
+    },
+    {
+      "epoch": 2.9130434782608696,
+      "loss_partial": 0.045166015625,
+      "step": 5561
+    },
+    {
+      "epoch": 2.9130434782608696,
+      "loss_partial": 0.041748046875,
+      "step": 5561
+    },
+    {
+      "epoch": 2.9135673127291777,
+      "grad_norm": 1.3552182912826538,
+      "learning_rate": 4.171817212245321e-06,
+      "loss": 0.1005,
+      "step": 5562
+    },
+    {
+      "epoch": 2.9135673127291777,
+      "loss_partial": 0.046630859375,
+      "step": 5562
+    },
+    {
+      "epoch": 2.9135673127291777,
+      "loss_partial": 0.047119140625,
+      "step": 5562
+    },
+    {
+      "epoch": 2.9140911471974857,
+      "grad_norm": 1.308115005493164,
+      "learning_rate": 4.168102940575172e-06,
+      "loss": 0.1038,
+      "step": 5563
+    },
+    {
+      "epoch": 2.9140911471974857,
+      "loss_partial": 0.057861328125,
+      "step": 5563
+    },
+    {
+      "epoch": 2.9140911471974857,
+      "loss_partial": 0.044189453125,
+      "step": 5563
+    },
+    {
+      "epoch": 2.9146149816657934,
+      "grad_norm": 1.3373713493347168,
+      "learning_rate": 4.16438988773545e-06,
+      "loss": 0.1042,
+      "step": 5564
+    },
+    {
+      "epoch": 2.9146149816657934,
+      "loss_partial": 0.0654296875,
+      "step": 5564
+    },
+    {
+      "epoch": 2.9146149816657934,
+      "loss_partial": 0.05615234375,
+      "step": 5564
+    },
+    {
+      "epoch": 2.9151388161341014,
+      "grad_norm": 1.595784068107605,
+      "learning_rate": 4.160678054502153e-06,
+      "loss": 0.1003,
+      "step": 5565
+    },
+    {
+      "epoch": 2.9151388161341014,
+      "loss_partial": 0.047607421875,
+      "step": 5565
+    },
+    {
+      "epoch": 2.9151388161341014,
+      "loss_partial": 0.04296875,
+      "step": 5565
+    },
+    {
+      "epoch": 2.9156626506024095,
+      "grad_norm": 1.749428629875183,
+      "learning_rate": 4.156967441651036e-06,
+      "loss": 0.0999,
+      "step": 5566
+    },
+    {
+      "epoch": 2.9156626506024095,
+      "loss_partial": 0.05615234375,
+      "step": 5566
+    },
+    {
+      "epoch": 2.9156626506024095,
+      "loss_partial": 0.046875,
+      "step": 5566
+    },
+    {
+      "epoch": 2.9161864850707175,
+      "grad_norm": 1.8993165493011475,
+      "learning_rate": 4.1532580499575854e-06,
+      "loss": 0.0988,
+      "step": 5567
+    },
+    {
+      "epoch": 2.9161864850707175,
+      "loss_partial": 0.05712890625,
+      "step": 5567
+    },
+    {
+      "epoch": 2.9161864850707175,
+      "loss_partial": 0.059814453125,
+      "step": 5567
+    },
+    {
+      "epoch": 2.9167103195390256,
+      "grad_norm": 1.404921531677246,
+      "learning_rate": 4.1495498801970516e-06,
+      "loss": 0.1027,
+      "step": 5568
+    },
+    {
+      "epoch": 2.9167103195390256,
+      "loss_partial": 0.042724609375,
+      "step": 5568
+    },
+    {
+      "epoch": 2.9167103195390256,
+      "loss_partial": 0.04736328125,
+      "step": 5568
+    },
+    {
+      "epoch": 2.9172341540073337,
+      "grad_norm": 1.3618348836898804,
+      "learning_rate": 4.145842933144415e-06,
+      "loss": 0.0974,
+      "step": 5569
+    },
+    {
+      "epoch": 2.9172341540073337,
+      "loss_partial": 0.0498046875,
+      "step": 5569
+    },
+    {
+      "epoch": 2.9172341540073337,
+      "loss_partial": 0.04541015625,
+      "step": 5569
+    },
+    {
+      "epoch": 2.9177579884756417,
+      "grad_norm": 1.3491408824920654,
+      "learning_rate": 4.142137209574408e-06,
+      "loss": 0.1045,
+      "step": 5570
+    },
+    {
+      "epoch": 2.9177579884756417,
+      "loss_partial": 0.047607421875,
+      "step": 5570
+    },
+    {
+      "epoch": 2.9177579884756417,
+      "loss_partial": 0.05078125,
+      "step": 5570
+    },
+    {
+      "epoch": 2.9182818229439498,
+      "grad_norm": 1.1148545742034912,
+      "learning_rate": 4.138432710261493e-06,
+      "loss": 0.1001,
+      "step": 5571
+    },
+    {
+      "epoch": 2.9182818229439498,
+      "loss_partial": 0.05322265625,
+      "step": 5571
+    },
+    {
+      "epoch": 2.9182818229439498,
+      "loss_partial": 0.04296875,
+      "step": 5571
+    },
+    {
+      "epoch": 2.918805657412258,
+      "grad_norm": 1.3174034357070923,
+      "learning_rate": 4.134729435979898e-06,
+      "loss": 0.098,
+      "step": 5572
+    },
+    {
+      "epoch": 2.918805657412258,
+      "loss_partial": 0.046875,
+      "step": 5572
+    },
+    {
+      "epoch": 2.918805657412258,
+      "loss_partial": 0.051513671875,
+      "step": 5572
+    },
+    {
+      "epoch": 2.919329491880566,
+      "grad_norm": 1.1170521974563599,
+      "learning_rate": 4.131027387503581e-06,
+      "loss": 0.0975,
+      "step": 5573
+    },
+    {
+      "epoch": 2.919329491880566,
+      "loss_partial": 0.0400390625,
+      "step": 5573
+    },
+    {
+      "epoch": 2.919329491880566,
+      "loss_partial": 0.04736328125,
+      "step": 5573
+    },
+    {
+      "epoch": 2.919853326348874,
+      "grad_norm": 1.4234049320220947,
+      "learning_rate": 4.1273265656062486e-06,
+      "loss": 0.0979,
+      "step": 5574
+    },
+    {
+      "epoch": 2.919853326348874,
+      "loss_partial": 0.04931640625,
+      "step": 5574
+    },
+    {
+      "epoch": 2.919853326348874,
+      "loss_partial": 0.049072265625,
+      "step": 5574
+    },
+    {
+      "epoch": 2.920377160817182,
+      "grad_norm": 1.0163103342056274,
+      "learning_rate": 4.123626971061344e-06,
+      "loss": 0.0997,
+      "step": 5575
+    },
+    {
+      "epoch": 2.920377160817182,
+      "loss_partial": 0.0439453125,
+      "step": 5575
+    },
+    {
+      "epoch": 2.920377160817182,
+      "loss_partial": 0.050537109375,
+      "step": 5575
+    },
+    {
+      "epoch": 2.9209009952854896,
+      "grad_norm": 1.1242023706436157,
+      "learning_rate": 4.119928604642067e-06,
+      "loss": 0.0992,
+      "step": 5576
+    },
+    {
+      "epoch": 2.9209009952854896,
+      "loss_partial": 0.052001953125,
+      "step": 5576
+    },
+    {
+      "epoch": 2.9209009952854896,
+      "loss_partial": 0.05419921875,
+      "step": 5576
+    },
+    {
+      "epoch": 2.9214248297537977,
+      "grad_norm": 1.2942355871200562,
+      "learning_rate": 4.116231467121352e-06,
+      "loss": 0.1035,
+      "step": 5577
+    },
+    {
+      "epoch": 2.9214248297537977,
+      "loss_partial": 0.0556640625,
+      "step": 5577
+    },
+    {
+      "epoch": 2.9214248297537977,
+      "loss_partial": 0.041259765625,
+      "step": 5577
+    },
+    {
+      "epoch": 2.9219486642221058,
+      "grad_norm": 1.0840075016021729,
+      "learning_rate": 4.1125355592718755e-06,
+      "loss": 0.1006,
+      "step": 5578
+    },
+    {
+      "epoch": 2.9219486642221058,
+      "loss_partial": 0.049560546875,
+      "step": 5578
+    },
+    {
+      "epoch": 2.9219486642221058,
+      "loss_partial": 0.0546875,
+      "step": 5578
+    },
+    {
+      "epoch": 2.922472498690414,
+      "grad_norm": 1.3612860441207886,
+      "learning_rate": 4.10884088186606e-06,
+      "loss": 0.0974,
+      "step": 5579
+    },
+    {
+      "epoch": 2.922472498690414,
+      "loss_partial": 0.04736328125,
+      "step": 5579
+    },
+    {
+      "epoch": 2.922472498690414,
+      "loss_partial": 0.044921875,
+      "step": 5579
+    },
+    {
+      "epoch": 2.922996333158722,
+      "grad_norm": 1.2237532138824463,
+      "learning_rate": 4.105147435676065e-06,
+      "loss": 0.0991,
+      "step": 5580
+    },
+    {
+      "epoch": 2.922996333158722,
+      "loss_partial": 0.04345703125,
+      "step": 5580
+    },
+    {
+      "epoch": 2.922996333158722,
+      "loss_partial": 0.05078125,
+      "step": 5580
+    },
+    {
+      "epoch": 2.92352016762703,
+      "grad_norm": 1.1645413637161255,
+      "learning_rate": 4.1014552214738104e-06,
+      "loss": 0.0966,
+      "step": 5581
+    },
+    {
+      "epoch": 2.92352016762703,
+      "loss_partial": 0.050048828125,
+      "step": 5581
+    },
+    {
+      "epoch": 2.92352016762703,
+      "loss_partial": 0.0537109375,
+      "step": 5581
+    },
+    {
+      "epoch": 2.924044002095338,
+      "grad_norm": 1.5365378856658936,
+      "learning_rate": 4.097764240030937e-06,
+      "loss": 0.1017,
+      "step": 5582
+    },
+    {
+      "epoch": 2.924044002095338,
+      "loss_partial": 0.051025390625,
+      "step": 5582
+    },
+    {
+      "epoch": 2.924044002095338,
+      "loss_partial": 0.04833984375,
+      "step": 5582
+    },
+    {
+      "epoch": 2.924567836563646,
+      "grad_norm": 1.244121789932251,
+      "learning_rate": 4.094074492118843e-06,
+      "loss": 0.104,
+      "step": 5583
+    },
+    {
+      "epoch": 2.924567836563646,
+      "loss_partial": 0.044921875,
+      "step": 5583
+    },
+    {
+      "epoch": 2.924567836563646,
+      "loss_partial": 0.05615234375,
+      "step": 5583
+    },
+    {
+      "epoch": 2.9250916710319537,
+      "grad_norm": 1.7401708364486694,
+      "learning_rate": 4.0903859785086505e-06,
+      "loss": 0.102,
+      "step": 5584
+    },
+    {
+      "epoch": 2.9250916710319537,
+      "loss_partial": 0.048583984375,
+      "step": 5584
+    },
+    {
+      "epoch": 2.9250916710319537,
+      "loss_partial": 0.05078125,
+      "step": 5584
+    },
+    {
+      "epoch": 2.9256155055002617,
+      "grad_norm": 1.3911484479904175,
+      "learning_rate": 4.08669869997125e-06,
+      "loss": 0.102,
+      "step": 5585
+    },
+    {
+      "epoch": 2.9256155055002617,
+      "loss_partial": 0.052978515625,
+      "step": 5585
+    },
+    {
+      "epoch": 2.9256155055002617,
+      "loss_partial": 0.04833984375,
+      "step": 5585
+    },
+    {
+      "epoch": 2.92613933996857,
+      "grad_norm": 1.4477146863937378,
+      "learning_rate": 4.0830126572772536e-06,
+      "loss": 0.0992,
+      "step": 5586
+    },
+    {
+      "epoch": 2.92613933996857,
+      "loss_partial": 0.0556640625,
+      "step": 5586
+    },
+    {
+      "epoch": 2.92613933996857,
+      "loss_partial": 0.04541015625,
+      "step": 5586
+    },
+    {
+      "epoch": 2.926663174436878,
+      "grad_norm": 1.3966633081436157,
+      "learning_rate": 4.079327851197023e-06,
+      "loss": 0.0988,
+      "step": 5587
+    },
+    {
+      "epoch": 2.926663174436878,
+      "loss_partial": 0.054443359375,
+      "step": 5587
+    },
+    {
+      "epoch": 2.926663174436878,
+      "loss_partial": 0.045166015625,
+      "step": 5587
+    },
+    {
+      "epoch": 2.927187008905186,
+      "grad_norm": 1.3405877351760864,
+      "learning_rate": 4.075644282500655e-06,
+      "loss": 0.1046,
+      "step": 5588
+    },
+    {
+      "epoch": 2.927187008905186,
+      "loss_partial": 0.0419921875,
+      "step": 5588
+    },
+    {
+      "epoch": 2.927187008905186,
+      "loss_partial": 0.053955078125,
+      "step": 5588
+    },
+    {
+      "epoch": 2.927710843373494,
+      "grad_norm": 1.329229712486267,
+      "learning_rate": 4.071961951958e-06,
+      "loss": 0.1,
+      "step": 5589
+    },
+    {
+      "epoch": 2.927710843373494,
+      "loss_partial": 0.047119140625,
+      "step": 5589
+    },
+    {
+      "epoch": 2.927710843373494,
+      "loss_partial": 0.04443359375,
+      "step": 5589
+    },
+    {
+      "epoch": 2.928234677841802,
+      "grad_norm": 1.6479226350784302,
+      "learning_rate": 4.0682808603386405e-06,
+      "loss": 0.0995,
+      "step": 5590
+    },
+    {
+      "epoch": 2.928234677841802,
+      "loss_partial": 0.05419921875,
+      "step": 5590
+    },
+    {
+      "epoch": 2.928234677841802,
+      "loss_partial": 0.054931640625,
+      "step": 5590
+    },
+    {
+      "epoch": 2.92875851231011,
+      "grad_norm": 1.3416553735733032,
+      "learning_rate": 4.064601008411899e-06,
+      "loss": 0.0926,
+      "step": 5591
+    },
+    {
+      "epoch": 2.92875851231011,
+      "loss_partial": 0.046630859375,
+      "step": 5591
+    },
+    {
+      "epoch": 2.92875851231011,
+      "loss_partial": 0.052978515625,
+      "step": 5591
+    },
+    {
+      "epoch": 2.929282346778418,
+      "grad_norm": 1.364940881729126,
+      "learning_rate": 4.06092239694684e-06,
+      "loss": 0.0976,
+      "step": 5592
+    },
+    {
+      "epoch": 2.929282346778418,
+      "loss_partial": 0.050537109375,
+      "step": 5592
+    },
+    {
+      "epoch": 2.929282346778418,
+      "loss_partial": 0.054443359375,
+      "step": 5592
+    },
+    {
+      "epoch": 2.929806181246726,
+      "grad_norm": 1.0396226644515991,
+      "learning_rate": 4.057245026712279e-06,
+      "loss": 0.0999,
+      "step": 5593
+    },
+    {
+      "epoch": 2.929806181246726,
+      "loss_partial": 0.041015625,
+      "step": 5593
+    },
+    {
+      "epoch": 2.929806181246726,
+      "loss_partial": 0.051025390625,
+      "step": 5593
+    },
+    {
+      "epoch": 2.9303300157150343,
+      "grad_norm": 1.001657485961914,
+      "learning_rate": 4.053568898476756e-06,
+      "loss": 0.0995,
+      "step": 5594
+    },
+    {
+      "epoch": 2.9303300157150343,
+      "loss_partial": 0.04833984375,
+      "step": 5594
+    },
+    {
+      "epoch": 2.9303300157150343,
+      "loss_partial": 0.043212890625,
+      "step": 5594
+    },
+    {
+      "epoch": 2.9308538501833423,
+      "grad_norm": 1.4030802249908447,
+      "learning_rate": 4.049894013008567e-06,
+      "loss": 0.0991,
+      "step": 5595
+    },
+    {
+      "epoch": 2.9308538501833423,
+      "loss_partial": 0.0458984375,
+      "step": 5595
+    },
+    {
+      "epoch": 2.9308538501833423,
+      "loss_partial": 0.06298828125,
+      "step": 5595
+    },
+    {
+      "epoch": 2.93137768465165,
+      "grad_norm": 1.0411245822906494,
+      "learning_rate": 4.046220371075728e-06,
+      "loss": 0.0988,
+      "step": 5596
+    },
+    {
+      "epoch": 2.93137768465165,
+      "loss_partial": 0.049072265625,
+      "step": 5596
+    },
+    {
+      "epoch": 2.93137768465165,
+      "loss_partial": 0.047119140625,
+      "step": 5596
+    },
+    {
+      "epoch": 2.931901519119958,
+      "grad_norm": 1.2027411460876465,
+      "learning_rate": 4.042547973446017e-06,
+      "loss": 0.0985,
+      "step": 5597
+    },
+    {
+      "epoch": 2.931901519119958,
+      "loss_partial": 0.047119140625,
+      "step": 5597
+    },
+    {
+      "epoch": 2.931901519119958,
+      "loss_partial": 0.044189453125,
+      "step": 5597
+    },
+    {
+      "epoch": 2.932425353588266,
+      "grad_norm": 1.124907374382019,
+      "learning_rate": 4.03887682088694e-06,
+      "loss": 0.0978,
+      "step": 5598
+    },
+    {
+      "epoch": 2.932425353588266,
+      "loss_partial": 0.048095703125,
+      "step": 5598
+    },
+    {
+      "epoch": 2.932425353588266,
+      "loss_partial": 0.0546875,
+      "step": 5598
+    },
+    {
+      "epoch": 2.932949188056574,
+      "grad_norm": 1.7166569232940674,
+      "learning_rate": 4.035206914165746e-06,
+      "loss": 0.0984,
+      "step": 5599
+    },
+    {
+      "epoch": 2.932949188056574,
+      "loss_partial": 0.05419921875,
+      "step": 5599
+    },
+    {
+      "epoch": 2.932949188056574,
+      "loss_partial": 0.052001953125,
+      "step": 5599
+    },
+    {
+      "epoch": 2.933473022524882,
+      "grad_norm": 1.1459600925445557,
+      "learning_rate": 4.031538254049416e-06,
+      "loss": 0.1042,
+      "step": 5600
+    },
+    {
+      "epoch": 2.933473022524882,
+      "loss_partial": 0.045166015625,
+      "step": 5600
+    },
+    {
+      "epoch": 2.933473022524882,
+      "loss_partial": 0.045166015625,
+      "step": 5600
+    },
+    {
+      "epoch": 2.9339968569931902,
+      "grad_norm": 1.7246391773223877,
+      "learning_rate": 4.027870841304688e-06,
+      "loss": 0.0964,
+      "step": 5601
+    },
+    {
+      "epoch": 2.9339968569931902,
+      "loss_partial": 0.044189453125,
+      "step": 5601
+    },
+    {
+      "epoch": 2.9339968569931902,
+      "loss_partial": 0.052978515625,
+      "step": 5601
+    },
+    {
+      "epoch": 2.9345206914614983,
+      "grad_norm": 1.093467116355896,
+      "learning_rate": 4.024204676698024e-06,
+      "loss": 0.1051,
+      "step": 5602
+    },
+    {
+      "epoch": 2.9345206914614983,
+      "loss_partial": 0.047607421875,
+      "step": 5602
+    },
+    {
+      "epoch": 2.9345206914614983,
+      "loss_partial": 0.053466796875,
+      "step": 5602
+    },
+    {
+      "epoch": 2.935044525929806,
+      "grad_norm": 1.4629626274108887,
+      "learning_rate": 4.020539760995629e-06,
+      "loss": 0.1039,
+      "step": 5603
+    },
+    {
+      "epoch": 2.935044525929806,
+      "loss_partial": 0.05078125,
+      "step": 5603
+    },
+    {
+      "epoch": 2.935044525929806,
+      "loss_partial": 0.0498046875,
+      "step": 5603
+    },
+    {
+      "epoch": 2.935568360398114,
+      "grad_norm": 1.2311314344406128,
+      "learning_rate": 4.016876094963444e-06,
+      "loss": 0.0965,
+      "step": 5604
+    },
+    {
+      "epoch": 2.935568360398114,
+      "loss_partial": 0.046875,
+      "step": 5604
+    },
+    {
+      "epoch": 2.935568360398114,
+      "loss_partial": 0.050537109375,
+      "step": 5604
+    },
+    {
+      "epoch": 2.936092194866422,
+      "grad_norm": 1.1579128503799438,
+      "learning_rate": 4.013213679367161e-06,
+      "loss": 0.0983,
+      "step": 5605
+    },
+    {
+      "epoch": 2.936092194866422,
+      "loss_partial": 0.04736328125,
+      "step": 5605
+    },
+    {
+      "epoch": 2.936092194866422,
+      "loss_partial": 0.050537109375,
+      "step": 5605
+    },
+    {
+      "epoch": 2.93661602933473,
+      "grad_norm": 1.0797035694122314,
+      "learning_rate": 4.009552514972197e-06,
+      "loss": 0.0967,
+      "step": 5606
+    },
+    {
+      "epoch": 2.93661602933473,
+      "loss_partial": 0.051513671875,
+      "step": 5606
+    },
+    {
+      "epoch": 2.93661602933473,
+      "loss_partial": 0.05322265625,
+      "step": 5606
+    },
+    {
+      "epoch": 2.937139863803038,
+      "grad_norm": 1.2678413391113281,
+      "learning_rate": 4.005892602543714e-06,
+      "loss": 0.0976,
+      "step": 5607
+    },
+    {
+      "epoch": 2.937139863803038,
+      "loss_partial": 0.05712890625,
+      "step": 5607
+    },
+    {
+      "epoch": 2.937139863803038,
+      "loss_partial": 0.046875,
+      "step": 5607
+    },
+    {
+      "epoch": 2.937663698271346,
+      "grad_norm": 1.4559329748153687,
+      "learning_rate": 4.00223394284661e-06,
+      "loss": 0.1013,
+      "step": 5608
+    },
+    {
+      "epoch": 2.937663698271346,
+      "loss_partial": 0.052490234375,
+      "step": 5608
+    },
+    {
+      "epoch": 2.937663698271346,
+      "loss_partial": 0.058349609375,
+      "step": 5608
+    },
+    {
+      "epoch": 2.9381875327396543,
+      "grad_norm": 1.4922901391983032,
+      "learning_rate": 3.998576536645524e-06,
+      "loss": 0.1018,
+      "step": 5609
+    },
+    {
+      "epoch": 2.9381875327396543,
+      "loss_partial": 0.050048828125,
+      "step": 5609
+    },
+    {
+      "epoch": 2.9381875327396543,
+      "loss_partial": 0.048828125,
+      "step": 5609
+    },
+    {
+      "epoch": 2.9387113672079623,
+      "grad_norm": 1.7453101873397827,
+      "learning_rate": 3.99492038470483e-06,
+      "loss": 0.0978,
+      "step": 5610
+    },
+    {
+      "epoch": 2.9387113672079623,
+      "loss_partial": 0.040771484375,
+      "step": 5610
+    },
+    {
+      "epoch": 2.9387113672079623,
+      "loss_partial": 0.050537109375,
+      "step": 5610
+    },
+    {
+      "epoch": 2.9392352016762704,
+      "grad_norm": 1.7174689769744873,
+      "learning_rate": 3.991265487788641e-06,
+      "loss": 0.0987,
+      "step": 5611
+    },
+    {
+      "epoch": 2.9392352016762704,
+      "loss_partial": 0.05029296875,
+      "step": 5611
+    },
+    {
+      "epoch": 2.9392352016762704,
+      "loss_partial": 0.05078125,
+      "step": 5611
+    },
+    {
+      "epoch": 2.9397590361445785,
+      "grad_norm": 1.7731425762176514,
+      "learning_rate": 3.987611846660805e-06,
+      "loss": 0.1007,
+      "step": 5612
+    },
+    {
+      "epoch": 2.9397590361445785,
+      "loss_partial": 0.048095703125,
+      "step": 5612
+    },
+    {
+      "epoch": 2.9397590361445785,
+      "loss_partial": 0.04345703125,
+      "step": 5612
+    },
+    {
+      "epoch": 2.9402828706128865,
+      "grad_norm": 1.1660525798797607,
+      "learning_rate": 3.983959462084917e-06,
+      "loss": 0.0984,
+      "step": 5613
+    },
+    {
+      "epoch": 2.9402828706128865,
+      "loss_partial": 0.050537109375,
+      "step": 5613
+    },
+    {
+      "epoch": 2.9402828706128865,
+      "loss_partial": 0.04541015625,
+      "step": 5613
+    },
+    {
+      "epoch": 2.9408067050811946,
+      "grad_norm": 1.2349376678466797,
+      "learning_rate": 3.9803083348242975e-06,
+      "loss": 0.0989,
+      "step": 5614
+    },
+    {
+      "epoch": 2.9408067050811946,
+      "loss_partial": 0.06201171875,
+      "step": 5614
+    },
+    {
+      "epoch": 2.9408067050811946,
+      "loss_partial": 0.05078125,
+      "step": 5614
+    },
+    {
+      "epoch": 2.9413305395495026,
+      "grad_norm": 1.478745937347412,
+      "learning_rate": 3.976658465642014e-06,
+      "loss": 0.1027,
+      "step": 5615
+    },
+    {
+      "epoch": 2.9413305395495026,
+      "loss_partial": 0.042724609375,
+      "step": 5615
+    },
+    {
+      "epoch": 2.9413305395495026,
+      "loss_partial": 0.043212890625,
+      "step": 5615
+    },
+    {
+      "epoch": 2.9418543740178102,
+      "grad_norm": 1.2313816547393799,
+      "learning_rate": 3.973009855300859e-06,
+      "loss": 0.0993,
+      "step": 5616
+    },
+    {
+      "epoch": 2.9418543740178102,
+      "loss_partial": 0.0400390625,
+      "step": 5616
+    },
+    {
+      "epoch": 2.9418543740178102,
+      "loss_partial": 0.061279296875,
+      "step": 5616
+    },
+    {
+      "epoch": 2.9423782084861183,
+      "grad_norm": 1.488586664199829,
+      "learning_rate": 3.969362504563378e-06,
+      "loss": 0.0998,
+      "step": 5617
+    },
+    {
+      "epoch": 2.9423782084861183,
+      "loss_partial": 0.04541015625,
+      "step": 5617
+    },
+    {
+      "epoch": 2.9423782084861183,
+      "loss_partial": 0.05908203125,
+      "step": 5617
+    },
+    {
+      "epoch": 2.9429020429544264,
+      "grad_norm": 1.8339930772781372,
+      "learning_rate": 3.9657164141918416e-06,
+      "loss": 0.0997,
+      "step": 5618
+    },
+    {
+      "epoch": 2.9429020429544264,
+      "loss_partial": 0.051025390625,
+      "step": 5618
+    },
+    {
+      "epoch": 2.9429020429544264,
+      "loss_partial": 0.048583984375,
+      "step": 5618
+    },
+    {
+      "epoch": 2.9434258774227344,
+      "grad_norm": 1.481052279472351,
+      "learning_rate": 3.962071584948259e-06,
+      "loss": 0.0999,
+      "step": 5619
+    },
+    {
+      "epoch": 2.9434258774227344,
+      "loss_partial": 0.0439453125,
+      "step": 5619
+    },
+    {
+      "epoch": 2.9434258774227344,
+      "loss_partial": 0.04541015625,
+      "step": 5619
+    },
+    {
+      "epoch": 2.9439497118910425,
+      "grad_norm": 1.4301509857177734,
+      "learning_rate": 3.958428017594379e-06,
+      "loss": 0.1009,
+      "step": 5620
+    },
+    {
+      "epoch": 2.9439497118910425,
+      "loss_partial": 0.041748046875,
+      "step": 5620
+    },
+    {
+      "epoch": 2.9439497118910425,
+      "loss_partial": 0.04638671875,
+      "step": 5620
+    },
+    {
+      "epoch": 2.9444735463593505,
+      "grad_norm": 1.262710690498352,
+      "learning_rate": 3.954785712891682e-06,
+      "loss": 0.0964,
+      "step": 5621
+    },
+    {
+      "epoch": 2.9444735463593505,
+      "loss_partial": 0.04345703125,
+      "step": 5621
+    },
+    {
+      "epoch": 2.9444735463593505,
+      "loss_partial": 0.044677734375,
+      "step": 5621
+    },
+    {
+      "epoch": 2.9449973808276586,
+      "grad_norm": 1.216662883758545,
+      "learning_rate": 3.95114467160139e-06,
+      "loss": 0.0965,
+      "step": 5622
+    },
+    {
+      "epoch": 2.9449973808276586,
+      "loss_partial": 0.05615234375,
+      "step": 5622
+    },
+    {
+      "epoch": 2.9449973808276586,
+      "loss_partial": 0.05029296875,
+      "step": 5622
+    },
+    {
+      "epoch": 2.945521215295966,
+      "grad_norm": 2.1866531372070312,
+      "learning_rate": 3.947504894484456e-06,
+      "loss": 0.1012,
+      "step": 5623
+    },
+    {
+      "epoch": 2.945521215295966,
+      "loss_partial": 0.044921875,
+      "step": 5623
+    },
+    {
+      "epoch": 2.945521215295966,
+      "loss_partial": 0.04150390625,
+      "step": 5623
+    },
+    {
+      "epoch": 2.9460450497642743,
+      "grad_norm": 1.5275750160217285,
+      "learning_rate": 3.943866382301567e-06,
+      "loss": 0.0964,
+      "step": 5624
+    },
+    {
+      "epoch": 2.9460450497642743,
+      "loss_partial": 0.05029296875,
+      "step": 5624
+    },
+    {
+      "epoch": 2.9460450497642743,
+      "loss_partial": 0.05126953125,
+      "step": 5624
+    },
+    {
+      "epoch": 2.9465688842325823,
+      "grad_norm": 1.5811079740524292,
+      "learning_rate": 3.9402291358131596e-06,
+      "loss": 0.0991,
+      "step": 5625
+    },
+    {
+      "epoch": 2.9465688842325823,
+      "loss_partial": 0.04931640625,
+      "step": 5625
+    },
+    {
+      "epoch": 2.9465688842325823,
+      "loss_partial": 0.049560546875,
+      "step": 5625
+    },
+    {
+      "epoch": 2.9470927187008904,
+      "grad_norm": 1.6443657875061035,
+      "learning_rate": 3.936593155779389e-06,
+      "loss": 0.0997,
+      "step": 5626
+    },
+    {
+      "epoch": 2.9470927187008904,
+      "loss_partial": 0.041015625,
+      "step": 5626
+    },
+    {
+      "epoch": 2.9470927187008904,
+      "loss_partial": 0.05419921875,
+      "step": 5626
+    },
+    {
+      "epoch": 2.9476165531691985,
+      "grad_norm": 1.1967543363571167,
+      "learning_rate": 3.932958442960154e-06,
+      "loss": 0.1042,
+      "step": 5627
+    },
+    {
+      "epoch": 2.9476165531691985,
+      "loss_partial": 0.050048828125,
+      "step": 5627
+    },
+    {
+      "epoch": 2.9476165531691985,
+      "loss_partial": 0.05126953125,
+      "step": 5627
+    },
+    {
+      "epoch": 2.9481403876375065,
+      "grad_norm": 1.7524864673614502,
+      "learning_rate": 3.929324998115084e-06,
+      "loss": 0.1011,
+      "step": 5628
+    },
+    {
+      "epoch": 2.9481403876375065,
+      "loss_partial": 0.048828125,
+      "step": 5628
+    },
+    {
+      "epoch": 2.9481403876375065,
+      "loss_partial": 0.04443359375,
+      "step": 5628
+    },
+    {
+      "epoch": 2.9486642221058146,
+      "grad_norm": 1.1128727197647095,
+      "learning_rate": 3.9256928220035516e-06,
+      "loss": 0.0974,
+      "step": 5629
+    },
+    {
+      "epoch": 2.9486642221058146,
+      "loss_partial": 0.053466796875,
+      "step": 5629
+    },
+    {
+      "epoch": 2.9486642221058146,
+      "loss_partial": 0.043212890625,
+      "step": 5629
+    },
+    {
+      "epoch": 2.9491880565741226,
+      "grad_norm": 1.2001734972000122,
+      "learning_rate": 3.922061915384658e-06,
+      "loss": 0.1,
+      "step": 5630
+    },
+    {
+      "epoch": 2.9491880565741226,
+      "loss_partial": 0.05078125,
+      "step": 5630
+    },
+    {
+      "epoch": 2.9491880565741226,
+      "loss_partial": 0.044921875,
+      "step": 5630
+    },
+    {
+      "epoch": 2.9497118910424307,
+      "grad_norm": 1.1180806159973145,
+      "learning_rate": 3.91843227901724e-06,
+      "loss": 0.0994,
+      "step": 5631
+    },
+    {
+      "epoch": 2.9497118910424307,
+      "loss_partial": 0.05029296875,
+      "step": 5631
+    },
+    {
+      "epoch": 2.9497118910424307,
+      "loss_partial": 0.048828125,
+      "step": 5631
+    },
+    {
+      "epoch": 2.9502357255107388,
+      "grad_norm": 1.2114554643630981,
+      "learning_rate": 3.914803913659867e-06,
+      "loss": 0.1035,
+      "step": 5632
+    },
+    {
+      "epoch": 2.9502357255107388,
+      "loss_partial": 0.06298828125,
+      "step": 5632
+    },
+    {
+      "epoch": 2.9502357255107388,
+      "loss_partial": 0.05126953125,
+      "step": 5632
+    },
+    {
+      "epoch": 2.950759559979047,
+      "grad_norm": 1.359020471572876,
+      "learning_rate": 3.911176820070848e-06,
+      "loss": 0.1063,
+      "step": 5633
+    },
+    {
+      "epoch": 2.950759559979047,
+      "loss_partial": 0.04541015625,
+      "step": 5633
+    },
+    {
+      "epoch": 2.950759559979047,
+      "loss_partial": 0.048095703125,
+      "step": 5633
+    },
+    {
+      "epoch": 2.951283394447355,
+      "grad_norm": 1.2854658365249634,
+      "learning_rate": 3.907550999008219e-06,
+      "loss": 0.1004,
+      "step": 5634
+    },
+    {
+      "epoch": 2.951283394447355,
+      "loss_partial": 0.05859375,
+      "step": 5634
+    },
+    {
+      "epoch": 2.951283394447355,
+      "loss_partial": 0.056640625,
+      "step": 5634
+    },
+    {
+      "epoch": 2.9518072289156625,
+      "grad_norm": 1.4742774963378906,
+      "learning_rate": 3.903926451229759e-06,
+      "loss": 0.1028,
+      "step": 5635
+    },
+    {
+      "epoch": 2.9518072289156625,
+      "loss_partial": 0.0517578125,
+      "step": 5635
+    },
+    {
+      "epoch": 2.9518072289156625,
+      "loss_partial": 0.04150390625,
+      "step": 5635
+    },
+    {
+      "epoch": 2.9523310633839706,
+      "grad_norm": 1.41044282913208,
+      "learning_rate": 3.900303177492972e-06,
+      "loss": 0.1006,
+      "step": 5636
+    },
+    {
+      "epoch": 2.9523310633839706,
+      "loss_partial": 0.04443359375,
+      "step": 5636
+    },
+    {
+      "epoch": 2.9523310633839706,
+      "loss_partial": 0.048583984375,
+      "step": 5636
+    },
+    {
+      "epoch": 2.9528548978522786,
+      "grad_norm": 1.2575864791870117,
+      "learning_rate": 3.896681178555099e-06,
+      "loss": 0.1,
+      "step": 5637
+    },
+    {
+      "epoch": 2.9528548978522786,
+      "loss_partial": 0.055419921875,
+      "step": 5637
+    },
+    {
+      "epoch": 2.9528548978522786,
+      "loss_partial": 0.053955078125,
+      "step": 5637
+    },
+    {
+      "epoch": 2.9533787323205867,
+      "grad_norm": 1.4739174842834473,
+      "learning_rate": 3.893060455173121e-06,
+      "loss": 0.1034,
+      "step": 5638
+    },
+    {
+      "epoch": 2.9533787323205867,
+      "loss_partial": 0.052001953125,
+      "step": 5638
+    },
+    {
+      "epoch": 2.9533787323205867,
+      "loss_partial": 0.04345703125,
+      "step": 5638
+    },
+    {
+      "epoch": 2.9539025667888947,
+      "grad_norm": 1.2438371181488037,
+      "learning_rate": 3.889441008103745e-06,
+      "loss": 0.099,
+      "step": 5639
+    },
+    {
+      "epoch": 2.9539025667888947,
+      "loss_partial": 0.044189453125,
+      "step": 5639
+    },
+    {
+      "epoch": 2.9539025667888947,
+      "loss_partial": 0.040771484375,
+      "step": 5639
+    },
+    {
+      "epoch": 2.954426401257203,
+      "grad_norm": 1.4761236906051636,
+      "learning_rate": 3.88582283810341e-06,
+      "loss": 0.1005,
+      "step": 5640
+    },
+    {
+      "epoch": 2.954426401257203,
+      "loss_partial": 0.05322265625,
+      "step": 5640
+    },
+    {
+      "epoch": 2.954426401257203,
+      "loss_partial": 0.05029296875,
+      "step": 5640
+    },
+    {
+      "epoch": 2.954950235725511,
+      "grad_norm": 1.4585599899291992,
+      "learning_rate": 3.88220594592829e-06,
+      "loss": 0.0995,
+      "step": 5641
+    },
+    {
+      "epoch": 2.954950235725511,
+      "loss_partial": 0.0556640625,
+      "step": 5641
+    },
+    {
+      "epoch": 2.954950235725511,
+      "loss_partial": 0.054443359375,
+      "step": 5641
+    },
+    {
+      "epoch": 2.955474070193819,
+      "grad_norm": 1.4376448392868042,
+      "learning_rate": 3.8785903323343e-06,
+      "loss": 0.0989,
+      "step": 5642
+    },
+    {
+      "epoch": 2.955474070193819,
+      "loss_partial": 0.0595703125,
+      "step": 5642
+    },
+    {
+      "epoch": 2.955474070193819,
+      "loss_partial": 0.053466796875,
+      "step": 5642
+    },
+    {
+      "epoch": 2.9559979046621265,
+      "grad_norm": 1.2762829065322876,
+      "learning_rate": 3.874975998077076e-06,
+      "loss": 0.1031,
+      "step": 5643
+    },
+    {
+      "epoch": 2.9559979046621265,
+      "loss_partial": 0.044921875,
+      "step": 5643
+    },
+    {
+      "epoch": 2.9559979046621265,
+      "loss_partial": 0.05517578125,
+      "step": 5643
+    },
+    {
+      "epoch": 2.9565217391304346,
+      "grad_norm": 1.214534044265747,
+      "learning_rate": 3.871362943911993e-06,
+      "loss": 0.1001,
+      "step": 5644
+    },
+    {
+      "epoch": 2.9565217391304346,
+      "loss_partial": 0.049072265625,
+      "step": 5644
+    },
+    {
+      "epoch": 2.9565217391304346,
+      "loss_partial": 0.0439453125,
+      "step": 5644
+    },
+    {
+      "epoch": 2.9570455735987426,
+      "grad_norm": 1.3580243587493896,
+      "learning_rate": 3.8677511705941565e-06,
+      "loss": 0.103,
+      "step": 5645
+    },
+    {
+      "epoch": 2.9570455735987426,
+      "loss_partial": 0.050537109375,
+      "step": 5645
+    },
+    {
+      "epoch": 2.9570455735987426,
+      "loss_partial": 0.04638671875,
+      "step": 5645
+    },
+    {
+      "epoch": 2.9575694080670507,
+      "grad_norm": 1.0593892335891724,
+      "learning_rate": 3.864140678878403e-06,
+      "loss": 0.0982,
+      "step": 5646
+    },
+    {
+      "epoch": 2.9575694080670507,
+      "loss_partial": 0.051025390625,
+      "step": 5646
+    },
+    {
+      "epoch": 2.9575694080670507,
+      "loss_partial": 0.04296875,
+      "step": 5646
+    },
+    {
+      "epoch": 2.9580932425353588,
+      "grad_norm": 1.0801112651824951,
+      "learning_rate": 3.860531469519306e-06,
+      "loss": 0.1011,
+      "step": 5647
+    },
+    {
+      "epoch": 2.9580932425353588,
+      "loss_partial": 0.048828125,
+      "step": 5647
+    },
+    {
+      "epoch": 2.9580932425353588,
+      "loss_partial": 0.046142578125,
+      "step": 5647
+    },
+    {
+      "epoch": 2.958617077003667,
+      "grad_norm": 1.2918466329574585,
+      "learning_rate": 3.856923543271167e-06,
+      "loss": 0.1008,
+      "step": 5648
+    },
+    {
+      "epoch": 2.958617077003667,
+      "loss_partial": 0.04931640625,
+      "step": 5648
+    },
+    {
+      "epoch": 2.958617077003667,
+      "loss_partial": 0.04541015625,
+      "step": 5648
+    },
+    {
+      "epoch": 2.959140911471975,
+      "grad_norm": 1.2543199062347412,
+      "learning_rate": 3.853316900888016e-06,
+      "loss": 0.0967,
+      "step": 5649
+    },
+    {
+      "epoch": 2.959140911471975,
+      "loss_partial": 0.05126953125,
+      "step": 5649
+    },
+    {
+      "epoch": 2.959140911471975,
+      "loss_partial": 0.044677734375,
+      "step": 5649
+    },
+    {
+      "epoch": 2.959664745940283,
+      "grad_norm": 1.465669870376587,
+      "learning_rate": 3.849711543123625e-06,
+      "loss": 0.1014,
+      "step": 5650
+    },
+    {
+      "epoch": 2.959664745940283,
+      "loss_partial": 0.04833984375,
+      "step": 5650
+    },
+    {
+      "epoch": 2.959664745940283,
+      "loss_partial": 0.048583984375,
+      "step": 5650
+    },
+    {
+      "epoch": 2.960188580408591,
+      "grad_norm": 1.331255555152893,
+      "learning_rate": 3.8461074707314906e-06,
+      "loss": 0.0994,
+      "step": 5651
+    },
+    {
+      "epoch": 2.960188580408591,
+      "loss_partial": 0.052490234375,
+      "step": 5651
+    },
+    {
+      "epoch": 2.960188580408591,
+      "loss_partial": 0.05224609375,
+      "step": 5651
+    },
+    {
+      "epoch": 2.960712414876899,
+      "grad_norm": 0.8799547553062439,
+      "learning_rate": 3.84250468446484e-06,
+      "loss": 0.0998,
+      "step": 5652
+    },
+    {
+      "epoch": 2.960712414876899,
+      "loss_partial": 0.0556640625,
+      "step": 5652
+    },
+    {
+      "epoch": 2.960712414876899,
+      "loss_partial": 0.04345703125,
+      "step": 5652
+    },
+    {
+      "epoch": 2.961236249345207,
+      "grad_norm": 1.0928773880004883,
+      "learning_rate": 3.83890318507663e-06,
+      "loss": 0.1035,
+      "step": 5653
+    },
+    {
+      "epoch": 2.961236249345207,
+      "loss_partial": 0.04736328125,
+      "step": 5653
+    },
+    {
+      "epoch": 2.961236249345207,
+      "loss_partial": 0.054931640625,
+      "step": 5653
+    },
+    {
+      "epoch": 2.961760083813515,
+      "grad_norm": 1.4402676820755005,
+      "learning_rate": 3.835302973319559e-06,
+      "loss": 0.1017,
+      "step": 5654
+    },
+    {
+      "epoch": 2.961760083813515,
+      "loss_partial": 0.04833984375,
+      "step": 5654
+    },
+    {
+      "epoch": 2.961760083813515,
+      "loss_partial": 0.060302734375,
+      "step": 5654
+    },
+    {
+      "epoch": 2.962283918281823,
+      "grad_norm": 1.1453982591629028,
+      "learning_rate": 3.831704049946044e-06,
+      "loss": 0.0988,
+      "step": 5655
+    },
+    {
+      "epoch": 2.962283918281823,
+      "loss_partial": 0.05029296875,
+      "step": 5655
+    },
+    {
+      "epoch": 2.962283918281823,
+      "loss_partial": 0.048095703125,
+      "step": 5655
+    },
+    {
+      "epoch": 2.962807752750131,
+      "grad_norm": 1.4811878204345703,
+      "learning_rate": 3.828106415708241e-06,
+      "loss": 0.0966,
+      "step": 5656
+    },
+    {
+      "epoch": 2.962807752750131,
+      "loss_partial": 0.0517578125,
+      "step": 5656
+    },
+    {
+      "epoch": 2.962807752750131,
+      "loss_partial": 0.052978515625,
+      "step": 5656
+    },
+    {
+      "epoch": 2.963331587218439,
+      "grad_norm": 1.2737287282943726,
+      "learning_rate": 3.824510071358031e-06,
+      "loss": 0.1,
+      "step": 5657
+    },
+    {
+      "epoch": 2.963331587218439,
+      "loss_partial": 0.04345703125,
+      "step": 5657
+    },
+    {
+      "epoch": 2.963331587218439,
+      "loss_partial": 0.054443359375,
+      "step": 5657
+    },
+    {
+      "epoch": 2.963855421686747,
+      "grad_norm": 1.2943540811538696,
+      "learning_rate": 3.820915017647026e-06,
+      "loss": 0.1006,
+      "step": 5658
+    },
+    {
+      "epoch": 2.963855421686747,
+      "loss_partial": 0.04931640625,
+      "step": 5658
+    },
+    {
+      "epoch": 2.963855421686747,
+      "loss_partial": 0.05712890625,
+      "step": 5658
+    },
+    {
+      "epoch": 2.964379256155055,
+      "grad_norm": 1.7521134614944458,
+      "learning_rate": 3.817321255326574e-06,
+      "loss": 0.0999,
+      "step": 5659
+    },
+    {
+      "epoch": 2.964379256155055,
+      "loss_partial": 0.04638671875,
+      "step": 5659
+    },
+    {
+      "epoch": 2.964379256155055,
+      "loss_partial": 0.053466796875,
+      "step": 5659
+    },
+    {
+      "epoch": 2.964903090623363,
+      "grad_norm": 1.1710323095321655,
+      "learning_rate": 3.8137287851477456e-06,
+      "loss": 0.1041,
+      "step": 5660
+    },
+    {
+      "epoch": 2.964903090623363,
+      "loss_partial": 0.041015625,
+      "step": 5660
+    },
+    {
+      "epoch": 2.964903090623363,
+      "loss_partial": 0.060302734375,
+      "step": 5660
+    },
+    {
+      "epoch": 2.965426925091671,
+      "grad_norm": 1.12892746925354,
+      "learning_rate": 3.810137607861344e-06,
+      "loss": 0.0986,
+      "step": 5661
+    },
+    {
+      "epoch": 2.965426925091671,
+      "loss_partial": 0.058837890625,
+      "step": 5661
+    },
+    {
+      "epoch": 2.965426925091671,
+      "loss_partial": 0.054443359375,
+      "step": 5661
+    },
+    {
+      "epoch": 2.9659507595599788,
+      "grad_norm": 1.4563299417495728,
+      "learning_rate": 3.8065477242179083e-06,
+      "loss": 0.1054,
+      "step": 5662
+    },
+    {
+      "epoch": 2.9659507595599788,
+      "loss_partial": 0.053955078125,
+      "step": 5662
+    },
+    {
+      "epoch": 2.9659507595599788,
+      "loss_partial": 0.050048828125,
+      "step": 5662
+    },
+    {
+      "epoch": 2.966474594028287,
+      "grad_norm": 1.2406679391860962,
+      "learning_rate": 3.8029591349676996e-06,
+      "loss": 0.0992,
+      "step": 5663
+    },
+    {
+      "epoch": 2.966474594028287,
+      "loss_partial": 0.05615234375,
+      "step": 5663
+    },
+    {
+      "epoch": 2.966474594028287,
+      "loss_partial": 0.05224609375,
+      "step": 5663
+    },
+    {
+      "epoch": 2.966998428496595,
+      "grad_norm": 1.4616328477859497,
+      "learning_rate": 3.7993718408607095e-06,
+      "loss": 0.1028,
+      "step": 5664
+    },
+    {
+      "epoch": 2.966998428496595,
+      "loss_partial": 0.04833984375,
+      "step": 5664
+    },
+    {
+      "epoch": 2.966998428496595,
+      "loss_partial": 0.050048828125,
+      "step": 5664
+    },
+    {
+      "epoch": 2.967522262964903,
+      "grad_norm": 1.2117220163345337,
+      "learning_rate": 3.7957858426466587e-06,
+      "loss": 0.1027,
+      "step": 5665
+    },
+    {
+      "epoch": 2.967522262964903,
+      "loss_partial": 0.04931640625,
+      "step": 5665
+    },
+    {
+      "epoch": 2.967522262964903,
+      "loss_partial": 0.041259765625,
+      "step": 5665
+    },
+    {
+      "epoch": 2.968046097433211,
+      "grad_norm": 1.1214725971221924,
+      "learning_rate": 3.7922011410750036e-06,
+      "loss": 0.0976,
+      "step": 5666
+    },
+    {
+      "epoch": 2.968046097433211,
+      "loss_partial": 0.04443359375,
+      "step": 5666
+    },
+    {
+      "epoch": 2.968046097433211,
+      "loss_partial": 0.05224609375,
+      "step": 5666
+    },
+    {
+      "epoch": 2.968569931901519,
+      "grad_norm": 1.363229751586914,
+      "learning_rate": 3.788617736894923e-06,
+      "loss": 0.0998,
+      "step": 5667
+    },
+    {
+      "epoch": 2.968569931901519,
+      "loss_partial": 0.050048828125,
+      "step": 5667
+    },
+    {
+      "epoch": 2.968569931901519,
+      "loss_partial": 0.06494140625,
+      "step": 5667
+    },
+    {
+      "epoch": 2.969093766369827,
+      "grad_norm": 0.8589926958084106,
+      "learning_rate": 3.785035630855325e-06,
+      "loss": 0.1002,
+      "step": 5668
+    },
+    {
+      "epoch": 2.969093766369827,
+      "loss_partial": 0.0576171875,
+      "step": 5668
+    },
+    {
+      "epoch": 2.969093766369827,
+      "loss_partial": 0.047119140625,
+      "step": 5668
+    },
+    {
+      "epoch": 2.969617600838135,
+      "grad_norm": 1.6645866632461548,
+      "learning_rate": 3.7814548237048477e-06,
+      "loss": 0.1013,
+      "step": 5669
+    },
+    {
+      "epoch": 2.969617600838135,
+      "loss_partial": 0.041015625,
+      "step": 5669
+    },
+    {
+      "epoch": 2.969617600838135,
+      "loss_partial": 0.046875,
+      "step": 5669
+    },
+    {
+      "epoch": 2.9701414353064433,
+      "grad_norm": 1.1901668310165405,
+      "learning_rate": 3.777875316191858e-06,
+      "loss": 0.1034,
+      "step": 5670
+    },
+    {
+      "epoch": 2.9701414353064433,
+      "loss_partial": 0.042236328125,
+      "step": 5670
+    },
+    {
+      "epoch": 2.9701414353064433,
+      "loss_partial": 0.052001953125,
+      "step": 5670
+    },
+    {
+      "epoch": 2.9706652697747513,
+      "grad_norm": 1.0543935298919678,
+      "learning_rate": 3.7742971090644507e-06,
+      "loss": 0.0977,
+      "step": 5671
+    },
+    {
+      "epoch": 2.9706652697747513,
+      "loss_partial": 0.048828125,
+      "step": 5671
+    },
+    {
+      "epoch": 2.9706652697747513,
+      "loss_partial": 0.041015625,
+      "step": 5671
+    },
+    {
+      "epoch": 2.9711891042430594,
+      "grad_norm": 1.322086215019226,
+      "learning_rate": 3.7707202030704493e-06,
+      "loss": 0.0973,
+      "step": 5672
+    },
+    {
+      "epoch": 2.9711891042430594,
+      "loss_partial": 0.05029296875,
+      "step": 5672
+    },
+    {
+      "epoch": 2.9711891042430594,
+      "loss_partial": 0.049560546875,
+      "step": 5672
+    },
+    {
+      "epoch": 2.9717129387113674,
+      "grad_norm": 1.1229047775268555,
+      "learning_rate": 3.767144598957402e-06,
+      "loss": 0.1008,
+      "step": 5673
+    },
+    {
+      "epoch": 2.9717129387113674,
+      "loss_partial": 0.048583984375,
+      "step": 5673
+    },
+    {
+      "epoch": 2.9717129387113674,
+      "loss_partial": 0.043701171875,
+      "step": 5673
+    },
+    {
+      "epoch": 2.9722367731796755,
+      "grad_norm": 1.443023443222046,
+      "learning_rate": 3.763570297472593e-06,
+      "loss": 0.0984,
+      "step": 5674
+    },
+    {
+      "epoch": 2.9722367731796755,
+      "loss_partial": 0.043212890625,
+      "step": 5674
+    },
+    {
+      "epoch": 2.9722367731796755,
+      "loss_partial": 0.047119140625,
+      "step": 5674
+    },
+    {
+      "epoch": 2.972760607647983,
+      "grad_norm": 1.2598220109939575,
+      "learning_rate": 3.7599972993630274e-06,
+      "loss": 0.0988,
+      "step": 5675
+    },
+    {
+      "epoch": 2.972760607647983,
+      "loss_partial": 0.048583984375,
+      "step": 5675
+    },
+    {
+      "epoch": 2.972760607647983,
+      "loss_partial": 0.05419921875,
+      "step": 5675
+    },
+    {
+      "epoch": 2.973284442116291,
+      "grad_norm": 1.4470411539077759,
+      "learning_rate": 3.756425605375439e-06,
+      "loss": 0.1013,
+      "step": 5676
+    },
+    {
+      "epoch": 2.973284442116291,
+      "loss_partial": 0.048583984375,
+      "step": 5676
+    },
+    {
+      "epoch": 2.973284442116291,
+      "loss_partial": 0.047607421875,
+      "step": 5676
+    },
+    {
+      "epoch": 2.9738082765845992,
+      "grad_norm": 1.261264681816101,
+      "learning_rate": 3.7528552162562858e-06,
+      "loss": 0.096,
+      "step": 5677
+    },
+    {
+      "epoch": 2.9738082765845992,
+      "loss_partial": 0.04638671875,
+      "step": 5677
+    },
+    {
+      "epoch": 2.9738082765845992,
+      "loss_partial": 0.05029296875,
+      "step": 5677
+    },
+    {
+      "epoch": 2.9743321110529073,
+      "grad_norm": 1.2505122423171997,
+      "learning_rate": 3.749286132751766e-06,
+      "loss": 0.0988,
+      "step": 5678
+    },
+    {
+      "epoch": 2.9743321110529073,
+      "loss_partial": 0.054443359375,
+      "step": 5678
+    },
+    {
+      "epoch": 2.9743321110529073,
+      "loss_partial": 0.0556640625,
+      "step": 5678
+    },
+    {
+      "epoch": 2.9748559455212154,
+      "grad_norm": 1.4673311710357666,
+      "learning_rate": 3.7457183556077904e-06,
+      "loss": 0.0994,
+      "step": 5679
+    },
+    {
+      "epoch": 2.9748559455212154,
+      "loss_partial": 0.04345703125,
+      "step": 5679
+    },
+    {
+      "epoch": 2.9748559455212154,
+      "loss_partial": 0.050537109375,
+      "step": 5679
+    },
+    {
+      "epoch": 2.9753797799895234,
+      "grad_norm": 1.2014025449752808,
+      "learning_rate": 3.7421518855700034e-06,
+      "loss": 0.1013,
+      "step": 5680
+    },
+    {
+      "epoch": 2.9753797799895234,
+      "loss_partial": 0.0478515625,
+      "step": 5680
+    },
+    {
+      "epoch": 2.9753797799895234,
+      "loss_partial": 0.06103515625,
+      "step": 5680
+    },
+    {
+      "epoch": 2.9759036144578315,
+      "grad_norm": 1.4237713813781738,
+      "learning_rate": 3.738586723383775e-06,
+      "loss": 0.0995,
+      "step": 5681
+    },
+    {
+      "epoch": 2.9759036144578315,
+      "loss_partial": 0.04345703125,
+      "step": 5681
+    },
+    {
+      "epoch": 2.9759036144578315,
+      "loss_partial": 0.058349609375,
+      "step": 5681
+    },
+    {
+      "epoch": 2.976427448926139,
+      "grad_norm": 1.409646987915039,
+      "learning_rate": 3.7350228697942027e-06,
+      "loss": 0.1024,
+      "step": 5682
+    },
+    {
+      "epoch": 2.976427448926139,
+      "loss_partial": 0.042236328125,
+      "step": 5682
+    },
+    {
+      "epoch": 2.976427448926139,
+      "loss_partial": 0.051513671875,
+      "step": 5682
+    },
+    {
+      "epoch": 2.976951283394447,
+      "grad_norm": 1.200814962387085,
+      "learning_rate": 3.731460325546108e-06,
+      "loss": 0.0979,
+      "step": 5683
+    },
+    {
+      "epoch": 2.976951283394447,
+      "loss_partial": 0.037353515625,
+      "step": 5683
+    },
+    {
+      "epoch": 2.976951283394447,
+      "loss_partial": 0.05078125,
+      "step": 5683
+    },
+    {
+      "epoch": 2.977475117862755,
+      "grad_norm": 1.510542392730713,
+      "learning_rate": 3.727899091384042e-06,
+      "loss": 0.1003,
+      "step": 5684
+    },
+    {
+      "epoch": 2.977475117862755,
+      "loss_partial": 0.05029296875,
+      "step": 5684
+    },
+    {
+      "epoch": 2.977475117862755,
+      "loss_partial": 0.045166015625,
+      "step": 5684
+    },
+    {
+      "epoch": 2.9779989523310633,
+      "grad_norm": 1.3466780185699463,
+      "learning_rate": 3.724339168052278e-06,
+      "loss": 0.0991,
+      "step": 5685
+    },
+    {
+      "epoch": 2.9779989523310633,
+      "loss_partial": 0.044921875,
+      "step": 5685
+    },
+    {
+      "epoch": 2.9779989523310633,
+      "loss_partial": 0.050048828125,
+      "step": 5685
+    },
+    {
+      "epoch": 2.9785227867993713,
+      "grad_norm": 1.2110824584960938,
+      "learning_rate": 3.720780556294822e-06,
+      "loss": 0.1029,
+      "step": 5686
+    },
+    {
+      "epoch": 2.9785227867993713,
+      "loss_partial": 0.05615234375,
+      "step": 5686
+    },
+    {
+      "epoch": 2.9785227867993713,
+      "loss_partial": 0.048583984375,
+      "step": 5686
+    },
+    {
+      "epoch": 2.9790466212676794,
+      "grad_norm": 1.7660260200500488,
+      "learning_rate": 3.717223256855401e-06,
+      "loss": 0.106,
+      "step": 5687
+    },
+    {
+      "epoch": 2.9790466212676794,
+      "loss_partial": 0.047119140625,
+      "step": 5687
+    },
+    {
+      "epoch": 2.9790466212676794,
+      "loss_partial": 0.0478515625,
+      "step": 5687
+    },
+    {
+      "epoch": 2.9795704557359874,
+      "grad_norm": 1.2298978567123413,
+      "learning_rate": 3.713667270477468e-06,
+      "loss": 0.1016,
+      "step": 5688
+    },
+    {
+      "epoch": 2.9795704557359874,
+      "loss_partial": 0.049560546875,
+      "step": 5688
+    },
+    {
+      "epoch": 2.9795704557359874,
+      "loss_partial": 0.045166015625,
+      "step": 5688
+    },
+    {
+      "epoch": 2.9800942902042955,
+      "grad_norm": 1.421622633934021,
+      "learning_rate": 3.7101125979041985e-06,
+      "loss": 0.1014,
+      "step": 5689
+    },
+    {
+      "epoch": 2.9800942902042955,
+      "loss_partial": 0.0478515625,
+      "step": 5689
+    },
+    {
+      "epoch": 2.9800942902042955,
+      "loss_partial": 0.05078125,
+      "step": 5689
+    },
+    {
+      "epoch": 2.9806181246726036,
+      "grad_norm": 1.439591646194458,
+      "learning_rate": 3.7065592398785043e-06,
+      "loss": 0.1034,
+      "step": 5690
+    },
+    {
+      "epoch": 2.9806181246726036,
+      "loss_partial": 0.048583984375,
+      "step": 5690
+    },
+    {
+      "epoch": 2.9806181246726036,
+      "loss_partial": 0.04638671875,
+      "step": 5690
+    },
+    {
+      "epoch": 2.9811419591409116,
+      "grad_norm": 1.459883689880371,
+      "learning_rate": 3.7030071971430115e-06,
+      "loss": 0.1028,
+      "step": 5691
+    },
+    {
+      "epoch": 2.9811419591409116,
+      "loss_partial": 0.03857421875,
+      "step": 5691
+    },
+    {
+      "epoch": 2.9811419591409116,
+      "loss_partial": 0.047119140625,
+      "step": 5691
+    },
+    {
+      "epoch": 2.9816657936092197,
+      "grad_norm": 1.5149266719818115,
+      "learning_rate": 3.699456470440075e-06,
+      "loss": 0.0951,
+      "step": 5692
+    },
+    {
+      "epoch": 2.9816657936092197,
+      "loss_partial": 0.05419921875,
+      "step": 5692
+    },
+    {
+      "epoch": 2.9816657936092197,
+      "loss_partial": 0.046142578125,
+      "step": 5692
+    },
+    {
+      "epoch": 2.9821896280775277,
+      "grad_norm": 1.2644846439361572,
+      "learning_rate": 3.695907060511774e-06,
+      "loss": 0.1044,
+      "step": 5693
+    },
+    {
+      "epoch": 2.9821896280775277,
+      "loss_partial": 0.048828125,
+      "step": 5693
+    },
+    {
+      "epoch": 2.9821896280775277,
+      "loss_partial": 0.04345703125,
+      "step": 5693
+    },
+    {
+      "epoch": 2.9827134625458354,
+      "grad_norm": 1.6697810888290405,
+      "learning_rate": 3.692358968099915e-06,
+      "loss": 0.0993,
+      "step": 5694
+    },
+    {
+      "epoch": 2.9827134625458354,
+      "loss_partial": 0.038818359375,
+      "step": 5694
+    },
+    {
+      "epoch": 2.9827134625458354,
+      "loss_partial": 0.05712890625,
+      "step": 5694
+    },
+    {
+      "epoch": 2.9832372970141434,
+      "grad_norm": 1.1181693077087402,
+      "learning_rate": 3.6888121939460255e-06,
+      "loss": 0.1035,
+      "step": 5695
+    },
+    {
+      "epoch": 2.9832372970141434,
+      "loss_partial": 0.049072265625,
+      "step": 5695
+    },
+    {
+      "epoch": 2.9832372970141434,
+      "loss_partial": 0.05615234375,
+      "step": 5695
+    },
+    {
+      "epoch": 2.9837611314824515,
+      "grad_norm": 1.4197624921798706,
+      "learning_rate": 3.685266738791361e-06,
+      "loss": 0.1061,
+      "step": 5696
+    },
+    {
+      "epoch": 2.9837611314824515,
+      "loss_partial": 0.048828125,
+      "step": 5696
+    },
+    {
+      "epoch": 2.9837611314824515,
+      "loss_partial": 0.04345703125,
+      "step": 5696
+    },
+    {
+      "epoch": 2.9842849659507595,
+      "grad_norm": 1.2998526096343994,
+      "learning_rate": 3.681722603376898e-06,
+      "loss": 0.1003,
+      "step": 5697
+    },
+    {
+      "epoch": 2.9842849659507595,
+      "loss_partial": 0.046875,
+      "step": 5697
+    },
+    {
+      "epoch": 2.9842849659507595,
+      "loss_partial": 0.0498046875,
+      "step": 5697
+    },
+    {
+      "epoch": 2.9848088004190676,
+      "grad_norm": 1.3897887468338013,
+      "learning_rate": 3.678179788443338e-06,
+      "loss": 0.0975,
+      "step": 5698
+    },
+    {
+      "epoch": 2.9848088004190676,
+      "loss_partial": 0.048828125,
+      "step": 5698
+    },
+    {
+      "epoch": 2.9848088004190676,
+      "loss_partial": 0.048828125,
+      "step": 5698
+    },
+    {
+      "epoch": 2.9853326348873757,
+      "grad_norm": 1.3403854370117188,
+      "learning_rate": 3.674638294731111e-06,
+      "loss": 0.1025,
+      "step": 5699
+    },
+    {
+      "epoch": 2.9853326348873757,
+      "loss_partial": 0.060302734375,
+      "step": 5699
+    },
+    {
+      "epoch": 2.9853326348873757,
+      "loss_partial": 0.0634765625,
+      "step": 5699
+    },
+    {
+      "epoch": 2.9858564693556837,
+      "grad_norm": 2.322143077850342,
+      "learning_rate": 3.6710981229803666e-06,
+      "loss": 0.099,
+      "step": 5700
+    },
+    {
+      "epoch": 2.9858564693556837,
+      "loss_partial": 0.05810546875,
+      "step": 5700
+    },
+    {
+      "epoch": 2.9858564693556837,
+      "loss_partial": 0.05078125,
+      "step": 5700
+    },
+    {
+      "epoch": 2.9863803038239918,
+      "grad_norm": 1.6647944450378418,
+      "learning_rate": 3.6675592739309775e-06,
+      "loss": 0.0999,
+      "step": 5701
+    },
+    {
+      "epoch": 2.9863803038239918,
+      "loss_partial": 0.053466796875,
+      "step": 5701
+    },
+    {
+      "epoch": 2.9863803038239918,
+      "loss_partial": 0.04345703125,
+      "step": 5701
+    },
+    {
+      "epoch": 2.9869041382922994,
+      "grad_norm": 1.1677286624908447,
+      "learning_rate": 3.6640217483225384e-06,
+      "loss": 0.0988,
+      "step": 5702
+    },
+    {
+      "epoch": 2.9869041382922994,
+      "loss_partial": 0.04296875,
+      "step": 5702
+    },
+    {
+      "epoch": 2.9869041382922994,
+      "loss_partial": 0.053955078125,
+      "step": 5702
+    },
+    {
+      "epoch": 2.9874279727606075,
+      "grad_norm": 1.5586947202682495,
+      "learning_rate": 3.6604855468943755e-06,
+      "loss": 0.1035,
+      "step": 5703
+    },
+    {
+      "epoch": 2.9874279727606075,
+      "loss_partial": 0.050048828125,
+      "step": 5703
+    },
+    {
+      "epoch": 2.9874279727606075,
+      "loss_partial": 0.04345703125,
+      "step": 5703
+    },
+    {
+      "epoch": 2.9879518072289155,
+      "grad_norm": 1.5991666316986084,
+      "learning_rate": 3.6569506703855316e-06,
+      "loss": 0.0984,
+      "step": 5704
+    },
+    {
+      "epoch": 2.9879518072289155,
+      "loss_partial": 0.045166015625,
+      "step": 5704
+    },
+    {
+      "epoch": 2.9879518072289155,
+      "loss_partial": 0.040283203125,
+      "step": 5704
+    },
+    {
+      "epoch": 2.9884756416972236,
+      "grad_norm": 1.4287947416305542,
+      "learning_rate": 3.6534171195347747e-06,
+      "loss": 0.0984,
+      "step": 5705
+    },
+    {
+      "epoch": 2.9884756416972236,
+      "loss_partial": 0.050537109375,
+      "step": 5705
+    },
+    {
+      "epoch": 2.9884756416972236,
+      "loss_partial": 0.053466796875,
+      "step": 5705
+    },
+    {
+      "epoch": 2.9889994761655316,
+      "grad_norm": 1.621971845626831,
+      "learning_rate": 3.649884895080592e-06,
+      "loss": 0.0973,
+      "step": 5706
+    },
+    {
+      "epoch": 2.9889994761655316,
+      "loss_partial": 0.045654296875,
+      "step": 5706
+    },
+    {
+      "epoch": 2.9889994761655316,
+      "loss_partial": 0.048583984375,
+      "step": 5706
+    },
+    {
+      "epoch": 2.9895233106338397,
+      "grad_norm": 1.1917537450790405,
+      "learning_rate": 3.6463539977611995e-06,
+      "loss": 0.0978,
+      "step": 5707
+    },
+    {
+      "epoch": 2.9895233106338397,
+      "loss_partial": 0.058837890625,
+      "step": 5707
+    },
+    {
+      "epoch": 2.9895233106338397,
+      "loss_partial": 0.051513671875,
+      "step": 5707
+    },
+    {
+      "epoch": 2.9900471451021478,
+      "grad_norm": 1.2623904943466187,
+      "learning_rate": 3.642824428314531e-06,
+      "loss": 0.1011,
+      "step": 5708
+    },
+    {
+      "epoch": 2.9900471451021478,
+      "loss_partial": 0.05126953125,
+      "step": 5708
+    },
+    {
+      "epoch": 2.9900471451021478,
+      "loss_partial": 0.046142578125,
+      "step": 5708
+    },
+    {
+      "epoch": 2.990570979570456,
+      "grad_norm": 1.2252347469329834,
+      "learning_rate": 3.6392961874782453e-06,
+      "loss": 0.0999,
+      "step": 5709
+    },
+    {
+      "epoch": 2.990570979570456,
+      "loss_partial": 0.042236328125,
+      "step": 5709
+    },
+    {
+      "epoch": 2.990570979570456,
+      "loss_partial": 0.056884765625,
+      "step": 5709
+    },
+    {
+      "epoch": 2.991094814038764,
+      "grad_norm": 1.5010716915130615,
+      "learning_rate": 3.63576927598972e-06,
+      "loss": 0.101,
+      "step": 5710
+    },
+    {
+      "epoch": 2.991094814038764,
+      "loss_partial": 0.051513671875,
+      "step": 5710
+    },
+    {
+      "epoch": 2.991094814038764,
+      "loss_partial": 0.052001953125,
+      "step": 5710
+    },
+    {
+      "epoch": 2.991618648507072,
+      "grad_norm": 1.5358468294143677,
+      "learning_rate": 3.632243694586064e-06,
+      "loss": 0.0969,
+      "step": 5711
+    },
+    {
+      "epoch": 2.991618648507072,
+      "loss_partial": 0.0458984375,
+      "step": 5711
+    },
+    {
+      "epoch": 2.991618648507072,
+      "loss_partial": 0.045654296875,
+      "step": 5711
+    },
+    {
+      "epoch": 2.99214248297538,
+      "grad_norm": 1.600608229637146,
+      "learning_rate": 3.6287194440040987e-06,
+      "loss": 0.0942,
+      "step": 5712
+    },
+    {
+      "epoch": 2.99214248297538,
+      "loss_partial": 0.051025390625,
+      "step": 5712
+    },
+    {
+      "epoch": 2.99214248297538,
+      "loss_partial": 0.044189453125,
+      "step": 5712
+    },
+    {
+      "epoch": 2.992666317443688,
+      "grad_norm": 1.2884390354156494,
+      "learning_rate": 3.6251965249803713e-06,
+      "loss": 0.1003,
+      "step": 5713
+    },
+    {
+      "epoch": 2.992666317443688,
+      "loss_partial": 0.040771484375,
+      "step": 5713
+    },
+    {
+      "epoch": 2.992666317443688,
+      "loss_partial": 0.0556640625,
+      "step": 5713
+    },
+    {
+      "epoch": 2.9931901519119957,
+      "grad_norm": 1.6168925762176514,
+      "learning_rate": 3.621674938251145e-06,
+      "loss": 0.1,
+      "step": 5714
+    },
+    {
+      "epoch": 2.9931901519119957,
+      "loss_partial": 0.0556640625,
+      "step": 5714
+    },
+    {
+      "epoch": 2.9931901519119957,
+      "loss_partial": 0.05908203125,
+      "step": 5714
+    },
+    {
+      "epoch": 2.9937139863803037,
+      "grad_norm": 1.710491418838501,
+      "learning_rate": 3.6181546845524187e-06,
+      "loss": 0.1018,
+      "step": 5715
+    },
+    {
+      "epoch": 2.9937139863803037,
+      "loss_partial": 0.048095703125,
+      "step": 5715
+    },
+    {
+      "epoch": 2.9937139863803037,
+      "loss_partial": 0.0419921875,
+      "step": 5715
+    },
+    {
+      "epoch": 2.994237820848612,
+      "grad_norm": 1.7324162721633911,
+      "learning_rate": 3.614635764619899e-06,
+      "loss": 0.1024,
+      "step": 5716
+    },
+    {
+      "epoch": 2.994237820848612,
+      "loss_partial": 0.05224609375,
+      "step": 5716
+    },
+    {
+      "epoch": 2.994237820848612,
+      "loss_partial": 0.0498046875,
+      "step": 5716
+    },
+    {
+      "epoch": 2.99476165531692,
+      "grad_norm": 1.4301984310150146,
+      "learning_rate": 3.6111181791890184e-06,
+      "loss": 0.0989,
+      "step": 5717
+    },
+    {
+      "epoch": 2.99476165531692,
+      "loss_partial": 0.04833984375,
+      "step": 5717
+    },
+    {
+      "epoch": 2.99476165531692,
+      "loss_partial": 0.0556640625,
+      "step": 5717
+    },
+    {
+      "epoch": 2.995285489785228,
+      "grad_norm": 1.5845729112625122,
+      "learning_rate": 3.6076019289949315e-06,
+      "loss": 0.0986,
+      "step": 5718
+    },
+    {
+      "epoch": 2.995285489785228,
+      "loss_partial": 0.04248046875,
+      "step": 5718
+    },
+    {
+      "epoch": 2.995285489785228,
+      "loss_partial": 0.0517578125,
+      "step": 5718
+    },
+    {
+      "epoch": 2.995809324253536,
+      "grad_norm": 1.3502379655838013,
+      "learning_rate": 3.6040870147725106e-06,
+      "loss": 0.0994,
+      "step": 5719
+    },
+    {
+      "epoch": 2.995809324253536,
+      "loss_partial": 0.042724609375,
+      "step": 5719
+    },
+    {
+      "epoch": 2.995809324253536,
+      "loss_partial": 0.050048828125,
+      "step": 5719
+    },
+    {
+      "epoch": 2.996333158721844,
+      "grad_norm": 1.4938329458236694,
+      "learning_rate": 3.600573437256354e-06,
+      "loss": 0.1002,
+      "step": 5720
+    },
+    {
+      "epoch": 2.996333158721844,
+      "loss_partial": 0.04931640625,
+      "step": 5720
+    },
+    {
+      "epoch": 2.996333158721844,
+      "loss_partial": 0.051513671875,
+      "step": 5720
+    },
+    {
+      "epoch": 2.9968569931901516,
+      "grad_norm": 1.5139473676681519,
+      "learning_rate": 3.5970611971807746e-06,
+      "loss": 0.1004,
+      "step": 5721
+    },
+    {
+      "epoch": 2.9968569931901516,
+      "loss_partial": 0.0439453125,
+      "step": 5721
+    },
+    {
+      "epoch": 2.9968569931901516,
+      "loss_partial": 0.04443359375,
+      "step": 5721
+    },
+    {
+      "epoch": 2.9973808276584597,
+      "grad_norm": 1.258803367614746,
+      "learning_rate": 3.5935502952798084e-06,
+      "loss": 0.0987,
+      "step": 5722
+    },
+    {
+      "epoch": 2.9973808276584597,
+      "loss_partial": 0.044921875,
+      "step": 5722
+    },
+    {
+      "epoch": 2.9973808276584597,
+      "loss_partial": 0.048828125,
+      "step": 5722
+    },
+    {
+      "epoch": 2.9979046621267678,
+      "grad_norm": 1.587805986404419,
+      "learning_rate": 3.5900407322872165e-06,
+      "loss": 0.0979,
+      "step": 5723
+    },
+    {
+      "epoch": 2.9979046621267678,
+      "loss_partial": 0.050048828125,
+      "step": 5723
+    },
+    {
+      "epoch": 2.9979046621267678,
+      "loss_partial": 0.0517578125,
+      "step": 5723
+    },
+    {
+      "epoch": 2.998428496595076,
+      "grad_norm": 1.617457389831543,
+      "learning_rate": 3.5865325089364746e-06,
+      "loss": 0.1017,
+      "step": 5724
+    },
+    {
+      "epoch": 2.998428496595076,
+      "loss_partial": 0.048828125,
+      "step": 5724
+    },
+    {
+      "epoch": 2.998428496595076,
+      "loss_partial": 0.05712890625,
+      "step": 5724
+    },
+    {
+      "epoch": 2.998952331063384,
+      "grad_norm": 1.5532355308532715,
+      "learning_rate": 3.5830256259607775e-06,
+      "loss": 0.1046,
+      "step": 5725
+    },
+    {
+      "epoch": 2.998952331063384,
+      "loss_partial": 0.052978515625,
+      "step": 5725
+    },
+    {
+      "epoch": 2.998952331063384,
+      "loss_partial": 0.0419921875,
+      "step": 5725
+    },
+    {
+      "epoch": 2.999476165531692,
+      "grad_norm": 1.3971742391586304,
+      "learning_rate": 3.5795200840930413e-06,
+      "loss": 0.1,
+      "step": 5726
+    },
+    {
+      "epoch": 2.999476165531692,
+      "loss_partial": 0.046630859375,
+      "step": 5726
+    },
+    {
+      "epoch": 2.999476165531692,
+      "loss_partial": 0.046142578125,
+      "step": 5726
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.4349571466445923,
+      "learning_rate": 3.5760158840659065e-06,
+      "loss": 0.0997,
+      "step": 5727
+    },
+    {
+      "epoch": 3.0,
+      "loss_partial": 0.04248046875,
+      "step": 5727
+    },
+    {
+      "epoch": 3.0,
+      "loss_partial": 0.06201171875,
+      "step": 5727
+    },
+    {
+      "epoch": 3.000523834468308,
+      "grad_norm": 1.4456571340560913,
+      "learning_rate": 3.572513026611727e-06,
+      "loss": 0.1046,
+      "step": 5728
+    },
+    {
+      "epoch": 3.000523834468308,
+      "loss_partial": 0.0458984375,
+      "step": 5728
+    },
+    {
+      "epoch": 3.000523834468308,
+      "loss_partial": 0.049560546875,
+      "step": 5728
+    },
+    {
+      "epoch": 3.001047668936616,
+      "grad_norm": 1.4295741319656372,
+      "learning_rate": 3.5690115124625802e-06,
+      "loss": 0.1001,
+      "step": 5729
+    },
+    {
+      "epoch": 3.001047668936616,
+      "loss_partial": 0.045166015625,
+      "step": 5729
+    },
+    {
+      "epoch": 3.001047668936616,
+      "loss_partial": 0.0537109375,
+      "step": 5729
+    },
+    {
+      "epoch": 3.001571503404924,
+      "grad_norm": 2.1191720962524414,
+      "learning_rate": 3.5655113423502573e-06,
+      "loss": 0.0987,
+      "step": 5730
+    },
+    {
+      "epoch": 3.001571503404924,
+      "loss_partial": 0.048828125,
+      "step": 5730
+    },
+    {
+      "epoch": 3.001571503404924,
+      "loss_partial": 0.051513671875,
+      "step": 5730
+    },
+    {
+      "epoch": 3.0020953378732322,
+      "grad_norm": 1.0702712535858154,
+      "learning_rate": 3.562012517006276e-06,
+      "loss": 0.0991,
+      "step": 5731
+    },
+    {
+      "epoch": 3.0020953378732322,
+      "loss_partial": 0.048583984375,
+      "step": 5731
+    },
+    {
+      "epoch": 3.0020953378732322,
+      "loss_partial": 0.046875,
+      "step": 5731
+    },
+    {
+      "epoch": 3.00261917234154,
+      "grad_norm": 1.264752745628357,
+      "learning_rate": 3.558515037161866e-06,
+      "loss": 0.1014,
+      "step": 5732
+    },
+    {
+      "epoch": 3.00261917234154,
+      "loss_partial": 0.0517578125,
+      "step": 5732
+    },
+    {
+      "epoch": 3.00261917234154,
+      "loss_partial": 0.050537109375,
+      "step": 5732
+    },
+    {
+      "epoch": 3.003143006809848,
+      "grad_norm": 1.6320362091064453,
+      "learning_rate": 3.555018903547981e-06,
+      "loss": 0.1009,
+      "step": 5733
+    },
+    {
+      "epoch": 3.003143006809848,
+      "loss_partial": 0.05078125,
+      "step": 5733
+    },
+    {
+      "epoch": 3.003143006809848,
+      "loss_partial": 0.046630859375,
+      "step": 5733
+    },
+    {
+      "epoch": 3.003666841278156,
+      "grad_norm": 1.291682481765747,
+      "learning_rate": 3.5515241168952863e-06,
+      "loss": 0.1015,
+      "step": 5734
+    },
+    {
+      "epoch": 3.003666841278156,
+      "loss_partial": 0.0537109375,
+      "step": 5734
+    },
+    {
+      "epoch": 3.003666841278156,
+      "loss_partial": 0.05908203125,
+      "step": 5734
+    },
+    {
+      "epoch": 3.004190675746464,
+      "grad_norm": 1.5923904180526733,
+      "learning_rate": 3.54803067793418e-06,
+      "loss": 0.1037,
+      "step": 5735
+    },
+    {
+      "epoch": 3.004190675746464,
+      "loss_partial": 0.04931640625,
+      "step": 5735
+    },
+    {
+      "epoch": 3.004190675746464,
+      "loss_partial": 0.05224609375,
+      "step": 5735
+    },
+    {
+      "epoch": 3.004714510214772,
+      "grad_norm": 1.313253402709961,
+      "learning_rate": 3.5445385873947623e-06,
+      "loss": 0.0951,
+      "step": 5736
+    },
+    {
+      "epoch": 3.004714510214772,
+      "loss_partial": 0.0458984375,
+      "step": 5736
+    },
+    {
+      "epoch": 3.004714510214772,
+      "loss_partial": 0.06689453125,
+      "step": 5736
+    },
+    {
+      "epoch": 3.00523834468308,
+      "grad_norm": 1.390165090560913,
+      "learning_rate": 3.541047846006862e-06,
+      "loss": 0.1017,
+      "step": 5737
+    },
+    {
+      "epoch": 3.00523834468308,
+      "loss_partial": 0.05322265625,
+      "step": 5737
+    },
+    {
+      "epoch": 3.00523834468308,
+      "loss_partial": 0.05517578125,
+      "step": 5737
+    },
+    {
+      "epoch": 3.005762179151388,
+      "grad_norm": 1.463139533996582,
+      "learning_rate": 3.5375584545000176e-06,
+      "loss": 0.0999,
+      "step": 5738
+    },
+    {
+      "epoch": 3.005762179151388,
+      "loss_partial": 0.056640625,
+      "step": 5738
+    },
+    {
+      "epoch": 3.005762179151388,
+      "loss_partial": 0.046142578125,
+      "step": 5738
+    },
+    {
+      "epoch": 3.0062860136196963,
+      "grad_norm": 1.3433510065078735,
+      "learning_rate": 3.534070413603495e-06,
+      "loss": 0.101,
+      "step": 5739
+    },
+    {
+      "epoch": 3.0062860136196963,
+      "loss_partial": 0.06298828125,
+      "step": 5739
+    },
+    {
+      "epoch": 3.0062860136196963,
+      "loss_partial": 0.044677734375,
+      "step": 5739
+    },
+    {
+      "epoch": 3.0068098480880043,
+      "grad_norm": 1.5008275508880615,
+      "learning_rate": 3.530583724046274e-06,
+      "loss": 0.1034,
+      "step": 5740
+    },
+    {
+      "epoch": 3.0068098480880043,
+      "loss_partial": 0.04052734375,
+      "step": 5740
+    },
+    {
+      "epoch": 3.0068098480880043,
+      "loss_partial": 0.043701171875,
+      "step": 5740
+    },
+    {
+      "epoch": 3.0073336825563124,
+      "grad_norm": 1.8957042694091797,
+      "learning_rate": 3.5270983865570486e-06,
+      "loss": 0.0978,
+      "step": 5741
+    },
+    {
+      "epoch": 3.0073336825563124,
+      "loss_partial": 0.04541015625,
+      "step": 5741
+    },
+    {
+      "epoch": 3.0073336825563124,
+      "loss_partial": 0.04833984375,
+      "step": 5741
+    },
+    {
+      "epoch": 3.00785751702462,
+      "grad_norm": 1.1544289588928223,
+      "learning_rate": 3.5236144018642324e-06,
+      "loss": 0.1019,
+      "step": 5742
+    },
+    {
+      "epoch": 3.00785751702462,
+      "loss_partial": 0.058837890625,
+      "step": 5742
+    },
+    {
+      "epoch": 3.00785751702462,
+      "loss_partial": 0.046630859375,
+      "step": 5742
+    },
+    {
+      "epoch": 3.008381351492928,
+      "grad_norm": 1.3608770370483398,
+      "learning_rate": 3.520131770695958e-06,
+      "loss": 0.1004,
+      "step": 5743
+    },
+    {
+      "epoch": 3.008381351492928,
+      "loss_partial": 0.041748046875,
+      "step": 5743
+    },
+    {
+      "epoch": 3.008381351492928,
+      "loss_partial": 0.04931640625,
+      "step": 5743
+    },
+    {
+      "epoch": 3.008905185961236,
+      "grad_norm": 1.4661916494369507,
+      "learning_rate": 3.5166504937800737e-06,
+      "loss": 0.0985,
+      "step": 5744
+    },
+    {
+      "epoch": 3.008905185961236,
+      "loss_partial": 0.0439453125,
+      "step": 5744
+    },
+    {
+      "epoch": 3.008905185961236,
+      "loss_partial": 0.06103515625,
+      "step": 5744
+    },
+    {
+      "epoch": 3.009429020429544,
+      "grad_norm": 1.1962473392486572,
+      "learning_rate": 3.5131705718441435e-06,
+      "loss": 0.1015,
+      "step": 5745
+    },
+    {
+      "epoch": 3.009429020429544,
+      "loss_partial": 0.04638671875,
+      "step": 5745
+    },
+    {
+      "epoch": 3.009429020429544,
+      "loss_partial": 0.048828125,
+      "step": 5745
+    },
+    {
+      "epoch": 3.0099528548978522,
+      "grad_norm": 1.3638882637023926,
+      "learning_rate": 3.509692005615447e-06,
+      "loss": 0.0976,
+      "step": 5746
+    },
+    {
+      "epoch": 3.0099528548978522,
+      "loss_partial": 0.048583984375,
+      "step": 5746
+    },
+    {
+      "epoch": 3.0099528548978522,
+      "loss_partial": 0.048095703125,
+      "step": 5746
+    },
+    {
+      "epoch": 3.0104766893661603,
+      "grad_norm": 1.0750157833099365,
+      "learning_rate": 3.5062147958209913e-06,
+      "loss": 0.1016,
+      "step": 5747
+    },
+    {
+      "epoch": 3.0104766893661603,
+      "loss_partial": 0.044921875,
+      "step": 5747
+    },
+    {
+      "epoch": 3.0104766893661603,
+      "loss_partial": 0.043212890625,
+      "step": 5747
+    },
+    {
+      "epoch": 3.0110005238344684,
+      "grad_norm": 1.6013003587722778,
+      "learning_rate": 3.502738943187486e-06,
+      "loss": 0.1003,
+      "step": 5748
+    },
+    {
+      "epoch": 3.0110005238344684,
+      "loss_partial": 0.047119140625,
+      "step": 5748
+    },
+    {
+      "epoch": 3.0110005238344684,
+      "loss_partial": 0.046142578125,
+      "step": 5748
+    },
+    {
+      "epoch": 3.0115243583027764,
+      "grad_norm": 1.5218666791915894,
+      "learning_rate": 3.499264448441363e-06,
+      "loss": 0.0992,
+      "step": 5749
+    },
+    {
+      "epoch": 3.0115243583027764,
+      "loss_partial": 0.048828125,
+      "step": 5749
+    },
+    {
+      "epoch": 3.0115243583027764,
+      "loss_partial": 0.046630859375,
+      "step": 5749
+    },
+    {
+      "epoch": 3.0120481927710845,
+      "grad_norm": 1.4683729410171509,
+      "learning_rate": 3.4957913123087718e-06,
+      "loss": 0.0988,
+      "step": 5750
+    },
+    {
+      "epoch": 3.0120481927710845,
+      "loss_partial": 0.046875,
+      "step": 5750
+    },
+    {
+      "epoch": 3.0120481927710845,
+      "loss_partial": 0.044921875,
+      "step": 5750
+    },
+    {
+      "epoch": 3.0125720272393925,
+      "grad_norm": 0.9339583516120911,
+      "learning_rate": 3.4923195355155714e-06,
+      "loss": 0.101,
+      "step": 5751
+    },
+    {
+      "epoch": 3.0125720272393925,
+      "loss_partial": 0.047607421875,
+      "step": 5751
+    },
+    {
+      "epoch": 3.0125720272393925,
+      "loss_partial": 0.05029296875,
+      "step": 5751
+    },
+    {
+      "epoch": 3.0130958617077,
+      "grad_norm": 1.4114975929260254,
+      "learning_rate": 3.4888491187873498e-06,
+      "loss": 0.1006,
+      "step": 5752
+    },
+    {
+      "epoch": 3.0130958617077,
+      "loss_partial": 0.059814453125,
+      "step": 5752
+    },
+    {
+      "epoch": 3.0130958617077,
+      "loss_partial": 0.043701171875,
+      "step": 5752
+    },
+    {
+      "epoch": 3.0136196961760082,
+      "grad_norm": 0.9463948607444763,
+      "learning_rate": 3.4853800628493983e-06,
+      "loss": 0.0996,
+      "step": 5753
+    },
+    {
+      "epoch": 3.0136196961760082,
+      "loss_partial": 0.0625,
+      "step": 5753
+    },
+    {
+      "epoch": 3.0136196961760082,
+      "loss_partial": 0.04736328125,
+      "step": 5753
+    },
+    {
+      "epoch": 3.0141435306443163,
+      "grad_norm": 1.118863582611084,
+      "learning_rate": 3.4819123684267285e-06,
+      "loss": 0.0999,
+      "step": 5754
+    },
+    {
+      "epoch": 3.0141435306443163,
+      "loss_partial": 0.0439453125,
+      "step": 5754
+    },
+    {
+      "epoch": 3.0141435306443163,
+      "loss_partial": 0.047119140625,
+      "step": 5754
+    },
+    {
+      "epoch": 3.0146673651126243,
+      "grad_norm": 1.3093836307525635,
+      "learning_rate": 3.4784460362440676e-06,
+      "loss": 0.1002,
+      "step": 5755
+    },
+    {
+      "epoch": 3.0146673651126243,
+      "loss_partial": 0.051025390625,
+      "step": 5755
+    },
+    {
+      "epoch": 3.0146673651126243,
+      "loss_partial": 0.044677734375,
+      "step": 5755
+    },
+    {
+      "epoch": 3.0151911995809324,
+      "grad_norm": 1.2284287214279175,
+      "learning_rate": 3.4749810670258568e-06,
+      "loss": 0.0979,
+      "step": 5756
+    },
+    {
+      "epoch": 3.0151911995809324,
+      "loss_partial": 0.052490234375,
+      "step": 5756
+    },
+    {
+      "epoch": 3.0151911995809324,
+      "loss_partial": 0.04833984375,
+      "step": 5756
+    },
+    {
+      "epoch": 3.0157150340492405,
+      "grad_norm": 1.2235424518585205,
+      "learning_rate": 3.471517461496253e-06,
+      "loss": 0.1015,
+      "step": 5757
+    },
+    {
+      "epoch": 3.0157150340492405,
+      "loss_partial": 0.046630859375,
+      "step": 5757
+    },
+    {
+      "epoch": 3.0157150340492405,
+      "loss_partial": 0.05126953125,
+      "step": 5757
+    },
+    {
+      "epoch": 3.0162388685175485,
+      "grad_norm": 1.5211739540100098,
+      "learning_rate": 3.4680552203791295e-06,
+      "loss": 0.1008,
+      "step": 5758
+    },
+    {
+      "epoch": 3.0162388685175485,
+      "loss_partial": 0.043212890625,
+      "step": 5758
+    },
+    {
+      "epoch": 3.0162388685175485,
+      "loss_partial": 0.05126953125,
+      "step": 5758
+    },
+    {
+      "epoch": 3.0167627029858566,
+      "grad_norm": 1.804889440536499,
+      "learning_rate": 3.464594344398069e-06,
+      "loss": 0.0985,
+      "step": 5759
+    },
+    {
+      "epoch": 3.0167627029858566,
+      "loss_partial": 0.056396484375,
+      "step": 5759
+    },
+    {
+      "epoch": 3.0167627029858566,
+      "loss_partial": 0.05419921875,
+      "step": 5759
+    },
+    {
+      "epoch": 3.0172865374541646,
+      "grad_norm": 1.455479621887207,
+      "learning_rate": 3.4611348342763794e-06,
+      "loss": 0.1005,
+      "step": 5760
+    },
+    {
+      "epoch": 3.0172865374541646,
+      "loss_partial": 0.0673828125,
+      "step": 5760
+    },
+    {
+      "epoch": 3.0172865374541646,
+      "loss_partial": 0.052734375,
+      "step": 5760
+    },
+    {
+      "epoch": 3.0178103719224727,
+      "grad_norm": 1.203415036201477,
+      "learning_rate": 3.4576766907370752e-06,
+      "loss": 0.1021,
+      "step": 5761
+    },
+    {
+      "epoch": 3.0178103719224727,
+      "loss_partial": 0.05126953125,
+      "step": 5761
+    },
+    {
+      "epoch": 3.0178103719224727,
+      "loss_partial": 0.041015625,
+      "step": 5761
+    },
+    {
+      "epoch": 3.0183342063907803,
+      "grad_norm": 1.5330480337142944,
+      "learning_rate": 3.4542199145028855e-06,
+      "loss": 0.0975,
+      "step": 5762
+    },
+    {
+      "epoch": 3.0183342063907803,
+      "loss_partial": 0.04541015625,
+      "step": 5762
+    },
+    {
+      "epoch": 3.0183342063907803,
+      "loss_partial": 0.046875,
+      "step": 5762
+    },
+    {
+      "epoch": 3.0188580408590884,
+      "grad_norm": 1.2708359956741333,
+      "learning_rate": 3.4507645062962534e-06,
+      "loss": 0.099,
+      "step": 5763
+    },
+    {
+      "epoch": 3.0188580408590884,
+      "loss_partial": 0.048583984375,
+      "step": 5763
+    },
+    {
+      "epoch": 3.0188580408590884,
+      "loss_partial": 0.056640625,
+      "step": 5763
+    },
+    {
+      "epoch": 3.0193818753273964,
+      "grad_norm": 1.4723576307296753,
+      "learning_rate": 3.4473104668393432e-06,
+      "loss": 0.0991,
+      "step": 5764
+    },
+    {
+      "epoch": 3.0193818753273964,
+      "loss_partial": 0.046142578125,
+      "step": 5764
+    },
+    {
+      "epoch": 3.0193818753273964,
+      "loss_partial": 0.0498046875,
+      "step": 5764
+    },
+    {
+      "epoch": 3.0199057097957045,
+      "grad_norm": 1.2517799139022827,
+      "learning_rate": 3.4438577968540243e-06,
+      "loss": 0.1008,
+      "step": 5765
+    },
+    {
+      "epoch": 3.0199057097957045,
+      "loss_partial": 0.0546875,
+      "step": 5765
+    },
+    {
+      "epoch": 3.0199057097957045,
+      "loss_partial": 0.05908203125,
+      "step": 5765
+    },
+    {
+      "epoch": 3.0204295442640126,
+      "grad_norm": 1.7186412811279297,
+      "learning_rate": 3.4404064970618845e-06,
+      "loss": 0.0999,
+      "step": 5766
+    },
+    {
+      "epoch": 3.0204295442640126,
+      "loss_partial": 0.039794921875,
+      "step": 5766
+    },
+    {
+      "epoch": 3.0204295442640126,
+      "loss_partial": 0.0517578125,
+      "step": 5766
+    },
+    {
+      "epoch": 3.0209533787323206,
+      "grad_norm": 1.5839005708694458,
+      "learning_rate": 3.436956568184223e-06,
+      "loss": 0.1048,
+      "step": 5767
+    },
+    {
+      "epoch": 3.0209533787323206,
+      "loss_partial": 0.05419921875,
+      "step": 5767
+    },
+    {
+      "epoch": 3.0209533787323206,
+      "loss_partial": 0.053955078125,
+      "step": 5767
+    },
+    {
+      "epoch": 3.0214772132006287,
+      "grad_norm": 1.5450721979141235,
+      "learning_rate": 3.433508010942054e-06,
+      "loss": 0.1041,
+      "step": 5768
+    },
+    {
+      "epoch": 3.0214772132006287,
+      "loss_partial": 0.046875,
+      "step": 5768
+    },
+    {
+      "epoch": 3.0214772132006287,
+      "loss_partial": 0.037841796875,
+      "step": 5768
+    },
+    {
+      "epoch": 3.0220010476689367,
+      "grad_norm": 1.6290931701660156,
+      "learning_rate": 3.4300608260561053e-06,
+      "loss": 0.1007,
+      "step": 5769
+    },
+    {
+      "epoch": 3.0220010476689367,
+      "loss_partial": 0.04931640625,
+      "step": 5769
+    },
+    {
+      "epoch": 3.0220010476689367,
+      "loss_partial": 0.056396484375,
+      "step": 5769
+    },
+    {
+      "epoch": 3.022524882137245,
+      "grad_norm": 1.461418867111206,
+      "learning_rate": 3.426615014246817e-06,
+      "loss": 0.0977,
+      "step": 5770
+    },
+    {
+      "epoch": 3.022524882137245,
+      "loss_partial": 0.060546875,
+      "step": 5770
+    },
+    {
+      "epoch": 3.022524882137245,
+      "loss_partial": 0.05908203125,
+      "step": 5770
+    },
+    {
+      "epoch": 3.023048716605553,
+      "grad_norm": 1.535744309425354,
+      "learning_rate": 3.423170576234338e-06,
+      "loss": 0.1033,
+      "step": 5771
+    },
+    {
+      "epoch": 3.023048716605553,
+      "loss_partial": 0.049072265625,
+      "step": 5771
+    },
+    {
+      "epoch": 3.023048716605553,
+      "loss_partial": 0.0458984375,
+      "step": 5771
+    },
+    {
+      "epoch": 3.0235725510738605,
+      "grad_norm": 1.7868789434432983,
+      "learning_rate": 3.419727512738542e-06,
+      "loss": 0.0977,
+      "step": 5772
+    },
+    {
+      "epoch": 3.0235725510738605,
+      "loss_partial": 0.04833984375,
+      "step": 5772
+    },
+    {
+      "epoch": 3.0235725510738605,
+      "loss_partial": 0.041748046875,
+      "step": 5772
+    },
+    {
+      "epoch": 3.0240963855421685,
+      "grad_norm": 1.516769528388977,
+      "learning_rate": 3.4162858244790044e-06,
+      "loss": 0.1,
+      "step": 5773
+    },
+    {
+      "epoch": 3.0240963855421685,
+      "loss_partial": 0.043212890625,
+      "step": 5773
+    },
+    {
+      "epoch": 3.0240963855421685,
+      "loss_partial": 0.0439453125,
+      "step": 5773
+    },
+    {
+      "epoch": 3.0246202200104766,
+      "grad_norm": 1.2680178880691528,
+      "learning_rate": 3.4128455121750183e-06,
+      "loss": 0.0957,
+      "step": 5774
+    },
+    {
+      "epoch": 3.0246202200104766,
+      "loss_partial": 0.061279296875,
+      "step": 5774
+    },
+    {
+      "epoch": 3.0246202200104766,
+      "loss_partial": 0.049072265625,
+      "step": 5774
+    },
+    {
+      "epoch": 3.0251440544787846,
+      "grad_norm": 1.5561717748641968,
+      "learning_rate": 3.409406576545582e-06,
+      "loss": 0.1029,
+      "step": 5775
+    },
+    {
+      "epoch": 3.0251440544787846,
+      "loss_partial": 0.051513671875,
+      "step": 5775
+    },
+    {
+      "epoch": 3.0251440544787846,
+      "loss_partial": 0.048828125,
+      "step": 5775
+    },
+    {
+      "epoch": 3.0256678889470927,
+      "grad_norm": 1.6042327880859375,
+      "learning_rate": 3.4059690183094196e-06,
+      "loss": 0.0981,
+      "step": 5776
+    },
+    {
+      "epoch": 3.0256678889470927,
+      "loss_partial": 0.0419921875,
+      "step": 5776
+    },
+    {
+      "epoch": 3.0256678889470927,
+      "loss_partial": 0.0546875,
+      "step": 5776
+    },
+    {
+      "epoch": 3.0261917234154008,
+      "grad_norm": 1.3215075731277466,
+      "learning_rate": 3.4025328381849564e-06,
+      "loss": 0.1028,
+      "step": 5777
+    },
+    {
+      "epoch": 3.0261917234154008,
+      "loss_partial": 0.047607421875,
+      "step": 5777
+    },
+    {
+      "epoch": 3.0261917234154008,
+      "loss_partial": 0.061279296875,
+      "step": 5777
+    },
+    {
+      "epoch": 3.026715557883709,
+      "grad_norm": 1.4443191289901733,
+      "learning_rate": 3.3990980368903327e-06,
+      "loss": 0.0996,
+      "step": 5778
+    },
+    {
+      "epoch": 3.026715557883709,
+      "loss_partial": 0.047119140625,
+      "step": 5778
+    },
+    {
+      "epoch": 3.026715557883709,
+      "loss_partial": 0.042724609375,
+      "step": 5778
+    },
+    {
+      "epoch": 3.027239392352017,
+      "grad_norm": 1.2646886110305786,
+      "learning_rate": 3.3956646151434e-06,
+      "loss": 0.0975,
+      "step": 5779
+    },
+    {
+      "epoch": 3.027239392352017,
+      "loss_partial": 0.053466796875,
+      "step": 5779
+    },
+    {
+      "epoch": 3.027239392352017,
+      "loss_partial": 0.052001953125,
+      "step": 5779
+    },
+    {
+      "epoch": 3.027763226820325,
+      "grad_norm": 1.4933665990829468,
+      "learning_rate": 3.3922325736617235e-06,
+      "loss": 0.0994,
+      "step": 5780
+    },
+    {
+      "epoch": 3.027763226820325,
+      "loss_partial": 0.046630859375,
+      "step": 5780
+    },
+    {
+      "epoch": 3.027763226820325,
+      "loss_partial": 0.049072265625,
+      "step": 5780
+    },
+    {
+      "epoch": 3.028287061288633,
+      "grad_norm": 1.2959725856781006,
+      "learning_rate": 3.388801913162577e-06,
+      "loss": 0.0969,
+      "step": 5781
+    },
+    {
+      "epoch": 3.028287061288633,
+      "loss_partial": 0.052978515625,
+      "step": 5781
+    },
+    {
+      "epoch": 3.028287061288633,
+      "loss_partial": 0.055419921875,
+      "step": 5781
+    },
+    {
+      "epoch": 3.0288108957569406,
+      "grad_norm": 1.383186936378479,
+      "learning_rate": 3.385372634362949e-06,
+      "loss": 0.1029,
+      "step": 5782
+    },
+    {
+      "epoch": 3.0288108957569406,
+      "loss_partial": 0.048828125,
+      "step": 5782
+    },
+    {
+      "epoch": 3.0288108957569406,
+      "loss_partial": 0.05126953125,
+      "step": 5782
+    },
+    {
+      "epoch": 3.0293347302252487,
+      "grad_norm": 1.078295111656189,
+      "learning_rate": 3.381944737979533e-06,
+      "loss": 0.0989,
+      "step": 5783
+    },
+    {
+      "epoch": 3.0293347302252487,
+      "loss_partial": 0.05712890625,
+      "step": 5783
+    },
+    {
+      "epoch": 3.0293347302252487,
+      "loss_partial": 0.0556640625,
+      "step": 5783
+    },
+    {
+      "epoch": 3.0298585646935567,
+      "grad_norm": 1.4886997938156128,
+      "learning_rate": 3.378518224728745e-06,
+      "loss": 0.1037,
+      "step": 5784
+    },
+    {
+      "epoch": 3.0298585646935567,
+      "loss_partial": 0.05810546875,
+      "step": 5784
+    },
+    {
+      "epoch": 3.0298585646935567,
+      "loss_partial": 0.045654296875,
+      "step": 5784
+    },
+    {
+      "epoch": 3.030382399161865,
+      "grad_norm": 0.9702770709991455,
+      "learning_rate": 3.3750930953267037e-06,
+      "loss": 0.097,
+      "step": 5785
+    },
+    {
+      "epoch": 3.030382399161865,
+      "loss_partial": 0.047607421875,
+      "step": 5785
+    },
+    {
+      "epoch": 3.030382399161865,
+      "loss_partial": 0.048583984375,
+      "step": 5785
+    },
+    {
+      "epoch": 3.030906233630173,
+      "grad_norm": 1.704056978225708,
+      "learning_rate": 3.3716693504892363e-06,
+      "loss": 0.0974,
+      "step": 5786
+    },
+    {
+      "epoch": 3.030906233630173,
+      "loss_partial": 0.053466796875,
+      "step": 5786
+    },
+    {
+      "epoch": 3.030906233630173,
+      "loss_partial": 0.04736328125,
+      "step": 5786
+    },
+    {
+      "epoch": 3.031430068098481,
+      "grad_norm": 0.9715403318405151,
+      "learning_rate": 3.368246990931884e-06,
+      "loss": 0.0988,
+      "step": 5787
+    },
+    {
+      "epoch": 3.031430068098481,
+      "loss_partial": 0.043212890625,
+      "step": 5787
+    },
+    {
+      "epoch": 3.031430068098481,
+      "loss_partial": 0.052734375,
+      "step": 5787
+    },
+    {
+      "epoch": 3.031953902566789,
+      "grad_norm": 0.9808283448219299,
+      "learning_rate": 3.364826017369904e-06,
+      "loss": 0.0986,
+      "step": 5788
+    },
+    {
+      "epoch": 3.031953902566789,
+      "loss_partial": 0.052001953125,
+      "step": 5788
+    },
+    {
+      "epoch": 3.031953902566789,
+      "loss_partial": 0.0537109375,
+      "step": 5788
+    },
+    {
+      "epoch": 3.032477737035097,
+      "grad_norm": 1.0355099439620972,
+      "learning_rate": 3.3614064305182558e-06,
+      "loss": 0.0989,
+      "step": 5789
+    },
+    {
+      "epoch": 3.032477737035097,
+      "loss_partial": 0.049560546875,
+      "step": 5789
+    },
+    {
+      "epoch": 3.032477737035097,
+      "loss_partial": 0.049072265625,
+      "step": 5789
+    },
+    {
+      "epoch": 3.033001571503405,
+      "grad_norm": 1.3490556478500366,
+      "learning_rate": 3.3579882310916113e-06,
+      "loss": 0.1035,
+      "step": 5790
+    },
+    {
+      "epoch": 3.033001571503405,
+      "loss_partial": 0.05126953125,
+      "step": 5790
+    },
+    {
+      "epoch": 3.033001571503405,
+      "loss_partial": 0.044921875,
+      "step": 5790
+    },
+    {
+      "epoch": 3.033525405971713,
+      "grad_norm": 1.101711392402649,
+      "learning_rate": 3.3545714198043546e-06,
+      "loss": 0.1027,
+      "step": 5791
+    },
+    {
+      "epoch": 3.033525405971713,
+      "loss_partial": 0.04345703125,
+      "step": 5791
+    },
+    {
+      "epoch": 3.033525405971713,
+      "loss_partial": 0.052734375,
+      "step": 5791
+    },
+    {
+      "epoch": 3.034049240440021,
+      "grad_norm": 1.180870771408081,
+      "learning_rate": 3.3511559973705766e-06,
+      "loss": 0.0999,
+      "step": 5792
+    },
+    {
+      "epoch": 3.034049240440021,
+      "loss_partial": 0.06005859375,
+      "step": 5792
+    },
+    {
+      "epoch": 3.034049240440021,
+      "loss_partial": 0.046875,
+      "step": 5792
+    },
+    {
+      "epoch": 3.034573074908329,
+      "grad_norm": 1.1020445823669434,
+      "learning_rate": 3.34774196450408e-06,
+      "loss": 0.0981,
+      "step": 5793
+    },
+    {
+      "epoch": 3.034573074908329,
+      "loss_partial": 0.042724609375,
+      "step": 5793
+    },
+    {
+      "epoch": 3.034573074908329,
+      "loss_partial": 0.0634765625,
+      "step": 5793
+    },
+    {
+      "epoch": 3.035096909376637,
+      "grad_norm": 1.4086366891860962,
+      "learning_rate": 3.344329321918376e-06,
+      "loss": 0.1024,
+      "step": 5794
+    },
+    {
+      "epoch": 3.035096909376637,
+      "loss_partial": 0.044921875,
+      "step": 5794
+    },
+    {
+      "epoch": 3.035096909376637,
+      "loss_partial": 0.046875,
+      "step": 5794
+    },
+    {
+      "epoch": 3.035620743844945,
+      "grad_norm": 1.2859348058700562,
+      "learning_rate": 3.340918070326683e-06,
+      "loss": 0.0954,
+      "step": 5795
+    },
+    {
+      "epoch": 3.035620743844945,
+      "loss_partial": 0.05859375,
+      "step": 5795
+    },
+    {
+      "epoch": 3.035620743844945,
+      "loss_partial": 0.043701171875,
+      "step": 5795
+    },
+    {
+      "epoch": 3.036144578313253,
+      "grad_norm": 1.0110479593276978,
+      "learning_rate": 3.3375082104419387e-06,
+      "loss": 0.0959,
+      "step": 5796
+    },
+    {
+      "epoch": 3.036144578313253,
+      "loss_partial": 0.048828125,
+      "step": 5796
+    },
+    {
+      "epoch": 3.036144578313253,
+      "loss_partial": 0.0439453125,
+      "step": 5796
+    },
+    {
+      "epoch": 3.036668412781561,
+      "grad_norm": 1.0553463697433472,
+      "learning_rate": 3.3340997429767786e-06,
+      "loss": 0.1004,
+      "step": 5797
+    },
+    {
+      "epoch": 3.036668412781561,
+      "loss_partial": 0.0419921875,
+      "step": 5797
+    },
+    {
+      "epoch": 3.036668412781561,
+      "loss_partial": 0.045654296875,
+      "step": 5797
+    },
+    {
+      "epoch": 3.037192247249869,
+      "grad_norm": 1.2856348752975464,
+      "learning_rate": 3.3306926686435513e-06,
+      "loss": 0.0987,
+      "step": 5798
+    },
+    {
+      "epoch": 3.037192247249869,
+      "loss_partial": 0.05810546875,
+      "step": 5798
+    },
+    {
+      "epoch": 3.037192247249869,
+      "loss_partial": 0.055419921875,
+      "step": 5798
+    },
+    {
+      "epoch": 3.037716081718177,
+      "grad_norm": 1.1560766696929932,
+      "learning_rate": 3.3272869881543113e-06,
+      "loss": 0.1001,
+      "step": 5799
+    },
+    {
+      "epoch": 3.037716081718177,
+      "loss_partial": 0.046875,
+      "step": 5799
+    },
+    {
+      "epoch": 3.037716081718177,
+      "loss_partial": 0.046630859375,
+      "step": 5799
+    },
+    {
+      "epoch": 3.0382399161864853,
+      "grad_norm": 1.5624244213104248,
+      "learning_rate": 3.323882702220831e-06,
+      "loss": 0.1004,
+      "step": 5800
+    },
+    {
+      "epoch": 3.0382399161864853,
+      "loss_partial": 0.048095703125,
+      "step": 5800
+    },
+    {
+      "epoch": 3.0382399161864853,
+      "loss_partial": 0.053466796875,
+      "step": 5800
+    },
+    {
+      "epoch": 3.038763750654793,
+      "grad_norm": 1.5995632410049438,
+      "learning_rate": 3.320479811554582e-06,
+      "loss": 0.0996,
+      "step": 5801
+    },
+    {
+      "epoch": 3.038763750654793,
+      "loss_partial": 0.044189453125,
+      "step": 5801
+    },
+    {
+      "epoch": 3.038763750654793,
+      "loss_partial": 0.048095703125,
+      "step": 5801
+    },
+    {
+      "epoch": 3.039287585123101,
+      "grad_norm": 1.2836487293243408,
+      "learning_rate": 3.3170783168667486e-06,
+      "loss": 0.1043,
+      "step": 5802
+    },
+    {
+      "epoch": 3.039287585123101,
+      "loss_partial": 0.0595703125,
+      "step": 5802
+    },
+    {
+      "epoch": 3.039287585123101,
+      "loss_partial": 0.0517578125,
+      "step": 5802
+    },
+    {
+      "epoch": 3.039811419591409,
+      "grad_norm": 1.323478102684021,
+      "learning_rate": 3.3136782188682203e-06,
+      "loss": 0.0971,
+      "step": 5803
+    },
+    {
+      "epoch": 3.039811419591409,
+      "loss_partial": 0.052734375,
+      "step": 5803
+    },
+    {
+      "epoch": 3.039811419591409,
+      "loss_partial": 0.044677734375,
+      "step": 5803
+    },
+    {
+      "epoch": 3.040335254059717,
+      "grad_norm": 1.0669782161712646,
+      "learning_rate": 3.310279518269597e-06,
+      "loss": 0.0979,
+      "step": 5804
+    },
+    {
+      "epoch": 3.040335254059717,
+      "loss_partial": 0.044189453125,
+      "step": 5804
+    },
+    {
+      "epoch": 3.040335254059717,
+      "loss_partial": 0.048095703125,
+      "step": 5804
+    },
+    {
+      "epoch": 3.040859088528025,
+      "grad_norm": 0.9751498103141785,
+      "learning_rate": 3.3068822157811863e-06,
+      "loss": 0.1029,
+      "step": 5805
+    },
+    {
+      "epoch": 3.040859088528025,
+      "loss_partial": 0.06591796875,
+      "step": 5805
+    },
+    {
+      "epoch": 3.040859088528025,
+      "loss_partial": 0.04345703125,
+      "step": 5805
+    },
+    {
+      "epoch": 3.041382922996333,
+      "grad_norm": 1.2411017417907715,
+      "learning_rate": 3.303486312113003e-06,
+      "loss": 0.1006,
+      "step": 5806
+    },
+    {
+      "epoch": 3.041382922996333,
+      "loss_partial": 0.056396484375,
+      "step": 5806
+    },
+    {
+      "epoch": 3.041382922996333,
+      "loss_partial": 0.051513671875,
+      "step": 5806
+    },
+    {
+      "epoch": 3.0419067574646412,
+      "grad_norm": 1.269156813621521,
+      "learning_rate": 3.3000918079747723e-06,
+      "loss": 0.1007,
+      "step": 5807
+    },
+    {
+      "epoch": 3.0419067574646412,
+      "loss_partial": 0.051025390625,
+      "step": 5807
+    },
+    {
+      "epoch": 3.0419067574646412,
+      "loss_partial": 0.04638671875,
+      "step": 5807
+    },
+    {
+      "epoch": 3.0424305919329493,
+      "grad_norm": 0.9070917367935181,
+      "learning_rate": 3.296698704075918e-06,
+      "loss": 0.0959,
+      "step": 5808
+    },
+    {
+      "epoch": 3.0424305919329493,
+      "loss_partial": 0.04638671875,
+      "step": 5808
+    },
+    {
+      "epoch": 3.0424305919329493,
+      "loss_partial": 0.048583984375,
+      "step": 5808
+    },
+    {
+      "epoch": 3.0429544264012574,
+      "grad_norm": 1.0894005298614502,
+      "learning_rate": 3.2933070011255864e-06,
+      "loss": 0.0989,
+      "step": 5809
+    },
+    {
+      "epoch": 3.0429544264012574,
+      "loss_partial": 0.0556640625,
+      "step": 5809
+    },
+    {
+      "epoch": 3.0429544264012574,
+      "loss_partial": 0.05126953125,
+      "step": 5809
+    },
+    {
+      "epoch": 3.0434782608695654,
+      "grad_norm": 1.193760871887207,
+      "learning_rate": 3.289916699832619e-06,
+      "loss": 0.103,
+      "step": 5810
+    },
+    {
+      "epoch": 3.0434782608695654,
+      "loss_partial": 0.05615234375,
+      "step": 5810
+    },
+    {
+      "epoch": 3.0434782608695654,
+      "loss_partial": 0.051025390625,
+      "step": 5810
+    },
+    {
+      "epoch": 3.044002095337873,
+      "grad_norm": 1.6502199172973633,
+      "learning_rate": 3.2865278009055666e-06,
+      "loss": 0.0988,
+      "step": 5811
+    },
+    {
+      "epoch": 3.044002095337873,
+      "loss_partial": 0.050048828125,
+      "step": 5811
+    },
+    {
+      "epoch": 3.044002095337873,
+      "loss_partial": 0.046875,
+      "step": 5811
+    },
+    {
+      "epoch": 3.044525929806181,
+      "grad_norm": 1.058337688446045,
+      "learning_rate": 3.2831403050526867e-06,
+      "loss": 0.0985,
+      "step": 5812
+    },
+    {
+      "epoch": 3.044525929806181,
+      "loss_partial": 0.04443359375,
+      "step": 5812
+    },
+    {
+      "epoch": 3.044525929806181,
+      "loss_partial": 0.051025390625,
+      "step": 5812
+    },
+    {
+      "epoch": 3.045049764274489,
+      "grad_norm": 1.4888832569122314,
+      "learning_rate": 3.2797542129819495e-06,
+      "loss": 0.1033,
+      "step": 5813
+    },
+    {
+      "epoch": 3.045049764274489,
+      "loss_partial": 0.05419921875,
+      "step": 5813
+    },
+    {
+      "epoch": 3.045049764274489,
+      "loss_partial": 0.049560546875,
+      "step": 5813
+    },
+    {
+      "epoch": 3.045573598742797,
+      "grad_norm": 1.1101255416870117,
+      "learning_rate": 3.276369525401025e-06,
+      "loss": 0.0984,
+      "step": 5814
+    },
+    {
+      "epoch": 3.045573598742797,
+      "loss_partial": 0.0556640625,
+      "step": 5814
+    },
+    {
+      "epoch": 3.045573598742797,
+      "loss_partial": 0.058837890625,
+      "step": 5814
+    },
+    {
+      "epoch": 3.0460974332111053,
+      "grad_norm": 1.6582194566726685,
+      "learning_rate": 3.272986243017293e-06,
+      "loss": 0.1038,
+      "step": 5815
+    },
+    {
+      "epoch": 3.0460974332111053,
+      "loss_partial": 0.053466796875,
+      "step": 5815
+    },
+    {
+      "epoch": 3.0460974332111053,
+      "loss_partial": 0.045654296875,
+      "step": 5815
+    },
+    {
+      "epoch": 3.0466212676794133,
+      "grad_norm": 1.205264687538147,
+      "learning_rate": 3.2696043665378375e-06,
+      "loss": 0.0985,
+      "step": 5816
+    },
+    {
+      "epoch": 3.0466212676794133,
+      "loss_partial": 0.048095703125,
+      "step": 5816
+    },
+    {
+      "epoch": 3.0466212676794133,
+      "loss_partial": 0.048828125,
+      "step": 5816
+    },
+    {
+      "epoch": 3.0471451021477214,
+      "grad_norm": 1.8146032094955444,
+      "learning_rate": 3.266223896669449e-06,
+      "loss": 0.0997,
+      "step": 5817
+    },
+    {
+      "epoch": 3.0471451021477214,
+      "loss_partial": 0.05126953125,
+      "step": 5817
+    },
+    {
+      "epoch": 3.0471451021477214,
+      "loss_partial": 0.043701171875,
+      "step": 5817
+    },
+    {
+      "epoch": 3.0476689366160294,
+      "grad_norm": 1.43179452419281,
+      "learning_rate": 3.262844834118626e-06,
+      "loss": 0.1002,
+      "step": 5818
+    },
+    {
+      "epoch": 3.0476689366160294,
+      "loss_partial": 0.0498046875,
+      "step": 5818
+    },
+    {
+      "epoch": 3.0476689366160294,
+      "loss_partial": 0.046142578125,
+      "step": 5818
+    },
+    {
+      "epoch": 3.0481927710843375,
+      "grad_norm": 1.1604812145233154,
+      "learning_rate": 3.259467179591571e-06,
+      "loss": 0.1002,
+      "step": 5819
+    },
+    {
+      "epoch": 3.0481927710843375,
+      "loss_partial": 0.041748046875,
+      "step": 5819
+    },
+    {
+      "epoch": 3.0481927710843375,
+      "loss_partial": 0.04052734375,
+      "step": 5819
+    },
+    {
+      "epoch": 3.0487166055526456,
+      "grad_norm": 1.2061165571212769,
+      "learning_rate": 3.2560909337941905e-06,
+      "loss": 0.0984,
+      "step": 5820
+    },
+    {
+      "epoch": 3.0487166055526456,
+      "loss_partial": 0.0498046875,
+      "step": 5820
+    },
+    {
+      "epoch": 3.0487166055526456,
+      "loss_partial": 0.046142578125,
+      "step": 5820
+    },
+    {
+      "epoch": 3.049240440020953,
+      "grad_norm": 1.4071234464645386,
+      "learning_rate": 3.252716097432105e-06,
+      "loss": 0.1029,
+      "step": 5821
+    },
+    {
+      "epoch": 3.049240440020953,
+      "loss_partial": 0.0458984375,
+      "step": 5821
+    },
+    {
+      "epoch": 3.049240440020953,
+      "loss_partial": 0.054443359375,
+      "step": 5821
+    },
+    {
+      "epoch": 3.0497642744892612,
+      "grad_norm": 1.5898096561431885,
+      "learning_rate": 3.249342671210631e-06,
+      "loss": 0.099,
+      "step": 5822
+    },
+    {
+      "epoch": 3.0497642744892612,
+      "loss_partial": 0.049072265625,
+      "step": 5822
+    },
+    {
+      "epoch": 3.0497642744892612,
+      "loss_partial": 0.043701171875,
+      "step": 5822
+    },
+    {
+      "epoch": 3.0502881089575693,
+      "grad_norm": 1.5129934549331665,
+      "learning_rate": 3.245970655834795e-06,
+      "loss": 0.1047,
+      "step": 5823
+    },
+    {
+      "epoch": 3.0502881089575693,
+      "loss_partial": 0.053466796875,
+      "step": 5823
+    },
+    {
+      "epoch": 3.0502881089575693,
+      "loss_partial": 0.0498046875,
+      "step": 5823
+    },
+    {
+      "epoch": 3.0508119434258774,
+      "grad_norm": 1.4519622325897217,
+      "learning_rate": 3.2426000520093215e-06,
+      "loss": 0.1006,
+      "step": 5824
+    },
+    {
+      "epoch": 3.0508119434258774,
+      "loss_partial": 0.0537109375,
+      "step": 5824
+    },
+    {
+      "epoch": 3.0508119434258774,
+      "loss_partial": 0.043212890625,
+      "step": 5824
+    },
+    {
+      "epoch": 3.0513357778941854,
+      "grad_norm": 1.5230635404586792,
+      "learning_rate": 3.2392308604386546e-06,
+      "loss": 0.0985,
+      "step": 5825
+    },
+    {
+      "epoch": 3.0513357778941854,
+      "loss_partial": 0.05419921875,
+      "step": 5825
+    },
+    {
+      "epoch": 3.0513357778941854,
+      "loss_partial": 0.06591796875,
+      "step": 5825
+    },
+    {
+      "epoch": 3.0518596123624935,
+      "grad_norm": 1.185141682624817,
+      "learning_rate": 3.23586308182693e-06,
+      "loss": 0.099,
+      "step": 5826
+    },
+    {
+      "epoch": 3.0518596123624935,
+      "loss_partial": 0.052490234375,
+      "step": 5826
+    },
+    {
+      "epoch": 3.0518596123624935,
+      "loss_partial": 0.047607421875,
+      "step": 5826
+    },
+    {
+      "epoch": 3.0523834468308015,
+      "grad_norm": 1.432131290435791,
+      "learning_rate": 3.232496716877994e-06,
+      "loss": 0.1005,
+      "step": 5827
+    },
+    {
+      "epoch": 3.0523834468308015,
+      "loss_partial": 0.05126953125,
+      "step": 5827
+    },
+    {
+      "epoch": 3.0523834468308015,
+      "loss_partial": 0.049072265625,
+      "step": 5827
+    },
+    {
+      "epoch": 3.0529072812991096,
+      "grad_norm": 1.4671123027801514,
+      "learning_rate": 3.2291317662953947e-06,
+      "loss": 0.1018,
+      "step": 5828
+    },
+    {
+      "epoch": 3.0529072812991096,
+      "loss_partial": 0.043212890625,
+      "step": 5828
+    },
+    {
+      "epoch": 3.0529072812991096,
+      "loss_partial": 0.04150390625,
+      "step": 5828
+    },
+    {
+      "epoch": 3.0534311157674177,
+      "grad_norm": 1.4123198986053467,
+      "learning_rate": 3.225768230782387e-06,
+      "loss": 0.1021,
+      "step": 5829
+    },
+    {
+      "epoch": 3.0534311157674177,
+      "loss_partial": 0.0517578125,
+      "step": 5829
+    },
+    {
+      "epoch": 3.0534311157674177,
+      "loss_partial": 0.05908203125,
+      "step": 5829
+    },
+    {
+      "epoch": 3.0539549502357257,
+      "grad_norm": 1.3676337003707886,
+      "learning_rate": 3.222406111041927e-06,
+      "loss": 0.099,
+      "step": 5830
+    },
+    {
+      "epoch": 3.0539549502357257,
+      "loss_partial": 0.056640625,
+      "step": 5830
+    },
+    {
+      "epoch": 3.0539549502357257,
+      "loss_partial": 0.0654296875,
+      "step": 5830
+    },
+    {
+      "epoch": 3.0544787847040333,
+      "grad_norm": 1.0753872394561768,
+      "learning_rate": 3.2190454077766797e-06,
+      "loss": 0.1047,
+      "step": 5831
+    },
+    {
+      "epoch": 3.0544787847040333,
+      "loss_partial": 0.05126953125,
+      "step": 5831
+    },
+    {
+      "epoch": 3.0544787847040333,
+      "loss_partial": 0.05224609375,
+      "step": 5831
+    },
+    {
+      "epoch": 3.0550026191723414,
+      "grad_norm": 1.1387852430343628,
+      "learning_rate": 3.2156861216890066e-06,
+      "loss": 0.1015,
+      "step": 5832
+    },
+    {
+      "epoch": 3.0550026191723414,
+      "loss_partial": 0.047119140625,
+      "step": 5832
+    },
+    {
+      "epoch": 3.0550026191723414,
+      "loss_partial": 0.0478515625,
+      "step": 5832
+    },
+    {
+      "epoch": 3.0555264536406495,
+      "grad_norm": 1.582694411277771,
+      "learning_rate": 3.212328253480984e-06,
+      "loss": 0.1016,
+      "step": 5833
+    },
+    {
+      "epoch": 3.0555264536406495,
+      "loss_partial": 0.0439453125,
+      "step": 5833
+    },
+    {
+      "epoch": 3.0555264536406495,
+      "loss_partial": 0.047119140625,
+      "step": 5833
+    },
+    {
+      "epoch": 3.0560502881089575,
+      "grad_norm": 1.405563473701477,
+      "learning_rate": 3.2089718038543826e-06,
+      "loss": 0.0984,
+      "step": 5834
+    },
+    {
+      "epoch": 3.0560502881089575,
+      "loss_partial": 0.049560546875,
+      "step": 5834
+    },
+    {
+      "epoch": 3.0560502881089575,
+      "loss_partial": 0.04736328125,
+      "step": 5834
+    },
+    {
+      "epoch": 3.0565741225772656,
+      "grad_norm": 1.7276989221572876,
+      "learning_rate": 3.2056167735106804e-06,
+      "loss": 0.0992,
+      "step": 5835
+    },
+    {
+      "epoch": 3.0565741225772656,
+      "loss_partial": 0.05712890625,
+      "step": 5835
+    },
+    {
+      "epoch": 3.0565741225772656,
+      "loss_partial": 0.051513671875,
+      "step": 5835
+    },
+    {
+      "epoch": 3.0570979570455736,
+      "grad_norm": 1.6833570003509521,
+      "learning_rate": 3.2022631631510536e-06,
+      "loss": 0.1,
+      "step": 5836
+    },
+    {
+      "epoch": 3.0570979570455736,
+      "loss_partial": 0.04931640625,
+      "step": 5836
+    },
+    {
+      "epoch": 3.0570979570455736,
+      "loss_partial": 0.0634765625,
+      "step": 5836
+    },
+    {
+      "epoch": 3.0576217915138817,
+      "grad_norm": 1.5114115476608276,
+      "learning_rate": 3.1989109734763936e-06,
+      "loss": 0.1011,
+      "step": 5837
+    },
+    {
+      "epoch": 3.0576217915138817,
+      "loss_partial": 0.05908203125,
+      "step": 5837
+    },
+    {
+      "epoch": 3.0576217915138817,
+      "loss_partial": 0.052734375,
+      "step": 5837
+    },
+    {
+      "epoch": 3.0581456259821898,
+      "grad_norm": 1.8203580379486084,
+      "learning_rate": 3.1955602051872827e-06,
+      "loss": 0.1044,
+      "step": 5838
+    },
+    {
+      "epoch": 3.0581456259821898,
+      "loss_partial": 0.051513671875,
+      "step": 5838
+    },
+    {
+      "epoch": 3.0581456259821898,
+      "loss_partial": 0.050537109375,
+      "step": 5838
+    },
+    {
+      "epoch": 3.058669460450498,
+      "grad_norm": 1.1783192157745361,
+      "learning_rate": 3.1922108589840116e-06,
+      "loss": 0.0964,
+      "step": 5839
+    },
+    {
+      "epoch": 3.058669460450498,
+      "loss_partial": 0.04931640625,
+      "step": 5839
+    },
+    {
+      "epoch": 3.058669460450498,
+      "loss_partial": 0.05126953125,
+      "step": 5839
+    },
+    {
+      "epoch": 3.059193294918806,
+      "grad_norm": 2.2286460399627686,
+      "learning_rate": 3.1888629355665735e-06,
+      "loss": 0.1042,
+      "step": 5840
+    },
+    {
+      "epoch": 3.059193294918806,
+      "loss_partial": 0.0419921875,
+      "step": 5840
+    },
+    {
+      "epoch": 3.059193294918806,
+      "loss_partial": 0.052001953125,
+      "step": 5840
+    },
+    {
+      "epoch": 3.0597171293871135,
+      "grad_norm": 1.0590623617172241,
+      "learning_rate": 3.1855164356346626e-06,
+      "loss": 0.0991,
+      "step": 5841
+    },
+    {
+      "epoch": 3.0597171293871135,
+      "loss_partial": 0.0546875,
+      "step": 5841
+    },
+    {
+      "epoch": 3.0597171293871135,
+      "loss_partial": 0.043212890625,
+      "step": 5841
+    },
+    {
+      "epoch": 3.0602409638554215,
+      "grad_norm": 1.7244913578033447,
+      "learning_rate": 3.1821713598876767e-06,
+      "loss": 0.0977,
+      "step": 5842
+    },
+    {
+      "epoch": 3.0602409638554215,
+      "loss_partial": 0.051513671875,
+      "step": 5842
+    },
+    {
+      "epoch": 3.0602409638554215,
+      "loss_partial": 0.052490234375,
+      "step": 5842
+    },
+    {
+      "epoch": 3.0607647983237296,
+      "grad_norm": 1.1386243104934692,
+      "learning_rate": 3.1788277090247177e-06,
+      "loss": 0.0966,
+      "step": 5843
+    },
+    {
+      "epoch": 3.0607647983237296,
+      "loss_partial": 0.049072265625,
+      "step": 5843
+    },
+    {
+      "epoch": 3.0607647983237296,
+      "loss_partial": 0.0537109375,
+      "step": 5843
+    },
+    {
+      "epoch": 3.0612886327920377,
+      "grad_norm": 1.6441296339035034,
+      "learning_rate": 3.1754854837445825e-06,
+      "loss": 0.1013,
+      "step": 5844
+    },
+    {
+      "epoch": 3.0612886327920377,
+      "loss_partial": 0.04248046875,
+      "step": 5844
+    },
+    {
+      "epoch": 3.0612886327920377,
+      "loss_partial": 0.049560546875,
+      "step": 5844
+    },
+    {
+      "epoch": 3.0618124672603457,
+      "grad_norm": 1.204049825668335,
+      "learning_rate": 3.172144684745783e-06,
+      "loss": 0.0999,
+      "step": 5845
+    },
+    {
+      "epoch": 3.0618124672603457,
+      "loss_partial": 0.050537109375,
+      "step": 5845
+    },
+    {
+      "epoch": 3.0618124672603457,
+      "loss_partial": 0.0419921875,
+      "step": 5845
+    },
+    {
+      "epoch": 3.062336301728654,
+      "grad_norm": 1.377307653427124,
+      "learning_rate": 3.168805312726523e-06,
+      "loss": 0.1003,
+      "step": 5846
+    },
+    {
+      "epoch": 3.062336301728654,
+      "loss_partial": 0.042724609375,
+      "step": 5846
+    },
+    {
+      "epoch": 3.062336301728654,
+      "loss_partial": 0.052490234375,
+      "step": 5846
+    },
+    {
+      "epoch": 3.062860136196962,
+      "grad_norm": 1.4386789798736572,
+      "learning_rate": 3.165467368384708e-06,
+      "loss": 0.1013,
+      "step": 5847
+    },
+    {
+      "epoch": 3.062860136196962,
+      "loss_partial": 0.05419921875,
+      "step": 5847
+    },
+    {
+      "epoch": 3.062860136196962,
+      "loss_partial": 0.04638671875,
+      "step": 5847
+    },
+    {
+      "epoch": 3.06338397066527,
+      "grad_norm": 1.1733065843582153,
+      "learning_rate": 3.1621308524179484e-06,
+      "loss": 0.0984,
+      "step": 5848
+    },
+    {
+      "epoch": 3.06338397066527,
+      "loss_partial": 0.046630859375,
+      "step": 5848
+    },
+    {
+      "epoch": 3.06338397066527,
+      "loss_partial": 0.0498046875,
+      "step": 5848
+    },
+    {
+      "epoch": 3.063907805133578,
+      "grad_norm": 1.0402899980545044,
+      "learning_rate": 3.1587957655235578e-06,
+      "loss": 0.0988,
+      "step": 5849
+    },
+    {
+      "epoch": 3.063907805133578,
+      "loss_partial": 0.051025390625,
+      "step": 5849
+    },
+    {
+      "epoch": 3.063907805133578,
+      "loss_partial": 0.052490234375,
+      "step": 5849
+    },
+    {
+      "epoch": 3.064431639601886,
+      "grad_norm": 1.2969036102294922,
+      "learning_rate": 3.155462108398549e-06,
+      "loss": 0.099,
+      "step": 5850
+    },
+    {
+      "epoch": 3.064431639601886,
+      "loss_partial": 0.04541015625,
+      "step": 5850
+    },
+    {
+      "epoch": 3.064431639601886,
+      "loss_partial": 0.05322265625,
+      "step": 5850
+    },
+    {
+      "epoch": 3.0649554740701936,
+      "grad_norm": 1.430191159248352,
+      "learning_rate": 3.1521298817396327e-06,
+      "loss": 0.1012,
+      "step": 5851
+    },
+    {
+      "epoch": 3.0649554740701936,
+      "loss_partial": 0.0576171875,
+      "step": 5851
+    },
+    {
+      "epoch": 3.0649554740701936,
+      "loss_partial": 0.041748046875,
+      "step": 5851
+    },
+    {
+      "epoch": 3.0654793085385017,
+      "grad_norm": 1.244375467300415,
+      "learning_rate": 3.1487990862432273e-06,
+      "loss": 0.1004,
+      "step": 5852
+    },
+    {
+      "epoch": 3.0654793085385017,
+      "loss_partial": 0.04931640625,
+      "step": 5852
+    },
+    {
+      "epoch": 3.0654793085385017,
+      "loss_partial": 0.05322265625,
+      "step": 5852
+    },
+    {
+      "epoch": 3.0660031430068098,
+      "grad_norm": 1.134286642074585,
+      "learning_rate": 3.145469722605444e-06,
+      "loss": 0.0979,
+      "step": 5853
+    },
+    {
+      "epoch": 3.0660031430068098,
+      "loss_partial": 0.050537109375,
+      "step": 5853
+    },
+    {
+      "epoch": 3.0660031430068098,
+      "loss_partial": 0.047119140625,
+      "step": 5853
+    },
+    {
+      "epoch": 3.066526977475118,
+      "grad_norm": 1.3110356330871582,
+      "learning_rate": 3.142141791522103e-06,
+      "loss": 0.1,
+      "step": 5854
+    },
+    {
+      "epoch": 3.066526977475118,
+      "loss_partial": 0.045166015625,
+      "step": 5854
+    },
+    {
+      "epoch": 3.066526977475118,
+      "loss_partial": 0.043701171875,
+      "step": 5854
+    },
+    {
+      "epoch": 3.067050811943426,
+      "grad_norm": 0.967409610748291,
+      "learning_rate": 3.1388152936887193e-06,
+      "loss": 0.0999,
+      "step": 5855
+    },
+    {
+      "epoch": 3.067050811943426,
+      "loss_partial": 0.05517578125,
+      "step": 5855
+    },
+    {
+      "epoch": 3.067050811943426,
+      "loss_partial": 0.05029296875,
+      "step": 5855
+    },
+    {
+      "epoch": 3.067574646411734,
+      "grad_norm": 1.3824814558029175,
+      "learning_rate": 3.1354902298005087e-06,
+      "loss": 0.1018,
+      "step": 5856
+    },
+    {
+      "epoch": 3.067574646411734,
+      "loss_partial": 0.056884765625,
+      "step": 5856
+    },
+    {
+      "epoch": 3.067574646411734,
+      "loss_partial": 0.050537109375,
+      "step": 5856
+    },
+    {
+      "epoch": 3.068098480880042,
+      "grad_norm": 1.2127480506896973,
+      "learning_rate": 3.132166600552393e-06,
+      "loss": 0.0993,
+      "step": 5857
+    },
+    {
+      "epoch": 3.068098480880042,
+      "loss_partial": 0.053466796875,
+      "step": 5857
+    },
+    {
+      "epoch": 3.068098480880042,
+      "loss_partial": 0.060791015625,
+      "step": 5857
+    },
+    {
+      "epoch": 3.06862231534835,
+      "grad_norm": 1.3392292261123657,
+      "learning_rate": 3.128844406638989e-06,
+      "loss": 0.0987,
+      "step": 5858
+    },
+    {
+      "epoch": 3.06862231534835,
+      "loss_partial": 0.0419921875,
+      "step": 5858
+    },
+    {
+      "epoch": 3.06862231534835,
+      "loss_partial": 0.060546875,
+      "step": 5858
+    },
+    {
+      "epoch": 3.069146149816658,
+      "grad_norm": 1.268536925315857,
+      "learning_rate": 3.1255236487546147e-06,
+      "loss": 0.098,
+      "step": 5859
+    },
+    {
+      "epoch": 3.069146149816658,
+      "loss_partial": 0.04833984375,
+      "step": 5859
+    },
+    {
+      "epoch": 3.069146149816658,
+      "loss_partial": 0.04443359375,
+      "step": 5859
+    },
+    {
+      "epoch": 3.0696699842849657,
+      "grad_norm": 1.160801887512207,
+      "learning_rate": 3.1222043275932835e-06,
+      "loss": 0.099,
+      "step": 5860
+    },
+    {
+      "epoch": 3.0696699842849657,
+      "loss_partial": 0.055419921875,
+      "step": 5860
+    },
+    {
+      "epoch": 3.0696699842849657,
+      "loss_partial": 0.048583984375,
+      "step": 5860
+    },
+    {
+      "epoch": 3.070193818753274,
+      "grad_norm": 1.2672470808029175,
+      "learning_rate": 3.1188864438487198e-06,
+      "loss": 0.101,
+      "step": 5861
+    },
+    {
+      "epoch": 3.070193818753274,
+      "loss_partial": 0.048095703125,
+      "step": 5861
+    },
+    {
+      "epoch": 3.070193818753274,
+      "loss_partial": 0.050048828125,
+      "step": 5861
+    },
+    {
+      "epoch": 3.070717653221582,
+      "grad_norm": 1.643795132637024,
+      "learning_rate": 3.1155699982143384e-06,
+      "loss": 0.0981,
+      "step": 5862
+    },
+    {
+      "epoch": 3.070717653221582,
+      "loss_partial": 0.045166015625,
+      "step": 5862
+    },
+    {
+      "epoch": 3.070717653221582,
+      "loss_partial": 0.05224609375,
+      "step": 5862
+    },
+    {
+      "epoch": 3.07124148768989,
+      "grad_norm": 1.2846013307571411,
+      "learning_rate": 3.1122549913832546e-06,
+      "loss": 0.0985,
+      "step": 5863
+    },
+    {
+      "epoch": 3.07124148768989,
+      "loss_partial": 0.047607421875,
+      "step": 5863
+    },
+    {
+      "epoch": 3.07124148768989,
+      "loss_partial": 0.061279296875,
+      "step": 5863
+    },
+    {
+      "epoch": 3.071765322158198,
+      "grad_norm": 1.6039953231811523,
+      "learning_rate": 3.108941424048285e-06,
+      "loss": 0.103,
+      "step": 5864
+    },
+    {
+      "epoch": 3.071765322158198,
+      "loss_partial": 0.05712890625,
+      "step": 5864
+    },
+    {
+      "epoch": 3.071765322158198,
+      "loss_partial": 0.064453125,
+      "step": 5864
+    },
+    {
+      "epoch": 3.072289156626506,
+      "grad_norm": 1.3886222839355469,
+      "learning_rate": 3.105629296901945e-06,
+      "loss": 0.0983,
+      "step": 5865
+    },
+    {
+      "epoch": 3.072289156626506,
+      "loss_partial": 0.047119140625,
+      "step": 5865
+    },
+    {
+      "epoch": 3.072289156626506,
+      "loss_partial": 0.05712890625,
+      "step": 5865
+    },
+    {
+      "epoch": 3.072812991094814,
+      "grad_norm": 1.2537219524383545,
+      "learning_rate": 3.102318610636449e-06,
+      "loss": 0.1013,
+      "step": 5866
+    },
+    {
+      "epoch": 3.072812991094814,
+      "loss_partial": 0.03857421875,
+      "step": 5866
+    },
+    {
+      "epoch": 3.072812991094814,
+      "loss_partial": 0.04638671875,
+      "step": 5866
+    },
+    {
+      "epoch": 3.073336825563122,
+      "grad_norm": 1.241923213005066,
+      "learning_rate": 3.0990093659437094e-06,
+      "loss": 0.099,
+      "step": 5867
+    },
+    {
+      "epoch": 3.073336825563122,
+      "loss_partial": 0.0595703125,
+      "step": 5867
+    },
+    {
+      "epoch": 3.073336825563122,
+      "loss_partial": 0.05126953125,
+      "step": 5867
+    },
+    {
+      "epoch": 3.07386066003143,
+      "grad_norm": 1.3195524215698242,
+      "learning_rate": 3.0957015635153375e-06,
+      "loss": 0.1019,
+      "step": 5868
+    },
+    {
+      "epoch": 3.07386066003143,
+      "loss_partial": 0.052490234375,
+      "step": 5868
+    },
+    {
+      "epoch": 3.07386066003143,
+      "loss_partial": 0.05322265625,
+      "step": 5868
+    },
+    {
+      "epoch": 3.0743844944997383,
+      "grad_norm": 1.0438655614852905,
+      "learning_rate": 3.0923952040426407e-06,
+      "loss": 0.099,
+      "step": 5869
+    },
+    {
+      "epoch": 3.0743844944997383,
+      "loss_partial": 0.04541015625,
+      "step": 5869
+    },
+    {
+      "epoch": 3.0743844944997383,
+      "loss_partial": 0.04736328125,
+      "step": 5869
+    },
+    {
+      "epoch": 3.074908328968046,
+      "grad_norm": 1.7359097003936768,
+      "learning_rate": 3.089090288216634e-06,
+      "loss": 0.1025,
+      "step": 5870
+    },
+    {
+      "epoch": 3.074908328968046,
+      "loss_partial": 0.051025390625,
+      "step": 5870
+    },
+    {
+      "epoch": 3.074908328968046,
+      "loss_partial": 0.0556640625,
+      "step": 5870
+    },
+    {
+      "epoch": 3.075432163436354,
+      "grad_norm": 1.2438336610794067,
+      "learning_rate": 3.085786816728021e-06,
+      "loss": 0.0996,
+      "step": 5871
+    },
+    {
+      "epoch": 3.075432163436354,
+      "loss_partial": 0.04833984375,
+      "step": 5871
+    },
+    {
+      "epoch": 3.075432163436354,
+      "loss_partial": 0.0498046875,
+      "step": 5871
+    },
+    {
+      "epoch": 3.075955997904662,
+      "grad_norm": 1.4549453258514404,
+      "learning_rate": 3.082484790267208e-06,
+      "loss": 0.0984,
+      "step": 5872
+    },
+    {
+      "epoch": 3.075955997904662,
+      "loss_partial": 0.05126953125,
+      "step": 5872
+    },
+    {
+      "epoch": 3.075955997904662,
+      "loss_partial": 0.0576171875,
+      "step": 5872
+    },
+    {
+      "epoch": 3.07647983237297,
+      "grad_norm": 1.5440056324005127,
+      "learning_rate": 3.0791842095242942e-06,
+      "loss": 0.0988,
+      "step": 5873
+    },
+    {
+      "epoch": 3.07647983237297,
+      "loss_partial": 0.045654296875,
+      "step": 5873
+    },
+    {
+      "epoch": 3.07647983237297,
+      "loss_partial": 0.0419921875,
+      "step": 5873
+    },
+    {
+      "epoch": 3.077003666841278,
+      "grad_norm": 1.1400889158248901,
+      "learning_rate": 3.0758850751890855e-06,
+      "loss": 0.096,
+      "step": 5874
+    },
+    {
+      "epoch": 3.077003666841278,
+      "loss_partial": 0.05810546875,
+      "step": 5874
+    },
+    {
+      "epoch": 3.077003666841278,
+      "loss_partial": 0.04345703125,
+      "step": 5874
+    },
+    {
+      "epoch": 3.077527501309586,
+      "grad_norm": 1.526100993156433,
+      "learning_rate": 3.072587387951079e-06,
+      "loss": 0.0988,
+      "step": 5875
+    },
+    {
+      "epoch": 3.077527501309586,
+      "loss_partial": 0.059326171875,
+      "step": 5875
+    },
+    {
+      "epoch": 3.077527501309586,
+      "loss_partial": 0.053955078125,
+      "step": 5875
+    },
+    {
+      "epoch": 3.0780513357778942,
+      "grad_norm": 1.3582582473754883,
+      "learning_rate": 3.0692911484994713e-06,
+      "loss": 0.1003,
+      "step": 5876
+    },
+    {
+      "epoch": 3.0780513357778942,
+      "loss_partial": 0.051025390625,
+      "step": 5876
+    },
+    {
+      "epoch": 3.0780513357778942,
+      "loss_partial": 0.052490234375,
+      "step": 5876
+    },
+    {
+      "epoch": 3.0785751702462023,
+      "grad_norm": 1.1934161186218262,
+      "learning_rate": 3.0659963575231544e-06,
+      "loss": 0.0984,
+      "step": 5877
+    },
+    {
+      "epoch": 3.0785751702462023,
+      "loss_partial": 0.051025390625,
+      "step": 5877
+    },
+    {
+      "epoch": 3.0785751702462023,
+      "loss_partial": 0.047607421875,
+      "step": 5877
+    },
+    {
+      "epoch": 3.0790990047145104,
+      "grad_norm": 1.6746490001678467,
+      "learning_rate": 3.0627030157107207e-06,
+      "loss": 0.1011,
+      "step": 5878
+    },
+    {
+      "epoch": 3.0790990047145104,
+      "loss_partial": 0.0498046875,
+      "step": 5878
+    },
+    {
+      "epoch": 3.0790990047145104,
+      "loss_partial": 0.04443359375,
+      "step": 5878
+    },
+    {
+      "epoch": 3.0796228391828184,
+      "grad_norm": 1.550561785697937,
+      "learning_rate": 3.0594111237504586e-06,
+      "loss": 0.0985,
+      "step": 5879
+    },
+    {
+      "epoch": 3.0796228391828184,
+      "loss_partial": 0.041259765625,
+      "step": 5879
+    },
+    {
+      "epoch": 3.0796228391828184,
+      "loss_partial": 0.05029296875,
+      "step": 5879
+    },
+    {
+      "epoch": 3.080146673651126,
+      "grad_norm": 1.3801409006118774,
+      "learning_rate": 3.0561206823303514e-06,
+      "loss": 0.0983,
+      "step": 5880
+    },
+    {
+      "epoch": 3.080146673651126,
+      "loss_partial": 0.052490234375,
+      "step": 5880
+    },
+    {
+      "epoch": 3.080146673651126,
+      "loss_partial": 0.05322265625,
+      "step": 5880
+    },
+    {
+      "epoch": 3.080670508119434,
+      "grad_norm": 1.1974163055419922,
+      "learning_rate": 3.0528316921380786e-06,
+      "loss": 0.1009,
+      "step": 5881
+    },
+    {
+      "epoch": 3.080670508119434,
+      "loss_partial": 0.052734375,
+      "step": 5881
+    },
+    {
+      "epoch": 3.080670508119434,
+      "loss_partial": 0.046630859375,
+      "step": 5881
+    },
+    {
+      "epoch": 3.081194342587742,
+      "grad_norm": 1.42418372631073,
+      "learning_rate": 3.049544153861026e-06,
+      "loss": 0.1028,
+      "step": 5882
+    },
+    {
+      "epoch": 3.081194342587742,
+      "loss_partial": 0.0625,
+      "step": 5882
+    },
+    {
+      "epoch": 3.081194342587742,
+      "loss_partial": 0.04443359375,
+      "step": 5882
+    },
+    {
+      "epoch": 3.0817181770560502,
+      "grad_norm": 1.128727674484253,
+      "learning_rate": 3.046258068186264e-06,
+      "loss": 0.1002,
+      "step": 5883
+    },
+    {
+      "epoch": 3.0817181770560502,
+      "loss_partial": 0.060302734375,
+      "step": 5883
+    },
+    {
+      "epoch": 3.0817181770560502,
+      "loss_partial": 0.045166015625,
+      "step": 5883
+    },
+    {
+      "epoch": 3.0822420115243583,
+      "grad_norm": 1.155806064605713,
+      "learning_rate": 3.042973435800566e-06,
+      "loss": 0.1042,
+      "step": 5884
+    },
+    {
+      "epoch": 3.0822420115243583,
+      "loss_partial": 0.05224609375,
+      "step": 5884
+    },
+    {
+      "epoch": 3.0822420115243583,
+      "loss_partial": 0.04296875,
+      "step": 5884
+    },
+    {
+      "epoch": 3.0827658459926663,
+      "grad_norm": 1.636552095413208,
+      "learning_rate": 3.0396902573903954e-06,
+      "loss": 0.1018,
+      "step": 5885
+    },
+    {
+      "epoch": 3.0827658459926663,
+      "loss_partial": 0.054443359375,
+      "step": 5885
+    },
+    {
+      "epoch": 3.0827658459926663,
+      "loss_partial": 0.044677734375,
+      "step": 5885
+    },
+    {
+      "epoch": 3.0832896804609744,
+      "grad_norm": 1.7257838249206543,
+      "learning_rate": 3.0364085336419224e-06,
+      "loss": 0.1004,
+      "step": 5886
+    },
+    {
+      "epoch": 3.0832896804609744,
+      "loss_partial": 0.058349609375,
+      "step": 5886
+    },
+    {
+      "epoch": 3.0832896804609744,
+      "loss_partial": 0.049560546875,
+      "step": 5886
+    },
+    {
+      "epoch": 3.0838135149292825,
+      "grad_norm": 1.4202756881713867,
+      "learning_rate": 3.0331282652410043e-06,
+      "loss": 0.1021,
+      "step": 5887
+    },
+    {
+      "epoch": 3.0838135149292825,
+      "loss_partial": 0.0419921875,
+      "step": 5887
+    },
+    {
+      "epoch": 3.0838135149292825,
+      "loss_partial": 0.046875,
+      "step": 5887
+    },
+    {
+      "epoch": 3.0843373493975905,
+      "grad_norm": 1.3793317079544067,
+      "learning_rate": 3.0298494528731982e-06,
+      "loss": 0.0955,
+      "step": 5888
+    },
+    {
+      "epoch": 3.0843373493975905,
+      "loss_partial": 0.049072265625,
+      "step": 5888
+    },
+    {
+      "epoch": 3.0843373493975905,
+      "loss_partial": 0.0556640625,
+      "step": 5888
+    },
+    {
+      "epoch": 3.0848611838658986,
+      "grad_norm": 1.4941030740737915,
+      "learning_rate": 3.0265720972237533e-06,
+      "loss": 0.1002,
+      "step": 5889
+    },
+    {
+      "epoch": 3.0848611838658986,
+      "loss_partial": 0.056884765625,
+      "step": 5889
+    },
+    {
+      "epoch": 3.0848611838658986,
+      "loss_partial": 0.05419921875,
+      "step": 5889
+    },
+    {
+      "epoch": 3.085385018334206,
+      "grad_norm": 1.9419857263565063,
+      "learning_rate": 3.023296198977618e-06,
+      "loss": 0.0983,
+      "step": 5890
+    },
+    {
+      "epoch": 3.085385018334206,
+      "loss_partial": 0.046630859375,
+      "step": 5890
+    },
+    {
+      "epoch": 3.085385018334206,
+      "loss_partial": 0.0537109375,
+      "step": 5890
+    },
+    {
+      "epoch": 3.0859088528025143,
+      "grad_norm": 1.1786861419677734,
+      "learning_rate": 3.0200217588194347e-06,
+      "loss": 0.1002,
+      "step": 5891
+    },
+    {
+      "epoch": 3.0859088528025143,
+      "loss_partial": 0.04345703125,
+      "step": 5891
+    },
+    {
+      "epoch": 3.0859088528025143,
+      "loss_partial": 0.042236328125,
+      "step": 5891
+    },
+    {
+      "epoch": 3.0864326872708223,
+      "grad_norm": 1.3697729110717773,
+      "learning_rate": 3.0167487774335403e-06,
+      "loss": 0.096,
+      "step": 5892
+    },
+    {
+      "epoch": 3.0864326872708223,
+      "loss_partial": 0.044677734375,
+      "step": 5892
+    },
+    {
+      "epoch": 3.0864326872708223,
+      "loss_partial": 0.05419921875,
+      "step": 5892
+    },
+    {
+      "epoch": 3.0869565217391304,
+      "grad_norm": 1.410597801208496,
+      "learning_rate": 3.0134772555039658e-06,
+      "loss": 0.1044,
+      "step": 5893
+    },
+    {
+      "epoch": 3.0869565217391304,
+      "loss_partial": 0.04541015625,
+      "step": 5893
+    },
+    {
+      "epoch": 3.0869565217391304,
+      "loss_partial": 0.045166015625,
+      "step": 5893
+    },
+    {
+      "epoch": 3.0874803562074384,
+      "grad_norm": 1.58199143409729,
+      "learning_rate": 3.010207193714444e-06,
+      "loss": 0.1007,
+      "step": 5894
+    },
+    {
+      "epoch": 3.0874803562074384,
+      "loss_partial": 0.046142578125,
+      "step": 5894
+    },
+    {
+      "epoch": 3.0874803562074384,
+      "loss_partial": 0.05615234375,
+      "step": 5894
+    },
+    {
+      "epoch": 3.0880041906757465,
+      "grad_norm": 1.241719365119934,
+      "learning_rate": 3.0069385927483953e-06,
+      "loss": 0.0989,
+      "step": 5895
+    },
+    {
+      "epoch": 3.0880041906757465,
+      "loss_partial": 0.046630859375,
+      "step": 5895
+    },
+    {
+      "epoch": 3.0880041906757465,
+      "loss_partial": 0.0498046875,
+      "step": 5895
+    },
+    {
+      "epoch": 3.0885280251440546,
+      "grad_norm": 1.0916725397109985,
+      "learning_rate": 3.0036714532889376e-06,
+      "loss": 0.0966,
+      "step": 5896
+    },
+    {
+      "epoch": 3.0885280251440546,
+      "loss_partial": 0.046630859375,
+      "step": 5896
+    },
+    {
+      "epoch": 3.0885280251440546,
+      "loss_partial": 0.05029296875,
+      "step": 5896
+    },
+    {
+      "epoch": 3.0890518596123626,
+      "grad_norm": 1.50862455368042,
+      "learning_rate": 3.0004057760188786e-06,
+      "loss": 0.1017,
+      "step": 5897
+    },
+    {
+      "epoch": 3.0890518596123626,
+      "loss_partial": 0.051513671875,
+      "step": 5897
+    },
+    {
+      "epoch": 3.0890518596123626,
+      "loss_partial": 0.04296875,
+      "step": 5897
+    },
+    {
+      "epoch": 3.0895756940806707,
+      "grad_norm": 1.0116640329360962,
+      "learning_rate": 2.997141561620731e-06,
+      "loss": 0.1017,
+      "step": 5898
+    },
+    {
+      "epoch": 3.0895756940806707,
+      "loss_partial": 0.041748046875,
+      "step": 5898
+    },
+    {
+      "epoch": 3.0895756940806707,
+      "loss_partial": 0.04736328125,
+      "step": 5898
+    },
+    {
+      "epoch": 3.0900995285489787,
+      "grad_norm": 1.461971640586853,
+      "learning_rate": 2.9938788107766935e-06,
+      "loss": 0.0978,
+      "step": 5899
+    },
+    {
+      "epoch": 3.0900995285489787,
+      "loss_partial": 0.046142578125,
+      "step": 5899
+    },
+    {
+      "epoch": 3.0900995285489787,
+      "loss_partial": 0.044677734375,
+      "step": 5899
+    },
+    {
+      "epoch": 3.0906233630172864,
+      "grad_norm": 1.3699743747711182,
+      "learning_rate": 2.9906175241686598e-06,
+      "loss": 0.0986,
+      "step": 5900
+    },
+    {
+      "epoch": 3.0906233630172864,
+      "loss_partial": 0.048095703125,
+      "step": 5900
+    },
+    {
+      "epoch": 3.0906233630172864,
+      "loss_partial": 0.04638671875,
+      "step": 5900
+    },
+    {
+      "epoch": 3.0911471974855944,
+      "grad_norm": 1.256796956062317,
+      "learning_rate": 2.9873577024782175e-06,
+      "loss": 0.0988,
+      "step": 5901
+    },
+    {
+      "epoch": 3.0911471974855944,
+      "loss_partial": 0.048095703125,
+      "step": 5901
+    },
+    {
+      "epoch": 3.0911471974855944,
+      "loss_partial": 0.052734375,
+      "step": 5901
+    },
+    {
+      "epoch": 3.0916710319539025,
+      "grad_norm": 1.6180444955825806,
+      "learning_rate": 2.984099346386652e-06,
+      "loss": 0.1,
+      "step": 5902
+    },
+    {
+      "epoch": 3.0916710319539025,
+      "loss_partial": 0.06005859375,
+      "step": 5902
+    },
+    {
+      "epoch": 3.0916710319539025,
+      "loss_partial": 0.05029296875,
+      "step": 5902
+    },
+    {
+      "epoch": 3.0921948664222105,
+      "grad_norm": 1.4300559759140015,
+      "learning_rate": 2.9808424565749373e-06,
+      "loss": 0.1035,
+      "step": 5903
+    },
+    {
+      "epoch": 3.0921948664222105,
+      "loss_partial": 0.0458984375,
+      "step": 5903
+    },
+    {
+      "epoch": 3.0921948664222105,
+      "loss_partial": 0.04931640625,
+      "step": 5903
+    },
+    {
+      "epoch": 3.0927187008905186,
+      "grad_norm": 1.0504059791564941,
+      "learning_rate": 2.9775870337237443e-06,
+      "loss": 0.0995,
+      "step": 5904
+    },
+    {
+      "epoch": 3.0927187008905186,
+      "loss_partial": 0.05615234375,
+      "step": 5904
+    },
+    {
+      "epoch": 3.0927187008905186,
+      "loss_partial": 0.040283203125,
+      "step": 5904
+    },
+    {
+      "epoch": 3.0932425353588267,
+      "grad_norm": 1.0993820428848267,
+      "learning_rate": 2.974333078513433e-06,
+      "loss": 0.1001,
+      "step": 5905
+    },
+    {
+      "epoch": 3.0932425353588267,
+      "loss_partial": 0.0517578125,
+      "step": 5905
+    },
+    {
+      "epoch": 3.0932425353588267,
+      "loss_partial": 0.044921875,
+      "step": 5905
+    },
+    {
+      "epoch": 3.0937663698271347,
+      "grad_norm": 1.1500036716461182,
+      "learning_rate": 2.9710805916240657e-06,
+      "loss": 0.1013,
+      "step": 5906
+    },
+    {
+      "epoch": 3.0937663698271347,
+      "loss_partial": 0.03759765625,
+      "step": 5906
+    },
+    {
+      "epoch": 3.0937663698271347,
+      "loss_partial": 0.0478515625,
+      "step": 5906
+    },
+    {
+      "epoch": 3.0942902042954428,
+      "grad_norm": 1.7500847578048706,
+      "learning_rate": 2.96782957373539e-06,
+      "loss": 0.1021,
+      "step": 5907
+    },
+    {
+      "epoch": 3.0942902042954428,
+      "loss_partial": 0.050537109375,
+      "step": 5907
+    },
+    {
+      "epoch": 3.0942902042954428,
+      "loss_partial": 0.043701171875,
+      "step": 5907
+    },
+    {
+      "epoch": 3.094814038763751,
+      "grad_norm": 1.1962602138519287,
+      "learning_rate": 2.9645800255268463e-06,
+      "loss": 0.0988,
+      "step": 5908
+    },
+    {
+      "epoch": 3.094814038763751,
+      "loss_partial": 0.052978515625,
+      "step": 5908
+    },
+    {
+      "epoch": 3.094814038763751,
+      "loss_partial": 0.059814453125,
+      "step": 5908
+    },
+    {
+      "epoch": 3.095337873232059,
+      "grad_norm": 1.0788260698318481,
+      "learning_rate": 2.961331947677569e-06,
+      "loss": 0.1033,
+      "step": 5909
+    },
+    {
+      "epoch": 3.095337873232059,
+      "loss_partial": 0.048583984375,
+      "step": 5909
+    },
+    {
+      "epoch": 3.095337873232059,
+      "loss_partial": 0.0546875,
+      "step": 5909
+    },
+    {
+      "epoch": 3.0958617077003665,
+      "grad_norm": 1.7174190282821655,
+      "learning_rate": 2.9580853408663934e-06,
+      "loss": 0.1025,
+      "step": 5910
+    },
+    {
+      "epoch": 3.0958617077003665,
+      "loss_partial": 0.04443359375,
+      "step": 5910
+    },
+    {
+      "epoch": 3.0958617077003665,
+      "loss_partial": 0.048828125,
+      "step": 5910
+    },
+    {
+      "epoch": 3.0963855421686746,
+      "grad_norm": 1.1620237827301025,
+      "learning_rate": 2.954840205771834e-06,
+      "loss": 0.098,
+      "step": 5911
+    },
+    {
+      "epoch": 3.0963855421686746,
+      "loss_partial": 0.0498046875,
+      "step": 5911
+    },
+    {
+      "epoch": 3.0963855421686746,
+      "loss_partial": 0.0439453125,
+      "step": 5911
+    },
+    {
+      "epoch": 3.0969093766369826,
+      "grad_norm": 0.9389757513999939,
+      "learning_rate": 2.951596543072106e-06,
+      "loss": 0.0992,
+      "step": 5912
+    },
+    {
+      "epoch": 3.0969093766369826,
+      "loss_partial": 0.04150390625,
+      "step": 5912
+    },
+    {
+      "epoch": 3.0969093766369826,
+      "loss_partial": 0.052734375,
+      "step": 5912
+    },
+    {
+      "epoch": 3.0974332111052907,
+      "grad_norm": 1.0949270725250244,
+      "learning_rate": 2.948354353445114e-06,
+      "loss": 0.0956,
+      "step": 5913
+    },
+    {
+      "epoch": 3.0974332111052907,
+      "loss_partial": 0.048095703125,
+      "step": 5913
+    },
+    {
+      "epoch": 3.0974332111052907,
+      "loss_partial": 0.0517578125,
+      "step": 5913
+    },
+    {
+      "epoch": 3.0979570455735987,
+      "grad_norm": 0.803616464138031,
+      "learning_rate": 2.9451136375684576e-06,
+      "loss": 0.0969,
+      "step": 5914
+    },
+    {
+      "epoch": 3.0979570455735987,
+      "loss_partial": 0.04052734375,
+      "step": 5914
+    },
+    {
+      "epoch": 3.0979570455735987,
+      "loss_partial": 0.046875,
+      "step": 5914
+    },
+    {
+      "epoch": 3.098480880041907,
+      "grad_norm": 1.0007383823394775,
+      "learning_rate": 2.941874396119424e-06,
+      "loss": 0.1006,
+      "step": 5915
+    },
+    {
+      "epoch": 3.098480880041907,
+      "loss_partial": 0.049560546875,
+      "step": 5915
+    },
+    {
+      "epoch": 3.098480880041907,
+      "loss_partial": 0.053955078125,
+      "step": 5915
+    },
+    {
+      "epoch": 3.099004714510215,
+      "grad_norm": 1.0592395067214966,
+      "learning_rate": 2.938636629774995e-06,
+      "loss": 0.0971,
+      "step": 5916
+    },
+    {
+      "epoch": 3.099004714510215,
+      "loss_partial": 0.0439453125,
+      "step": 5916
+    },
+    {
+      "epoch": 3.099004714510215,
+      "loss_partial": 0.051513671875,
+      "step": 5916
+    },
+    {
+      "epoch": 3.099528548978523,
+      "grad_norm": 1.1586215496063232,
+      "learning_rate": 2.935400339211841e-06,
+      "loss": 0.0989,
+      "step": 5917
+    },
+    {
+      "epoch": 3.099528548978523,
+      "loss_partial": 0.046875,
+      "step": 5917
+    },
+    {
+      "epoch": 3.099528548978523,
+      "loss_partial": 0.0478515625,
+      "step": 5917
+    },
+    {
+      "epoch": 3.100052383446831,
+      "grad_norm": 1.0582518577575684,
+      "learning_rate": 2.9321655251063342e-06,
+      "loss": 0.1013,
+      "step": 5918
+    },
+    {
+      "epoch": 3.100052383446831,
+      "loss_partial": 0.055419921875,
+      "step": 5918
+    },
+    {
+      "epoch": 3.100052383446831,
+      "loss_partial": 0.044921875,
+      "step": 5918
+    },
+    {
+      "epoch": 3.1005762179151386,
+      "grad_norm": 1.2576261758804321,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 0.0994,
+      "step": 5919
+    },
+    {
+      "epoch": 3.1005762179151386,
+      "loss_partial": 0.04541015625,
+      "step": 5919
+    },
+    {
+      "epoch": 3.1005762179151386,
+      "loss_partial": 0.04443359375,
+      "step": 5919
+    },
+    {
+      "epoch": 3.1011000523834467,
+      "grad_norm": 1.2862504720687866,
+      "learning_rate": 2.925700328972164e-06,
+      "loss": 0.1003,
+      "step": 5920
+    },
+    {
+      "epoch": 3.1011000523834467,
+      "loss_partial": 0.046875,
+      "step": 5920
+    },
+    {
+      "epoch": 3.1011000523834467,
+      "loss_partial": 0.05078125,
+      "step": 5920
+    },
+    {
+      "epoch": 3.1016238868517547,
+      "grad_norm": 1.3057221174240112,
+      "learning_rate": 2.9224699482946872e-06,
+      "loss": 0.1006,
+      "step": 5921
+    },
+    {
+      "epoch": 3.1016238868517547,
+      "loss_partial": 0.0439453125,
+      "step": 5921
+    },
+    {
+      "epoch": 3.1016238868517547,
+      "loss_partial": 0.050537109375,
+      "step": 5921
+    },
+    {
+      "epoch": 3.102147721320063,
+      "grad_norm": 1.3625824451446533,
+      "learning_rate": 2.9192410467772213e-06,
+      "loss": 0.1039,
+      "step": 5922
+    },
+    {
+      "epoch": 3.102147721320063,
+      "loss_partial": 0.044677734375,
+      "step": 5922
+    },
+    {
+      "epoch": 3.102147721320063,
+      "loss_partial": 0.044921875,
+      "step": 5922
+    },
+    {
+      "epoch": 3.102671555788371,
+      "grad_norm": 1.4644343852996826,
+      "learning_rate": 2.916013625094595e-06,
+      "loss": 0.097,
+      "step": 5923
+    },
+    {
+      "epoch": 3.102671555788371,
+      "loss_partial": 0.052001953125,
+      "step": 5923
+    },
+    {
+      "epoch": 3.102671555788371,
+      "loss_partial": 0.052490234375,
+      "step": 5923
+    },
+    {
+      "epoch": 3.103195390256679,
+      "grad_norm": 1.1109156608581543,
+      "learning_rate": 2.912787683921314e-06,
+      "loss": 0.1007,
+      "step": 5924
+    },
+    {
+      "epoch": 3.103195390256679,
+      "loss_partial": 0.046142578125,
+      "step": 5924
+    },
+    {
+      "epoch": 3.103195390256679,
+      "loss_partial": 0.054443359375,
+      "step": 5924
+    },
+    {
+      "epoch": 3.103719224724987,
+      "grad_norm": 1.1446176767349243,
+      "learning_rate": 2.909563223931582e-06,
+      "loss": 0.1008,
+      "step": 5925
+    },
+    {
+      "epoch": 3.103719224724987,
+      "loss_partial": 0.046142578125,
+      "step": 5925
+    },
+    {
+      "epoch": 3.103719224724987,
+      "loss_partial": 0.0576171875,
+      "step": 5925
+    },
+    {
+      "epoch": 3.104243059193295,
+      "grad_norm": 1.1217745542526245,
+      "learning_rate": 2.9063402457992897e-06,
+      "loss": 0.0991,
+      "step": 5926
+    },
+    {
+      "epoch": 3.104243059193295,
+      "loss_partial": 0.05029296875,
+      "step": 5926
+    },
+    {
+      "epoch": 3.104243059193295,
+      "loss_partial": 0.056884765625,
+      "step": 5926
+    },
+    {
+      "epoch": 3.104766893661603,
+      "grad_norm": 1.2728118896484375,
+      "learning_rate": 2.9031187501980196e-06,
+      "loss": 0.0983,
+      "step": 5927
+    },
+    {
+      "epoch": 3.104766893661603,
+      "loss_partial": 0.047119140625,
+      "step": 5927
+    },
+    {
+      "epoch": 3.104766893661603,
+      "loss_partial": 0.05419921875,
+      "step": 5927
+    },
+    {
+      "epoch": 3.105290728129911,
+      "grad_norm": 1.0111467838287354,
+      "learning_rate": 2.8998987378010458e-06,
+      "loss": 0.1009,
+      "step": 5928
+    },
+    {
+      "epoch": 3.105290728129911,
+      "loss_partial": 0.04541015625,
+      "step": 5928
+    },
+    {
+      "epoch": 3.105290728129911,
+      "loss_partial": 0.06103515625,
+      "step": 5928
+    },
+    {
+      "epoch": 3.1058145625982188,
+      "grad_norm": 1.4658600091934204,
+      "learning_rate": 2.8966802092813295e-06,
+      "loss": 0.0997,
+      "step": 5929
+    },
+    {
+      "epoch": 3.1058145625982188,
+      "loss_partial": 0.048095703125,
+      "step": 5929
+    },
+    {
+      "epoch": 3.1058145625982188,
+      "loss_partial": 0.05615234375,
+      "step": 5929
+    },
+    {
+      "epoch": 3.106338397066527,
+      "grad_norm": 1.3081051111221313,
+      "learning_rate": 2.893463165311521e-06,
+      "loss": 0.1018,
+      "step": 5930
+    },
+    {
+      "epoch": 3.106338397066527,
+      "loss_partial": 0.052734375,
+      "step": 5930
+    },
+    {
+      "epoch": 3.106338397066527,
+      "loss_partial": 0.050048828125,
+      "step": 5930
+    },
+    {
+      "epoch": 3.106862231534835,
+      "grad_norm": 1.3287512063980103,
+      "learning_rate": 2.8902476065639685e-06,
+      "loss": 0.1021,
+      "step": 5931
+    },
+    {
+      "epoch": 3.106862231534835,
+      "loss_partial": 0.0478515625,
+      "step": 5931
+    },
+    {
+      "epoch": 3.106862231534835,
+      "loss_partial": 0.04150390625,
+      "step": 5931
+    },
+    {
+      "epoch": 3.107386066003143,
+      "grad_norm": 1.3415642976760864,
+      "learning_rate": 2.8870335337107e-06,
+      "loss": 0.0999,
+      "step": 5932
+    },
+    {
+      "epoch": 3.107386066003143,
+      "loss_partial": 0.05126953125,
+      "step": 5932
+    },
+    {
+      "epoch": 3.107386066003143,
+      "loss_partial": 0.049072265625,
+      "step": 5932
+    },
+    {
+      "epoch": 3.107909900471451,
+      "grad_norm": 0.9190629124641418,
+      "learning_rate": 2.883820947423439e-06,
+      "loss": 0.0998,
+      "step": 5933
+    },
+    {
+      "epoch": 3.107909900471451,
+      "loss_partial": 0.046875,
+      "step": 5933
+    },
+    {
+      "epoch": 3.107909900471451,
+      "loss_partial": 0.045166015625,
+      "step": 5933
+    },
+    {
+      "epoch": 3.108433734939759,
+      "grad_norm": 1.4485024213790894,
+      "learning_rate": 2.8806098483735912e-06,
+      "loss": 0.101,
+      "step": 5934
+    },
+    {
+      "epoch": 3.108433734939759,
+      "loss_partial": 0.05078125,
+      "step": 5934
+    },
+    {
+      "epoch": 3.108433734939759,
+      "loss_partial": 0.05029296875,
+      "step": 5934
+    },
+    {
+      "epoch": 3.108957569408067,
+      "grad_norm": 1.1848152875900269,
+      "learning_rate": 2.877400237232263e-06,
+      "loss": 0.1031,
+      "step": 5935
+    },
+    {
+      "epoch": 3.108957569408067,
+      "loss_partial": 0.0517578125,
+      "step": 5935
+    },
+    {
+      "epoch": 3.108957569408067,
+      "loss_partial": 0.04638671875,
+      "step": 5935
+    },
+    {
+      "epoch": 3.109481403876375,
+      "grad_norm": 1.1598036289215088,
+      "learning_rate": 2.874192114670241e-06,
+      "loss": 0.0969,
+      "step": 5936
+    },
+    {
+      "epoch": 3.109481403876375,
+      "loss_partial": 0.05859375,
+      "step": 5936
+    },
+    {
+      "epoch": 3.109481403876375,
+      "loss_partial": 0.055419921875,
+      "step": 5936
+    },
+    {
+      "epoch": 3.1100052383446832,
+      "grad_norm": 1.642563819885254,
+      "learning_rate": 2.8709854813580017e-06,
+      "loss": 0.1003,
+      "step": 5937
+    },
+    {
+      "epoch": 3.1100052383446832,
+      "loss_partial": 0.060791015625,
+      "step": 5937
+    },
+    {
+      "epoch": 3.1100052383446832,
+      "loss_partial": 0.044189453125,
+      "step": 5937
+    },
+    {
+      "epoch": 3.1105290728129913,
+      "grad_norm": 1.3482730388641357,
+      "learning_rate": 2.8677803379657133e-06,
+      "loss": 0.0988,
+      "step": 5938
+    },
+    {
+      "epoch": 3.1105290728129913,
+      "loss_partial": 0.043212890625,
+      "step": 5938
+    },
+    {
+      "epoch": 3.1105290728129913,
+      "loss_partial": 0.045166015625,
+      "step": 5938
+    },
+    {
+      "epoch": 3.111052907281299,
+      "grad_norm": 1.1043779850006104,
+      "learning_rate": 2.8645766851632297e-06,
+      "loss": 0.0973,
+      "step": 5939
+    },
+    {
+      "epoch": 3.111052907281299,
+      "loss_partial": 0.052490234375,
+      "step": 5939
+    },
+    {
+      "epoch": 3.111052907281299,
+      "loss_partial": 0.050048828125,
+      "step": 5939
+    },
+    {
+      "epoch": 3.111576741749607,
+      "grad_norm": 1.4784151315689087,
+      "learning_rate": 2.861374523620094e-06,
+      "loss": 0.099,
+      "step": 5940
+    },
+    {
+      "epoch": 3.111576741749607,
+      "loss_partial": 0.046142578125,
+      "step": 5940
+    },
+    {
+      "epoch": 3.111576741749607,
+      "loss_partial": 0.038818359375,
+      "step": 5940
+    },
+    {
+      "epoch": 3.112100576217915,
+      "grad_norm": 1.635640025138855,
+      "learning_rate": 2.8581738540055394e-06,
+      "loss": 0.0975,
+      "step": 5941
+    },
+    {
+      "epoch": 3.112100576217915,
+      "loss_partial": 0.050048828125,
+      "step": 5941
+    },
+    {
+      "epoch": 3.112100576217915,
+      "loss_partial": 0.057861328125,
+      "step": 5941
+    },
+    {
+      "epoch": 3.112624410686223,
+      "grad_norm": 2.037623882293701,
+      "learning_rate": 2.854974676988482e-06,
+      "loss": 0.1035,
+      "step": 5942
+    },
+    {
+      "epoch": 3.112624410686223,
+      "loss_partial": 0.044189453125,
+      "step": 5942
+    },
+    {
+      "epoch": 3.112624410686223,
+      "loss_partial": 0.044677734375,
+      "step": 5942
+    },
+    {
+      "epoch": 3.113148245154531,
+      "grad_norm": 1.3720577955245972,
+      "learning_rate": 2.851776993237537e-06,
+      "loss": 0.0977,
+      "step": 5943
+    },
+    {
+      "epoch": 3.113148245154531,
+      "loss_partial": 0.0556640625,
+      "step": 5943
+    },
+    {
+      "epoch": 3.113148245154531,
+      "loss_partial": 0.05078125,
+      "step": 5943
+    },
+    {
+      "epoch": 3.113672079622839,
+      "grad_norm": 1.456466555595398,
+      "learning_rate": 2.848580803420997e-06,
+      "loss": 0.0984,
+      "step": 5944
+    },
+    {
+      "epoch": 3.113672079622839,
+      "loss_partial": 0.041259765625,
+      "step": 5944
+    },
+    {
+      "epoch": 3.113672079622839,
+      "loss_partial": 0.04638671875,
+      "step": 5944
+    },
+    {
+      "epoch": 3.1141959140911473,
+      "grad_norm": 1.4175643920898438,
+      "learning_rate": 2.8453861082068466e-06,
+      "loss": 0.102,
+      "step": 5945
+    },
+    {
+      "epoch": 3.1141959140911473,
+      "loss_partial": 0.046875,
+      "step": 5945
+    },
+    {
+      "epoch": 3.1141959140911473,
+      "loss_partial": 0.0556640625,
+      "step": 5945
+    },
+    {
+      "epoch": 3.1147197485594553,
+      "grad_norm": 1.6113470792770386,
+      "learning_rate": 2.842192908262753e-06,
+      "loss": 0.0988,
+      "step": 5946
+    },
+    {
+      "epoch": 3.1147197485594553,
+      "loss_partial": 0.057373046875,
+      "step": 5946
+    },
+    {
+      "epoch": 3.1147197485594553,
+      "loss_partial": 0.0556640625,
+      "step": 5946
+    },
+    {
+      "epoch": 3.1152435830277634,
+      "grad_norm": 1.1070983409881592,
+      "learning_rate": 2.8390012042560834e-06,
+      "loss": 0.1031,
+      "step": 5947
+    },
+    {
+      "epoch": 3.1152435830277634,
+      "loss_partial": 0.0478515625,
+      "step": 5947
+    },
+    {
+      "epoch": 3.1152435830277634,
+      "loss_partial": 0.05078125,
+      "step": 5947
+    },
+    {
+      "epoch": 3.1157674174960714,
+      "grad_norm": 1.357068657875061,
+      "learning_rate": 2.8358109968538783e-06,
+      "loss": 0.0984,
+      "step": 5948
+    },
+    {
+      "epoch": 3.1157674174960714,
+      "loss_partial": 0.049560546875,
+      "step": 5948
+    },
+    {
+      "epoch": 3.1157674174960714,
+      "loss_partial": 0.037841796875,
+      "step": 5948
+    },
+    {
+      "epoch": 3.116291251964379,
+      "grad_norm": 1.3400734663009644,
+      "learning_rate": 2.8326222867228735e-06,
+      "loss": 0.1,
+      "step": 5949
+    },
+    {
+      "epoch": 3.116291251964379,
+      "loss_partial": 0.039794921875,
+      "step": 5949
+    },
+    {
+      "epoch": 3.116291251964379,
+      "loss_partial": 0.047607421875,
+      "step": 5949
+    },
+    {
+      "epoch": 3.116815086432687,
+      "grad_norm": 1.3116815090179443,
+      "learning_rate": 2.8294350745294887e-06,
+      "loss": 0.1,
+      "step": 5950
+    },
+    {
+      "epoch": 3.116815086432687,
+      "loss_partial": 0.04052734375,
+      "step": 5950
+    },
+    {
+      "epoch": 3.116815086432687,
+      "loss_partial": 0.0595703125,
+      "step": 5950
+    },
+    {
+      "epoch": 3.117338920900995,
+      "grad_norm": 1.0829237699508667,
+      "learning_rate": 2.8262493609398322e-06,
+      "loss": 0.1014,
+      "step": 5951
+    },
+    {
+      "epoch": 3.117338920900995,
+      "loss_partial": 0.042724609375,
+      "step": 5951
+    },
+    {
+      "epoch": 3.117338920900995,
+      "loss_partial": 0.046875,
+      "step": 5951
+    },
+    {
+      "epoch": 3.1178627553693032,
+      "grad_norm": 0.8894127607345581,
+      "learning_rate": 2.823065146619698e-06,
+      "loss": 0.1005,
+      "step": 5952
+    },
+    {
+      "epoch": 3.1178627553693032,
+      "loss_partial": 0.05126953125,
+      "step": 5952
+    },
+    {
+      "epoch": 3.1178627553693032,
+      "loss_partial": 0.0439453125,
+      "step": 5952
+    },
+    {
+      "epoch": 3.1183865898376113,
+      "grad_norm": 1.44095778465271,
+      "learning_rate": 2.8198824322345677e-06,
+      "loss": 0.1006,
+      "step": 5953
+    },
+    {
+      "epoch": 3.1183865898376113,
+      "loss_partial": 0.042236328125,
+      "step": 5953
+    },
+    {
+      "epoch": 3.1183865898376113,
+      "loss_partial": 0.04296875,
+      "step": 5953
+    },
+    {
+      "epoch": 3.1189104243059194,
+      "grad_norm": 0.9270811080932617,
+      "learning_rate": 2.8167012184496056e-06,
+      "loss": 0.0935,
+      "step": 5954
+    },
+    {
+      "epoch": 3.1189104243059194,
+      "loss_partial": 0.058349609375,
+      "step": 5954
+    },
+    {
+      "epoch": 3.1189104243059194,
+      "loss_partial": 0.049072265625,
+      "step": 5954
+    },
+    {
+      "epoch": 3.1194342587742274,
+      "grad_norm": 1.433241367340088,
+      "learning_rate": 2.8135215059296706e-06,
+      "loss": 0.1016,
+      "step": 5955
+    },
+    {
+      "epoch": 3.1194342587742274,
+      "loss_partial": 0.052001953125,
+      "step": 5955
+    },
+    {
+      "epoch": 3.1194342587742274,
+      "loss_partial": 0.04541015625,
+      "step": 5955
+    },
+    {
+      "epoch": 3.1199580932425355,
+      "grad_norm": 1.0559567213058472,
+      "learning_rate": 2.8103432953393015e-06,
+      "loss": 0.0975,
+      "step": 5956
+    },
+    {
+      "epoch": 3.1199580932425355,
+      "loss_partial": 0.044921875,
+      "step": 5956
+    },
+    {
+      "epoch": 3.1199580932425355,
+      "loss_partial": 0.04345703125,
+      "step": 5956
+    },
+    {
+      "epoch": 3.1204819277108435,
+      "grad_norm": 0.9582811594009399,
+      "learning_rate": 2.8071665873427244e-06,
+      "loss": 0.0982,
+      "step": 5957
+    },
+    {
+      "epoch": 3.1204819277108435,
+      "loss_partial": 0.052001953125,
+      "step": 5957
+    },
+    {
+      "epoch": 3.1204819277108435,
+      "loss_partial": 0.05615234375,
+      "step": 5957
+    },
+    {
+      "epoch": 3.1210057621791516,
+      "grad_norm": 1.117547869682312,
+      "learning_rate": 2.803991382603849e-06,
+      "loss": 0.0969,
+      "step": 5958
+    },
+    {
+      "epoch": 3.1210057621791516,
+      "loss_partial": 0.05126953125,
+      "step": 5958
+    },
+    {
+      "epoch": 3.1210057621791516,
+      "loss_partial": 0.046875,
+      "step": 5958
+    },
+    {
+      "epoch": 3.121529596647459,
+      "grad_norm": 1.3509294986724854,
+      "learning_rate": 2.8008176817862787e-06,
+      "loss": 0.0981,
+      "step": 5959
+    },
+    {
+      "epoch": 3.121529596647459,
+      "loss_partial": 0.05078125,
+      "step": 5959
+    },
+    {
+      "epoch": 3.121529596647459,
+      "loss_partial": 0.048828125,
+      "step": 5959
+    },
+    {
+      "epoch": 3.1220534311157673,
+      "grad_norm": 1.196889042854309,
+      "learning_rate": 2.797645485553294e-06,
+      "loss": 0.0999,
+      "step": 5960
+    },
+    {
+      "epoch": 3.1220534311157673,
+      "loss_partial": 0.0439453125,
+      "step": 5960
+    },
+    {
+      "epoch": 3.1220534311157673,
+      "loss_partial": 0.05712890625,
+      "step": 5960
+    },
+    {
+      "epoch": 3.1225772655840753,
+      "grad_norm": 1.3299256563186646,
+      "learning_rate": 2.794474794567864e-06,
+      "loss": 0.0996,
+      "step": 5961
+    },
+    {
+      "epoch": 3.1225772655840753,
+      "loss_partial": 0.0478515625,
+      "step": 5961
+    },
+    {
+      "epoch": 3.1225772655840753,
+      "loss_partial": 0.047119140625,
+      "step": 5961
+    },
+    {
+      "epoch": 3.1231011000523834,
+      "grad_norm": 1.4433915615081787,
+      "learning_rate": 2.791305609492646e-06,
+      "loss": 0.1008,
+      "step": 5962
+    },
+    {
+      "epoch": 3.1231011000523834,
+      "loss_partial": 0.05615234375,
+      "step": 5962
+    },
+    {
+      "epoch": 3.1231011000523834,
+      "loss_partial": 0.04150390625,
+      "step": 5962
+    },
+    {
+      "epoch": 3.1236249345206915,
+      "grad_norm": 1.2110313177108765,
+      "learning_rate": 2.788137930989977e-06,
+      "loss": 0.0998,
+      "step": 5963
+    },
+    {
+      "epoch": 3.1236249345206915,
+      "loss_partial": 0.04833984375,
+      "step": 5963
+    },
+    {
+      "epoch": 3.1236249345206915,
+      "loss_partial": 0.05615234375,
+      "step": 5963
+    },
+    {
+      "epoch": 3.1241487689889995,
+      "grad_norm": 1.643395185470581,
+      "learning_rate": 2.784971759721885e-06,
+      "loss": 0.105,
+      "step": 5964
+    },
+    {
+      "epoch": 3.1241487689889995,
+      "loss_partial": 0.057373046875,
+      "step": 5964
+    },
+    {
+      "epoch": 3.1241487689889995,
+      "loss_partial": 0.048583984375,
+      "step": 5964
+    },
+    {
+      "epoch": 3.1246726034573076,
+      "grad_norm": 1.7796010971069336,
+      "learning_rate": 2.7818070963500776e-06,
+      "loss": 0.0989,
+      "step": 5965
+    },
+    {
+      "epoch": 3.1246726034573076,
+      "loss_partial": 0.04296875,
+      "step": 5965
+    },
+    {
+      "epoch": 3.1246726034573076,
+      "loss_partial": 0.0517578125,
+      "step": 5965
+    },
+    {
+      "epoch": 3.1251964379256156,
+      "grad_norm": 1.1671156883239746,
+      "learning_rate": 2.778643941535949e-06,
+      "loss": 0.1005,
+      "step": 5966
+    },
+    {
+      "epoch": 3.1251964379256156,
+      "loss_partial": 0.0517578125,
+      "step": 5966
+    },
+    {
+      "epoch": 3.1251964379256156,
+      "loss_partial": 0.05712890625,
+      "step": 5966
+    },
+    {
+      "epoch": 3.1257202723939237,
+      "grad_norm": 1.300260305404663,
+      "learning_rate": 2.7754822959405824e-06,
+      "loss": 0.1045,
+      "step": 5967
+    },
+    {
+      "epoch": 3.1257202723939237,
+      "loss_partial": 0.052978515625,
+      "step": 5967
+    },
+    {
+      "epoch": 3.1257202723939237,
+      "loss_partial": 0.04833984375,
+      "step": 5967
+    },
+    {
+      "epoch": 3.1262441068622318,
+      "grad_norm": 1.2396925687789917,
+      "learning_rate": 2.772322160224743e-06,
+      "loss": 0.101,
+      "step": 5968
+    },
+    {
+      "epoch": 3.1262441068622318,
+      "loss_partial": 0.05078125,
+      "step": 5968
+    },
+    {
+      "epoch": 3.1262441068622318,
+      "loss_partial": 0.0576171875,
+      "step": 5968
+    },
+    {
+      "epoch": 3.1267679413305394,
+      "grad_norm": 1.7618263959884644,
+      "learning_rate": 2.7691635350488764e-06,
+      "loss": 0.1036,
+      "step": 5969
+    },
+    {
+      "epoch": 3.1267679413305394,
+      "loss_partial": 0.047119140625,
+      "step": 5969
+    },
+    {
+      "epoch": 3.1267679413305394,
+      "loss_partial": 0.04736328125,
+      "step": 5969
+    },
+    {
+      "epoch": 3.1272917757988474,
+      "grad_norm": 1.4191079139709473,
+      "learning_rate": 2.766006421073114e-06,
+      "loss": 0.097,
+      "step": 5970
+    },
+    {
+      "epoch": 3.1272917757988474,
+      "loss_partial": 0.05810546875,
+      "step": 5970
+    },
+    {
+      "epoch": 3.1272917757988474,
+      "loss_partial": 0.05322265625,
+      "step": 5970
+    },
+    {
+      "epoch": 3.1278156102671555,
+      "grad_norm": 1.0952529907226562,
+      "learning_rate": 2.762850818957279e-06,
+      "loss": 0.1001,
+      "step": 5971
+    },
+    {
+      "epoch": 3.1278156102671555,
+      "loss_partial": 0.044189453125,
+      "step": 5971
+    },
+    {
+      "epoch": 3.1278156102671555,
+      "loss_partial": 0.0478515625,
+      "step": 5971
+    },
+    {
+      "epoch": 3.1283394447354635,
+      "grad_norm": 1.3386069536209106,
+      "learning_rate": 2.759696729360869e-06,
+      "loss": 0.0999,
+      "step": 5972
+    },
+    {
+      "epoch": 3.1283394447354635,
+      "loss_partial": 0.04248046875,
+      "step": 5972
+    },
+    {
+      "epoch": 3.1283394447354635,
+      "loss_partial": 0.052001953125,
+      "step": 5972
+    },
+    {
+      "epoch": 3.1288632792037716,
+      "grad_norm": 1.0301016569137573,
+      "learning_rate": 2.7565441529430715e-06,
+      "loss": 0.1011,
+      "step": 5973
+    },
+    {
+      "epoch": 3.1288632792037716,
+      "loss_partial": 0.042724609375,
+      "step": 5973
+    },
+    {
+      "epoch": 3.1288632792037716,
+      "loss_partial": 0.054931640625,
+      "step": 5973
+    },
+    {
+      "epoch": 3.1293871136720797,
+      "grad_norm": 1.3812965154647827,
+      "learning_rate": 2.7533930903627528e-06,
+      "loss": 0.0961,
+      "step": 5974
+    },
+    {
+      "epoch": 3.1293871136720797,
+      "loss_partial": 0.048828125,
+      "step": 5974
+    },
+    {
+      "epoch": 3.1293871136720797,
+      "loss_partial": 0.060791015625,
+      "step": 5974
+    },
+    {
+      "epoch": 3.1299109481403877,
+      "grad_norm": 1.162880539894104,
+      "learning_rate": 2.7502435422784656e-06,
+      "loss": 0.1008,
+      "step": 5975
+    },
+    {
+      "epoch": 3.1299109481403877,
+      "loss_partial": 0.04443359375,
+      "step": 5975
+    },
+    {
+      "epoch": 3.1299109481403877,
+      "loss_partial": 0.037841796875,
+      "step": 5975
+    },
+    {
+      "epoch": 3.130434782608696,
+      "grad_norm": 1.4304616451263428,
+      "learning_rate": 2.7470955093484486e-06,
+      "loss": 0.0983,
+      "step": 5976
+    },
+    {
+      "epoch": 3.130434782608696,
+      "loss_partial": 0.04052734375,
+      "step": 5976
+    },
+    {
+      "epoch": 3.130434782608696,
+      "loss_partial": 0.042236328125,
+      "step": 5976
+    },
+    {
+      "epoch": 3.130958617077004,
+      "grad_norm": 1.479559302330017,
+      "learning_rate": 2.7439489922306182e-06,
+      "loss": 0.0994,
+      "step": 5977
+    },
+    {
+      "epoch": 3.130958617077004,
+      "loss_partial": 0.054931640625,
+      "step": 5977
+    },
+    {
+      "epoch": 3.130958617077004,
+      "loss_partial": 0.05712890625,
+      "step": 5977
+    },
+    {
+      "epoch": 3.1314824515453115,
+      "grad_norm": 1.5398499965667725,
+      "learning_rate": 2.740803991582579e-06,
+      "loss": 0.1001,
+      "step": 5978
+    },
+    {
+      "epoch": 3.1314824515453115,
+      "loss_partial": 0.047119140625,
+      "step": 5978
+    },
+    {
+      "epoch": 3.1314824515453115,
+      "loss_partial": 0.041748046875,
+      "step": 5978
+    },
+    {
+      "epoch": 3.1320062860136195,
+      "grad_norm": 1.0989423990249634,
+      "learning_rate": 2.7376605080616136e-06,
+      "loss": 0.0952,
+      "step": 5979
+    },
+    {
+      "epoch": 3.1320062860136195,
+      "loss_partial": 0.04345703125,
+      "step": 5979
+    },
+    {
+      "epoch": 3.1320062860136195,
+      "loss_partial": 0.041748046875,
+      "step": 5979
+    },
+    {
+      "epoch": 3.1325301204819276,
+      "grad_norm": 1.3269985914230347,
+      "learning_rate": 2.7345185423246956e-06,
+      "loss": 0.1036,
+      "step": 5980
+    },
+    {
+      "epoch": 3.1325301204819276,
+      "loss_partial": 0.040771484375,
+      "step": 5980
+    },
+    {
+      "epoch": 3.1325301204819276,
+      "loss_partial": 0.05126953125,
+      "step": 5980
+    },
+    {
+      "epoch": 3.1330539549502356,
+      "grad_norm": 1.234334111213684,
+      "learning_rate": 2.731378095028473e-06,
+      "loss": 0.1028,
+      "step": 5981
+    },
+    {
+      "epoch": 3.1330539549502356,
+      "loss_partial": 0.046142578125,
+      "step": 5981
+    },
+    {
+      "epoch": 3.1330539549502356,
+      "loss_partial": 0.0517578125,
+      "step": 5981
+    },
+    {
+      "epoch": 3.1335777894185437,
+      "grad_norm": 1.5545971393585205,
+      "learning_rate": 2.7282391668292816e-06,
+      "loss": 0.1025,
+      "step": 5982
+    },
+    {
+      "epoch": 3.1335777894185437,
+      "loss_partial": 0.060546875,
+      "step": 5982
+    },
+    {
+      "epoch": 3.1335777894185437,
+      "loss_partial": 0.04638671875,
+      "step": 5982
+    },
+    {
+      "epoch": 3.1341016238868518,
+      "grad_norm": 1.2432366609573364,
+      "learning_rate": 2.725101758383134e-06,
+      "loss": 0.1017,
+      "step": 5983
+    },
+    {
+      "epoch": 3.1341016238868518,
+      "loss_partial": 0.050048828125,
+      "step": 5983
+    },
+    {
+      "epoch": 3.1341016238868518,
+      "loss_partial": 0.052978515625,
+      "step": 5983
+    },
+    {
+      "epoch": 3.13462545835516,
+      "grad_norm": 1.3854490518569946,
+      "learning_rate": 2.721965870345735e-06,
+      "loss": 0.0988,
+      "step": 5984
+    },
+    {
+      "epoch": 3.13462545835516,
+      "loss_partial": 0.047119140625,
+      "step": 5984
+    },
+    {
+      "epoch": 3.13462545835516,
+      "loss_partial": 0.05908203125,
+      "step": 5984
+    },
+    {
+      "epoch": 3.135149292823468,
+      "grad_norm": 1.5799238681793213,
+      "learning_rate": 2.7188315033724643e-06,
+      "loss": 0.1037,
+      "step": 5985
+    },
+    {
+      "epoch": 3.135149292823468,
+      "loss_partial": 0.046875,
+      "step": 5985
+    },
+    {
+      "epoch": 3.135149292823468,
+      "loss_partial": 0.05126953125,
+      "step": 5985
+    },
+    {
+      "epoch": 3.135673127291776,
+      "grad_norm": 1.175999641418457,
+      "learning_rate": 2.715698658118383e-06,
+      "loss": 0.0991,
+      "step": 5986
+    },
+    {
+      "epoch": 3.135673127291776,
+      "loss_partial": 0.058349609375,
+      "step": 5986
+    },
+    {
+      "epoch": 3.135673127291776,
+      "loss_partial": 0.0400390625,
+      "step": 5986
+    },
+    {
+      "epoch": 3.136196961760084,
+      "grad_norm": 1.3531726598739624,
+      "learning_rate": 2.7125673352382376e-06,
+      "loss": 0.0952,
+      "step": 5987
+    },
+    {
+      "epoch": 3.136196961760084,
+      "loss_partial": 0.045166015625,
+      "step": 5987
+    },
+    {
+      "epoch": 3.136196961760084,
+      "loss_partial": 0.05029296875,
+      "step": 5987
+    },
+    {
+      "epoch": 3.136720796228392,
+      "grad_norm": 1.4283615350723267,
+      "learning_rate": 2.7094375353864543e-06,
+      "loss": 0.101,
+      "step": 5988
+    },
+    {
+      "epoch": 3.136720796228392,
+      "loss_partial": 0.042236328125,
+      "step": 5988
+    },
+    {
+      "epoch": 3.136720796228392,
+      "loss_partial": 0.044189453125,
+      "step": 5988
+    },
+    {
+      "epoch": 3.1372446306966997,
+      "grad_norm": 1.1038031578063965,
+      "learning_rate": 2.7063092592171437e-06,
+      "loss": 0.1005,
+      "step": 5989
+    },
+    {
+      "epoch": 3.1372446306966997,
+      "loss_partial": 0.048583984375,
+      "step": 5989
+    },
+    {
+      "epoch": 3.1372446306966997,
+      "loss_partial": 0.0537109375,
+      "step": 5989
+    },
+    {
+      "epoch": 3.1377684651650077,
+      "grad_norm": 1.0984630584716797,
+      "learning_rate": 2.7031825073840947e-06,
+      "loss": 0.1,
+      "step": 5990
+    },
+    {
+      "epoch": 3.1377684651650077,
+      "loss_partial": 0.045654296875,
+      "step": 5990
+    },
+    {
+      "epoch": 3.1377684651650077,
+      "loss_partial": 0.057373046875,
+      "step": 5990
+    },
+    {
+      "epoch": 3.138292299633316,
+      "grad_norm": 1.2625099420547485,
+      "learning_rate": 2.7000572805407753e-06,
+      "loss": 0.102,
+      "step": 5991
+    },
+    {
+      "epoch": 3.138292299633316,
+      "loss_partial": 0.058837890625,
+      "step": 5991
+    },
+    {
+      "epoch": 3.138292299633316,
+      "loss_partial": 0.0537109375,
+      "step": 5991
+    },
+    {
+      "epoch": 3.138816134101624,
+      "grad_norm": 1.0609500408172607,
+      "learning_rate": 2.6969335793403475e-06,
+      "loss": 0.1013,
+      "step": 5992
+    },
+    {
+      "epoch": 3.138816134101624,
+      "loss_partial": 0.048828125,
+      "step": 5992
+    },
+    {
+      "epoch": 3.138816134101624,
+      "loss_partial": 0.043701171875,
+      "step": 5992
+    },
+    {
+      "epoch": 3.139339968569932,
+      "grad_norm": 0.9822592735290527,
+      "learning_rate": 2.69381140443564e-06,
+      "loss": 0.1006,
+      "step": 5993
+    },
+    {
+      "epoch": 3.139339968569932,
+      "loss_partial": 0.048095703125,
+      "step": 5993
+    },
+    {
+      "epoch": 3.139339968569932,
+      "loss_partial": 0.052734375,
+      "step": 5993
+    },
+    {
+      "epoch": 3.13986380303824,
+      "grad_norm": 1.121717929840088,
+      "learning_rate": 2.6906907564791693e-06,
+      "loss": 0.0996,
+      "step": 5994
+    },
+    {
+      "epoch": 3.13986380303824,
+      "loss_partial": 0.046875,
+      "step": 5994
+    },
+    {
+      "epoch": 3.13986380303824,
+      "loss_partial": 0.050048828125,
+      "step": 5994
+    },
+    {
+      "epoch": 3.140387637506548,
+      "grad_norm": 1.157188057899475,
+      "learning_rate": 2.6875716361231276e-06,
+      "loss": 0.0988,
+      "step": 5995
+    },
+    {
+      "epoch": 3.140387637506548,
+      "loss_partial": 0.059814453125,
+      "step": 5995
+    },
+    {
+      "epoch": 3.140387637506548,
+      "loss_partial": 0.04833984375,
+      "step": 5995
+    },
+    {
+      "epoch": 3.140911471974856,
+      "grad_norm": 0.9196450710296631,
+      "learning_rate": 2.6844540440193988e-06,
+      "loss": 0.101,
+      "step": 5996
+    },
+    {
+      "epoch": 3.140911471974856,
+      "loss_partial": 0.04931640625,
+      "step": 5996
+    },
+    {
+      "epoch": 3.140911471974856,
+      "loss_partial": 0.047119140625,
+      "step": 5996
+    },
+    {
+      "epoch": 3.141435306443164,
+      "grad_norm": 1.1844104528427124,
+      "learning_rate": 2.681337980819536e-06,
+      "loss": 0.1039,
+      "step": 5997
+    },
+    {
+      "epoch": 3.141435306443164,
+      "loss_partial": 0.047607421875,
+      "step": 5997
+    },
+    {
+      "epoch": 3.141435306443164,
+      "loss_partial": 0.045654296875,
+      "step": 5997
+    },
+    {
+      "epoch": 3.1419591409114718,
+      "grad_norm": 1.3187583684921265,
+      "learning_rate": 2.6782234471747794e-06,
+      "loss": 0.0996,
+      "step": 5998
+    },
+    {
+      "epoch": 3.1419591409114718,
+      "loss_partial": 0.059814453125,
+      "step": 5998
+    },
+    {
+      "epoch": 3.1419591409114718,
+      "loss_partial": 0.0400390625,
+      "step": 5998
+    },
+    {
+      "epoch": 3.14248297537978,
+      "grad_norm": 1.4739878177642822,
+      "learning_rate": 2.6751104437360444e-06,
+      "loss": 0.0981,
+      "step": 5999
+    },
+    {
+      "epoch": 3.14248297537978,
+      "loss_partial": 0.05224609375,
+      "step": 5999
+    },
+    {
+      "epoch": 3.14248297537978,
+      "loss_partial": 0.05029296875,
+      "step": 5999
+    },
+    {
+      "epoch": 3.143006809848088,
+      "grad_norm": 1.4163508415222168,
+      "learning_rate": 2.671998971153932e-06,
+      "loss": 0.0981,
+      "step": 6000
+    },
+    {
+      "epoch": 3.143006809848088,
+      "loss_partial": 0.05712890625,
+      "step": 6000
+    },
+    {
+      "epoch": 3.143006809848088,
+      "loss_partial": 0.04541015625,
+      "step": 6000
+    },
+    {
+      "epoch": 3.143530644316396,
+      "grad_norm": 1.3671752214431763,
+      "learning_rate": 2.6688890300787184e-06,
+      "loss": 0.1006,
+      "step": 6001
+    },
+    {
+      "epoch": 3.143530644316396,
+      "loss_partial": 0.0419921875,
+      "step": 6001
+    },
+    {
+      "epoch": 3.143530644316396,
+      "loss_partial": 0.051513671875,
+      "step": 6001
+    },
+    {
+      "epoch": 3.144054478784704,
+      "grad_norm": 1.453458547592163,
+      "learning_rate": 2.6657806211603644e-06,
+      "loss": 0.101,
+      "step": 6002
+    },
+    {
+      "epoch": 3.144054478784704,
+      "loss_partial": 0.05029296875,
+      "step": 6002
+    },
+    {
+      "epoch": 3.144054478784704,
+      "loss_partial": 0.044921875,
+      "step": 6002
+    },
+    {
+      "epoch": 3.144578313253012,
+      "grad_norm": 1.5930088758468628,
+      "learning_rate": 2.662673745048503e-06,
+      "loss": 0.0981,
+      "step": 6003
+    },
+    {
+      "epoch": 3.144578313253012,
+      "loss_partial": 0.053955078125,
+      "step": 6003
+    },
+    {
+      "epoch": 3.144578313253012,
+      "loss_partial": 0.05322265625,
+      "step": 6003
+    },
+    {
+      "epoch": 3.14510214772132,
+      "grad_norm": 1.0809069871902466,
+      "learning_rate": 2.6595684023924607e-06,
+      "loss": 0.1012,
+      "step": 6004
+    },
+    {
+      "epoch": 3.14510214772132,
+      "loss_partial": 0.053466796875,
+      "step": 6004
+    },
+    {
+      "epoch": 3.14510214772132,
+      "loss_partial": 0.051513671875,
+      "step": 6004
+    },
+    {
+      "epoch": 3.145625982189628,
+      "grad_norm": 1.5290286540985107,
+      "learning_rate": 2.656464593841229e-06,
+      "loss": 0.0997,
+      "step": 6005
+    },
+    {
+      "epoch": 3.145625982189628,
+      "loss_partial": 0.052978515625,
+      "step": 6005
+    },
+    {
+      "epoch": 3.145625982189628,
+      "loss_partial": 0.046630859375,
+      "step": 6005
+    },
+    {
+      "epoch": 3.1461498166579362,
+      "grad_norm": 1.154188871383667,
+      "learning_rate": 2.6533623200434867e-06,
+      "loss": 0.0979,
+      "step": 6006
+    },
+    {
+      "epoch": 3.1461498166579362,
+      "loss_partial": 0.055419921875,
+      "step": 6006
+    },
+    {
+      "epoch": 3.1461498166579362,
+      "loss_partial": 0.043212890625,
+      "step": 6006
+    },
+    {
+      "epoch": 3.1466736511262443,
+      "grad_norm": 1.0074206590652466,
+      "learning_rate": 2.6502615816475865e-06,
+      "loss": 0.0982,
+      "step": 6007
+    },
+    {
+      "epoch": 3.1466736511262443,
+      "loss_partial": 0.05078125,
+      "step": 6007
+    },
+    {
+      "epoch": 3.1466736511262443,
+      "loss_partial": 0.053466796875,
+      "step": 6007
+    },
+    {
+      "epoch": 3.147197485594552,
+      "grad_norm": 1.384070873260498,
+      "learning_rate": 2.647162379301569e-06,
+      "loss": 0.1001,
+      "step": 6008
+    },
+    {
+      "epoch": 3.147197485594552,
+      "loss_partial": 0.047607421875,
+      "step": 6008
+    },
+    {
+      "epoch": 3.147197485594552,
+      "loss_partial": 0.057373046875,
+      "step": 6008
+    },
+    {
+      "epoch": 3.14772132006286,
+      "grad_norm": 1.3260118961334229,
+      "learning_rate": 2.6440647136531463e-06,
+      "loss": 0.1045,
+      "step": 6009
+    },
+    {
+      "epoch": 3.14772132006286,
+      "loss_partial": 0.051025390625,
+      "step": 6009
+    },
+    {
+      "epoch": 3.14772132006286,
+      "loss_partial": 0.05078125,
+      "step": 6009
+    },
+    {
+      "epoch": 3.148245154531168,
+      "grad_norm": 1.3414721488952637,
+      "learning_rate": 2.6409685853497103e-06,
+      "loss": 0.0992,
+      "step": 6010
+    },
+    {
+      "epoch": 3.148245154531168,
+      "loss_partial": 0.043212890625,
+      "step": 6010
+    },
+    {
+      "epoch": 3.148245154531168,
+      "loss_partial": 0.0419921875,
+      "step": 6010
+    },
+    {
+      "epoch": 3.148768988999476,
+      "grad_norm": 1.2792372703552246,
+      "learning_rate": 2.6378739950383325e-06,
+      "loss": 0.1009,
+      "step": 6011
+    },
+    {
+      "epoch": 3.148768988999476,
+      "loss_partial": 0.050048828125,
+      "step": 6011
+    },
+    {
+      "epoch": 3.148768988999476,
+      "loss_partial": 0.051513671875,
+      "step": 6011
+    },
+    {
+      "epoch": 3.149292823467784,
+      "grad_norm": 1.18464994430542,
+      "learning_rate": 2.6347809433657647e-06,
+      "loss": 0.0985,
+      "step": 6012
+    },
+    {
+      "epoch": 3.149292823467784,
+      "loss_partial": 0.054443359375,
+      "step": 6012
+    },
+    {
+      "epoch": 3.149292823467784,
+      "loss_partial": 0.044189453125,
+      "step": 6012
+    },
+    {
+      "epoch": 3.1498166579360922,
+      "grad_norm": 1.8651676177978516,
+      "learning_rate": 2.6316894309784337e-06,
+      "loss": 0.1046,
+      "step": 6013
+    },
+    {
+      "epoch": 3.1498166579360922,
+      "loss_partial": 0.052490234375,
+      "step": 6013
+    },
+    {
+      "epoch": 3.1498166579360922,
+      "loss_partial": 0.048095703125,
+      "step": 6013
+    },
+    {
+      "epoch": 3.1503404924044003,
+      "grad_norm": 1.7502732276916504,
+      "learning_rate": 2.6285994585224474e-06,
+      "loss": 0.0988,
+      "step": 6014
+    },
+    {
+      "epoch": 3.1503404924044003,
+      "loss_partial": 0.048828125,
+      "step": 6014
+    },
+    {
+      "epoch": 3.1503404924044003,
+      "loss_partial": 0.0439453125,
+      "step": 6014
+    },
+    {
+      "epoch": 3.1508643268727083,
+      "grad_norm": 1.2539247274398804,
+      "learning_rate": 2.625511026643589e-06,
+      "loss": 0.0979,
+      "step": 6015
+    },
+    {
+      "epoch": 3.1508643268727083,
+      "loss_partial": 0.046630859375,
+      "step": 6015
+    },
+    {
+      "epoch": 3.1508643268727083,
+      "loss_partial": 0.0419921875,
+      "step": 6015
+    },
+    {
+      "epoch": 3.1513881613410164,
+      "grad_norm": 1.7227716445922852,
+      "learning_rate": 2.622424135987326e-06,
+      "loss": 0.0963,
+      "step": 6016
+    },
+    {
+      "epoch": 3.1513881613410164,
+      "loss_partial": 0.04638671875,
+      "step": 6016
+    },
+    {
+      "epoch": 3.1513881613410164,
+      "loss_partial": 0.046142578125,
+      "step": 6016
+    },
+    {
+      "epoch": 3.151911995809324,
+      "grad_norm": 1.546772837638855,
+      "learning_rate": 2.6193387871987964e-06,
+      "loss": 0.1024,
+      "step": 6017
+    },
+    {
+      "epoch": 3.151911995809324,
+      "loss_partial": 0.052978515625,
+      "step": 6017
+    },
+    {
+      "epoch": 3.151911995809324,
+      "loss_partial": 0.05419921875,
+      "step": 6017
+    },
+    {
+      "epoch": 3.152435830277632,
+      "grad_norm": 1.1223431825637817,
+      "learning_rate": 2.61625498092282e-06,
+      "loss": 0.0986,
+      "step": 6018
+    },
+    {
+      "epoch": 3.152435830277632,
+      "loss_partial": 0.05126953125,
+      "step": 6018
+    },
+    {
+      "epoch": 3.152435830277632,
+      "loss_partial": 0.04443359375,
+      "step": 6018
+    },
+    {
+      "epoch": 3.15295966474594,
+      "grad_norm": 1.4924914836883545,
+      "learning_rate": 2.61317271780389e-06,
+      "loss": 0.0992,
+      "step": 6019
+    },
+    {
+      "epoch": 3.15295966474594,
+      "loss_partial": 0.04931640625,
+      "step": 6019
+    },
+    {
+      "epoch": 3.15295966474594,
+      "loss_partial": 0.04736328125,
+      "step": 6019
+    },
+    {
+      "epoch": 3.153483499214248,
+      "grad_norm": 1.1682357788085938,
+      "learning_rate": 2.610091998486186e-06,
+      "loss": 0.1011,
+      "step": 6020
+    },
+    {
+      "epoch": 3.153483499214248,
+      "loss_partial": 0.056884765625,
+      "step": 6020
+    },
+    {
+      "epoch": 3.153483499214248,
+      "loss_partial": 0.046630859375,
+      "step": 6020
+    },
+    {
+      "epoch": 3.1540073336825563,
+      "grad_norm": 1.2353403568267822,
+      "learning_rate": 2.6070128236135574e-06,
+      "loss": 0.0996,
+      "step": 6021
+    },
+    {
+      "epoch": 3.1540073336825563,
+      "loss_partial": 0.053466796875,
+      "step": 6021
+    },
+    {
+      "epoch": 3.1540073336825563,
+      "loss_partial": 0.049072265625,
+      "step": 6021
+    },
+    {
+      "epoch": 3.1545311681508643,
+      "grad_norm": 1.408676266670227,
+      "learning_rate": 2.603935193829531e-06,
+      "loss": 0.1031,
+      "step": 6022
+    },
+    {
+      "epoch": 3.1545311681508643,
+      "loss_partial": 0.058837890625,
+      "step": 6022
+    },
+    {
+      "epoch": 3.1545311681508643,
+      "loss_partial": 0.042724609375,
+      "step": 6022
+    },
+    {
+      "epoch": 3.1550550026191724,
+      "grad_norm": 1.0622886419296265,
+      "learning_rate": 2.600859109777313e-06,
+      "loss": 0.0976,
+      "step": 6023
+    },
+    {
+      "epoch": 3.1550550026191724,
+      "loss_partial": 0.049560546875,
+      "step": 6023
+    },
+    {
+      "epoch": 3.1550550026191724,
+      "loss_partial": 0.05126953125,
+      "step": 6023
+    },
+    {
+      "epoch": 3.1555788370874804,
+      "grad_norm": 1.2387129068374634,
+      "learning_rate": 2.5977845720997863e-06,
+      "loss": 0.1017,
+      "step": 6024
+    },
+    {
+      "epoch": 3.1555788370874804,
+      "loss_partial": 0.05859375,
+      "step": 6024
+    },
+    {
+      "epoch": 3.1555788370874804,
+      "loss_partial": 0.04638671875,
+      "step": 6024
+    },
+    {
+      "epoch": 3.1561026715557885,
+      "grad_norm": 1.391801357269287,
+      "learning_rate": 2.594711581439511e-06,
+      "loss": 0.1004,
+      "step": 6025
+    },
+    {
+      "epoch": 3.1561026715557885,
+      "loss_partial": 0.046630859375,
+      "step": 6025
+    },
+    {
+      "epoch": 3.1561026715557885,
+      "loss_partial": 0.04345703125,
+      "step": 6025
+    },
+    {
+      "epoch": 3.1566265060240966,
+      "grad_norm": 1.3221577405929565,
+      "learning_rate": 2.591640138438721e-06,
+      "loss": 0.1013,
+      "step": 6026
+    },
+    {
+      "epoch": 3.1566265060240966,
+      "loss_partial": 0.05810546875,
+      "step": 6026
+    },
+    {
+      "epoch": 3.1566265060240966,
+      "loss_partial": 0.0478515625,
+      "step": 6026
+    },
+    {
+      "epoch": 3.1571503404924046,
+      "grad_norm": 1.0809216499328613,
+      "learning_rate": 2.588570243739328e-06,
+      "loss": 0.0992,
+      "step": 6027
+    },
+    {
+      "epoch": 3.1571503404924046,
+      "loss_partial": 0.043212890625,
+      "step": 6027
+    },
+    {
+      "epoch": 3.1571503404924046,
+      "loss_partial": 0.044677734375,
+      "step": 6027
+    },
+    {
+      "epoch": 3.1576741749607122,
+      "grad_norm": 1.289344310760498,
+      "learning_rate": 2.585501897982925e-06,
+      "loss": 0.0987,
+      "step": 6028
+    },
+    {
+      "epoch": 3.1576741749607122,
+      "loss_partial": 0.04443359375,
+      "step": 6028
+    },
+    {
+      "epoch": 3.1576741749607122,
+      "loss_partial": 0.059326171875,
+      "step": 6028
+    },
+    {
+      "epoch": 3.1581980094290203,
+      "grad_norm": 1.2801799774169922,
+      "learning_rate": 2.5824351018107763e-06,
+      "loss": 0.1012,
+      "step": 6029
+    },
+    {
+      "epoch": 3.1581980094290203,
+      "loss_partial": 0.045654296875,
+      "step": 6029
+    },
+    {
+      "epoch": 3.1581980094290203,
+      "loss_partial": 0.04248046875,
+      "step": 6029
+    },
+    {
+      "epoch": 3.1587218438973284,
+      "grad_norm": 1.2109559774398804,
+      "learning_rate": 2.5793698558638227e-06,
+      "loss": 0.097,
+      "step": 6030
+    },
+    {
+      "epoch": 3.1587218438973284,
+      "loss_partial": 0.049560546875,
+      "step": 6030
+    },
+    {
+      "epoch": 3.1587218438973284,
+      "loss_partial": 0.0556640625,
+      "step": 6030
+    },
+    {
+      "epoch": 3.1592456783656364,
+      "grad_norm": 1.1886215209960938,
+      "learning_rate": 2.5763061607826776e-06,
+      "loss": 0.1038,
+      "step": 6031
+    },
+    {
+      "epoch": 3.1592456783656364,
+      "loss_partial": 0.05322265625,
+      "step": 6031
+    },
+    {
+      "epoch": 3.1592456783656364,
+      "loss_partial": 0.049560546875,
+      "step": 6031
+    },
+    {
+      "epoch": 3.1597695128339445,
+      "grad_norm": 1.3629937171936035,
+      "learning_rate": 2.573244017207642e-06,
+      "loss": 0.1,
+      "step": 6032
+    },
+    {
+      "epoch": 3.1597695128339445,
+      "loss_partial": 0.045166015625,
+      "step": 6032
+    },
+    {
+      "epoch": 3.1597695128339445,
+      "loss_partial": 0.04296875,
+      "step": 6032
+    },
+    {
+      "epoch": 3.1602933473022525,
+      "grad_norm": 1.2046239376068115,
+      "learning_rate": 2.5701834257786807e-06,
+      "loss": 0.0967,
+      "step": 6033
+    },
+    {
+      "epoch": 3.1602933473022525,
+      "loss_partial": 0.049560546875,
+      "step": 6033
+    },
+    {
+      "epoch": 3.1602933473022525,
+      "loss_partial": 0.06298828125,
+      "step": 6033
+    },
+    {
+      "epoch": 3.1608171817705606,
+      "grad_norm": 1.0651904344558716,
+      "learning_rate": 2.567124387135439e-06,
+      "loss": 0.1003,
+      "step": 6034
+    },
+    {
+      "epoch": 3.1608171817705606,
+      "loss_partial": 0.06201171875,
+      "step": 6034
+    },
+    {
+      "epoch": 3.1608171817705606,
+      "loss_partial": 0.04150390625,
+      "step": 6034
+    },
+    {
+      "epoch": 3.1613410162388687,
+      "grad_norm": 1.2524135112762451,
+      "learning_rate": 2.564066901917237e-06,
+      "loss": 0.1016,
+      "step": 6035
+    },
+    {
+      "epoch": 3.1613410162388687,
+      "loss_partial": 0.044677734375,
+      "step": 6035
+    },
+    {
+      "epoch": 3.1613410162388687,
+      "loss_partial": 0.04736328125,
+      "step": 6035
+    },
+    {
+      "epoch": 3.1618648507071767,
+      "grad_norm": 1.813895344734192,
+      "learning_rate": 2.5610109707630703e-06,
+      "loss": 0.1051,
+      "step": 6036
+    },
+    {
+      "epoch": 3.1618648507071767,
+      "loss_partial": 0.053955078125,
+      "step": 6036
+    },
+    {
+      "epoch": 3.1618648507071767,
+      "loss_partial": 0.05322265625,
+      "step": 6036
+    },
+    {
+      "epoch": 3.1623886851754843,
+      "grad_norm": 1.0627257823944092,
+      "learning_rate": 2.5579565943116092e-06,
+      "loss": 0.0979,
+      "step": 6037
+    },
+    {
+      "epoch": 3.1623886851754843,
+      "loss_partial": 0.04736328125,
+      "step": 6037
+    },
+    {
+      "epoch": 3.1623886851754843,
+      "loss_partial": 0.05712890625,
+      "step": 6037
+    },
+    {
+      "epoch": 3.1629125196437924,
+      "grad_norm": 1.3087798357009888,
+      "learning_rate": 2.5549037732011995e-06,
+      "loss": 0.1012,
+      "step": 6038
+    },
+    {
+      "epoch": 3.1629125196437924,
+      "loss_partial": 0.053466796875,
+      "step": 6038
+    },
+    {
+      "epoch": 3.1629125196437924,
+      "loss_partial": 0.0546875,
+      "step": 6038
+    },
+    {
+      "epoch": 3.1634363541121004,
+      "grad_norm": 1.602848768234253,
+      "learning_rate": 2.5518525080698618e-06,
+      "loss": 0.1011,
+      "step": 6039
+    },
+    {
+      "epoch": 3.1634363541121004,
+      "loss_partial": 0.05029296875,
+      "step": 6039
+    },
+    {
+      "epoch": 3.1634363541121004,
+      "loss_partial": 0.046875,
+      "step": 6039
+    },
+    {
+      "epoch": 3.1639601885804085,
+      "grad_norm": 0.9688125252723694,
+      "learning_rate": 2.5488027995552887e-06,
+      "loss": 0.1018,
+      "step": 6040
+    },
+    {
+      "epoch": 3.1639601885804085,
+      "loss_partial": 0.047119140625,
+      "step": 6040
+    },
+    {
+      "epoch": 3.1639601885804085,
+      "loss_partial": 0.05126953125,
+      "step": 6040
+    },
+    {
+      "epoch": 3.1644840230487166,
+      "grad_norm": 1.3746036291122437,
+      "learning_rate": 2.545754648294858e-06,
+      "loss": 0.1032,
+      "step": 6041
+    },
+    {
+      "epoch": 3.1644840230487166,
+      "loss_partial": 0.0498046875,
+      "step": 6041
+    },
+    {
+      "epoch": 3.1644840230487166,
+      "loss_partial": 0.048583984375,
+      "step": 6041
+    },
+    {
+      "epoch": 3.1650078575170246,
+      "grad_norm": 1.6205936670303345,
+      "learning_rate": 2.542708054925609e-06,
+      "loss": 0.095,
+      "step": 6042
+    },
+    {
+      "epoch": 3.1650078575170246,
+      "loss_partial": 0.053955078125,
+      "step": 6042
+    },
+    {
+      "epoch": 3.1650078575170246,
+      "loss_partial": 0.047119140625,
+      "step": 6042
+    },
+    {
+      "epoch": 3.1655316919853327,
+      "grad_norm": 0.8193942904472351,
+      "learning_rate": 2.539663020084262e-06,
+      "loss": 0.1034,
+      "step": 6043
+    },
+    {
+      "epoch": 3.1655316919853327,
+      "loss_partial": 0.048583984375,
+      "step": 6043
+    },
+    {
+      "epoch": 3.1655316919853327,
+      "loss_partial": 0.045654296875,
+      "step": 6043
+    },
+    {
+      "epoch": 3.1660555264536407,
+      "grad_norm": 1.346250057220459,
+      "learning_rate": 2.5366195444072083e-06,
+      "loss": 0.0977,
+      "step": 6044
+    },
+    {
+      "epoch": 3.1660555264536407,
+      "loss_partial": 0.0439453125,
+      "step": 6044
+    },
+    {
+      "epoch": 3.1660555264536407,
+      "loss_partial": 0.051025390625,
+      "step": 6044
+    },
+    {
+      "epoch": 3.166579360921949,
+      "grad_norm": 1.1226085424423218,
+      "learning_rate": 2.5335776285305204e-06,
+      "loss": 0.0975,
+      "step": 6045
+    },
+    {
+      "epoch": 3.166579360921949,
+      "loss_partial": 0.0576171875,
+      "step": 6045
+    },
+    {
+      "epoch": 3.166579360921949,
+      "loss_partial": 0.038818359375,
+      "step": 6045
+    },
+    {
+      "epoch": 3.167103195390257,
+      "grad_norm": 1.332741141319275,
+      "learning_rate": 2.530537273089937e-06,
+      "loss": 0.0981,
+      "step": 6046
+    },
+    {
+      "epoch": 3.167103195390257,
+      "loss_partial": 0.05712890625,
+      "step": 6046
+    },
+    {
+      "epoch": 3.167103195390257,
+      "loss_partial": 0.0498046875,
+      "step": 6046
+    },
+    {
+      "epoch": 3.167627029858565,
+      "grad_norm": 1.336945652961731,
+      "learning_rate": 2.5274984787208735e-06,
+      "loss": 0.1035,
+      "step": 6047
+    },
+    {
+      "epoch": 3.167627029858565,
+      "loss_partial": 0.061279296875,
+      "step": 6047
+    },
+    {
+      "epoch": 3.167627029858565,
+      "loss_partial": 0.056640625,
+      "step": 6047
+    },
+    {
+      "epoch": 3.1681508643268725,
+      "grad_norm": 1.325383186340332,
+      "learning_rate": 2.5244612460584194e-06,
+      "loss": 0.102,
+      "step": 6048
+    },
+    {
+      "epoch": 3.1681508643268725,
+      "loss_partial": 0.058349609375,
+      "step": 6048
+    },
+    {
+      "epoch": 3.1681508643268725,
+      "loss_partial": 0.05517578125,
+      "step": 6048
+    },
+    {
+      "epoch": 3.1686746987951806,
+      "grad_norm": 1.25852632522583,
+      "learning_rate": 2.5214255757373373e-06,
+      "loss": 0.1013,
+      "step": 6049
+    },
+    {
+      "epoch": 3.1686746987951806,
+      "loss_partial": 0.054931640625,
+      "step": 6049
+    },
+    {
+      "epoch": 3.1686746987951806,
+      "loss_partial": 0.042724609375,
+      "step": 6049
+    },
+    {
+      "epoch": 3.1691985332634887,
+      "grad_norm": 1.861568808555603,
+      "learning_rate": 2.518391468392065e-06,
+      "loss": 0.0995,
+      "step": 6050
+    },
+    {
+      "epoch": 3.1691985332634887,
+      "loss_partial": 0.05029296875,
+      "step": 6050
+    },
+    {
+      "epoch": 3.1691985332634887,
+      "loss_partial": 0.045166015625,
+      "step": 6050
+    },
+    {
+      "epoch": 3.1697223677317967,
+      "grad_norm": 1.632797122001648,
+      "learning_rate": 2.51535892465671e-06,
+      "loss": 0.1001,
+      "step": 6051
+    },
+    {
+      "epoch": 3.1697223677317967,
+      "loss_partial": 0.053955078125,
+      "step": 6051
+    },
+    {
+      "epoch": 3.1697223677317967,
+      "loss_partial": 0.05908203125,
+      "step": 6051
+    },
+    {
+      "epoch": 3.170246202200105,
+      "grad_norm": 1.507678747177124,
+      "learning_rate": 2.512327945165053e-06,
+      "loss": 0.1006,
+      "step": 6052
+    },
+    {
+      "epoch": 3.170246202200105,
+      "loss_partial": 0.05322265625,
+      "step": 6052
+    },
+    {
+      "epoch": 3.170246202200105,
+      "loss_partial": 0.04638671875,
+      "step": 6052
+    },
+    {
+      "epoch": 3.170770036668413,
+      "grad_norm": 1.5381664037704468,
+      "learning_rate": 2.509298530550557e-06,
+      "loss": 0.1024,
+      "step": 6053
+    },
+    {
+      "epoch": 3.170770036668413,
+      "loss_partial": 0.054931640625,
+      "step": 6053
+    },
+    {
+      "epoch": 3.170770036668413,
+      "loss_partial": 0.044921875,
+      "step": 6053
+    },
+    {
+      "epoch": 3.171293871136721,
+      "grad_norm": 1.1464011669158936,
+      "learning_rate": 2.5062706814463456e-06,
+      "loss": 0.1011,
+      "step": 6054
+    },
+    {
+      "epoch": 3.171293871136721,
+      "loss_partial": 0.042236328125,
+      "step": 6054
+    },
+    {
+      "epoch": 3.171293871136721,
+      "loss_partial": 0.046142578125,
+      "step": 6054
+    },
+    {
+      "epoch": 3.171817705605029,
+      "grad_norm": 0.9610854983329773,
+      "learning_rate": 2.5032443984852228e-06,
+      "loss": 0.1009,
+      "step": 6055
+    },
+    {
+      "epoch": 3.171817705605029,
+      "loss_partial": 0.04931640625,
+      "step": 6055
+    },
+    {
+      "epoch": 3.171817705605029,
+      "loss_partial": 0.049072265625,
+      "step": 6055
+    },
+    {
+      "epoch": 3.172341540073337,
+      "grad_norm": 1.2943073511123657,
+      "learning_rate": 2.5002196822996593e-06,
+      "loss": 0.1012,
+      "step": 6056
+    },
+    {
+      "epoch": 3.172341540073337,
+      "loss_partial": 0.038818359375,
+      "step": 6056
+    },
+    {
+      "epoch": 3.172341540073337,
+      "loss_partial": 0.049072265625,
+      "step": 6056
+    },
+    {
+      "epoch": 3.1728653745416446,
+      "grad_norm": 1.1687554121017456,
+      "learning_rate": 2.4971965335218075e-06,
+      "loss": 0.0982,
+      "step": 6057
+    },
+    {
+      "epoch": 3.1728653745416446,
+      "loss_partial": 0.041748046875,
+      "step": 6057
+    },
+    {
+      "epoch": 3.1728653745416446,
+      "loss_partial": 0.04833984375,
+      "step": 6057
+    },
+    {
+      "epoch": 3.1733892090099527,
+      "grad_norm": 1.2321125268936157,
+      "learning_rate": 2.4941749527834847e-06,
+      "loss": 0.1038,
+      "step": 6058
+    },
+    {
+      "epoch": 3.1733892090099527,
+      "loss_partial": 0.06298828125,
+      "step": 6058
+    },
+    {
+      "epoch": 3.1733892090099527,
+      "loss_partial": 0.046875,
+      "step": 6058
+    },
+    {
+      "epoch": 3.1739130434782608,
+      "grad_norm": 1.232833981513977,
+      "learning_rate": 2.4911549407161816e-06,
+      "loss": 0.1019,
+      "step": 6059
+    },
+    {
+      "epoch": 3.1739130434782608,
+      "loss_partial": 0.052734375,
+      "step": 6059
+    },
+    {
+      "epoch": 3.1739130434782608,
+      "loss_partial": 0.047119140625,
+      "step": 6059
+    },
+    {
+      "epoch": 3.174436877946569,
+      "grad_norm": 1.0264990329742432,
+      "learning_rate": 2.4881364979510635e-06,
+      "loss": 0.0976,
+      "step": 6060
+    },
+    {
+      "epoch": 3.174436877946569,
+      "loss_partial": 0.056396484375,
+      "step": 6060
+    },
+    {
+      "epoch": 3.174436877946569,
+      "loss_partial": 0.05078125,
+      "step": 6060
+    },
+    {
+      "epoch": 3.174960712414877,
+      "grad_norm": 1.0709972381591797,
+      "learning_rate": 2.4851196251189646e-06,
+      "loss": 0.1031,
+      "step": 6061
+    },
+    {
+      "epoch": 3.174960712414877,
+      "loss_partial": 0.044677734375,
+      "step": 6061
+    },
+    {
+      "epoch": 3.174960712414877,
+      "loss_partial": 0.04736328125,
+      "step": 6061
+    },
+    {
+      "epoch": 3.175484546883185,
+      "grad_norm": 1.2874646186828613,
+      "learning_rate": 2.4821043228503937e-06,
+      "loss": 0.1025,
+      "step": 6062
+    },
+    {
+      "epoch": 3.175484546883185,
+      "loss_partial": 0.047119140625,
+      "step": 6062
+    },
+    {
+      "epoch": 3.175484546883185,
+      "loss_partial": 0.051513671875,
+      "step": 6062
+    },
+    {
+      "epoch": 3.176008381351493,
+      "grad_norm": 1.0934619903564453,
+      "learning_rate": 2.47909059177553e-06,
+      "loss": 0.1013,
+      "step": 6063
+    },
+    {
+      "epoch": 3.176008381351493,
+      "loss_partial": 0.04638671875,
+      "step": 6063
+    },
+    {
+      "epoch": 3.176008381351493,
+      "loss_partial": 0.045654296875,
+      "step": 6063
+    },
+    {
+      "epoch": 3.176532215819801,
+      "grad_norm": 1.172136902809143,
+      "learning_rate": 2.4760784325242215e-06,
+      "loss": 0.0946,
+      "step": 6064
+    },
+    {
+      "epoch": 3.176532215819801,
+      "loss_partial": 0.04833984375,
+      "step": 6064
+    },
+    {
+      "epoch": 3.176532215819801,
+      "loss_partial": 0.05322265625,
+      "step": 6064
+    },
+    {
+      "epoch": 3.177056050288109,
+      "grad_norm": 1.0702658891677856,
+      "learning_rate": 2.4730678457259973e-06,
+      "loss": 0.0995,
+      "step": 6065
+    },
+    {
+      "epoch": 3.177056050288109,
+      "loss_partial": 0.05224609375,
+      "step": 6065
+    },
+    {
+      "epoch": 3.177056050288109,
+      "loss_partial": 0.048095703125,
+      "step": 6065
+    },
+    {
+      "epoch": 3.177579884756417,
+      "grad_norm": 1.1645699739456177,
+      "learning_rate": 2.4700588320100484e-06,
+      "loss": 0.0987,
+      "step": 6066
+    },
+    {
+      "epoch": 3.177579884756417,
+      "loss_partial": 0.052734375,
+      "step": 6066
+    },
+    {
+      "epoch": 3.177579884756417,
+      "loss_partial": 0.045654296875,
+      "step": 6066
+    },
+    {
+      "epoch": 3.178103719224725,
+      "grad_norm": 1.2149543762207031,
+      "learning_rate": 2.4670513920052397e-06,
+      "loss": 0.0989,
+      "step": 6067
+    },
+    {
+      "epoch": 3.178103719224725,
+      "loss_partial": 0.04736328125,
+      "step": 6067
+    },
+    {
+      "epoch": 3.178103719224725,
+      "loss_partial": 0.050048828125,
+      "step": 6067
+    },
+    {
+      "epoch": 3.178627553693033,
+      "grad_norm": 1.3197388648986816,
+      "learning_rate": 2.4640455263401043e-06,
+      "loss": 0.1017,
+      "step": 6068
+    },
+    {
+      "epoch": 3.178627553693033,
+      "loss_partial": 0.053955078125,
+      "step": 6068
+    },
+    {
+      "epoch": 3.178627553693033,
+      "loss_partial": 0.044921875,
+      "step": 6068
+    },
+    {
+      "epoch": 3.179151388161341,
+      "grad_norm": 1.13032066822052,
+      "learning_rate": 2.461041235642856e-06,
+      "loss": 0.0992,
+      "step": 6069
+    },
+    {
+      "epoch": 3.179151388161341,
+      "loss_partial": 0.046630859375,
+      "step": 6069
+    },
+    {
+      "epoch": 3.179151388161341,
+      "loss_partial": 0.058349609375,
+      "step": 6069
+    },
+    {
+      "epoch": 3.179675222629649,
+      "grad_norm": 1.136338710784912,
+      "learning_rate": 2.4580385205413705e-06,
+      "loss": 0.1007,
+      "step": 6070
+    },
+    {
+      "epoch": 3.179675222629649,
+      "loss_partial": 0.058837890625,
+      "step": 6070
+    },
+    {
+      "epoch": 3.179675222629649,
+      "loss_partial": 0.049560546875,
+      "step": 6070
+    },
+    {
+      "epoch": 3.180199057097957,
+      "grad_norm": 1.1122711896896362,
+      "learning_rate": 2.4550373816631956e-06,
+      "loss": 0.1004,
+      "step": 6071
+    },
+    {
+      "epoch": 3.180199057097957,
+      "loss_partial": 0.042236328125,
+      "step": 6071
+    },
+    {
+      "epoch": 3.180199057097957,
+      "loss_partial": 0.044921875,
+      "step": 6071
+    },
+    {
+      "epoch": 3.180722891566265,
+      "grad_norm": 1.0494354963302612,
+      "learning_rate": 2.452037819635552e-06,
+      "loss": 0.099,
+      "step": 6072
+    },
+    {
+      "epoch": 3.180722891566265,
+      "loss_partial": 0.04638671875,
+      "step": 6072
+    },
+    {
+      "epoch": 3.180722891566265,
+      "loss_partial": 0.0478515625,
+      "step": 6072
+    },
+    {
+      "epoch": 3.181246726034573,
+      "grad_norm": 1.2492609024047852,
+      "learning_rate": 2.4490398350853284e-06,
+      "loss": 0.0991,
+      "step": 6073
+    },
+    {
+      "epoch": 3.181246726034573,
+      "loss_partial": 0.046875,
+      "step": 6073
+    },
+    {
+      "epoch": 3.181246726034573,
+      "loss_partial": 0.049072265625,
+      "step": 6073
+    },
+    {
+      "epoch": 3.181770560502881,
+      "grad_norm": 1.0975221395492554,
+      "learning_rate": 2.4460434286390855e-06,
+      "loss": 0.0974,
+      "step": 6074
+    },
+    {
+      "epoch": 3.181770560502881,
+      "loss_partial": 0.044189453125,
+      "step": 6074
+    },
+    {
+      "epoch": 3.181770560502881,
+      "loss_partial": 0.045654296875,
+      "step": 6074
+    },
+    {
+      "epoch": 3.1822943949711893,
+      "grad_norm": 1.1606491804122925,
+      "learning_rate": 2.4430486009230536e-06,
+      "loss": 0.0975,
+      "step": 6075
+    },
+    {
+      "epoch": 3.1822943949711893,
+      "loss_partial": 0.048828125,
+      "step": 6075
+    },
+    {
+      "epoch": 3.1822943949711893,
+      "loss_partial": 0.04541015625,
+      "step": 6075
+    },
+    {
+      "epoch": 3.1828182294394973,
+      "grad_norm": 1.4323786497116089,
+      "learning_rate": 2.4400553525631308e-06,
+      "loss": 0.0993,
+      "step": 6076
+    },
+    {
+      "epoch": 3.1828182294394973,
+      "loss_partial": 0.048095703125,
+      "step": 6076
+    },
+    {
+      "epoch": 3.1828182294394973,
+      "loss_partial": 0.0439453125,
+      "step": 6076
+    },
+    {
+      "epoch": 3.183342063907805,
+      "grad_norm": 1.5956512689590454,
+      "learning_rate": 2.437063684184893e-06,
+      "loss": 0.1004,
+      "step": 6077
+    },
+    {
+      "epoch": 3.183342063907805,
+      "loss_partial": 0.046142578125,
+      "step": 6077
+    },
+    {
+      "epoch": 3.183342063907805,
+      "loss_partial": 0.05615234375,
+      "step": 6077
+    },
+    {
+      "epoch": 3.183865898376113,
+      "grad_norm": 1.6313143968582153,
+      "learning_rate": 2.4340735964135766e-06,
+      "loss": 0.1014,
+      "step": 6078
+    },
+    {
+      "epoch": 3.183865898376113,
+      "loss_partial": 0.0537109375,
+      "step": 6078
+    },
+    {
+      "epoch": 3.183865898376113,
+      "loss_partial": 0.048583984375,
+      "step": 6078
+    },
+    {
+      "epoch": 3.184389732844421,
+      "grad_norm": 1.4177100658416748,
+      "learning_rate": 2.4310850898740934e-06,
+      "loss": 0.1015,
+      "step": 6079
+    },
+    {
+      "epoch": 3.184389732844421,
+      "loss_partial": 0.053466796875,
+      "step": 6079
+    },
+    {
+      "epoch": 3.184389732844421,
+      "loss_partial": 0.05712890625,
+      "step": 6079
+    },
+    {
+      "epoch": 3.184913567312729,
+      "grad_norm": 1.4825613498687744,
+      "learning_rate": 2.4280981651910183e-06,
+      "loss": 0.1013,
+      "step": 6080
+    },
+    {
+      "epoch": 3.184913567312729,
+      "loss_partial": 0.05078125,
+      "step": 6080
+    },
+    {
+      "epoch": 3.184913567312729,
+      "loss_partial": 0.0546875,
+      "step": 6080
+    },
+    {
+      "epoch": 3.185437401781037,
+      "grad_norm": 1.4066853523254395,
+      "learning_rate": 2.4251128229886056e-06,
+      "loss": 0.0986,
+      "step": 6081
+    },
+    {
+      "epoch": 3.185437401781037,
+      "loss_partial": 0.04736328125,
+      "step": 6081
+    },
+    {
+      "epoch": 3.185437401781037,
+      "loss_partial": 0.052001953125,
+      "step": 6081
+    },
+    {
+      "epoch": 3.1859612362493452,
+      "grad_norm": 0.9966062903404236,
+      "learning_rate": 2.4221290638907726e-06,
+      "loss": 0.1005,
+      "step": 6082
+    },
+    {
+      "epoch": 3.1859612362493452,
+      "loss_partial": 0.046875,
+      "step": 6082
+    },
+    {
+      "epoch": 3.1859612362493452,
+      "loss_partial": 0.05078125,
+      "step": 6082
+    },
+    {
+      "epoch": 3.1864850707176533,
+      "grad_norm": 1.4384052753448486,
+      "learning_rate": 2.4191468885211032e-06,
+      "loss": 0.0988,
+      "step": 6083
+    },
+    {
+      "epoch": 3.1864850707176533,
+      "loss_partial": 0.042724609375,
+      "step": 6083
+    },
+    {
+      "epoch": 3.1864850707176533,
+      "loss_partial": 0.042724609375,
+      "step": 6083
+    },
+    {
+      "epoch": 3.1870089051859614,
+      "grad_norm": 1.1989418268203735,
+      "learning_rate": 2.4161662975028575e-06,
+      "loss": 0.0994,
+      "step": 6084
+    },
+    {
+      "epoch": 3.1870089051859614,
+      "loss_partial": 0.056396484375,
+      "step": 6084
+    },
+    {
+      "epoch": 3.1870089051859614,
+      "loss_partial": 0.046142578125,
+      "step": 6084
+    },
+    {
+      "epoch": 3.1875327396542694,
+      "grad_norm": 1.4475876092910767,
+      "learning_rate": 2.413187291458957e-06,
+      "loss": 0.1032,
+      "step": 6085
+    },
+    {
+      "epoch": 3.1875327396542694,
+      "loss_partial": 0.046142578125,
+      "step": 6085
+    },
+    {
+      "epoch": 3.1875327396542694,
+      "loss_partial": 0.0478515625,
+      "step": 6085
+    },
+    {
+      "epoch": 3.1880565741225775,
+      "grad_norm": 1.1013391017913818,
+      "learning_rate": 2.410209871011998e-06,
+      "loss": 0.0985,
+      "step": 6086
+    },
+    {
+      "epoch": 3.1880565741225775,
+      "loss_partial": 0.048095703125,
+      "step": 6086
+    },
+    {
+      "epoch": 3.1880565741225775,
+      "loss_partial": 0.04736328125,
+      "step": 6086
+    },
+    {
+      "epoch": 3.188580408590885,
+      "grad_norm": 1.7690268754959106,
+      "learning_rate": 2.4072340367842426e-06,
+      "loss": 0.1002,
+      "step": 6087
+    },
+    {
+      "epoch": 3.188580408590885,
+      "loss_partial": 0.051513671875,
+      "step": 6087
+    },
+    {
+      "epoch": 3.188580408590885,
+      "loss_partial": 0.05615234375,
+      "step": 6087
+    },
+    {
+      "epoch": 3.189104243059193,
+      "grad_norm": 1.6559724807739258,
+      "learning_rate": 2.404259789397618e-06,
+      "loss": 0.1042,
+      "step": 6088
+    },
+    {
+      "epoch": 3.189104243059193,
+      "loss_partial": 0.04931640625,
+      "step": 6088
+    },
+    {
+      "epoch": 3.189104243059193,
+      "loss_partial": 0.05322265625,
+      "step": 6088
+    },
+    {
+      "epoch": 3.189628077527501,
+      "grad_norm": 1.3748916387557983,
+      "learning_rate": 2.4012871294737294e-06,
+      "loss": 0.0993,
+      "step": 6089
+    },
+    {
+      "epoch": 3.189628077527501,
+      "loss_partial": 0.04931640625,
+      "step": 6089
+    },
+    {
+      "epoch": 3.189628077527501,
+      "loss_partial": 0.05615234375,
+      "step": 6089
+    },
+    {
+      "epoch": 3.1901519119958093,
+      "grad_norm": 1.2999852895736694,
+      "learning_rate": 2.3983160576338416e-06,
+      "loss": 0.0974,
+      "step": 6090
+    },
+    {
+      "epoch": 3.1901519119958093,
+      "loss_partial": 0.040283203125,
+      "step": 6090
+    },
+    {
+      "epoch": 3.1901519119958093,
+      "loss_partial": 0.044677734375,
+      "step": 6090
+    },
+    {
+      "epoch": 3.1906757464641173,
+      "grad_norm": 1.4420589208602905,
+      "learning_rate": 2.39534657449889e-06,
+      "loss": 0.0982,
+      "step": 6091
+    },
+    {
+      "epoch": 3.1906757464641173,
+      "loss_partial": 0.057861328125,
+      "step": 6091
+    },
+    {
+      "epoch": 3.1906757464641173,
+      "loss_partial": 0.051025390625,
+      "step": 6091
+    },
+    {
+      "epoch": 3.1911995809324254,
+      "grad_norm": 1.2219464778900146,
+      "learning_rate": 2.39237868068948e-06,
+      "loss": 0.0991,
+      "step": 6092
+    },
+    {
+      "epoch": 3.1911995809324254,
+      "loss_partial": 0.050537109375,
+      "step": 6092
+    },
+    {
+      "epoch": 3.1911995809324254,
+      "loss_partial": 0.049560546875,
+      "step": 6092
+    },
+    {
+      "epoch": 3.1917234154007335,
+      "grad_norm": 1.0844429731369019,
+      "learning_rate": 2.389412376825877e-06,
+      "loss": 0.1018,
+      "step": 6093
+    },
+    {
+      "epoch": 3.1917234154007335,
+      "loss_partial": 0.04638671875,
+      "step": 6093
+    },
+    {
+      "epoch": 3.1917234154007335,
+      "loss_partial": 0.047607421875,
+      "step": 6093
+    },
+    {
+      "epoch": 3.1922472498690415,
+      "grad_norm": 1.3854016065597534,
+      "learning_rate": 2.386447663528029e-06,
+      "loss": 0.0966,
+      "step": 6094
+    },
+    {
+      "epoch": 3.1922472498690415,
+      "loss_partial": 0.052490234375,
+      "step": 6094
+    },
+    {
+      "epoch": 3.1922472498690415,
+      "loss_partial": 0.0498046875,
+      "step": 6094
+    },
+    {
+      "epoch": 3.1927710843373496,
+      "grad_norm": 1.4907557964324951,
+      "learning_rate": 2.383484541415538e-06,
+      "loss": 0.0969,
+      "step": 6095
+    },
+    {
+      "epoch": 3.1927710843373496,
+      "loss_partial": 0.053466796875,
+      "step": 6095
+    },
+    {
+      "epoch": 3.1927710843373496,
+      "loss_partial": 0.055908203125,
+      "step": 6095
+    },
+    {
+      "epoch": 3.193294918805657,
+      "grad_norm": 1.4287786483764648,
+      "learning_rate": 2.3805230111076783e-06,
+      "loss": 0.1006,
+      "step": 6096
+    },
+    {
+      "epoch": 3.193294918805657,
+      "loss_partial": 0.04931640625,
+      "step": 6096
+    },
+    {
+      "epoch": 3.193294918805657,
+      "loss_partial": 0.05419921875,
+      "step": 6096
+    },
+    {
+      "epoch": 3.1938187532739652,
+      "grad_norm": 1.1951297521591187,
+      "learning_rate": 2.3775630732233923e-06,
+      "loss": 0.101,
+      "step": 6097
+    },
+    {
+      "epoch": 3.1938187532739652,
+      "loss_partial": 0.048095703125,
+      "step": 6097
+    },
+    {
+      "epoch": 3.1938187532739652,
+      "loss_partial": 0.0498046875,
+      "step": 6097
+    },
+    {
+      "epoch": 3.1943425877422733,
+      "grad_norm": 1.3997176885604858,
+      "learning_rate": 2.374604728381289e-06,
+      "loss": 0.0999,
+      "step": 6098
+    },
+    {
+      "epoch": 3.1943425877422733,
+      "loss_partial": 0.053466796875,
+      "step": 6098
+    },
+    {
+      "epoch": 3.1943425877422733,
+      "loss_partial": 0.061279296875,
+      "step": 6098
+    },
+    {
+      "epoch": 3.1948664222105814,
+      "grad_norm": 1.840118169784546,
+      "learning_rate": 2.3716479771996424e-06,
+      "loss": 0.0984,
+      "step": 6099
+    },
+    {
+      "epoch": 3.1948664222105814,
+      "loss_partial": 0.048095703125,
+      "step": 6099
+    },
+    {
+      "epoch": 3.1948664222105814,
+      "loss_partial": 0.044677734375,
+      "step": 6099
+    },
+    {
+      "epoch": 3.1953902566788894,
+      "grad_norm": 1.3424919843673706,
+      "learning_rate": 2.3686928202963964e-06,
+      "loss": 0.1037,
+      "step": 6100
+    },
+    {
+      "epoch": 3.1953902566788894,
+      "loss_partial": 0.05224609375,
+      "step": 6100
+    },
+    {
+      "epoch": 3.1953902566788894,
+      "loss_partial": 0.054931640625,
+      "step": 6100
+    },
+    {
+      "epoch": 3.1959140911471975,
+      "grad_norm": 1.5774413347244263,
+      "learning_rate": 2.365739258289158e-06,
+      "loss": 0.1001,
+      "step": 6101
+    },
+    {
+      "epoch": 3.1959140911471975,
+      "loss_partial": 0.044189453125,
+      "step": 6101
+    },
+    {
+      "epoch": 3.1959140911471975,
+      "loss_partial": 0.044189453125,
+      "step": 6101
+    },
+    {
+      "epoch": 3.1964379256155055,
+      "grad_norm": 1.186048150062561,
+      "learning_rate": 2.362787291795208e-06,
+      "loss": 0.0962,
+      "step": 6102
+    },
+    {
+      "epoch": 3.1964379256155055,
+      "loss_partial": 0.04833984375,
+      "step": 6102
+    },
+    {
+      "epoch": 3.1964379256155055,
+      "loss_partial": 0.05078125,
+      "step": 6102
+    },
+    {
+      "epoch": 3.1969617600838136,
+      "grad_norm": 1.3083367347717285,
+      "learning_rate": 2.3598369214314877e-06,
+      "loss": 0.0997,
+      "step": 6103
+    },
+    {
+      "epoch": 3.1969617600838136,
+      "loss_partial": 0.05078125,
+      "step": 6103
+    },
+    {
+      "epoch": 3.1969617600838136,
+      "loss_partial": 0.05322265625,
+      "step": 6103
+    },
+    {
+      "epoch": 3.1974855945521217,
+      "grad_norm": 1.8515959978103638,
+      "learning_rate": 2.3568881478146044e-06,
+      "loss": 0.101,
+      "step": 6104
+    },
+    {
+      "epoch": 3.1974855945521217,
+      "loss_partial": 0.048583984375,
+      "step": 6104
+    },
+    {
+      "epoch": 3.1974855945521217,
+      "loss_partial": 0.04296875,
+      "step": 6104
+    },
+    {
+      "epoch": 3.1980094290204297,
+      "grad_norm": 1.2150336503982544,
+      "learning_rate": 2.353940971560833e-06,
+      "loss": 0.0966,
+      "step": 6105
+    },
+    {
+      "epoch": 3.1980094290204297,
+      "loss_partial": 0.053955078125,
+      "step": 6105
+    },
+    {
+      "epoch": 3.1980094290204297,
+      "loss_partial": 0.048095703125,
+      "step": 6105
+    },
+    {
+      "epoch": 3.198533263488738,
+      "grad_norm": 1.4477386474609375,
+      "learning_rate": 2.3509953932861186e-06,
+      "loss": 0.104,
+      "step": 6106
+    },
+    {
+      "epoch": 3.198533263488738,
+      "loss_partial": 0.04833984375,
+      "step": 6106
+    },
+    {
+      "epoch": 3.198533263488738,
+      "loss_partial": 0.04248046875,
+      "step": 6106
+    },
+    {
+      "epoch": 3.1990570979570454,
+      "grad_norm": 1.3169368505477905,
+      "learning_rate": 2.348051413606067e-06,
+      "loss": 0.1017,
+      "step": 6107
+    },
+    {
+      "epoch": 3.1990570979570454,
+      "loss_partial": 0.042724609375,
+      "step": 6107
+    },
+    {
+      "epoch": 3.1990570979570454,
+      "loss_partial": 0.042724609375,
+      "step": 6107
+    },
+    {
+      "epoch": 3.1995809324253535,
+      "grad_norm": 1.2499964237213135,
+      "learning_rate": 2.345109033135953e-06,
+      "loss": 0.0958,
+      "step": 6108
+    },
+    {
+      "epoch": 3.1995809324253535,
+      "loss_partial": 0.047607421875,
+      "step": 6108
+    },
+    {
+      "epoch": 3.1995809324253535,
+      "loss_partial": 0.0439453125,
+      "step": 6108
+    },
+    {
+      "epoch": 3.2001047668936615,
+      "grad_norm": 1.2353434562683105,
+      "learning_rate": 2.342168252490713e-06,
+      "loss": 0.0988,
+      "step": 6109
+    },
+    {
+      "epoch": 3.2001047668936615,
+      "loss_partial": 0.051513671875,
+      "step": 6109
+    },
+    {
+      "epoch": 3.2001047668936615,
+      "loss_partial": 0.043212890625,
+      "step": 6109
+    },
+    {
+      "epoch": 3.2006286013619696,
+      "grad_norm": 1.4679713249206543,
+      "learning_rate": 2.339229072284954e-06,
+      "loss": 0.0962,
+      "step": 6110
+    },
+    {
+      "epoch": 3.2006286013619696,
+      "loss_partial": 0.044189453125,
+      "step": 6110
+    },
+    {
+      "epoch": 3.2006286013619696,
+      "loss_partial": 0.048828125,
+      "step": 6110
+    },
+    {
+      "epoch": 3.2011524358302776,
+      "grad_norm": 1.1227021217346191,
+      "learning_rate": 2.3362914931329473e-06,
+      "loss": 0.0955,
+      "step": 6111
+    },
+    {
+      "epoch": 3.2011524358302776,
+      "loss_partial": 0.05126953125,
+      "step": 6111
+    },
+    {
+      "epoch": 3.2011524358302776,
+      "loss_partial": 0.048095703125,
+      "step": 6111
+    },
+    {
+      "epoch": 3.2016762702985857,
+      "grad_norm": 0.9367218017578125,
+      "learning_rate": 2.3333555156486266e-06,
+      "loss": 0.0961,
+      "step": 6112
+    },
+    {
+      "epoch": 3.2016762702985857,
+      "loss_partial": 0.050048828125,
+      "step": 6112
+    },
+    {
+      "epoch": 3.2016762702985857,
+      "loss_partial": 0.047607421875,
+      "step": 6112
+    },
+    {
+      "epoch": 3.2022001047668938,
+      "grad_norm": 1.244836449623108,
+      "learning_rate": 2.330421140445591e-06,
+      "loss": 0.0992,
+      "step": 6113
+    },
+    {
+      "epoch": 3.2022001047668938,
+      "loss_partial": 0.044921875,
+      "step": 6113
+    },
+    {
+      "epoch": 3.2022001047668938,
+      "loss_partial": 0.04296875,
+      "step": 6113
+    },
+    {
+      "epoch": 3.202723939235202,
+      "grad_norm": 1.374553918838501,
+      "learning_rate": 2.327488368137113e-06,
+      "loss": 0.1006,
+      "step": 6114
+    },
+    {
+      "epoch": 3.202723939235202,
+      "loss_partial": 0.05078125,
+      "step": 6114
+    },
+    {
+      "epoch": 3.202723939235202,
+      "loss_partial": 0.05078125,
+      "step": 6114
+    },
+    {
+      "epoch": 3.20324777370351,
+      "grad_norm": 0.9654310941696167,
+      "learning_rate": 2.3245571993361205e-06,
+      "loss": 0.0995,
+      "step": 6115
+    },
+    {
+      "epoch": 3.20324777370351,
+      "loss_partial": 0.052978515625,
+      "step": 6115
+    },
+    {
+      "epoch": 3.20324777370351,
+      "loss_partial": 0.0419921875,
+      "step": 6115
+    },
+    {
+      "epoch": 3.2037716081718175,
+      "grad_norm": 1.1929035186767578,
+      "learning_rate": 2.32162763465521e-06,
+      "loss": 0.1016,
+      "step": 6116
+    },
+    {
+      "epoch": 3.2037716081718175,
+      "loss_partial": 0.046630859375,
+      "step": 6116
+    },
+    {
+      "epoch": 3.2037716081718175,
+      "loss_partial": 0.039306640625,
+      "step": 6116
+    },
+    {
+      "epoch": 3.2042954426401256,
+      "grad_norm": 1.2546393871307373,
+      "learning_rate": 2.318699674706639e-06,
+      "loss": 0.1002,
+      "step": 6117
+    },
+    {
+      "epoch": 3.2042954426401256,
+      "loss_partial": 0.049560546875,
+      "step": 6117
+    },
+    {
+      "epoch": 3.2042954426401256,
+      "loss_partial": 0.0390625,
+      "step": 6117
+    },
+    {
+      "epoch": 3.2048192771084336,
+      "grad_norm": 1.5966440439224243,
+      "learning_rate": 2.3157733201023403e-06,
+      "loss": 0.0964,
+      "step": 6118
+    },
+    {
+      "epoch": 3.2048192771084336,
+      "loss_partial": 0.0478515625,
+      "step": 6118
+    },
+    {
+      "epoch": 3.2048192771084336,
+      "loss_partial": 0.0576171875,
+      "step": 6118
+    },
+    {
+      "epoch": 3.2053431115767417,
+      "grad_norm": 1.1719532012939453,
+      "learning_rate": 2.3128485714538985e-06,
+      "loss": 0.1026,
+      "step": 6119
+    },
+    {
+      "epoch": 3.2053431115767417,
+      "loss_partial": 0.05126953125,
+      "step": 6119
+    },
+    {
+      "epoch": 3.2053431115767417,
+      "loss_partial": 0.0595703125,
+      "step": 6119
+    },
+    {
+      "epoch": 3.2058669460450497,
+      "grad_norm": 0.9470285177230835,
+      "learning_rate": 2.3099254293725703e-06,
+      "loss": 0.1044,
+      "step": 6120
+    },
+    {
+      "epoch": 3.2058669460450497,
+      "loss_partial": 0.05419921875,
+      "step": 6120
+    },
+    {
+      "epoch": 3.2058669460450497,
+      "loss_partial": 0.04345703125,
+      "step": 6120
+    },
+    {
+      "epoch": 3.206390780513358,
+      "grad_norm": 1.1745613813400269,
+      "learning_rate": 2.307003894469273e-06,
+      "loss": 0.099,
+      "step": 6121
+    },
+    {
+      "epoch": 3.206390780513358,
+      "loss_partial": 0.049072265625,
+      "step": 6121
+    },
+    {
+      "epoch": 3.206390780513358,
+      "loss_partial": 0.048828125,
+      "step": 6121
+    },
+    {
+      "epoch": 3.206914614981666,
+      "grad_norm": 1.0322625637054443,
+      "learning_rate": 2.30408396735459e-06,
+      "loss": 0.1002,
+      "step": 6122
+    },
+    {
+      "epoch": 3.206914614981666,
+      "loss_partial": 0.04443359375,
+      "step": 6122
+    },
+    {
+      "epoch": 3.206914614981666,
+      "loss_partial": 0.045654296875,
+      "step": 6122
+    },
+    {
+      "epoch": 3.207438449449974,
+      "grad_norm": 1.1260234117507935,
+      "learning_rate": 2.3011656486387666e-06,
+      "loss": 0.0984,
+      "step": 6123
+    },
+    {
+      "epoch": 3.207438449449974,
+      "loss_partial": 0.050048828125,
+      "step": 6123
+    },
+    {
+      "epoch": 3.207438449449974,
+      "loss_partial": 0.052978515625,
+      "step": 6123
+    },
+    {
+      "epoch": 3.207962283918282,
+      "grad_norm": 1.0192402601242065,
+      "learning_rate": 2.2982489389317152e-06,
+      "loss": 0.0998,
+      "step": 6124
+    },
+    {
+      "epoch": 3.207962283918282,
+      "loss_partial": 0.0576171875,
+      "step": 6124
+    },
+    {
+      "epoch": 3.207962283918282,
+      "loss_partial": 0.0439453125,
+      "step": 6124
+    },
+    {
+      "epoch": 3.20848611838659,
+      "grad_norm": 1.3929738998413086,
+      "learning_rate": 2.295333838843007e-06,
+      "loss": 0.1007,
+      "step": 6125
+    },
+    {
+      "epoch": 3.20848611838659,
+      "loss_partial": 0.04296875,
+      "step": 6125
+    },
+    {
+      "epoch": 3.20848611838659,
+      "loss_partial": 0.0498046875,
+      "step": 6125
+    },
+    {
+      "epoch": 3.2090099528548977,
+      "grad_norm": 1.2972850799560547,
+      "learning_rate": 2.292420348981884e-06,
+      "loss": 0.1011,
+      "step": 6126
+    },
+    {
+      "epoch": 3.2090099528548977,
+      "loss_partial": 0.056396484375,
+      "step": 6126
+    },
+    {
+      "epoch": 3.2090099528548977,
+      "loss_partial": 0.04736328125,
+      "step": 6126
+    },
+    {
+      "epoch": 3.2095337873232057,
+      "grad_norm": 1.1303126811981201,
+      "learning_rate": 2.2895084699572467e-06,
+      "loss": 0.1,
+      "step": 6127
+    },
+    {
+      "epoch": 3.2095337873232057,
+      "loss_partial": 0.046630859375,
+      "step": 6127
+    },
+    {
+      "epoch": 3.2095337873232057,
+      "loss_partial": 0.054443359375,
+      "step": 6127
+    },
+    {
+      "epoch": 3.2100576217915138,
+      "grad_norm": 1.2138830423355103,
+      "learning_rate": 2.286598202377659e-06,
+      "loss": 0.0991,
+      "step": 6128
+    },
+    {
+      "epoch": 3.2100576217915138,
+      "loss_partial": 0.051025390625,
+      "step": 6128
+    },
+    {
+      "epoch": 3.2100576217915138,
+      "loss_partial": 0.05029296875,
+      "step": 6128
+    },
+    {
+      "epoch": 3.210581456259822,
+      "grad_norm": 1.8028236627578735,
+      "learning_rate": 2.2836895468513454e-06,
+      "loss": 0.0967,
+      "step": 6129
+    },
+    {
+      "epoch": 3.210581456259822,
+      "loss_partial": 0.046142578125,
+      "step": 6129
+    },
+    {
+      "epoch": 3.210581456259822,
+      "loss_partial": 0.0458984375,
+      "step": 6129
+    },
+    {
+      "epoch": 3.21110529072813,
+      "grad_norm": 1.2329435348510742,
+      "learning_rate": 2.2807825039862043e-06,
+      "loss": 0.1003,
+      "step": 6130
+    },
+    {
+      "epoch": 3.21110529072813,
+      "loss_partial": 0.0478515625,
+      "step": 6130
+    },
+    {
+      "epoch": 3.21110529072813,
+      "loss_partial": 0.046875,
+      "step": 6130
+    },
+    {
+      "epoch": 3.211629125196438,
+      "grad_norm": 1.2356363534927368,
+      "learning_rate": 2.2778770743897837e-06,
+      "loss": 0.1008,
+      "step": 6131
+    },
+    {
+      "epoch": 3.211629125196438,
+      "loss_partial": 0.0703125,
+      "step": 6131
+    },
+    {
+      "epoch": 3.211629125196438,
+      "loss_partial": 0.05322265625,
+      "step": 6131
+    },
+    {
+      "epoch": 3.212152959664746,
+      "grad_norm": 1.353224277496338,
+      "learning_rate": 2.2749732586693043e-06,
+      "loss": 0.1011,
+      "step": 6132
+    },
+    {
+      "epoch": 3.212152959664746,
+      "loss_partial": 0.06884765625,
+      "step": 6132
+    },
+    {
+      "epoch": 3.212152959664746,
+      "loss_partial": 0.050048828125,
+      "step": 6132
+    },
+    {
+      "epoch": 3.212676794133054,
+      "grad_norm": 1.1675978899002075,
+      "learning_rate": 2.272071057431643e-06,
+      "loss": 0.1038,
+      "step": 6133
+    },
+    {
+      "epoch": 3.212676794133054,
+      "loss_partial": 0.051513671875,
+      "step": 6133
+    },
+    {
+      "epoch": 3.212676794133054,
+      "loss_partial": 0.041015625,
+      "step": 6133
+    },
+    {
+      "epoch": 3.213200628601362,
+      "grad_norm": 1.041731834411621,
+      "learning_rate": 2.2691704712833418e-06,
+      "loss": 0.1012,
+      "step": 6134
+    },
+    {
+      "epoch": 3.213200628601362,
+      "loss_partial": 0.054443359375,
+      "step": 6134
+    },
+    {
+      "epoch": 3.213200628601362,
+      "loss_partial": 0.051513671875,
+      "step": 6134
+    },
+    {
+      "epoch": 3.21372446306967,
+      "grad_norm": 1.2918846607208252,
+      "learning_rate": 2.266271500830607e-06,
+      "loss": 0.0981,
+      "step": 6135
+    },
+    {
+      "epoch": 3.21372446306967,
+      "loss_partial": 0.042724609375,
+      "step": 6135
+    },
+    {
+      "epoch": 3.21372446306967,
+      "loss_partial": 0.044189453125,
+      "step": 6135
+    },
+    {
+      "epoch": 3.214248297537978,
+      "grad_norm": 1.0720378160476685,
+      "learning_rate": 2.2633741466793046e-06,
+      "loss": 0.0972,
+      "step": 6136
+    },
+    {
+      "epoch": 3.214248297537978,
+      "loss_partial": 0.04443359375,
+      "step": 6136
+    },
+    {
+      "epoch": 3.214248297537978,
+      "loss_partial": 0.05224609375,
+      "step": 6136
+    },
+    {
+      "epoch": 3.214772132006286,
+      "grad_norm": 1.1319535970687866,
+      "learning_rate": 2.2604784094349596e-06,
+      "loss": 0.1005,
+      "step": 6137
+    },
+    {
+      "epoch": 3.214772132006286,
+      "loss_partial": 0.052734375,
+      "step": 6137
+    },
+    {
+      "epoch": 3.214772132006286,
+      "loss_partial": 0.05224609375,
+      "step": 6137
+    },
+    {
+      "epoch": 3.215295966474594,
+      "grad_norm": 1.1458765268325806,
+      "learning_rate": 2.257584289702772e-06,
+      "loss": 0.1027,
+      "step": 6138
+    },
+    {
+      "epoch": 3.215295966474594,
+      "loss_partial": 0.052734375,
+      "step": 6138
+    },
+    {
+      "epoch": 3.215295966474594,
+      "loss_partial": 0.052001953125,
+      "step": 6138
+    },
+    {
+      "epoch": 3.215819800942902,
+      "grad_norm": 1.3052998781204224,
+      "learning_rate": 2.2546917880875883e-06,
+      "loss": 0.0985,
+      "step": 6139
+    },
+    {
+      "epoch": 3.215819800942902,
+      "loss_partial": 0.0498046875,
+      "step": 6139
+    },
+    {
+      "epoch": 3.215819800942902,
+      "loss_partial": 0.039794921875,
+      "step": 6139
+    },
+    {
+      "epoch": 3.21634363541121,
+      "grad_norm": 1.192751169204712,
+      "learning_rate": 2.2518009051939273e-06,
+      "loss": 0.0967,
+      "step": 6140
+    },
+    {
+      "epoch": 3.21634363541121,
+      "loss_partial": 0.054931640625,
+      "step": 6140
+    },
+    {
+      "epoch": 3.21634363541121,
+      "loss_partial": 0.038330078125,
+      "step": 6140
+    },
+    {
+      "epoch": 3.216867469879518,
+      "grad_norm": 1.229817271232605,
+      "learning_rate": 2.2489116416259594e-06,
+      "loss": 0.0956,
+      "step": 6141
+    },
+    {
+      "epoch": 3.216867469879518,
+      "loss_partial": 0.04541015625,
+      "step": 6141
+    },
+    {
+      "epoch": 3.216867469879518,
+      "loss_partial": 0.06103515625,
+      "step": 6141
+    },
+    {
+      "epoch": 3.217391304347826,
+      "grad_norm": 1.5855164527893066,
+      "learning_rate": 2.246023997987531e-06,
+      "loss": 0.0999,
+      "step": 6142
+    },
+    {
+      "epoch": 3.217391304347826,
+      "loss_partial": 0.054931640625,
+      "step": 6142
+    },
+    {
+      "epoch": 3.217391304347826,
+      "loss_partial": 0.049560546875,
+      "step": 6142
+    },
+    {
+      "epoch": 3.2179151388161342,
+      "grad_norm": 1.442771077156067,
+      "learning_rate": 2.243137974882137e-06,
+      "loss": 0.1052,
+      "step": 6143
+    },
+    {
+      "epoch": 3.2179151388161342,
+      "loss_partial": 0.046142578125,
+      "step": 6143
+    },
+    {
+      "epoch": 3.2179151388161342,
+      "loss_partial": 0.0498046875,
+      "step": 6143
+    },
+    {
+      "epoch": 3.2184389732844423,
+      "grad_norm": 1.160969614982605,
+      "learning_rate": 2.2402535729129394e-06,
+      "loss": 0.1,
+      "step": 6144
+    },
+    {
+      "epoch": 3.2184389732844423,
+      "loss_partial": 0.04443359375,
+      "step": 6144
+    },
+    {
+      "epoch": 3.2184389732844423,
+      "loss_partial": 0.05126953125,
+      "step": 6144
+    },
+    {
+      "epoch": 3.2189628077527503,
+      "grad_norm": 1.5454561710357666,
+      "learning_rate": 2.2373707926827605e-06,
+      "loss": 0.1039,
+      "step": 6145
+    },
+    {
+      "epoch": 3.2189628077527503,
+      "loss_partial": 0.046875,
+      "step": 6145
+    },
+    {
+      "epoch": 3.2189628077527503,
+      "loss_partial": 0.05029296875,
+      "step": 6145
+    },
+    {
+      "epoch": 3.219486642221058,
+      "grad_norm": 1.2103486061096191,
+      "learning_rate": 2.2344896347940835e-06,
+      "loss": 0.0957,
+      "step": 6146
+    },
+    {
+      "epoch": 3.219486642221058,
+      "loss_partial": 0.042236328125,
+      "step": 6146
+    },
+    {
+      "epoch": 3.219486642221058,
+      "loss_partial": 0.048583984375,
+      "step": 6146
+    },
+    {
+      "epoch": 3.220010476689366,
+      "grad_norm": 0.882927417755127,
+      "learning_rate": 2.231610099849052e-06,
+      "loss": 0.0982,
+      "step": 6147
+    },
+    {
+      "epoch": 3.220010476689366,
+      "loss_partial": 0.04248046875,
+      "step": 6147
+    },
+    {
+      "epoch": 3.220010476689366,
+      "loss_partial": 0.04736328125,
+      "step": 6147
+    },
+    {
+      "epoch": 3.220534311157674,
+      "grad_norm": 1.3961615562438965,
+      "learning_rate": 2.2287321884494706e-06,
+      "loss": 0.1019,
+      "step": 6148
+    },
+    {
+      "epoch": 3.220534311157674,
+      "loss_partial": 0.05126953125,
+      "step": 6148
+    },
+    {
+      "epoch": 3.220534311157674,
+      "loss_partial": 0.052978515625,
+      "step": 6148
+    },
+    {
+      "epoch": 3.221058145625982,
+      "grad_norm": 1.1281050443649292,
+      "learning_rate": 2.2258559011968074e-06,
+      "loss": 0.0999,
+      "step": 6149
+    },
+    {
+      "epoch": 3.221058145625982,
+      "loss_partial": 0.054931640625,
+      "step": 6149
+    },
+    {
+      "epoch": 3.221058145625982,
+      "loss_partial": 0.052734375,
+      "step": 6149
+    },
+    {
+      "epoch": 3.22158198009429,
+      "grad_norm": 1.2886184453964233,
+      "learning_rate": 2.2229812386921835e-06,
+      "loss": 0.0995,
+      "step": 6150
+    },
+    {
+      "epoch": 3.22158198009429,
+      "loss_partial": 0.055419921875,
+      "step": 6150
+    },
+    {
+      "epoch": 3.22158198009429,
+      "loss_partial": 0.043212890625,
+      "step": 6150
+    },
+    {
+      "epoch": 3.2221058145625983,
+      "grad_norm": 1.1808542013168335,
+      "learning_rate": 2.2201082015363905e-06,
+      "loss": 0.1026,
+      "step": 6151
+    },
+    {
+      "epoch": 3.2221058145625983,
+      "loss_partial": 0.04150390625,
+      "step": 6151
+    },
+    {
+      "epoch": 3.2221058145625983,
+      "loss_partial": 0.050537109375,
+      "step": 6151
+    },
+    {
+      "epoch": 3.2226296490309063,
+      "grad_norm": 1.1107938289642334,
+      "learning_rate": 2.2172367903298753e-06,
+      "loss": 0.0988,
+      "step": 6152
+    },
+    {
+      "epoch": 3.2226296490309063,
+      "loss_partial": 0.049072265625,
+      "step": 6152
+    },
+    {
+      "epoch": 3.2226296490309063,
+      "loss_partial": 0.0615234375,
+      "step": 6152
+    },
+    {
+      "epoch": 3.2231534834992144,
+      "grad_norm": 1.2477798461914062,
+      "learning_rate": 2.214367005672744e-06,
+      "loss": 0.1031,
+      "step": 6153
+    },
+    {
+      "epoch": 3.2231534834992144,
+      "loss_partial": 0.04248046875,
+      "step": 6153
+    },
+    {
+      "epoch": 3.2231534834992144,
+      "loss_partial": 0.0498046875,
+      "step": 6153
+    },
+    {
+      "epoch": 3.2236773179675224,
+      "grad_norm": 1.1362265348434448,
+      "learning_rate": 2.21149884816476e-06,
+      "loss": 0.0998,
+      "step": 6154
+    },
+    {
+      "epoch": 3.2236773179675224,
+      "loss_partial": 0.06201171875,
+      "step": 6154
+    },
+    {
+      "epoch": 3.2236773179675224,
+      "loss_partial": 0.054443359375,
+      "step": 6154
+    },
+    {
+      "epoch": 3.22420115243583,
+      "grad_norm": 1.1308966875076294,
+      "learning_rate": 2.2086323184053582e-06,
+      "loss": 0.1019,
+      "step": 6155
+    },
+    {
+      "epoch": 3.22420115243583,
+      "loss_partial": 0.04931640625,
+      "step": 6155
+    },
+    {
+      "epoch": 3.22420115243583,
+      "loss_partial": 0.051513671875,
+      "step": 6155
+    },
+    {
+      "epoch": 3.224724986904138,
+      "grad_norm": 1.0540584325790405,
+      "learning_rate": 2.2057674169936206e-06,
+      "loss": 0.0999,
+      "step": 6156
+    },
+    {
+      "epoch": 3.224724986904138,
+      "loss_partial": 0.047607421875,
+      "step": 6156
+    },
+    {
+      "epoch": 3.224724986904138,
+      "loss_partial": 0.044677734375,
+      "step": 6156
+    },
+    {
+      "epoch": 3.225248821372446,
+      "grad_norm": 1.4578391313552856,
+      "learning_rate": 2.202904144528295e-06,
+      "loss": 0.0989,
+      "step": 6157
+    },
+    {
+      "epoch": 3.225248821372446,
+      "loss_partial": 0.0537109375,
+      "step": 6157
+    },
+    {
+      "epoch": 3.225248821372446,
+      "loss_partial": 0.06396484375,
+      "step": 6157
+    },
+    {
+      "epoch": 3.2257726558407542,
+      "grad_norm": 1.5259912014007568,
+      "learning_rate": 2.2000425016077864e-06,
+      "loss": 0.1003,
+      "step": 6158
+    },
+    {
+      "epoch": 3.2257726558407542,
+      "loss_partial": 0.060302734375,
+      "step": 6158
+    },
+    {
+      "epoch": 3.2257726558407542,
+      "loss_partial": 0.044189453125,
+      "step": 6158
+    },
+    {
+      "epoch": 3.2262964903090623,
+      "grad_norm": 1.0220983028411865,
+      "learning_rate": 2.197182488830163e-06,
+      "loss": 0.1024,
+      "step": 6159
+    },
+    {
+      "epoch": 3.2262964903090623,
+      "loss_partial": 0.050048828125,
+      "step": 6159
+    },
+    {
+      "epoch": 3.2262964903090623,
+      "loss_partial": 0.050048828125,
+      "step": 6159
+    },
+    {
+      "epoch": 3.2268203247773704,
+      "grad_norm": 1.9203587770462036,
+      "learning_rate": 2.1943241067931476e-06,
+      "loss": 0.1,
+      "step": 6160
+    },
+    {
+      "epoch": 3.2268203247773704,
+      "loss_partial": 0.052978515625,
+      "step": 6160
+    },
+    {
+      "epoch": 3.2268203247773704,
+      "loss_partial": 0.041259765625,
+      "step": 6160
+    },
+    {
+      "epoch": 3.2273441592456784,
+      "grad_norm": 1.065316915512085,
+      "learning_rate": 2.1914673560941247e-06,
+      "loss": 0.0978,
+      "step": 6161
+    },
+    {
+      "epoch": 3.2273441592456784,
+      "loss_partial": 0.059326171875,
+      "step": 6161
+    },
+    {
+      "epoch": 3.2273441592456784,
+      "loss_partial": 0.053466796875,
+      "step": 6161
+    },
+    {
+      "epoch": 3.2278679937139865,
+      "grad_norm": 1.272895097732544,
+      "learning_rate": 2.1886122373301357e-06,
+      "loss": 0.0987,
+      "step": 6162
+    },
+    {
+      "epoch": 3.2278679937139865,
+      "loss_partial": 0.043701171875,
+      "step": 6162
+    },
+    {
+      "epoch": 3.2278679937139865,
+      "loss_partial": 0.050048828125,
+      "step": 6162
+    },
+    {
+      "epoch": 3.2283918281822945,
+      "grad_norm": 1.2977148294448853,
+      "learning_rate": 2.185758751097886e-06,
+      "loss": 0.1021,
+      "step": 6163
+    },
+    {
+      "epoch": 3.2283918281822945,
+      "loss_partial": 0.055908203125,
+      "step": 6163
+    },
+    {
+      "epoch": 3.2283918281822945,
+      "loss_partial": 0.0537109375,
+      "step": 6163
+    },
+    {
+      "epoch": 3.2289156626506026,
+      "grad_norm": 1.1634724140167236,
+      "learning_rate": 2.182906897993737e-06,
+      "loss": 0.1034,
+      "step": 6164
+    },
+    {
+      "epoch": 3.2289156626506026,
+      "loss_partial": 0.05322265625,
+      "step": 6164
+    },
+    {
+      "epoch": 3.2289156626506026,
+      "loss_partial": 0.050048828125,
+      "step": 6164
+    },
+    {
+      "epoch": 3.2294394971189107,
+      "grad_norm": 1.145604133605957,
+      "learning_rate": 2.180056678613708e-06,
+      "loss": 0.1038,
+      "step": 6165
+    },
+    {
+      "epoch": 3.2294394971189107,
+      "loss_partial": 0.04443359375,
+      "step": 6165
+    },
+    {
+      "epoch": 3.2294394971189107,
+      "loss_partial": 0.03759765625,
+      "step": 6165
+    },
+    {
+      "epoch": 3.2299633315872183,
+      "grad_norm": 1.3160301446914673,
+      "learning_rate": 2.1772080935534714e-06,
+      "loss": 0.096,
+      "step": 6166
+    },
+    {
+      "epoch": 3.2299633315872183,
+      "loss_partial": 0.046630859375,
+      "step": 6166
+    },
+    {
+      "epoch": 3.2299633315872183,
+      "loss_partial": 0.041259765625,
+      "step": 6166
+    },
+    {
+      "epoch": 3.2304871660555263,
+      "grad_norm": 0.9321683645248413,
+      "learning_rate": 2.1743611434083724e-06,
+      "loss": 0.0953,
+      "step": 6167
+    },
+    {
+      "epoch": 3.2304871660555263,
+      "loss_partial": 0.05517578125,
+      "step": 6167
+    },
+    {
+      "epoch": 3.2304871660555263,
+      "loss_partial": 0.052978515625,
+      "step": 6167
+    },
+    {
+      "epoch": 3.2310110005238344,
+      "grad_norm": 1.501633882522583,
+      "learning_rate": 2.1715158287734018e-06,
+      "loss": 0.0999,
+      "step": 6168
+    },
+    {
+      "epoch": 3.2310110005238344,
+      "loss_partial": 0.05322265625,
+      "step": 6168
+    },
+    {
+      "epoch": 3.2310110005238344,
+      "loss_partial": 0.049072265625,
+      "step": 6168
+    },
+    {
+      "epoch": 3.2315348349921424,
+      "grad_norm": 1.1830344200134277,
+      "learning_rate": 2.168672150243213e-06,
+      "loss": 0.1017,
+      "step": 6169
+    },
+    {
+      "epoch": 3.2315348349921424,
+      "loss_partial": 0.048583984375,
+      "step": 6169
+    },
+    {
+      "epoch": 3.2315348349921424,
+      "loss_partial": 0.0517578125,
+      "step": 6169
+    },
+    {
+      "epoch": 3.2320586694604505,
+      "grad_norm": 1.0933722257614136,
+      "learning_rate": 2.165830108412117e-06,
+      "loss": 0.1015,
+      "step": 6170
+    },
+    {
+      "epoch": 3.2320586694604505,
+      "loss_partial": 0.04052734375,
+      "step": 6170
+    },
+    {
+      "epoch": 3.2320586694604505,
+      "loss_partial": 0.052734375,
+      "step": 6170
+    },
+    {
+      "epoch": 3.2325825039287586,
+      "grad_norm": 1.4287306070327759,
+      "learning_rate": 2.1629897038740833e-06,
+      "loss": 0.0981,
+      "step": 6171
+    },
+    {
+      "epoch": 3.2325825039287586,
+      "loss_partial": 0.04443359375,
+      "step": 6171
+    },
+    {
+      "epoch": 3.2325825039287586,
+      "loss_partial": 0.037841796875,
+      "step": 6171
+    },
+    {
+      "epoch": 3.2331063383970666,
+      "grad_norm": 0.8973186612129211,
+      "learning_rate": 2.160150937222739e-06,
+      "loss": 0.099,
+      "step": 6172
+    },
+    {
+      "epoch": 3.2331063383970666,
+      "loss_partial": 0.064453125,
+      "step": 6172
+    },
+    {
+      "epoch": 3.2331063383970666,
+      "loss_partial": 0.042236328125,
+      "step": 6172
+    },
+    {
+      "epoch": 3.2336301728653747,
+      "grad_norm": 0.9575915336608887,
+      "learning_rate": 2.157313809051369e-06,
+      "loss": 0.1016,
+      "step": 6173
+    },
+    {
+      "epoch": 3.2336301728653747,
+      "loss_partial": 0.050048828125,
+      "step": 6173
+    },
+    {
+      "epoch": 3.2336301728653747,
+      "loss_partial": 0.042724609375,
+      "step": 6173
+    },
+    {
+      "epoch": 3.2341540073336827,
+      "grad_norm": 1.1344605684280396,
+      "learning_rate": 2.154478319952913e-06,
+      "loss": 0.1019,
+      "step": 6174
+    },
+    {
+      "epoch": 3.2341540073336827,
+      "loss_partial": 0.0537109375,
+      "step": 6174
+    },
+    {
+      "epoch": 3.2341540073336827,
+      "loss_partial": 0.04736328125,
+      "step": 6174
+    },
+    {
+      "epoch": 3.2346778418019904,
+      "grad_norm": 1.2494295835494995,
+      "learning_rate": 2.1516444705199745e-06,
+      "loss": 0.0984,
+      "step": 6175
+    },
+    {
+      "epoch": 3.2346778418019904,
+      "loss_partial": 0.04296875,
+      "step": 6175
+    },
+    {
+      "epoch": 3.2346778418019904,
+      "loss_partial": 0.054443359375,
+      "step": 6175
+    },
+    {
+      "epoch": 3.2352016762702984,
+      "grad_norm": 1.3386069536209106,
+      "learning_rate": 2.1488122613448093e-06,
+      "loss": 0.0995,
+      "step": 6176
+    },
+    {
+      "epoch": 3.2352016762702984,
+      "loss_partial": 0.056884765625,
+      "step": 6176
+    },
+    {
+      "epoch": 3.2352016762702984,
+      "loss_partial": 0.054443359375,
+      "step": 6176
+    },
+    {
+      "epoch": 3.2357255107386065,
+      "grad_norm": 1.0497392416000366,
+      "learning_rate": 2.1459816930193313e-06,
+      "loss": 0.0987,
+      "step": 6177
+    },
+    {
+      "epoch": 3.2357255107386065,
+      "loss_partial": 0.058349609375,
+      "step": 6177
+    },
+    {
+      "epoch": 3.2357255107386065,
+      "loss_partial": 0.0537109375,
+      "step": 6177
+    },
+    {
+      "epoch": 3.2362493452069145,
+      "grad_norm": 1.2858645915985107,
+      "learning_rate": 2.1431527661351094e-06,
+      "loss": 0.1014,
+      "step": 6178
+    },
+    {
+      "epoch": 3.2362493452069145,
+      "loss_partial": 0.048583984375,
+      "step": 6178
+    },
+    {
+      "epoch": 3.2362493452069145,
+      "loss_partial": 0.053466796875,
+      "step": 6178
+    },
+    {
+      "epoch": 3.2367731796752226,
+      "grad_norm": 1.2267699241638184,
+      "learning_rate": 2.140325481283375e-06,
+      "loss": 0.1013,
+      "step": 6179
+    },
+    {
+      "epoch": 3.2367731796752226,
+      "loss_partial": 0.0556640625,
+      "step": 6179
+    },
+    {
+      "epoch": 3.2367731796752226,
+      "loss_partial": 0.05322265625,
+      "step": 6179
+    },
+    {
+      "epoch": 3.2372970141435307,
+      "grad_norm": 1.2765928506851196,
+      "learning_rate": 2.1374998390550138e-06,
+      "loss": 0.101,
+      "step": 6180
+    },
+    {
+      "epoch": 3.2372970141435307,
+      "loss_partial": 0.054931640625,
+      "step": 6180
+    },
+    {
+      "epoch": 3.2372970141435307,
+      "loss_partial": 0.052734375,
+      "step": 6180
+    },
+    {
+      "epoch": 3.2378208486118387,
+      "grad_norm": 1.0903571844100952,
+      "learning_rate": 2.1346758400405666e-06,
+      "loss": 0.1025,
+      "step": 6181
+    },
+    {
+      "epoch": 3.2378208486118387,
+      "loss_partial": 0.03662109375,
+      "step": 6181
+    },
+    {
+      "epoch": 3.2378208486118387,
+      "loss_partial": 0.054931640625,
+      "step": 6181
+    },
+    {
+      "epoch": 3.238344683080147,
+      "grad_norm": 1.3835034370422363,
+      "learning_rate": 2.1318534848302296e-06,
+      "loss": 0.0946,
+      "step": 6182
+    },
+    {
+      "epoch": 3.238344683080147,
+      "loss_partial": 0.046630859375,
+      "step": 6182
+    },
+    {
+      "epoch": 3.238344683080147,
+      "loss_partial": 0.047119140625,
+      "step": 6182
+    },
+    {
+      "epoch": 3.238868517548455,
+      "grad_norm": 1.4058772325515747,
+      "learning_rate": 2.1290327740138595e-06,
+      "loss": 0.0975,
+      "step": 6183
+    },
+    {
+      "epoch": 3.238868517548455,
+      "loss_partial": 0.04833984375,
+      "step": 6183
+    },
+    {
+      "epoch": 3.238868517548455,
+      "loss_partial": 0.053955078125,
+      "step": 6183
+    },
+    {
+      "epoch": 3.239392352016763,
+      "grad_norm": 0.9743680357933044,
+      "learning_rate": 2.1262137081809676e-06,
+      "loss": 0.1007,
+      "step": 6184
+    },
+    {
+      "epoch": 3.239392352016763,
+      "loss_partial": 0.0634765625,
+      "step": 6184
+    },
+    {
+      "epoch": 3.239392352016763,
+      "loss_partial": 0.046875,
+      "step": 6184
+    },
+    {
+      "epoch": 3.2399161864850705,
+      "grad_norm": 1.4998466968536377,
+      "learning_rate": 2.12339628792072e-06,
+      "loss": 0.0988,
+      "step": 6185
+    },
+    {
+      "epoch": 3.2399161864850705,
+      "loss_partial": 0.04345703125,
+      "step": 6185
+    },
+    {
+      "epoch": 3.2399161864850705,
+      "loss_partial": 0.05859375,
+      "step": 6185
+    },
+    {
+      "epoch": 3.2404400209533786,
+      "grad_norm": 1.5773024559020996,
+      "learning_rate": 2.120580513821939e-06,
+      "loss": 0.1016,
+      "step": 6186
+    },
+    {
+      "epoch": 3.2404400209533786,
+      "loss_partial": 0.052734375,
+      "step": 6186
+    },
+    {
+      "epoch": 3.2404400209533786,
+      "loss_partial": 0.047607421875,
+      "step": 6186
+    },
+    {
+      "epoch": 3.2409638554216866,
+      "grad_norm": 1.4396940469741821,
+      "learning_rate": 2.117766386473108e-06,
+      "loss": 0.106,
+      "step": 6187
+    },
+    {
+      "epoch": 3.2409638554216866,
+      "loss_partial": 0.048095703125,
+      "step": 6187
+    },
+    {
+      "epoch": 3.2409638554216866,
+      "loss_partial": 0.0498046875,
+      "step": 6187
+    },
+    {
+      "epoch": 3.2414876898899947,
+      "grad_norm": 1.3293250799179077,
+      "learning_rate": 2.1149539064623592e-06,
+      "loss": 0.1054,
+      "step": 6188
+    },
+    {
+      "epoch": 3.2414876898899947,
+      "loss_partial": 0.04345703125,
+      "step": 6188
+    },
+    {
+      "epoch": 3.2414876898899947,
+      "loss_partial": 0.052978515625,
+      "step": 6188
+    },
+    {
+      "epoch": 3.2420115243583028,
+      "grad_norm": 1.4724053144454956,
+      "learning_rate": 2.112143074377485e-06,
+      "loss": 0.0981,
+      "step": 6189
+    },
+    {
+      "epoch": 3.2420115243583028,
+      "loss_partial": 0.0537109375,
+      "step": 6189
+    },
+    {
+      "epoch": 3.2420115243583028,
+      "loss_partial": 0.050048828125,
+      "step": 6189
+    },
+    {
+      "epoch": 3.242535358826611,
+      "grad_norm": 1.3407734632492065,
+      "learning_rate": 2.109333890805929e-06,
+      "loss": 0.1042,
+      "step": 6190
+    },
+    {
+      "epoch": 3.242535358826611,
+      "loss_partial": 0.050537109375,
+      "step": 6190
+    },
+    {
+      "epoch": 3.242535358826611,
+      "loss_partial": 0.048583984375,
+      "step": 6190
+    },
+    {
+      "epoch": 3.243059193294919,
+      "grad_norm": 1.1448488235473633,
+      "learning_rate": 2.106526356334797e-06,
+      "loss": 0.1011,
+      "step": 6191
+    },
+    {
+      "epoch": 3.243059193294919,
+      "loss_partial": 0.05224609375,
+      "step": 6191
+    },
+    {
+      "epoch": 3.243059193294919,
+      "loss_partial": 0.05322265625,
+      "step": 6191
+    },
+    {
+      "epoch": 3.243583027763227,
+      "grad_norm": 1.3070900440216064,
+      "learning_rate": 2.1037204715508464e-06,
+      "loss": 0.1036,
+      "step": 6192
+    },
+    {
+      "epoch": 3.243583027763227,
+      "loss_partial": 0.046630859375,
+      "step": 6192
+    },
+    {
+      "epoch": 3.243583027763227,
+      "loss_partial": 0.04296875,
+      "step": 6192
+    },
+    {
+      "epoch": 3.244106862231535,
+      "grad_norm": 1.212805986404419,
+      "learning_rate": 2.100916237040487e-06,
+      "loss": 0.0966,
+      "step": 6193
+    },
+    {
+      "epoch": 3.244106862231535,
+      "loss_partial": 0.0419921875,
+      "step": 6193
+    },
+    {
+      "epoch": 3.244106862231535,
+      "loss_partial": 0.046630859375,
+      "step": 6193
+    },
+    {
+      "epoch": 3.244630696699843,
+      "grad_norm": 1.4364488124847412,
+      "learning_rate": 2.0981136533897874e-06,
+      "loss": 0.1011,
+      "step": 6194
+    },
+    {
+      "epoch": 3.244630696699843,
+      "loss_partial": 0.040771484375,
+      "step": 6194
+    },
+    {
+      "epoch": 3.244630696699843,
+      "loss_partial": 0.047119140625,
+      "step": 6194
+    },
+    {
+      "epoch": 3.2451545311681507,
+      "grad_norm": 1.3931083679199219,
+      "learning_rate": 2.09531272118447e-06,
+      "loss": 0.1021,
+      "step": 6195
+    },
+    {
+      "epoch": 3.2451545311681507,
+      "loss_partial": 0.050048828125,
+      "step": 6195
+    },
+    {
+      "epoch": 3.2451545311681507,
+      "loss_partial": 0.05029296875,
+      "step": 6195
+    },
+    {
+      "epoch": 3.2456783656364587,
+      "grad_norm": 1.1616501808166504,
+      "learning_rate": 2.0925134410099135e-06,
+      "loss": 0.1011,
+      "step": 6196
+    },
+    {
+      "epoch": 3.2456783656364587,
+      "loss_partial": 0.045654296875,
+      "step": 6196
+    },
+    {
+      "epoch": 3.2456783656364587,
+      "loss_partial": 0.054931640625,
+      "step": 6196
+    },
+    {
+      "epoch": 3.246202200104767,
+      "grad_norm": 1.2334022521972656,
+      "learning_rate": 2.08971581345115e-06,
+      "loss": 0.1023,
+      "step": 6197
+    },
+    {
+      "epoch": 3.246202200104767,
+      "loss_partial": 0.041748046875,
+      "step": 6197
+    },
+    {
+      "epoch": 3.246202200104767,
+      "loss_partial": 0.046142578125,
+      "step": 6197
+    },
+    {
+      "epoch": 3.246726034573075,
+      "grad_norm": 1.3553599119186401,
+      "learning_rate": 2.0869198390928623e-06,
+      "loss": 0.0985,
+      "step": 6198
+    },
+    {
+      "epoch": 3.246726034573075,
+      "loss_partial": 0.05029296875,
+      "step": 6198
+    },
+    {
+      "epoch": 3.246726034573075,
+      "loss_partial": 0.059326171875,
+      "step": 6198
+    },
+    {
+      "epoch": 3.247249869041383,
+      "grad_norm": 1.2360873222351074,
+      "learning_rate": 2.0841255185193978e-06,
+      "loss": 0.101,
+      "step": 6199
+    },
+    {
+      "epoch": 3.247249869041383,
+      "loss_partial": 0.0556640625,
+      "step": 6199
+    },
+    {
+      "epoch": 3.247249869041383,
+      "loss_partial": 0.042724609375,
+      "step": 6199
+    },
+    {
+      "epoch": 3.247773703509691,
+      "grad_norm": 1.3105449676513672,
+      "learning_rate": 2.0813328523147503e-06,
+      "loss": 0.1011,
+      "step": 6200
+    },
+    {
+      "epoch": 3.247773703509691,
+      "loss_partial": 0.047607421875,
+      "step": 6200
+    },
+    {
+      "epoch": 3.247773703509691,
+      "loss_partial": 0.060791015625,
+      "step": 6200
+    },
+    {
+      "epoch": 3.248297537977999,
+      "grad_norm": 1.1219321489334106,
+      "learning_rate": 2.078541841062568e-06,
+      "loss": 0.0999,
+      "step": 6201
+    },
+    {
+      "epoch": 3.248297537977999,
+      "loss_partial": 0.05029296875,
+      "step": 6201
+    },
+    {
+      "epoch": 3.248297537977999,
+      "loss_partial": 0.052490234375,
+      "step": 6201
+    },
+    {
+      "epoch": 3.248821372446307,
+      "grad_norm": 0.9706894159317017,
+      "learning_rate": 2.075752485346154e-06,
+      "loss": 0.1008,
+      "step": 6202
+    },
+    {
+      "epoch": 3.248821372446307,
+      "loss_partial": 0.0458984375,
+      "step": 6202
+    },
+    {
+      "epoch": 3.248821372446307,
+      "loss_partial": 0.04931640625,
+      "step": 6202
+    },
+    {
+      "epoch": 3.249345206914615,
+      "grad_norm": 1.0958584547042847,
+      "learning_rate": 2.0729647857484713e-06,
+      "loss": 0.1028,
+      "step": 6203
+    },
+    {
+      "epoch": 3.249345206914615,
+      "loss_partial": 0.05078125,
+      "step": 6203
+    },
+    {
+      "epoch": 3.249345206914615,
+      "loss_partial": 0.048583984375,
+      "step": 6203
+    },
+    {
+      "epoch": 3.249869041382923,
+      "grad_norm": 0.9453843832015991,
+      "learning_rate": 2.070178742852129e-06,
+      "loss": 0.0987,
+      "step": 6204
+    },
+    {
+      "epoch": 3.249869041382923,
+      "loss_partial": 0.051513671875,
+      "step": 6204
+    },
+    {
+      "epoch": 3.249869041382923,
+      "loss_partial": 0.05224609375,
+      "step": 6204
+    },
+    {
+      "epoch": 3.250392875851231,
+      "grad_norm": 0.9625800251960754,
+      "learning_rate": 2.0673943572393928e-06,
+      "loss": 0.0986,
+      "step": 6205
+    },
+    {
+      "epoch": 3.250392875851231,
+      "loss_partial": 0.041748046875,
+      "step": 6205
+    },
+    {
+      "epoch": 3.250392875851231,
+      "loss_partial": 0.047607421875,
+      "step": 6205
+    },
+    {
+      "epoch": 3.250916710319539,
+      "grad_norm": 1.1887949705123901,
+      "learning_rate": 2.0646116294921827e-06,
+      "loss": 0.1,
+      "step": 6206
+    },
+    {
+      "epoch": 3.250916710319539,
+      "loss_partial": 0.05224609375,
+      "step": 6206
+    },
+    {
+      "epoch": 3.250916710319539,
+      "loss_partial": 0.041748046875,
+      "step": 6206
+    },
+    {
+      "epoch": 3.251440544787847,
+      "grad_norm": 1.123940110206604,
+      "learning_rate": 2.061830560192071e-06,
+      "loss": 0.0988,
+      "step": 6207
+    },
+    {
+      "epoch": 3.251440544787847,
+      "loss_partial": 0.050048828125,
+      "step": 6207
+    },
+    {
+      "epoch": 3.251440544787847,
+      "loss_partial": 0.043701171875,
+      "step": 6207
+    },
+    {
+      "epoch": 3.251964379256155,
+      "grad_norm": 0.9340064525604248,
+      "learning_rate": 2.0590511499202836e-06,
+      "loss": 0.102,
+      "step": 6208
+    },
+    {
+      "epoch": 3.251964379256155,
+      "loss_partial": 0.046630859375,
+      "step": 6208
+    },
+    {
+      "epoch": 3.251964379256155,
+      "loss_partial": 0.045654296875,
+      "step": 6208
+    },
+    {
+      "epoch": 3.252488213724463,
+      "grad_norm": 1.1382721662521362,
+      "learning_rate": 2.056273399257701e-06,
+      "loss": 0.0989,
+      "step": 6209
+    },
+    {
+      "epoch": 3.252488213724463,
+      "loss_partial": 0.05029296875,
+      "step": 6209
+    },
+    {
+      "epoch": 3.252488213724463,
+      "loss_partial": 0.04541015625,
+      "step": 6209
+    },
+    {
+      "epoch": 3.253012048192771,
+      "grad_norm": 1.1087794303894043,
+      "learning_rate": 2.0534973087848543e-06,
+      "loss": 0.0992,
+      "step": 6210
+    },
+    {
+      "epoch": 3.253012048192771,
+      "loss_partial": 0.054443359375,
+      "step": 6210
+    },
+    {
+      "epoch": 3.253012048192771,
+      "loss_partial": 0.052490234375,
+      "step": 6210
+    },
+    {
+      "epoch": 3.253535882661079,
+      "grad_norm": 0.9613993167877197,
+      "learning_rate": 2.050722879081929e-06,
+      "loss": 0.1015,
+      "step": 6211
+    },
+    {
+      "epoch": 3.253535882661079,
+      "loss_partial": 0.04296875,
+      "step": 6211
+    },
+    {
+      "epoch": 3.253535882661079,
+      "loss_partial": 0.0625,
+      "step": 6211
+    },
+    {
+      "epoch": 3.2540597171293872,
+      "grad_norm": 1.2349612712860107,
+      "learning_rate": 2.047950110728767e-06,
+      "loss": 0.0988,
+      "step": 6212
+    },
+    {
+      "epoch": 3.2540597171293872,
+      "loss_partial": 0.04931640625,
+      "step": 6212
+    },
+    {
+      "epoch": 3.2540597171293872,
+      "loss_partial": 0.0458984375,
+      "step": 6212
+    },
+    {
+      "epoch": 3.2545835515976953,
+      "grad_norm": 1.5139193534851074,
+      "learning_rate": 2.045179004304858e-06,
+      "loss": 0.0984,
+      "step": 6213
+    },
+    {
+      "epoch": 3.2545835515976953,
+      "loss_partial": 0.045654296875,
+      "step": 6213
+    },
+    {
+      "epoch": 3.2545835515976953,
+      "loss_partial": 0.0556640625,
+      "step": 6213
+    },
+    {
+      "epoch": 3.255107386066003,
+      "grad_norm": 1.1824339628219604,
+      "learning_rate": 2.042409560389347e-06,
+      "loss": 0.1009,
+      "step": 6214
+    },
+    {
+      "epoch": 3.255107386066003,
+      "loss_partial": 0.056640625,
+      "step": 6214
+    },
+    {
+      "epoch": 3.255107386066003,
+      "loss_partial": 0.0517578125,
+      "step": 6214
+    },
+    {
+      "epoch": 3.255631220534311,
+      "grad_norm": 0.9998104572296143,
+      "learning_rate": 2.039641779561026e-06,
+      "loss": 0.1036,
+      "step": 6215
+    },
+    {
+      "epoch": 3.255631220534311,
+      "loss_partial": 0.047119140625,
+      "step": 6215
+    },
+    {
+      "epoch": 3.255631220534311,
+      "loss_partial": 0.054443359375,
+      "step": 6215
+    },
+    {
+      "epoch": 3.256155055002619,
+      "grad_norm": 1.279046893119812,
+      "learning_rate": 2.03687566239835e-06,
+      "loss": 0.1046,
+      "step": 6216
+    },
+    {
+      "epoch": 3.256155055002619,
+      "loss_partial": 0.048828125,
+      "step": 6216
+    },
+    {
+      "epoch": 3.256155055002619,
+      "loss_partial": 0.049072265625,
+      "step": 6216
+    },
+    {
+      "epoch": 3.256678889470927,
+      "grad_norm": 0.956575334072113,
+      "learning_rate": 2.034111209479418e-06,
+      "loss": 0.1035,
+      "step": 6217
+    },
+    {
+      "epoch": 3.256678889470927,
+      "loss_partial": 0.047119140625,
+      "step": 6217
+    },
+    {
+      "epoch": 3.256678889470927,
+      "loss_partial": 0.0458984375,
+      "step": 6217
+    },
+    {
+      "epoch": 3.257202723939235,
+      "grad_norm": 1.0336121320724487,
+      "learning_rate": 2.0313484213819845e-06,
+      "loss": 0.0972,
+      "step": 6218
+    },
+    {
+      "epoch": 3.257202723939235,
+      "loss_partial": 0.0546875,
+      "step": 6218
+    },
+    {
+      "epoch": 3.257202723939235,
+      "loss_partial": 0.053955078125,
+      "step": 6218
+    },
+    {
+      "epoch": 3.257726558407543,
+      "grad_norm": 0.9757483005523682,
+      "learning_rate": 2.028587298683454e-06,
+      "loss": 0.0977,
+      "step": 6219
+    },
+    {
+      "epoch": 3.257726558407543,
+      "loss_partial": 0.04736328125,
+      "step": 6219
+    },
+    {
+      "epoch": 3.257726558407543,
+      "loss_partial": 0.046630859375,
+      "step": 6219
+    },
+    {
+      "epoch": 3.2582503928758513,
+      "grad_norm": 1.4578951597213745,
+      "learning_rate": 2.0258278419608846e-06,
+      "loss": 0.1004,
+      "step": 6220
+    },
+    {
+      "epoch": 3.2582503928758513,
+      "loss_partial": 0.045654296875,
+      "step": 6220
+    },
+    {
+      "epoch": 3.2582503928758513,
+      "loss_partial": 0.04541015625,
+      "step": 6220
+    },
+    {
+      "epoch": 3.2587742273441593,
+      "grad_norm": 1.1314877271652222,
+      "learning_rate": 2.0230700517909852e-06,
+      "loss": 0.1013,
+      "step": 6221
+    },
+    {
+      "epoch": 3.2587742273441593,
+      "loss_partial": 0.04248046875,
+      "step": 6221
+    },
+    {
+      "epoch": 3.2587742273441593,
+      "loss_partial": 0.053466796875,
+      "step": 6221
+    },
+    {
+      "epoch": 3.2592980618124674,
+      "grad_norm": 1.1434749364852905,
+      "learning_rate": 2.020313928750116e-06,
+      "loss": 0.1021,
+      "step": 6222
+    },
+    {
+      "epoch": 3.2592980618124674,
+      "loss_partial": 0.05615234375,
+      "step": 6222
+    },
+    {
+      "epoch": 3.2592980618124674,
+      "loss_partial": 0.05322265625,
+      "step": 6222
+    },
+    {
+      "epoch": 3.2598218962807755,
+      "grad_norm": 1.261680006980896,
+      "learning_rate": 2.01755947341429e-06,
+      "loss": 0.1015,
+      "step": 6223
+    },
+    {
+      "epoch": 3.2598218962807755,
+      "loss_partial": 0.060546875,
+      "step": 6223
+    },
+    {
+      "epoch": 3.2598218962807755,
+      "loss_partial": 0.044677734375,
+      "step": 6223
+    },
+    {
+      "epoch": 3.2603457307490835,
+      "grad_norm": 1.4181121587753296,
+      "learning_rate": 2.014806686359173e-06,
+      "loss": 0.1002,
+      "step": 6224
+    },
+    {
+      "epoch": 3.2603457307490835,
+      "loss_partial": 0.046875,
+      "step": 6224
+    },
+    {
+      "epoch": 3.2603457307490835,
+      "loss_partial": 0.049560546875,
+      "step": 6224
+    },
+    {
+      "epoch": 3.260869565217391,
+      "grad_norm": 1.2438528537750244,
+      "learning_rate": 2.0120555681600807e-06,
+      "loss": 0.1002,
+      "step": 6225
+    },
+    {
+      "epoch": 3.260869565217391,
+      "loss_partial": 0.0517578125,
+      "step": 6225
+    },
+    {
+      "epoch": 3.260869565217391,
+      "loss_partial": 0.0478515625,
+      "step": 6225
+    },
+    {
+      "epoch": 3.261393399685699,
+      "grad_norm": 1.1088773012161255,
+      "learning_rate": 2.0093061193919796e-06,
+      "loss": 0.1011,
+      "step": 6226
+    },
+    {
+      "epoch": 3.261393399685699,
+      "loss_partial": 0.058349609375,
+      "step": 6226
+    },
+    {
+      "epoch": 3.261393399685699,
+      "loss_partial": 0.054931640625,
+      "step": 6226
+    },
+    {
+      "epoch": 3.2619172341540073,
+      "grad_norm": 1.2703663110733032,
+      "learning_rate": 2.006558340629482e-06,
+      "loss": 0.1025,
+      "step": 6227
+    },
+    {
+      "epoch": 3.2619172341540073,
+      "loss_partial": 0.05224609375,
+      "step": 6227
+    },
+    {
+      "epoch": 3.2619172341540073,
+      "loss_partial": 0.053955078125,
+      "step": 6227
+    },
+    {
+      "epoch": 3.2624410686223153,
+      "grad_norm": 1.3439868688583374,
+      "learning_rate": 2.0038122324468646e-06,
+      "loss": 0.1001,
+      "step": 6228
+    },
+    {
+      "epoch": 3.2624410686223153,
+      "loss_partial": 0.04833984375,
+      "step": 6228
+    },
+    {
+      "epoch": 3.2624410686223153,
+      "loss_partial": 0.05224609375,
+      "step": 6228
+    },
+    {
+      "epoch": 3.2629649030906234,
+      "grad_norm": 1.0141215324401855,
+      "learning_rate": 2.0010677954180446e-06,
+      "loss": 0.1003,
+      "step": 6229
+    },
+    {
+      "epoch": 3.2629649030906234,
+      "loss_partial": 0.0498046875,
+      "step": 6229
+    },
+    {
+      "epoch": 3.2629649030906234,
+      "loss_partial": 0.050048828125,
+      "step": 6229
+    },
+    {
+      "epoch": 3.2634887375589314,
+      "grad_norm": 1.1610990762710571,
+      "learning_rate": 1.9983250301165902e-06,
+      "loss": 0.0985,
+      "step": 6230
+    },
+    {
+      "epoch": 3.2634887375589314,
+      "loss_partial": 0.048828125,
+      "step": 6230
+    },
+    {
+      "epoch": 3.2634887375589314,
+      "loss_partial": 0.052734375,
+      "step": 6230
+    },
+    {
+      "epoch": 3.2640125720272395,
+      "grad_norm": 1.1193506717681885,
+      "learning_rate": 1.9955839371157247e-06,
+      "loss": 0.0998,
+      "step": 6231
+    },
+    {
+      "epoch": 3.2640125720272395,
+      "loss_partial": 0.04248046875,
+      "step": 6231
+    },
+    {
+      "epoch": 3.2640125720272395,
+      "loss_partial": 0.0458984375,
+      "step": 6231
+    },
+    {
+      "epoch": 3.2645364064955475,
+      "grad_norm": 0.8987656235694885,
+      "learning_rate": 1.992844516988318e-06,
+      "loss": 0.0984,
+      "step": 6232
+    },
+    {
+      "epoch": 3.2645364064955475,
+      "loss_partial": 0.046142578125,
+      "step": 6232
+    },
+    {
+      "epoch": 3.2645364064955475,
+      "loss_partial": 0.0400390625,
+      "step": 6232
+    },
+    {
+      "epoch": 3.2650602409638556,
+      "grad_norm": 0.9985066056251526,
+      "learning_rate": 1.9901067703068923e-06,
+      "loss": 0.0976,
+      "step": 6233
+    },
+    {
+      "epoch": 3.2650602409638556,
+      "loss_partial": 0.053466796875,
+      "step": 6233
+    },
+    {
+      "epoch": 3.2650602409638556,
+      "loss_partial": 0.052490234375,
+      "step": 6233
+    },
+    {
+      "epoch": 3.2655840754321632,
+      "grad_norm": 1.1471022367477417,
+      "learning_rate": 1.9873706976436202e-06,
+      "loss": 0.1034,
+      "step": 6234
+    },
+    {
+      "epoch": 3.2655840754321632,
+      "loss_partial": 0.050537109375,
+      "step": 6234
+    },
+    {
+      "epoch": 3.2655840754321632,
+      "loss_partial": 0.052734375,
+      "step": 6234
+    },
+    {
+      "epoch": 3.2661079099004713,
+      "grad_norm": 1.3362877368927002,
+      "learning_rate": 1.9846362995703205e-06,
+      "loss": 0.102,
+      "step": 6235
+    },
+    {
+      "epoch": 3.2661079099004713,
+      "loss_partial": 0.049072265625,
+      "step": 6235
+    },
+    {
+      "epoch": 3.2661079099004713,
+      "loss_partial": 0.04443359375,
+      "step": 6235
+    },
+    {
+      "epoch": 3.2666317443687793,
+      "grad_norm": 1.1374397277832031,
+      "learning_rate": 1.9819035766584714e-06,
+      "loss": 0.1012,
+      "step": 6236
+    },
+    {
+      "epoch": 3.2666317443687793,
+      "loss_partial": 0.047607421875,
+      "step": 6236
+    },
+    {
+      "epoch": 3.2666317443687793,
+      "loss_partial": 0.05078125,
+      "step": 6236
+    },
+    {
+      "epoch": 3.2671555788370874,
+      "grad_norm": 0.9187967777252197,
+      "learning_rate": 1.979172529479193e-06,
+      "loss": 0.098,
+      "step": 6237
+    },
+    {
+      "epoch": 3.2671555788370874,
+      "loss_partial": 0.05322265625,
+      "step": 6237
+    },
+    {
+      "epoch": 3.2671555788370874,
+      "loss_partial": 0.0498046875,
+      "step": 6237
+    },
+    {
+      "epoch": 3.2676794133053955,
+      "grad_norm": 1.3759546279907227,
+      "learning_rate": 1.9764431586032563e-06,
+      "loss": 0.1006,
+      "step": 6238
+    },
+    {
+      "epoch": 3.2676794133053955,
+      "loss_partial": 0.051513671875,
+      "step": 6238
+    },
+    {
+      "epoch": 3.2676794133053955,
+      "loss_partial": 0.0498046875,
+      "step": 6238
+    },
+    {
+      "epoch": 3.2682032477737035,
+      "grad_norm": 1.170305609703064,
+      "learning_rate": 1.9737154646010804e-06,
+      "loss": 0.1024,
+      "step": 6239
+    },
+    {
+      "epoch": 3.2682032477737035,
+      "loss_partial": 0.051025390625,
+      "step": 6239
+    },
+    {
+      "epoch": 3.2682032477737035,
+      "loss_partial": 0.0615234375,
+      "step": 6239
+    },
+    {
+      "epoch": 3.2687270822420116,
+      "grad_norm": 1.6106146574020386,
+      "learning_rate": 1.9709894480427427e-06,
+      "loss": 0.1027,
+      "step": 6240
+    },
+    {
+      "epoch": 3.2687270822420116,
+      "loss_partial": 0.05224609375,
+      "step": 6240
+    },
+    {
+      "epoch": 3.2687270822420116,
+      "loss_partial": 0.039306640625,
+      "step": 6240
+    },
+    {
+      "epoch": 3.2692509167103196,
+      "grad_norm": 1.3277153968811035,
+      "learning_rate": 1.96826510949796e-06,
+      "loss": 0.1019,
+      "step": 6241
+    },
+    {
+      "epoch": 3.2692509167103196,
+      "loss_partial": 0.042724609375,
+      "step": 6241
+    },
+    {
+      "epoch": 3.2692509167103196,
+      "loss_partial": 0.060546875,
+      "step": 6241
+    },
+    {
+      "epoch": 3.2697747511786277,
+      "grad_norm": 1.0574637651443481,
+      "learning_rate": 1.965542449536103e-06,
+      "loss": 0.0995,
+      "step": 6242
+    },
+    {
+      "epoch": 3.2697747511786277,
+      "loss_partial": 0.057373046875,
+      "step": 6242
+    },
+    {
+      "epoch": 3.2697747511786277,
+      "loss_partial": 0.0498046875,
+      "step": 6242
+    },
+    {
+      "epoch": 3.2702985856469358,
+      "grad_norm": 1.7751520872116089,
+      "learning_rate": 1.962821468726187e-06,
+      "loss": 0.0961,
+      "step": 6243
+    },
+    {
+      "epoch": 3.2702985856469358,
+      "loss_partial": 0.04296875,
+      "step": 6243
+    },
+    {
+      "epoch": 3.2702985856469358,
+      "loss_partial": 0.0517578125,
+      "step": 6243
+    },
+    {
+      "epoch": 3.270822420115244,
+      "grad_norm": 1.6776759624481201,
+      "learning_rate": 1.960102167636889e-06,
+      "loss": 0.0975,
+      "step": 6244
+    },
+    {
+      "epoch": 3.270822420115244,
+      "loss_partial": 0.056640625,
+      "step": 6244
+    },
+    {
+      "epoch": 3.270822420115244,
+      "loss_partial": 0.050048828125,
+      "step": 6244
+    },
+    {
+      "epoch": 3.2713462545835514,
+      "grad_norm": 1.140762209892273,
+      "learning_rate": 1.9573845468365193e-06,
+      "loss": 0.0988,
+      "step": 6245
+    },
+    {
+      "epoch": 3.2713462545835514,
+      "loss_partial": 0.04248046875,
+      "step": 6245
+    },
+    {
+      "epoch": 3.2713462545835514,
+      "loss_partial": 0.04638671875,
+      "step": 6245
+    },
+    {
+      "epoch": 3.2718700890518595,
+      "grad_norm": 1.576521873474121,
+      "learning_rate": 1.954668606893044e-06,
+      "loss": 0.0992,
+      "step": 6246
+    },
+    {
+      "epoch": 3.2718700890518595,
+      "loss_partial": 0.058349609375,
+      "step": 6246
+    },
+    {
+      "epoch": 3.2718700890518595,
+      "loss_partial": 0.054931640625,
+      "step": 6246
+    },
+    {
+      "epoch": 3.2723939235201676,
+      "grad_norm": 1.2263470888137817,
+      "learning_rate": 1.9519543483740765e-06,
+      "loss": 0.0997,
+      "step": 6247
+    },
+    {
+      "epoch": 3.2723939235201676,
+      "loss_partial": 0.050048828125,
+      "step": 6247
+    },
+    {
+      "epoch": 3.2723939235201676,
+      "loss_partial": 0.054931640625,
+      "step": 6247
+    },
+    {
+      "epoch": 3.2729177579884756,
+      "grad_norm": 1.5402799844741821,
+      "learning_rate": 1.949241771846886e-06,
+      "loss": 0.1,
+      "step": 6248
+    },
+    {
+      "epoch": 3.2729177579884756,
+      "loss_partial": 0.04833984375,
+      "step": 6248
+    },
+    {
+      "epoch": 3.2729177579884756,
+      "loss_partial": 0.048095703125,
+      "step": 6248
+    },
+    {
+      "epoch": 3.2734415924567837,
+      "grad_norm": 1.4215505123138428,
+      "learning_rate": 1.94653087787838e-06,
+      "loss": 0.0977,
+      "step": 6249
+    },
+    {
+      "epoch": 3.2734415924567837,
+      "loss_partial": 0.04541015625,
+      "step": 6249
+    },
+    {
+      "epoch": 3.2734415924567837,
+      "loss_partial": 0.0439453125,
+      "step": 6249
+    },
+    {
+      "epoch": 3.2739654269250917,
+      "grad_norm": 1.0203436613082886,
+      "learning_rate": 1.9438216670351195e-06,
+      "loss": 0.0972,
+      "step": 6250
+    },
+    {
+      "epoch": 3.2739654269250917,
+      "loss_partial": 0.044677734375,
+      "step": 6250
+    },
+    {
+      "epoch": 3.2739654269250917,
+      "loss_partial": 0.07958984375,
+      "step": 6250
+    },
+    {
+      "epoch": 3.2744892613934,
+      "grad_norm": 1.614823341369629,
+      "learning_rate": 1.941114139883311e-06,
+      "loss": 0.1009,
+      "step": 6251
+    },
+    {
+      "epoch": 3.2744892613934,
+      "loss_partial": 0.0517578125,
+      "step": 6251
+    },
+    {
+      "epoch": 3.2744892613934,
+      "loss_partial": 0.05029296875,
+      "step": 6251
+    },
+    {
+      "epoch": 3.275013095861708,
+      "grad_norm": 1.5145360231399536,
+      "learning_rate": 1.938408296988814e-06,
+      "loss": 0.105,
+      "step": 6252
+    },
+    {
+      "epoch": 3.275013095861708,
+      "loss_partial": 0.049560546875,
+      "step": 6252
+    },
+    {
+      "epoch": 3.275013095861708,
+      "loss_partial": 0.051025390625,
+      "step": 6252
+    },
+    {
+      "epoch": 3.2755369303300155,
+      "grad_norm": 1.1151278018951416,
+      "learning_rate": 1.935704138917133e-06,
+      "loss": 0.1012,
+      "step": 6253
+    },
+    {
+      "epoch": 3.2755369303300155,
+      "loss_partial": 0.0546875,
+      "step": 6253
+    },
+    {
+      "epoch": 3.2755369303300155,
+      "loss_partial": 0.052001953125,
+      "step": 6253
+    },
+    {
+      "epoch": 3.2760607647983235,
+      "grad_norm": 1.2780184745788574,
+      "learning_rate": 1.933001666233417e-06,
+      "loss": 0.101,
+      "step": 6254
+    },
+    {
+      "epoch": 3.2760607647983235,
+      "loss_partial": 0.04248046875,
+      "step": 6254
+    },
+    {
+      "epoch": 3.2760607647983235,
+      "loss_partial": 0.044921875,
+      "step": 6254
+    },
+    {
+      "epoch": 3.2765845992666316,
+      "grad_norm": 1.2072012424468994,
+      "learning_rate": 1.930300879502467e-06,
+      "loss": 0.0982,
+      "step": 6255
+    },
+    {
+      "epoch": 3.2765845992666316,
+      "loss_partial": 0.043701171875,
+      "step": 6255
+    },
+    {
+      "epoch": 3.2765845992666316,
+      "loss_partial": 0.06640625,
+      "step": 6255
+    },
+    {
+      "epoch": 3.2771084337349397,
+      "grad_norm": 1.0963724851608276,
+      "learning_rate": 1.927601779288736e-06,
+      "loss": 0.1023,
+      "step": 6256
+    },
+    {
+      "epoch": 3.2771084337349397,
+      "loss_partial": 0.060546875,
+      "step": 6256
+    },
+    {
+      "epoch": 3.2771084337349397,
+      "loss_partial": 0.052001953125,
+      "step": 6256
+    },
+    {
+      "epoch": 3.2776322682032477,
+      "grad_norm": 1.21784245967865,
+      "learning_rate": 1.9249043661563114e-06,
+      "loss": 0.0975,
+      "step": 6257
+    },
+    {
+      "epoch": 3.2776322682032477,
+      "loss_partial": 0.052978515625,
+      "step": 6257
+    },
+    {
+      "epoch": 3.2776322682032477,
+      "loss_partial": 0.044921875,
+      "step": 6257
+    },
+    {
+      "epoch": 3.2781561026715558,
+      "grad_norm": 1.1201103925704956,
+      "learning_rate": 1.922208640668941e-06,
+      "loss": 0.101,
+      "step": 6258
+    },
+    {
+      "epoch": 3.2781561026715558,
+      "loss_partial": 0.0615234375,
+      "step": 6258
+    },
+    {
+      "epoch": 3.2781561026715558,
+      "loss_partial": 0.054931640625,
+      "step": 6258
+    },
+    {
+      "epoch": 3.278679937139864,
+      "grad_norm": 0.9788439273834229,
+      "learning_rate": 1.91951460339001e-06,
+      "loss": 0.0986,
+      "step": 6259
+    },
+    {
+      "epoch": 3.278679937139864,
+      "loss_partial": 0.04345703125,
+      "step": 6259
+    },
+    {
+      "epoch": 3.278679937139864,
+      "loss_partial": 0.053955078125,
+      "step": 6259
+    },
+    {
+      "epoch": 3.279203771608172,
+      "grad_norm": 0.9261265397071838,
+      "learning_rate": 1.916822254882561e-06,
+      "loss": 0.1005,
+      "step": 6260
+    },
+    {
+      "epoch": 3.279203771608172,
+      "loss_partial": 0.04931640625,
+      "step": 6260
+    },
+    {
+      "epoch": 3.279203771608172,
+      "loss_partial": 0.048583984375,
+      "step": 6260
+    },
+    {
+      "epoch": 3.27972760607648,
+      "grad_norm": 1.0844011306762695,
+      "learning_rate": 1.9141315957092766e-06,
+      "loss": 0.099,
+      "step": 6261
+    },
+    {
+      "epoch": 3.27972760607648,
+      "loss_partial": 0.043212890625,
+      "step": 6261
+    },
+    {
+      "epoch": 3.27972760607648,
+      "loss_partial": 0.0546875,
+      "step": 6261
+    },
+    {
+      "epoch": 3.280251440544788,
+      "grad_norm": 1.1150869131088257,
+      "learning_rate": 1.911442626432486e-06,
+      "loss": 0.1018,
+      "step": 6262
+    },
+    {
+      "epoch": 3.280251440544788,
+      "loss_partial": 0.04296875,
+      "step": 6262
+    },
+    {
+      "epoch": 3.280251440544788,
+      "loss_partial": 0.05078125,
+      "step": 6262
+    },
+    {
+      "epoch": 3.280775275013096,
+      "grad_norm": 1.1586556434631348,
+      "learning_rate": 1.9087553476141698e-06,
+      "loss": 0.1001,
+      "step": 6263
+    },
+    {
+      "epoch": 3.280775275013096,
+      "loss_partial": 0.0419921875,
+      "step": 6263
+    },
+    {
+      "epoch": 3.280775275013096,
+      "loss_partial": 0.040771484375,
+      "step": 6263
+    },
+    {
+      "epoch": 3.2812991094814037,
+      "grad_norm": 1.0611424446105957,
+      "learning_rate": 1.9060697598159472e-06,
+      "loss": 0.0975,
+      "step": 6264
+    },
+    {
+      "epoch": 3.2812991094814037,
+      "loss_partial": 0.045654296875,
+      "step": 6264
+    },
+    {
+      "epoch": 3.2812991094814037,
+      "loss_partial": 0.055419921875,
+      "step": 6264
+    },
+    {
+      "epoch": 3.2818229439497117,
+      "grad_norm": 0.8563669323921204,
+      "learning_rate": 1.9033858635990966e-06,
+      "loss": 0.1005,
+      "step": 6265
+    },
+    {
+      "epoch": 3.2818229439497117,
+      "loss_partial": 0.056640625,
+      "step": 6265
+    },
+    {
+      "epoch": 3.2818229439497117,
+      "loss_partial": 0.038330078125,
+      "step": 6265
+    },
+    {
+      "epoch": 3.28234677841802,
+      "grad_norm": 1.1310127973556519,
+      "learning_rate": 1.900703659524533e-06,
+      "loss": 0.0965,
+      "step": 6266
+    },
+    {
+      "epoch": 3.28234677841802,
+      "loss_partial": 0.04345703125,
+      "step": 6266
+    },
+    {
+      "epoch": 3.28234677841802,
+      "loss_partial": 0.0576171875,
+      "step": 6266
+    },
+    {
+      "epoch": 3.282870612886328,
+      "grad_norm": 1.2938486337661743,
+      "learning_rate": 1.89802314815282e-06,
+      "loss": 0.0958,
+      "step": 6267
+    },
+    {
+      "epoch": 3.282870612886328,
+      "loss_partial": 0.044921875,
+      "step": 6267
+    },
+    {
+      "epoch": 3.282870612886328,
+      "loss_partial": 0.05078125,
+      "step": 6267
+    },
+    {
+      "epoch": 3.283394447354636,
+      "grad_norm": 1.1821938753128052,
+      "learning_rate": 1.8953443300441643e-06,
+      "loss": 0.0984,
+      "step": 6268
+    },
+    {
+      "epoch": 3.283394447354636,
+      "loss_partial": 0.053466796875,
+      "step": 6268
+    },
+    {
+      "epoch": 3.283394447354636,
+      "loss_partial": 0.0546875,
+      "step": 6268
+    },
+    {
+      "epoch": 3.283918281822944,
+      "grad_norm": 1.2344121932983398,
+      "learning_rate": 1.8926672057584317e-06,
+      "loss": 0.0989,
+      "step": 6269
+    },
+    {
+      "epoch": 3.283918281822944,
+      "loss_partial": 0.054443359375,
+      "step": 6269
+    },
+    {
+      "epoch": 3.283918281822944,
+      "loss_partial": 0.0498046875,
+      "step": 6269
+    },
+    {
+      "epoch": 3.284442116291252,
+      "grad_norm": 1.1259756088256836,
+      "learning_rate": 1.8899917758551166e-06,
+      "loss": 0.1027,
+      "step": 6270
+    },
+    {
+      "epoch": 3.284442116291252,
+      "loss_partial": 0.04052734375,
+      "step": 6270
+    },
+    {
+      "epoch": 3.284442116291252,
+      "loss_partial": 0.0576171875,
+      "step": 6270
+    },
+    {
+      "epoch": 3.28496595075956,
+      "grad_norm": 1.0042095184326172,
+      "learning_rate": 1.887318040893369e-06,
+      "loss": 0.0981,
+      "step": 6271
+    },
+    {
+      "epoch": 3.28496595075956,
+      "loss_partial": 0.048828125,
+      "step": 6271
+    },
+    {
+      "epoch": 3.28496595075956,
+      "loss_partial": 0.064453125,
+      "step": 6271
+    },
+    {
+      "epoch": 3.285489785227868,
+      "grad_norm": 1.0245558023452759,
+      "learning_rate": 1.8846460014319811e-06,
+      "loss": 0.1007,
+      "step": 6272
+    },
+    {
+      "epoch": 3.285489785227868,
+      "loss_partial": 0.062255859375,
+      "step": 6272
+    },
+    {
+      "epoch": 3.285489785227868,
+      "loss_partial": 0.05126953125,
+      "step": 6272
+    },
+    {
+      "epoch": 3.286013619696176,
+      "grad_norm": 1.1284250020980835,
+      "learning_rate": 1.881975658029398e-06,
+      "loss": 0.1012,
+      "step": 6273
+    },
+    {
+      "epoch": 3.286013619696176,
+      "loss_partial": 0.04833984375,
+      "step": 6273
+    },
+    {
+      "epoch": 3.286013619696176,
+      "loss_partial": 0.05078125,
+      "step": 6273
+    },
+    {
+      "epoch": 3.286537454164484,
+      "grad_norm": 1.4166203737258911,
+      "learning_rate": 1.8793070112437018e-06,
+      "loss": 0.0994,
+      "step": 6274
+    },
+    {
+      "epoch": 3.286537454164484,
+      "loss_partial": 0.055419921875,
+      "step": 6274
+    },
+    {
+      "epoch": 3.286537454164484,
+      "loss_partial": 0.040771484375,
+      "step": 6274
+    },
+    {
+      "epoch": 3.287061288632792,
+      "grad_norm": 0.9766794443130493,
+      "learning_rate": 1.876640061632624e-06,
+      "loss": 0.0995,
+      "step": 6275
+    },
+    {
+      "epoch": 3.287061288632792,
+      "loss_partial": 0.056640625,
+      "step": 6275
+    },
+    {
+      "epoch": 3.287061288632792,
+      "loss_partial": 0.043701171875,
+      "step": 6275
+    },
+    {
+      "epoch": 3.2875851231011,
+      "grad_norm": 1.633286714553833,
+      "learning_rate": 1.873974809753537e-06,
+      "loss": 0.1024,
+      "step": 6276
+    },
+    {
+      "epoch": 3.2875851231011,
+      "loss_partial": 0.0546875,
+      "step": 6276
+    },
+    {
+      "epoch": 3.2875851231011,
+      "loss_partial": 0.04931640625,
+      "step": 6276
+    },
+    {
+      "epoch": 3.288108957569408,
+      "grad_norm": 1.0570807456970215,
+      "learning_rate": 1.8713112561634671e-06,
+      "loss": 0.1,
+      "step": 6277
+    },
+    {
+      "epoch": 3.288108957569408,
+      "loss_partial": 0.047119140625,
+      "step": 6277
+    },
+    {
+      "epoch": 3.288108957569408,
+      "loss_partial": 0.040771484375,
+      "step": 6277
+    },
+    {
+      "epoch": 3.288632792037716,
+      "grad_norm": 0.8270609974861145,
+      "learning_rate": 1.8686494014190782e-06,
+      "loss": 0.1004,
+      "step": 6278
+    },
+    {
+      "epoch": 3.288632792037716,
+      "loss_partial": 0.047119140625,
+      "step": 6278
+    },
+    {
+      "epoch": 3.288632792037716,
+      "loss_partial": 0.04638671875,
+      "step": 6278
+    },
+    {
+      "epoch": 3.289156626506024,
+      "grad_norm": 1.3154935836791992,
+      "learning_rate": 1.8659892460766827e-06,
+      "loss": 0.0996,
+      "step": 6279
+    },
+    {
+      "epoch": 3.289156626506024,
+      "loss_partial": 0.0478515625,
+      "step": 6279
+    },
+    {
+      "epoch": 3.289156626506024,
+      "loss_partial": 0.053466796875,
+      "step": 6279
+    },
+    {
+      "epoch": 3.289680460974332,
+      "grad_norm": 1.5014888048171997,
+      "learning_rate": 1.8633307906922316e-06,
+      "loss": 0.0995,
+      "step": 6280
+    },
+    {
+      "epoch": 3.289680460974332,
+      "loss_partial": 0.050537109375,
+      "step": 6280
+    },
+    {
+      "epoch": 3.289680460974332,
+      "loss_partial": 0.05029296875,
+      "step": 6280
+    },
+    {
+      "epoch": 3.2902042954426403,
+      "grad_norm": 1.089076042175293,
+      "learning_rate": 1.860674035821336e-06,
+      "loss": 0.1013,
+      "step": 6281
+    },
+    {
+      "epoch": 3.2902042954426403,
+      "loss_partial": 0.051025390625,
+      "step": 6281
+    },
+    {
+      "epoch": 3.2902042954426403,
+      "loss_partial": 0.057861328125,
+      "step": 6281
+    },
+    {
+      "epoch": 3.2907281299109483,
+      "grad_norm": 1.2996855974197388,
+      "learning_rate": 1.8580189820192317e-06,
+      "loss": 0.0991,
+      "step": 6282
+    },
+    {
+      "epoch": 3.2907281299109483,
+      "loss_partial": 0.050537109375,
+      "step": 6282
+    },
+    {
+      "epoch": 3.2907281299109483,
+      "loss_partial": 0.0478515625,
+      "step": 6282
+    },
+    {
+      "epoch": 3.2912519643792564,
+      "grad_norm": 1.1956380605697632,
+      "learning_rate": 1.855365629840812e-06,
+      "loss": 0.0981,
+      "step": 6283
+    },
+    {
+      "epoch": 3.2912519643792564,
+      "loss_partial": 0.045654296875,
+      "step": 6283
+    },
+    {
+      "epoch": 3.2912519643792564,
+      "loss_partial": 0.054931640625,
+      "step": 6283
+    },
+    {
+      "epoch": 3.291775798847564,
+      "grad_norm": 1.5772234201431274,
+      "learning_rate": 1.852713979840608e-06,
+      "loss": 0.0992,
+      "step": 6284
+    },
+    {
+      "epoch": 3.291775798847564,
+      "loss_partial": 0.04833984375,
+      "step": 6284
+    },
+    {
+      "epoch": 3.291775798847564,
+      "loss_partial": 0.049072265625,
+      "step": 6284
+    },
+    {
+      "epoch": 3.292299633315872,
+      "grad_norm": 1.3008034229278564,
+      "learning_rate": 1.8500640325728025e-06,
+      "loss": 0.1024,
+      "step": 6285
+    },
+    {
+      "epoch": 3.292299633315872,
+      "loss_partial": 0.047607421875,
+      "step": 6285
+    },
+    {
+      "epoch": 3.292299633315872,
+      "loss_partial": 0.0537109375,
+      "step": 6285
+    },
+    {
+      "epoch": 3.29282346778418,
+      "grad_norm": 1.1638585329055786,
+      "learning_rate": 1.847415788591218e-06,
+      "loss": 0.1018,
+      "step": 6286
+    },
+    {
+      "epoch": 3.29282346778418,
+      "loss_partial": 0.052001953125,
+      "step": 6286
+    },
+    {
+      "epoch": 3.29282346778418,
+      "loss_partial": 0.056884765625,
+      "step": 6286
+    },
+    {
+      "epoch": 3.293347302252488,
+      "grad_norm": 1.0611361265182495,
+      "learning_rate": 1.8447692484493185e-06,
+      "loss": 0.1004,
+      "step": 6287
+    },
+    {
+      "epoch": 3.293347302252488,
+      "loss_partial": 0.048583984375,
+      "step": 6287
+    },
+    {
+      "epoch": 3.293347302252488,
+      "loss_partial": 0.0439453125,
+      "step": 6287
+    },
+    {
+      "epoch": 3.2938711367207962,
+      "grad_norm": 1.0151615142822266,
+      "learning_rate": 1.8421244127002135e-06,
+      "loss": 0.1008,
+      "step": 6288
+    },
+    {
+      "epoch": 3.2938711367207962,
+      "loss_partial": 0.049072265625,
+      "step": 6288
+    },
+    {
+      "epoch": 3.2938711367207962,
+      "loss_partial": 0.046630859375,
+      "step": 6288
+    },
+    {
+      "epoch": 3.2943949711891043,
+      "grad_norm": 1.1295379400253296,
+      "learning_rate": 1.8394812818966624e-06,
+      "loss": 0.1004,
+      "step": 6289
+    },
+    {
+      "epoch": 3.2943949711891043,
+      "loss_partial": 0.0458984375,
+      "step": 6289
+    },
+    {
+      "epoch": 3.2943949711891043,
+      "loss_partial": 0.051025390625,
+      "step": 6289
+    },
+    {
+      "epoch": 3.2949188056574124,
+      "grad_norm": 1.4369608163833618,
+      "learning_rate": 1.8368398565910605e-06,
+      "loss": 0.1018,
+      "step": 6290
+    },
+    {
+      "epoch": 3.2949188056574124,
+      "loss_partial": 0.04736328125,
+      "step": 6290
+    },
+    {
+      "epoch": 3.2949188056574124,
+      "loss_partial": 0.048095703125,
+      "step": 6290
+    },
+    {
+      "epoch": 3.2954426401257204,
+      "grad_norm": 1.253970980644226,
+      "learning_rate": 1.8342001373354478e-06,
+      "loss": 0.1015,
+      "step": 6291
+    },
+    {
+      "epoch": 3.2954426401257204,
+      "loss_partial": 0.046875,
+      "step": 6291
+    },
+    {
+      "epoch": 3.2954426401257204,
+      "loss_partial": 0.052001953125,
+      "step": 6291
+    },
+    {
+      "epoch": 3.2959664745940285,
+      "grad_norm": 1.0790249109268188,
+      "learning_rate": 1.8315621246815096e-06,
+      "loss": 0.1002,
+      "step": 6292
+    },
+    {
+      "epoch": 3.2959664745940285,
+      "loss_partial": 0.053955078125,
+      "step": 6292
+    },
+    {
+      "epoch": 3.2959664745940285,
+      "loss_partial": 0.052978515625,
+      "step": 6292
+    },
+    {
+      "epoch": 3.296490309062336,
+      "grad_norm": 1.1372427940368652,
+      "learning_rate": 1.8289258191805793e-06,
+      "loss": 0.0993,
+      "step": 6293
+    },
+    {
+      "epoch": 3.296490309062336,
+      "loss_partial": 0.05908203125,
+      "step": 6293
+    },
+    {
+      "epoch": 3.296490309062336,
+      "loss_partial": 0.0634765625,
+      "step": 6293
+    },
+    {
+      "epoch": 3.297014143530644,
+      "grad_norm": 1.1485182046890259,
+      "learning_rate": 1.8262912213836214e-06,
+      "loss": 0.0968,
+      "step": 6294
+    },
+    {
+      "epoch": 3.297014143530644,
+      "loss_partial": 0.06298828125,
+      "step": 6294
+    },
+    {
+      "epoch": 3.297014143530644,
+      "loss_partial": 0.04345703125,
+      "step": 6294
+    },
+    {
+      "epoch": 3.297537977998952,
+      "grad_norm": 1.168144702911377,
+      "learning_rate": 1.8236583318412549e-06,
+      "loss": 0.0998,
+      "step": 6295
+    },
+    {
+      "epoch": 3.297537977998952,
+      "loss_partial": 0.0400390625,
+      "step": 6295
+    },
+    {
+      "epoch": 3.297537977998952,
+      "loss_partial": 0.046630859375,
+      "step": 6295
+    },
+    {
+      "epoch": 3.2980618124672603,
+      "grad_norm": 0.9230632781982422,
+      "learning_rate": 1.8210271511037325e-06,
+      "loss": 0.0955,
+      "step": 6296
+    },
+    {
+      "epoch": 3.2980618124672603,
+      "loss_partial": 0.051513671875,
+      "step": 6296
+    },
+    {
+      "epoch": 3.2980618124672603,
+      "loss_partial": 0.05419921875,
+      "step": 6296
+    },
+    {
+      "epoch": 3.2985856469355683,
+      "grad_norm": 1.413541555404663,
+      "learning_rate": 1.8183976797209612e-06,
+      "loss": 0.1025,
+      "step": 6297
+    },
+    {
+      "epoch": 3.2985856469355683,
+      "loss_partial": 0.050537109375,
+      "step": 6297
+    },
+    {
+      "epoch": 3.2985856469355683,
+      "loss_partial": 0.044189453125,
+      "step": 6297
+    },
+    {
+      "epoch": 3.2991094814038764,
+      "grad_norm": 1.020051121711731,
+      "learning_rate": 1.8157699182424816e-06,
+      "loss": 0.1039,
+      "step": 6298
+    },
+    {
+      "epoch": 3.2991094814038764,
+      "loss_partial": 0.0439453125,
+      "step": 6298
+    },
+    {
+      "epoch": 3.2991094814038764,
+      "loss_partial": 0.047119140625,
+      "step": 6298
+    },
+    {
+      "epoch": 3.2996333158721844,
+      "grad_norm": 1.126955509185791,
+      "learning_rate": 1.8131438672174796e-06,
+      "loss": 0.0987,
+      "step": 6299
+    },
+    {
+      "epoch": 3.2996333158721844,
+      "loss_partial": 0.05224609375,
+      "step": 6299
+    },
+    {
+      "epoch": 3.2996333158721844,
+      "loss_partial": 0.046875,
+      "step": 6299
+    },
+    {
+      "epoch": 3.3001571503404925,
+      "grad_norm": 1.2491909265518188,
+      "learning_rate": 1.8105195271947795e-06,
+      "loss": 0.1007,
+      "step": 6300
+    },
+    {
+      "epoch": 3.3001571503404925,
+      "loss_partial": 0.050537109375,
+      "step": 6300
+    },
+    {
+      "epoch": 3.3001571503404925,
+      "loss_partial": 0.05712890625,
+      "step": 6300
+    },
+    {
+      "epoch": 3.3006809848088006,
+      "grad_norm": 0.9299445748329163,
+      "learning_rate": 1.8078968987228595e-06,
+      "loss": 0.1017,
+      "step": 6301
+    },
+    {
+      "epoch": 3.3006809848088006,
+      "loss_partial": 0.0439453125,
+      "step": 6301
+    },
+    {
+      "epoch": 3.3006809848088006,
+      "loss_partial": 0.046630859375,
+      "step": 6301
+    },
+    {
+      "epoch": 3.3012048192771086,
+      "grad_norm": 0.9706397652626038,
+      "learning_rate": 1.8052759823498289e-06,
+      "loss": 0.0983,
+      "step": 6302
+    },
+    {
+      "epoch": 3.3012048192771086,
+      "loss_partial": 0.0498046875,
+      "step": 6302
+    },
+    {
+      "epoch": 3.3012048192771086,
+      "loss_partial": 0.0498046875,
+      "step": 6302
+    },
+    {
+      "epoch": 3.3017286537454167,
+      "grad_norm": 1.3219958543777466,
+      "learning_rate": 1.8026567786234438e-06,
+      "loss": 0.1003,
+      "step": 6303
+    },
+    {
+      "epoch": 3.3017286537454167,
+      "loss_partial": 0.049560546875,
+      "step": 6303
+    },
+    {
+      "epoch": 3.3017286537454167,
+      "loss_partial": 0.054443359375,
+      "step": 6303
+    },
+    {
+      "epoch": 3.3022524882137243,
+      "grad_norm": 1.137734293937683,
+      "learning_rate": 1.8000392880910977e-06,
+      "loss": 0.0982,
+      "step": 6304
+    },
+    {
+      "epoch": 3.3022524882137243,
+      "loss_partial": 0.04150390625,
+      "step": 6304
+    },
+    {
+      "epoch": 3.3022524882137243,
+      "loss_partial": 0.05712890625,
+      "step": 6304
+    },
+    {
+      "epoch": 3.3027763226820324,
+      "grad_norm": 1.303082823753357,
+      "learning_rate": 1.7974235112998374e-06,
+      "loss": 0.1042,
+      "step": 6305
+    },
+    {
+      "epoch": 3.3027763226820324,
+      "loss_partial": 0.05078125,
+      "step": 6305
+    },
+    {
+      "epoch": 3.3027763226820324,
+      "loss_partial": 0.0546875,
+      "step": 6305
+    },
+    {
+      "epoch": 3.3033001571503404,
+      "grad_norm": 1.365654706954956,
+      "learning_rate": 1.7948094487963407e-06,
+      "loss": 0.1002,
+      "step": 6306
+    },
+    {
+      "epoch": 3.3033001571503404,
+      "loss_partial": 0.0458984375,
+      "step": 6306
+    },
+    {
+      "epoch": 3.3033001571503404,
+      "loss_partial": 0.051025390625,
+      "step": 6306
+    },
+    {
+      "epoch": 3.3038239916186485,
+      "grad_norm": 1.242458462715149,
+      "learning_rate": 1.792197101126929e-06,
+      "loss": 0.0991,
+      "step": 6307
+    },
+    {
+      "epoch": 3.3038239916186485,
+      "loss_partial": 0.0419921875,
+      "step": 6307
+    },
+    {
+      "epoch": 3.3038239916186485,
+      "loss_partial": 0.05419921875,
+      "step": 6307
+    },
+    {
+      "epoch": 3.3043478260869565,
+      "grad_norm": 1.3800677061080933,
+      "learning_rate": 1.7895864688375642e-06,
+      "loss": 0.1001,
+      "step": 6308
+    },
+    {
+      "epoch": 3.3043478260869565,
+      "loss_partial": 0.043701171875,
+      "step": 6308
+    },
+    {
+      "epoch": 3.3043478260869565,
+      "loss_partial": 0.046142578125,
+      "step": 6308
+    },
+    {
+      "epoch": 3.3048716605552646,
+      "grad_norm": 1.2100414037704468,
+      "learning_rate": 1.7869775524738587e-06,
+      "loss": 0.0986,
+      "step": 6309
+    },
+    {
+      "epoch": 3.3048716605552646,
+      "loss_partial": 0.053466796875,
+      "step": 6309
+    },
+    {
+      "epoch": 3.3048716605552646,
+      "loss_partial": 0.06396484375,
+      "step": 6309
+    },
+    {
+      "epoch": 3.3053954950235727,
+      "grad_norm": 1.015973448753357,
+      "learning_rate": 1.7843703525810564e-06,
+      "loss": 0.0985,
+      "step": 6310
+    },
+    {
+      "epoch": 3.3053954950235727,
+      "loss_partial": 0.04833984375,
+      "step": 6310
+    },
+    {
+      "epoch": 3.3053954950235727,
+      "loss_partial": 0.05810546875,
+      "step": 6310
+    },
+    {
+      "epoch": 3.3059193294918807,
+      "grad_norm": 1.2633296251296997,
+      "learning_rate": 1.7817648697040468e-06,
+      "loss": 0.1014,
+      "step": 6311
+    },
+    {
+      "epoch": 3.3059193294918807,
+      "loss_partial": 0.056396484375,
+      "step": 6311
+    },
+    {
+      "epoch": 3.3059193294918807,
+      "loss_partial": 0.0419921875,
+      "step": 6311
+    },
+    {
+      "epoch": 3.3064431639601883,
+      "grad_norm": 1.1373884677886963,
+      "learning_rate": 1.7791611043873558e-06,
+      "loss": 0.1003,
+      "step": 6312
+    },
+    {
+      "epoch": 3.3064431639601883,
+      "loss_partial": 0.046142578125,
+      "step": 6312
+    },
+    {
+      "epoch": 3.3064431639601883,
+      "loss_partial": 0.035400390625,
+      "step": 6312
+    },
+    {
+      "epoch": 3.3069669984284964,
+      "grad_norm": 1.2177455425262451,
+      "learning_rate": 1.7765590571751623e-06,
+      "loss": 0.1011,
+      "step": 6313
+    },
+    {
+      "epoch": 3.3069669984284964,
+      "loss_partial": 0.05224609375,
+      "step": 6313
+    },
+    {
+      "epoch": 3.3069669984284964,
+      "loss_partial": 0.041259765625,
+      "step": 6313
+    },
+    {
+      "epoch": 3.3074908328968045,
+      "grad_norm": 1.5558112859725952,
+      "learning_rate": 1.7739587286112724e-06,
+      "loss": 0.1006,
+      "step": 6314
+    },
+    {
+      "epoch": 3.3074908328968045,
+      "loss_partial": 0.055419921875,
+      "step": 6314
+    },
+    {
+      "epoch": 3.3074908328968045,
+      "loss_partial": 0.04638671875,
+      "step": 6314
+    },
+    {
+      "epoch": 3.3080146673651125,
+      "grad_norm": 1.1722605228424072,
+      "learning_rate": 1.7713601192391404e-06,
+      "loss": 0.1042,
+      "step": 6315
+    },
+    {
+      "epoch": 3.3080146673651125,
+      "loss_partial": 0.04248046875,
+      "step": 6315
+    },
+    {
+      "epoch": 3.3080146673651125,
+      "loss_partial": 0.061767578125,
+      "step": 6315
+    },
+    {
+      "epoch": 3.3085385018334206,
+      "grad_norm": 1.2426583766937256,
+      "learning_rate": 1.7687632296018563e-06,
+      "loss": 0.0989,
+      "step": 6316
+    },
+    {
+      "epoch": 3.3085385018334206,
+      "loss_partial": 0.04541015625,
+      "step": 6316
+    },
+    {
+      "epoch": 3.3085385018334206,
+      "loss_partial": 0.061767578125,
+      "step": 6316
+    },
+    {
+      "epoch": 3.3090623363017286,
+      "grad_norm": 0.8946720361709595,
+      "learning_rate": 1.7661680602421594e-06,
+      "loss": 0.0992,
+      "step": 6317
+    },
+    {
+      "epoch": 3.3090623363017286,
+      "loss_partial": 0.04541015625,
+      "step": 6317
+    },
+    {
+      "epoch": 3.3090623363017286,
+      "loss_partial": 0.050048828125,
+      "step": 6317
+    },
+    {
+      "epoch": 3.3095861707700367,
+      "grad_norm": 1.0065715312957764,
+      "learning_rate": 1.7635746117024233e-06,
+      "loss": 0.0967,
+      "step": 6318
+    },
+    {
+      "epoch": 3.3095861707700367,
+      "loss_partial": 0.042724609375,
+      "step": 6318
+    },
+    {
+      "epoch": 3.3095861707700367,
+      "loss_partial": 0.048828125,
+      "step": 6318
+    },
+    {
+      "epoch": 3.3101100052383448,
+      "grad_norm": 1.0464650392532349,
+      "learning_rate": 1.760982884524659e-06,
+      "loss": 0.1014,
+      "step": 6319
+    },
+    {
+      "epoch": 3.3101100052383448,
+      "loss_partial": 0.041748046875,
+      "step": 6319
+    },
+    {
+      "epoch": 3.3101100052383448,
+      "loss_partial": 0.06298828125,
+      "step": 6319
+    },
+    {
+      "epoch": 3.310633839706653,
+      "grad_norm": 0.9719547629356384,
+      "learning_rate": 1.7583928792505234e-06,
+      "loss": 0.1004,
+      "step": 6320
+    },
+    {
+      "epoch": 3.310633839706653,
+      "loss_partial": 0.050048828125,
+      "step": 6320
+    },
+    {
+      "epoch": 3.310633839706653,
+      "loss_partial": 0.04931640625,
+      "step": 6320
+    },
+    {
+      "epoch": 3.311157674174961,
+      "grad_norm": 0.8456039428710938,
+      "learning_rate": 1.7558045964213088e-06,
+      "loss": 0.1003,
+      "step": 6321
+    },
+    {
+      "epoch": 3.311157674174961,
+      "loss_partial": 0.0634765625,
+      "step": 6321
+    },
+    {
+      "epoch": 3.311157674174961,
+      "loss_partial": 0.05126953125,
+      "step": 6321
+    },
+    {
+      "epoch": 3.311681508643269,
+      "grad_norm": 1.091727614402771,
+      "learning_rate": 1.753218036577955e-06,
+      "loss": 0.0985,
+      "step": 6322
+    },
+    {
+      "epoch": 3.311681508643269,
+      "loss_partial": 0.05322265625,
+      "step": 6322
+    },
+    {
+      "epoch": 3.311681508643269,
+      "loss_partial": 0.0556640625,
+      "step": 6322
+    },
+    {
+      "epoch": 3.3122053431115765,
+      "grad_norm": 1.022923231124878,
+      "learning_rate": 1.750633200261036e-06,
+      "loss": 0.1014,
+      "step": 6323
+    },
+    {
+      "epoch": 3.3122053431115765,
+      "loss_partial": 0.046142578125,
+      "step": 6323
+    },
+    {
+      "epoch": 3.3122053431115765,
+      "loss_partial": 0.052978515625,
+      "step": 6323
+    },
+    {
+      "epoch": 3.3127291775798846,
+      "grad_norm": 0.8811379075050354,
+      "learning_rate": 1.7480500880107654e-06,
+      "loss": 0.103,
+      "step": 6324
+    },
+    {
+      "epoch": 3.3127291775798846,
+      "loss_partial": 0.04150390625,
+      "step": 6324
+    },
+    {
+      "epoch": 3.3127291775798846,
+      "loss_partial": 0.044921875,
+      "step": 6324
+    },
+    {
+      "epoch": 3.3132530120481927,
+      "grad_norm": 1.0094445943832397,
+      "learning_rate": 1.7454687003669946e-06,
+      "loss": 0.0979,
+      "step": 6325
+    },
+    {
+      "epoch": 3.3132530120481927,
+      "loss_partial": 0.058837890625,
+      "step": 6325
+    },
+    {
+      "epoch": 3.3132530120481927,
+      "loss_partial": 0.04736328125,
+      "step": 6325
+    },
+    {
+      "epoch": 3.3137768465165007,
+      "grad_norm": 1.0379531383514404,
+      "learning_rate": 1.7428890378692242e-06,
+      "loss": 0.1006,
+      "step": 6326
+    },
+    {
+      "epoch": 3.3137768465165007,
+      "loss_partial": 0.05322265625,
+      "step": 6326
+    },
+    {
+      "epoch": 3.3137768465165007,
+      "loss_partial": 0.0498046875,
+      "step": 6326
+    },
+    {
+      "epoch": 3.314300680984809,
+      "grad_norm": 0.9724178314208984,
+      "learning_rate": 1.7403111010565844e-06,
+      "loss": 0.0986,
+      "step": 6327
+    },
+    {
+      "epoch": 3.314300680984809,
+      "loss_partial": 0.04345703125,
+      "step": 6327
+    },
+    {
+      "epoch": 3.314300680984809,
+      "loss_partial": 0.052001953125,
+      "step": 6327
+    },
+    {
+      "epoch": 3.314824515453117,
+      "grad_norm": 1.042884349822998,
+      "learning_rate": 1.737734890467847e-06,
+      "loss": 0.0995,
+      "step": 6328
+    },
+    {
+      "epoch": 3.314824515453117,
+      "loss_partial": 0.05517578125,
+      "step": 6328
+    },
+    {
+      "epoch": 3.314824515453117,
+      "loss_partial": 0.044677734375,
+      "step": 6328
+    },
+    {
+      "epoch": 3.315348349921425,
+      "grad_norm": 0.8761633038520813,
+      "learning_rate": 1.735160406641424e-06,
+      "loss": 0.1032,
+      "step": 6329
+    },
+    {
+      "epoch": 3.315348349921425,
+      "loss_partial": 0.052001953125,
+      "step": 6329
+    },
+    {
+      "epoch": 3.315348349921425,
+      "loss_partial": 0.050537109375,
+      "step": 6329
+    },
+    {
+      "epoch": 3.315872184389733,
+      "grad_norm": 1.0562744140625,
+      "learning_rate": 1.7325876501153716e-06,
+      "loss": 0.1005,
+      "step": 6330
+    },
+    {
+      "epoch": 3.315872184389733,
+      "loss_partial": 0.060546875,
+      "step": 6330
+    },
+    {
+      "epoch": 3.315872184389733,
+      "loss_partial": 0.05126953125,
+      "step": 6330
+    },
+    {
+      "epoch": 3.316396018858041,
+      "grad_norm": 1.2685847282409668,
+      "learning_rate": 1.7300166214273751e-06,
+      "loss": 0.1014,
+      "step": 6331
+    },
+    {
+      "epoch": 3.316396018858041,
+      "loss_partial": 0.04931640625,
+      "step": 6331
+    },
+    {
+      "epoch": 3.316396018858041,
+      "loss_partial": 0.043701171875,
+      "step": 6331
+    },
+    {
+      "epoch": 3.3169198533263486,
+      "grad_norm": 1.4031473398208618,
+      "learning_rate": 1.7274473211147636e-06,
+      "loss": 0.0996,
+      "step": 6332
+    },
+    {
+      "epoch": 3.3169198533263486,
+      "loss_partial": 0.04833984375,
+      "step": 6332
+    },
+    {
+      "epoch": 3.3169198533263486,
+      "loss_partial": 0.046630859375,
+      "step": 6332
+    },
+    {
+      "epoch": 3.3174436877946567,
+      "grad_norm": 0.8584535121917725,
+      "learning_rate": 1.7248797497145042e-06,
+      "loss": 0.1033,
+      "step": 6333
+    },
+    {
+      "epoch": 3.3174436877946567,
+      "loss_partial": 0.04052734375,
+      "step": 6333
+    },
+    {
+      "epoch": 3.3174436877946567,
+      "loss_partial": 0.039306640625,
+      "step": 6333
+    },
+    {
+      "epoch": 3.3179675222629648,
+      "grad_norm": 1.0606982707977295,
+      "learning_rate": 1.7223139077632067e-06,
+      "loss": 0.0979,
+      "step": 6334
+    },
+    {
+      "epoch": 3.3179675222629648,
+      "loss_partial": 0.056640625,
+      "step": 6334
+    },
+    {
+      "epoch": 3.3179675222629648,
+      "loss_partial": 0.044921875,
+      "step": 6334
+    },
+    {
+      "epoch": 3.318491356731273,
+      "grad_norm": 1.4427021741867065,
+      "learning_rate": 1.7197497957971166e-06,
+      "loss": 0.1023,
+      "step": 6335
+    },
+    {
+      "epoch": 3.318491356731273,
+      "loss_partial": 0.051513671875,
+      "step": 6335
+    },
+    {
+      "epoch": 3.318491356731273,
+      "loss_partial": 0.0439453125,
+      "step": 6335
+    },
+    {
+      "epoch": 3.319015191199581,
+      "grad_norm": 1.2130082845687866,
+      "learning_rate": 1.7171874143521138e-06,
+      "loss": 0.0992,
+      "step": 6336
+    },
+    {
+      "epoch": 3.319015191199581,
+      "loss_partial": 0.049560546875,
+      "step": 6336
+    },
+    {
+      "epoch": 3.319015191199581,
+      "loss_partial": 0.0439453125,
+      "step": 6336
+    },
+    {
+      "epoch": 3.319539025667889,
+      "grad_norm": 1.142221212387085,
+      "learning_rate": 1.7146267639637205e-06,
+      "loss": 0.0965,
+      "step": 6337
+    },
+    {
+      "epoch": 3.319539025667889,
+      "loss_partial": 0.04541015625,
+      "step": 6337
+    },
+    {
+      "epoch": 3.319539025667889,
+      "loss_partial": 0.06494140625,
+      "step": 6337
+    },
+    {
+      "epoch": 3.320062860136197,
+      "grad_norm": 1.3588074445724487,
+      "learning_rate": 1.7120678451671014e-06,
+      "loss": 0.0976,
+      "step": 6338
+    },
+    {
+      "epoch": 3.320062860136197,
+      "loss_partial": 0.051513671875,
+      "step": 6338
+    },
+    {
+      "epoch": 3.320062860136197,
+      "loss_partial": 0.051025390625,
+      "step": 6338
+    },
+    {
+      "epoch": 3.320586694604505,
+      "grad_norm": 1.1820887327194214,
+      "learning_rate": 1.709510658497051e-06,
+      "loss": 0.0981,
+      "step": 6339
+    },
+    {
+      "epoch": 3.320586694604505,
+      "loss_partial": 0.04296875,
+      "step": 6339
+    },
+    {
+      "epoch": 3.320586694604505,
+      "loss_partial": 0.05126953125,
+      "step": 6339
+    },
+    {
+      "epoch": 3.321110529072813,
+      "grad_norm": 1.0290173292160034,
+      "learning_rate": 1.7069552044880056e-06,
+      "loss": 0.0988,
+      "step": 6340
+    },
+    {
+      "epoch": 3.321110529072813,
+      "loss_partial": 0.048828125,
+      "step": 6340
+    },
+    {
+      "epoch": 3.321110529072813,
+      "loss_partial": 0.04296875,
+      "step": 6340
+    },
+    {
+      "epoch": 3.321634363541121,
+      "grad_norm": 1.3409943580627441,
+      "learning_rate": 1.7044014836740374e-06,
+      "loss": 0.0974,
+      "step": 6341
+    },
+    {
+      "epoch": 3.321634363541121,
+      "loss_partial": 0.049072265625,
+      "step": 6341
+    },
+    {
+      "epoch": 3.321634363541121,
+      "loss_partial": 0.046875,
+      "step": 6341
+    },
+    {
+      "epoch": 3.3221581980094292,
+      "grad_norm": 1.5226353406906128,
+      "learning_rate": 1.7018494965888644e-06,
+      "loss": 0.1016,
+      "step": 6342
+    },
+    {
+      "epoch": 3.3221581980094292,
+      "loss_partial": 0.05078125,
+      "step": 6342
+    },
+    {
+      "epoch": 3.3221581980094292,
+      "loss_partial": 0.043212890625,
+      "step": 6342
+    },
+    {
+      "epoch": 3.322682032477737,
+      "grad_norm": 0.9288185834884644,
+      "learning_rate": 1.699299243765834e-06,
+      "loss": 0.0997,
+      "step": 6343
+    },
+    {
+      "epoch": 3.322682032477737,
+      "loss_partial": 0.04833984375,
+      "step": 6343
+    },
+    {
+      "epoch": 3.322682032477737,
+      "loss_partial": 0.04833984375,
+      "step": 6343
+    },
+    {
+      "epoch": 3.323205866946045,
+      "grad_norm": 1.1564791202545166,
+      "learning_rate": 1.696750725737929e-06,
+      "loss": 0.0972,
+      "step": 6344
+    },
+    {
+      "epoch": 3.323205866946045,
+      "loss_partial": 0.05224609375,
+      "step": 6344
+    },
+    {
+      "epoch": 3.323205866946045,
+      "loss_partial": 0.04296875,
+      "step": 6344
+    },
+    {
+      "epoch": 3.323729701414353,
+      "grad_norm": 1.3135111331939697,
+      "learning_rate": 1.6942039430377743e-06,
+      "loss": 0.0976,
+      "step": 6345
+    },
+    {
+      "epoch": 3.323729701414353,
+      "loss_partial": 0.054443359375,
+      "step": 6345
+    },
+    {
+      "epoch": 3.323729701414353,
+      "loss_partial": 0.0439453125,
+      "step": 6345
+    },
+    {
+      "epoch": 3.324253535882661,
+      "grad_norm": 1.089076280593872,
+      "learning_rate": 1.6916588961976365e-06,
+      "loss": 0.096,
+      "step": 6346
+    },
+    {
+      "epoch": 3.324253535882661,
+      "loss_partial": 0.046875,
+      "step": 6346
+    },
+    {
+      "epoch": 3.324253535882661,
+      "loss_partial": 0.0458984375,
+      "step": 6346
+    },
+    {
+      "epoch": 3.324777370350969,
+      "grad_norm": 1.3543572425842285,
+      "learning_rate": 1.6891155857494123e-06,
+      "loss": 0.0996,
+      "step": 6347
+    },
+    {
+      "epoch": 3.324777370350969,
+      "loss_partial": 0.046875,
+      "step": 6347
+    },
+    {
+      "epoch": 3.324777370350969,
+      "loss_partial": 0.045654296875,
+      "step": 6347
+    },
+    {
+      "epoch": 3.325301204819277,
+      "grad_norm": 1.8032249212265015,
+      "learning_rate": 1.6865740122246376e-06,
+      "loss": 0.0985,
+      "step": 6348
+    },
+    {
+      "epoch": 3.325301204819277,
+      "loss_partial": 0.050537109375,
+      "step": 6348
+    },
+    {
+      "epoch": 3.325301204819277,
+      "loss_partial": 0.060302734375,
+      "step": 6348
+    },
+    {
+      "epoch": 3.325825039287585,
+      "grad_norm": 1.5830491781234741,
+      "learning_rate": 1.6840341761544831e-06,
+      "loss": 0.1028,
+      "step": 6349
+    },
+    {
+      "epoch": 3.325825039287585,
+      "loss_partial": 0.04638671875,
+      "step": 6349
+    },
+    {
+      "epoch": 3.325825039287585,
+      "loss_partial": 0.05419921875,
+      "step": 6349
+    },
+    {
+      "epoch": 3.3263488737558933,
+      "grad_norm": 1.3535228967666626,
+      "learning_rate": 1.6814960780697643e-06,
+      "loss": 0.1022,
+      "step": 6350
+    },
+    {
+      "epoch": 3.3263488737558933,
+      "loss_partial": 0.046630859375,
+      "step": 6350
+    },
+    {
+      "epoch": 3.3263488737558933,
+      "loss_partial": 0.05322265625,
+      "step": 6350
+    },
+    {
+      "epoch": 3.3268727082242013,
+      "grad_norm": 1.5631239414215088,
+      "learning_rate": 1.6789597185009244e-06,
+      "loss": 0.1003,
+      "step": 6351
+    },
+    {
+      "epoch": 3.3268727082242013,
+      "loss_partial": 0.044677734375,
+      "step": 6351
+    },
+    {
+      "epoch": 3.3268727082242013,
+      "loss_partial": 0.058349609375,
+      "step": 6351
+    },
+    {
+      "epoch": 3.327396542692509,
+      "grad_norm": 1.0220164060592651,
+      "learning_rate": 1.6764250979780482e-06,
+      "loss": 0.0952,
+      "step": 6352
+    },
+    {
+      "epoch": 3.327396542692509,
+      "loss_partial": 0.045166015625,
+      "step": 6352
+    },
+    {
+      "epoch": 3.327396542692509,
+      "loss_partial": 0.04638671875,
+      "step": 6352
+    },
+    {
+      "epoch": 3.327920377160817,
+      "grad_norm": 1.1107265949249268,
+      "learning_rate": 1.6738922170308526e-06,
+      "loss": 0.0985,
+      "step": 6353
+    },
+    {
+      "epoch": 3.327920377160817,
+      "loss_partial": 0.044189453125,
+      "step": 6353
+    },
+    {
+      "epoch": 3.327920377160817,
+      "loss_partial": 0.051025390625,
+      "step": 6353
+    },
+    {
+      "epoch": 3.328444211629125,
+      "grad_norm": 0.972989022731781,
+      "learning_rate": 1.6713610761886978e-06,
+      "loss": 0.0995,
+      "step": 6354
+    },
+    {
+      "epoch": 3.328444211629125,
+      "loss_partial": 0.046630859375,
+      "step": 6354
+    },
+    {
+      "epoch": 3.328444211629125,
+      "loss_partial": 0.049072265625,
+      "step": 6354
+    },
+    {
+      "epoch": 3.328968046097433,
+      "grad_norm": 1.0717713832855225,
+      "learning_rate": 1.668831675980579e-06,
+      "loss": 0.0997,
+      "step": 6355
+    },
+    {
+      "epoch": 3.328968046097433,
+      "loss_partial": 0.043701171875,
+      "step": 6355
+    },
+    {
+      "epoch": 3.328968046097433,
+      "loss_partial": 0.052490234375,
+      "step": 6355
+    },
+    {
+      "epoch": 3.329491880565741,
+      "grad_norm": 1.144199013710022,
+      "learning_rate": 1.6663040169351187e-06,
+      "loss": 0.0989,
+      "step": 6356
+    },
+    {
+      "epoch": 3.329491880565741,
+      "loss_partial": 0.046142578125,
+      "step": 6356
+    },
+    {
+      "epoch": 3.329491880565741,
+      "loss_partial": 0.054931640625,
+      "step": 6356
+    },
+    {
+      "epoch": 3.3300157150340493,
+      "grad_norm": 1.3226748704910278,
+      "learning_rate": 1.663778099580583e-06,
+      "loss": 0.1005,
+      "step": 6357
+    },
+    {
+      "epoch": 3.3300157150340493,
+      "loss_partial": 0.05322265625,
+      "step": 6357
+    },
+    {
+      "epoch": 3.3300157150340493,
+      "loss_partial": 0.047607421875,
+      "step": 6357
+    },
+    {
+      "epoch": 3.3305395495023573,
+      "grad_norm": 1.0146359205245972,
+      "learning_rate": 1.6612539244448765e-06,
+      "loss": 0.1043,
+      "step": 6358
+    },
+    {
+      "epoch": 3.3305395495023573,
+      "loss_partial": 0.045654296875,
+      "step": 6358
+    },
+    {
+      "epoch": 3.3305395495023573,
+      "loss_partial": 0.053466796875,
+      "step": 6358
+    },
+    {
+      "epoch": 3.3310633839706654,
+      "grad_norm": 1.0235328674316406,
+      "learning_rate": 1.6587314920555353e-06,
+      "loss": 0.0961,
+      "step": 6359
+    },
+    {
+      "epoch": 3.3310633839706654,
+      "loss_partial": 0.047119140625,
+      "step": 6359
+    },
+    {
+      "epoch": 3.3310633839706654,
+      "loss_partial": 0.0419921875,
+      "step": 6359
+    },
+    {
+      "epoch": 3.3315872184389734,
+      "grad_norm": 1.1822669506072998,
+      "learning_rate": 1.6562108029397307e-06,
+      "loss": 0.1014,
+      "step": 6360
+    },
+    {
+      "epoch": 3.3315872184389734,
+      "loss_partial": 0.04931640625,
+      "step": 6360
+    },
+    {
+      "epoch": 3.3315872184389734,
+      "loss_partial": 0.049072265625,
+      "step": 6360
+    },
+    {
+      "epoch": 3.3321110529072815,
+      "grad_norm": 0.8956887722015381,
+      "learning_rate": 1.65369185762427e-06,
+      "loss": 0.0976,
+      "step": 6361
+    },
+    {
+      "epoch": 3.3321110529072815,
+      "loss_partial": 0.044921875,
+      "step": 6361
+    },
+    {
+      "epoch": 3.3321110529072815,
+      "loss_partial": 0.04931640625,
+      "step": 6361
+    },
+    {
+      "epoch": 3.3326348873755895,
+      "grad_norm": 1.060939073562622,
+      "learning_rate": 1.6511746566356001e-06,
+      "loss": 0.1007,
+      "step": 6362
+    },
+    {
+      "epoch": 3.3326348873755895,
+      "loss_partial": 0.052978515625,
+      "step": 6362
+    },
+    {
+      "epoch": 3.3326348873755895,
+      "loss_partial": 0.05517578125,
+      "step": 6362
+    },
+    {
+      "epoch": 3.333158721843897,
+      "grad_norm": 1.2054086923599243,
+      "learning_rate": 1.6486592004998003e-06,
+      "loss": 0.1025,
+      "step": 6363
+    },
+    {
+      "epoch": 3.333158721843897,
+      "loss_partial": 0.052734375,
+      "step": 6363
+    },
+    {
+      "epoch": 3.333158721843897,
+      "loss_partial": 0.043701171875,
+      "step": 6363
+    },
+    {
+      "epoch": 3.3336825563122052,
+      "grad_norm": 1.3388986587524414,
+      "learning_rate": 1.6461454897425855e-06,
+      "loss": 0.0995,
+      "step": 6364
+    },
+    {
+      "epoch": 3.3336825563122052,
+      "loss_partial": 0.045654296875,
+      "step": 6364
+    },
+    {
+      "epoch": 3.3336825563122052,
+      "loss_partial": 0.04541015625,
+      "step": 6364
+    },
+    {
+      "epoch": 3.3342063907805133,
+      "grad_norm": 1.1992439031600952,
+      "learning_rate": 1.643633524889301e-06,
+      "loss": 0.0998,
+      "step": 6365
+    },
+    {
+      "epoch": 3.3342063907805133,
+      "loss_partial": 0.0419921875,
+      "step": 6365
+    },
+    {
+      "epoch": 3.3342063907805133,
+      "loss_partial": 0.043212890625,
+      "step": 6365
+    },
+    {
+      "epoch": 3.3347302252488213,
+      "grad_norm": 1.0377955436706543,
+      "learning_rate": 1.641123306464938e-06,
+      "loss": 0.1027,
+      "step": 6366
+    },
+    {
+      "epoch": 3.3347302252488213,
+      "loss_partial": 0.056640625,
+      "step": 6366
+    },
+    {
+      "epoch": 3.3347302252488213,
+      "loss_partial": 0.050048828125,
+      "step": 6366
+    },
+    {
+      "epoch": 3.3352540597171294,
+      "grad_norm": 1.084071159362793,
+      "learning_rate": 1.6386148349941144e-06,
+      "loss": 0.1024,
+      "step": 6367
+    },
+    {
+      "epoch": 3.3352540597171294,
+      "loss_partial": 0.052978515625,
+      "step": 6367
+    },
+    {
+      "epoch": 3.3352540597171294,
+      "loss_partial": 0.0478515625,
+      "step": 6367
+    },
+    {
+      "epoch": 3.3357778941854375,
+      "grad_norm": 1.0787360668182373,
+      "learning_rate": 1.636108111001088e-06,
+      "loss": 0.0993,
+      "step": 6368
+    },
+    {
+      "epoch": 3.3357778941854375,
+      "loss_partial": 0.049072265625,
+      "step": 6368
+    },
+    {
+      "epoch": 3.3357778941854375,
+      "loss_partial": 0.04052734375,
+      "step": 6368
+    },
+    {
+      "epoch": 3.3363017286537455,
+      "grad_norm": 1.075821042060852,
+      "learning_rate": 1.6336031350097404e-06,
+      "loss": 0.101,
+      "step": 6369
+    },
+    {
+      "epoch": 3.3363017286537455,
+      "loss_partial": 0.0634765625,
+      "step": 6369
+    },
+    {
+      "epoch": 3.3363017286537455,
+      "loss_partial": 0.052978515625,
+      "step": 6369
+    },
+    {
+      "epoch": 3.3368255631220536,
+      "grad_norm": 1.4785735607147217,
+      "learning_rate": 1.631099907543604e-06,
+      "loss": 0.1018,
+      "step": 6370
+    },
+    {
+      "epoch": 3.3368255631220536,
+      "loss_partial": 0.047607421875,
+      "step": 6370
+    },
+    {
+      "epoch": 3.3368255631220536,
+      "loss_partial": 0.04443359375,
+      "step": 6370
+    },
+    {
+      "epoch": 3.337349397590361,
+      "grad_norm": 1.4782739877700806,
+      "learning_rate": 1.628598429125835e-06,
+      "loss": 0.0999,
+      "step": 6371
+    },
+    {
+      "epoch": 3.337349397590361,
+      "loss_partial": 0.0537109375,
+      "step": 6371
+    },
+    {
+      "epoch": 3.337349397590361,
+      "loss_partial": 0.053955078125,
+      "step": 6371
+    },
+    {
+      "epoch": 3.3378732320586693,
+      "grad_norm": 1.2574119567871094,
+      "learning_rate": 1.626098700279226e-06,
+      "loss": 0.0963,
+      "step": 6372
+    },
+    {
+      "epoch": 3.3378732320586693,
+      "loss_partial": 0.05419921875,
+      "step": 6372
+    },
+    {
+      "epoch": 3.3378732320586693,
+      "loss_partial": 0.04736328125,
+      "step": 6372
+    },
+    {
+      "epoch": 3.3383970665269773,
+      "grad_norm": 1.1869004964828491,
+      "learning_rate": 1.623600721526205e-06,
+      "loss": 0.1017,
+      "step": 6373
+    },
+    {
+      "epoch": 3.3383970665269773,
+      "loss_partial": 0.044921875,
+      "step": 6373
+    },
+    {
+      "epoch": 3.3383970665269773,
+      "loss_partial": 0.0546875,
+      "step": 6373
+    },
+    {
+      "epoch": 3.3389209009952854,
+      "grad_norm": 1.1825686693191528,
+      "learning_rate": 1.6211044933888376e-06,
+      "loss": 0.0973,
+      "step": 6374
+    },
+    {
+      "epoch": 3.3389209009952854,
+      "loss_partial": 0.043701171875,
+      "step": 6374
+    },
+    {
+      "epoch": 3.3389209009952854,
+      "loss_partial": 0.060546875,
+      "step": 6374
+    },
+    {
+      "epoch": 3.3394447354635934,
+      "grad_norm": 1.3427197933197021,
+      "learning_rate": 1.6186100163888163e-06,
+      "loss": 0.0988,
+      "step": 6375
+    },
+    {
+      "epoch": 3.3394447354635934,
+      "loss_partial": 0.054443359375,
+      "step": 6375
+    },
+    {
+      "epoch": 3.3394447354635934,
+      "loss_partial": 0.044677734375,
+      "step": 6375
+    },
+    {
+      "epoch": 3.3399685699319015,
+      "grad_norm": 1.0608408451080322,
+      "learning_rate": 1.6161172910474743e-06,
+      "loss": 0.0982,
+      "step": 6376
+    },
+    {
+      "epoch": 3.3399685699319015,
+      "loss_partial": 0.03759765625,
+      "step": 6376
+    },
+    {
+      "epoch": 3.3399685699319015,
+      "loss_partial": 0.050537109375,
+      "step": 6376
+    },
+    {
+      "epoch": 3.3404924044002096,
+      "grad_norm": 1.088329553604126,
+      "learning_rate": 1.6136263178857703e-06,
+      "loss": 0.097,
+      "step": 6377
+    },
+    {
+      "epoch": 3.3404924044002096,
+      "loss_partial": 0.044921875,
+      "step": 6377
+    },
+    {
+      "epoch": 3.3404924044002096,
+      "loss_partial": 0.0537109375,
+      "step": 6377
+    },
+    {
+      "epoch": 3.3410162388685176,
+      "grad_norm": 1.434821367263794,
+      "learning_rate": 1.611137097424309e-06,
+      "loss": 0.0981,
+      "step": 6378
+    },
+    {
+      "epoch": 3.3410162388685176,
+      "loss_partial": 0.04541015625,
+      "step": 6378
+    },
+    {
+      "epoch": 3.3410162388685176,
+      "loss_partial": 0.048095703125,
+      "step": 6378
+    },
+    {
+      "epoch": 3.3415400733368257,
+      "grad_norm": 1.2435526847839355,
+      "learning_rate": 1.60864963018332e-06,
+      "loss": 0.0972,
+      "step": 6379
+    },
+    {
+      "epoch": 3.3415400733368257,
+      "loss_partial": 0.06005859375,
+      "step": 6379
+    },
+    {
+      "epoch": 3.3415400733368257,
+      "loss_partial": 0.0546875,
+      "step": 6379
+    },
+    {
+      "epoch": 3.3420639078051337,
+      "grad_norm": 1.0999025106430054,
+      "learning_rate": 1.6061639166826693e-06,
+      "loss": 0.1002,
+      "step": 6380
+    },
+    {
+      "epoch": 3.3420639078051337,
+      "loss_partial": 0.04296875,
+      "step": 6380
+    },
+    {
+      "epoch": 3.3420639078051337,
+      "loss_partial": 0.046142578125,
+      "step": 6380
+    },
+    {
+      "epoch": 3.342587742273442,
+      "grad_norm": 1.042954683303833,
+      "learning_rate": 1.6036799574418516e-06,
+      "loss": 0.0969,
+      "step": 6381
+    },
+    {
+      "epoch": 3.342587742273442,
+      "loss_partial": 0.04833984375,
+      "step": 6381
+    },
+    {
+      "epoch": 3.342587742273442,
+      "loss_partial": 0.04931640625,
+      "step": 6381
+    },
+    {
+      "epoch": 3.3431115767417494,
+      "grad_norm": 1.2509825229644775,
+      "learning_rate": 1.6011977529799995e-06,
+      "loss": 0.1041,
+      "step": 6382
+    },
+    {
+      "epoch": 3.3431115767417494,
+      "loss_partial": 0.04296875,
+      "step": 6382
+    },
+    {
+      "epoch": 3.3431115767417494,
+      "loss_partial": 0.050537109375,
+      "step": 6382
+    },
+    {
+      "epoch": 3.3436354112100575,
+      "grad_norm": 1.0085954666137695,
+      "learning_rate": 1.598717303815882e-06,
+      "loss": 0.0989,
+      "step": 6383
+    },
+    {
+      "epoch": 3.3436354112100575,
+      "loss_partial": 0.06103515625,
+      "step": 6383
+    },
+    {
+      "epoch": 3.3436354112100575,
+      "loss_partial": 0.047607421875,
+      "step": 6383
+    },
+    {
+      "epoch": 3.3441592456783655,
+      "grad_norm": 1.0012966394424438,
+      "learning_rate": 1.5962386104678963e-06,
+      "loss": 0.0986,
+      "step": 6384
+    },
+    {
+      "epoch": 3.3441592456783655,
+      "loss_partial": 0.0478515625,
+      "step": 6384
+    },
+    {
+      "epoch": 3.3441592456783655,
+      "loss_partial": 0.057861328125,
+      "step": 6384
+    },
+    {
+      "epoch": 3.3446830801466736,
+      "grad_norm": 1.322428584098816,
+      "learning_rate": 1.593761673454074e-06,
+      "loss": 0.098,
+      "step": 6385
+    },
+    {
+      "epoch": 3.3446830801466736,
+      "loss_partial": 0.047119140625,
+      "step": 6385
+    },
+    {
+      "epoch": 3.3446830801466736,
+      "loss_partial": 0.056396484375,
+      "step": 6385
+    },
+    {
+      "epoch": 3.3452069146149817,
+      "grad_norm": 0.9848566651344299,
+      "learning_rate": 1.5912864932920768e-06,
+      "loss": 0.097,
+      "step": 6386
+    },
+    {
+      "epoch": 3.3452069146149817,
+      "loss_partial": 0.056640625,
+      "step": 6386
+    },
+    {
+      "epoch": 3.3452069146149817,
+      "loss_partial": 0.049560546875,
+      "step": 6386
+    },
+    {
+      "epoch": 3.3457307490832897,
+      "grad_norm": 1.1162084341049194,
+      "learning_rate": 1.5888130704992077e-06,
+      "loss": 0.1054,
+      "step": 6387
+    },
+    {
+      "epoch": 3.3457307490832897,
+      "loss_partial": 0.055908203125,
+      "step": 6387
+    },
+    {
+      "epoch": 3.3457307490832897,
+      "loss_partial": 0.052978515625,
+      "step": 6387
+    },
+    {
+      "epoch": 3.3462545835515978,
+      "grad_norm": 0.9651115536689758,
+      "learning_rate": 1.5863414055923932e-06,
+      "loss": 0.1017,
+      "step": 6388
+    },
+    {
+      "epoch": 3.3462545835515978,
+      "loss_partial": 0.05029296875,
+      "step": 6388
+    },
+    {
+      "epoch": 3.3462545835515978,
+      "loss_partial": 0.048095703125,
+      "step": 6388
+    },
+    {
+      "epoch": 3.346778418019906,
+      "grad_norm": 1.2519968748092651,
+      "learning_rate": 1.5838714990881954e-06,
+      "loss": 0.1,
+      "step": 6389
+    },
+    {
+      "epoch": 3.346778418019906,
+      "loss_partial": 0.049560546875,
+      "step": 6389
+    },
+    {
+      "epoch": 3.346778418019906,
+      "loss_partial": 0.043701171875,
+      "step": 6389
+    },
+    {
+      "epoch": 3.347302252488214,
+      "grad_norm": 1.0014945268630981,
+      "learning_rate": 1.5814033515028093e-06,
+      "loss": 0.0969,
+      "step": 6390
+    },
+    {
+      "epoch": 3.347302252488214,
+      "loss_partial": 0.051513671875,
+      "step": 6390
+    },
+    {
+      "epoch": 3.347302252488214,
+      "loss_partial": 0.047119140625,
+      "step": 6390
+    },
+    {
+      "epoch": 3.3478260869565215,
+      "grad_norm": 1.0724815130233765,
+      "learning_rate": 1.5789369633520656e-06,
+      "loss": 0.0948,
+      "step": 6391
+    },
+    {
+      "epoch": 3.3478260869565215,
+      "loss_partial": 0.049072265625,
+      "step": 6391
+    },
+    {
+      "epoch": 3.3478260869565215,
+      "loss_partial": 0.04296875,
+      "step": 6391
+    },
+    {
+      "epoch": 3.3483499214248296,
+      "grad_norm": 1.1110976934432983,
+      "learning_rate": 1.5764723351514255e-06,
+      "loss": 0.0964,
+      "step": 6392
+    },
+    {
+      "epoch": 3.3483499214248296,
+      "loss_partial": 0.054931640625,
+      "step": 6392
+    },
+    {
+      "epoch": 3.3483499214248296,
+      "loss_partial": 0.055908203125,
+      "step": 6392
+    },
+    {
+      "epoch": 3.3488737558931376,
+      "grad_norm": 1.1432855129241943,
+      "learning_rate": 1.5740094674159745e-06,
+      "loss": 0.1018,
+      "step": 6393
+    },
+    {
+      "epoch": 3.3488737558931376,
+      "loss_partial": 0.04296875,
+      "step": 6393
+    },
+    {
+      "epoch": 3.3488737558931376,
+      "loss_partial": 0.04345703125,
+      "step": 6393
+    },
+    {
+      "epoch": 3.3493975903614457,
+      "grad_norm": 1.7619816064834595,
+      "learning_rate": 1.5715483606604375e-06,
+      "loss": 0.101,
+      "step": 6394
+    },
+    {
+      "epoch": 3.3493975903614457,
+      "loss_partial": 0.04736328125,
+      "step": 6394
+    },
+    {
+      "epoch": 3.3493975903614457,
+      "loss_partial": 0.044921875,
+      "step": 6394
+    },
+    {
+      "epoch": 3.3499214248297537,
+      "grad_norm": 1.2984092235565186,
+      "learning_rate": 1.5690890153991768e-06,
+      "loss": 0.101,
+      "step": 6395
+    },
+    {
+      "epoch": 3.3499214248297537,
+      "loss_partial": 0.055908203125,
+      "step": 6395
+    },
+    {
+      "epoch": 3.3499214248297537,
+      "loss_partial": 0.05859375,
+      "step": 6395
+    },
+    {
+      "epoch": 3.350445259298062,
+      "grad_norm": 1.4810086488723755,
+      "learning_rate": 1.566631432146175e-06,
+      "loss": 0.0989,
+      "step": 6396
+    },
+    {
+      "epoch": 3.350445259298062,
+      "loss_partial": 0.04638671875,
+      "step": 6396
+    },
+    {
+      "epoch": 3.350445259298062,
+      "loss_partial": 0.058349609375,
+      "step": 6396
+    },
+    {
+      "epoch": 3.35096909376637,
+      "grad_norm": 1.3721286058425903,
+      "learning_rate": 1.5641756114150552e-06,
+      "loss": 0.0988,
+      "step": 6397
+    },
+    {
+      "epoch": 3.35096909376637,
+      "loss_partial": 0.044921875,
+      "step": 6397
+    },
+    {
+      "epoch": 3.35096909376637,
+      "loss_partial": 0.046142578125,
+      "step": 6397
+    },
+    {
+      "epoch": 3.351492928234678,
+      "grad_norm": 1.0645062923431396,
+      "learning_rate": 1.5617215537190634e-06,
+      "loss": 0.0975,
+      "step": 6398
+    },
+    {
+      "epoch": 3.351492928234678,
+      "loss_partial": 0.0498046875,
+      "step": 6398
+    },
+    {
+      "epoch": 3.351492928234678,
+      "loss_partial": 0.04736328125,
+      "step": 6398
+    },
+    {
+      "epoch": 3.352016762702986,
+      "grad_norm": 1.0864300727844238,
+      "learning_rate": 1.559269259571089e-06,
+      "loss": 0.0977,
+      "step": 6399
+    },
+    {
+      "epoch": 3.352016762702986,
+      "loss_partial": 0.047607421875,
+      "step": 6399
+    },
+    {
+      "epoch": 3.352016762702986,
+      "loss_partial": 0.0419921875,
+      "step": 6399
+    },
+    {
+      "epoch": 3.352540597171294,
+      "grad_norm": 1.3176264762878418,
+      "learning_rate": 1.5568187294836435e-06,
+      "loss": 0.0981,
+      "step": 6400
+    },
+    {
+      "epoch": 3.352540597171294,
+      "loss_partial": 0.05078125,
+      "step": 6400
+    },
+    {
+      "epoch": 3.352540597171294,
+      "loss_partial": 0.057373046875,
+      "step": 6400
+    },
+    {
+      "epoch": 3.353064431639602,
+      "grad_norm": 1.106199860572815,
+      "learning_rate": 1.5543699639688726e-06,
+      "loss": 0.0999,
+      "step": 6401
+    },
+    {
+      "epoch": 3.353064431639602,
+      "loss_partial": 0.038330078125,
+      "step": 6401
+    },
+    {
+      "epoch": 3.353064431639602,
+      "loss_partial": 0.060546875,
+      "step": 6401
+    },
+    {
+      "epoch": 3.3535882661079097,
+      "grad_norm": 0.8886693120002747,
+      "learning_rate": 1.5519229635385502e-06,
+      "loss": 0.1006,
+      "step": 6402
+    },
+    {
+      "epoch": 3.3535882661079097,
+      "loss_partial": 0.052490234375,
+      "step": 6402
+    },
+    {
+      "epoch": 3.3535882661079097,
+      "loss_partial": 0.050048828125,
+      "step": 6402
+    },
+    {
+      "epoch": 3.354112100576218,
+      "grad_norm": 1.1662317514419556,
+      "learning_rate": 1.5494777287040896e-06,
+      "loss": 0.0947,
+      "step": 6403
+    },
+    {
+      "epoch": 3.354112100576218,
+      "loss_partial": 0.053466796875,
+      "step": 6403
+    },
+    {
+      "epoch": 3.354112100576218,
+      "loss_partial": 0.04345703125,
+      "step": 6403
+    },
+    {
+      "epoch": 3.354635935044526,
+      "grad_norm": 1.2725991010665894,
+      "learning_rate": 1.5470342599765276e-06,
+      "loss": 0.1006,
+      "step": 6404
+    },
+    {
+      "epoch": 3.354635935044526,
+      "loss_partial": 0.052978515625,
+      "step": 6404
+    },
+    {
+      "epoch": 3.354635935044526,
+      "loss_partial": 0.05029296875,
+      "step": 6404
+    },
+    {
+      "epoch": 3.355159769512834,
+      "grad_norm": 0.9194716215133667,
+      "learning_rate": 1.5445925578665355e-06,
+      "loss": 0.0993,
+      "step": 6405
+    },
+    {
+      "epoch": 3.355159769512834,
+      "loss_partial": 0.04052734375,
+      "step": 6405
+    },
+    {
+      "epoch": 3.355159769512834,
+      "loss_partial": 0.06005859375,
+      "step": 6405
+    },
+    {
+      "epoch": 3.355683603981142,
+      "grad_norm": 1.0363683700561523,
+      "learning_rate": 1.5421526228844075e-06,
+      "loss": 0.1015,
+      "step": 6406
+    },
+    {
+      "epoch": 3.355683603981142,
+      "loss_partial": 0.05517578125,
+      "step": 6406
+    },
+    {
+      "epoch": 3.355683603981142,
+      "loss_partial": 0.049072265625,
+      "step": 6406
+    },
+    {
+      "epoch": 3.35620743844945,
+      "grad_norm": 1.2384823560714722,
+      "learning_rate": 1.5397144555400823e-06,
+      "loss": 0.1015,
+      "step": 6407
+    },
+    {
+      "epoch": 3.35620743844945,
+      "loss_partial": 0.058349609375,
+      "step": 6407
+    },
+    {
+      "epoch": 3.35620743844945,
+      "loss_partial": 0.0517578125,
+      "step": 6407
+    },
+    {
+      "epoch": 3.356731272917758,
+      "grad_norm": 1.6317775249481201,
+      "learning_rate": 1.537278056343119e-06,
+      "loss": 0.0986,
+      "step": 6408
+    },
+    {
+      "epoch": 3.356731272917758,
+      "loss_partial": 0.0546875,
+      "step": 6408
+    },
+    {
+      "epoch": 3.356731272917758,
+      "loss_partial": 0.046142578125,
+      "step": 6408
+    },
+    {
+      "epoch": 3.357255107386066,
+      "grad_norm": 1.1907358169555664,
+      "learning_rate": 1.53484342580271e-06,
+      "loss": 0.1014,
+      "step": 6409
+    },
+    {
+      "epoch": 3.357255107386066,
+      "loss_partial": 0.052978515625,
+      "step": 6409
+    },
+    {
+      "epoch": 3.357255107386066,
+      "loss_partial": 0.04931640625,
+      "step": 6409
+    },
+    {
+      "epoch": 3.357778941854374,
+      "grad_norm": 0.986816942691803,
+      "learning_rate": 1.5324105644276755e-06,
+      "loss": 0.1018,
+      "step": 6410
+    },
+    {
+      "epoch": 3.357778941854374,
+      "loss_partial": 0.046142578125,
+      "step": 6410
+    },
+    {
+      "epoch": 3.357778941854374,
+      "loss_partial": 0.052490234375,
+      "step": 6410
+    },
+    {
+      "epoch": 3.358302776322682,
+      "grad_norm": 1.2221436500549316,
+      "learning_rate": 1.529979472726474e-06,
+      "loss": 0.0995,
+      "step": 6411
+    },
+    {
+      "epoch": 3.358302776322682,
+      "loss_partial": 0.044677734375,
+      "step": 6411
+    },
+    {
+      "epoch": 3.358302776322682,
+      "loss_partial": 0.052001953125,
+      "step": 6411
+    },
+    {
+      "epoch": 3.35882661079099,
+      "grad_norm": 1.1553326845169067,
+      "learning_rate": 1.5275501512071854e-06,
+      "loss": 0.1014,
+      "step": 6412
+    },
+    {
+      "epoch": 3.35882661079099,
+      "loss_partial": 0.048583984375,
+      "step": 6412
+    },
+    {
+      "epoch": 3.35882661079099,
+      "loss_partial": 0.04736328125,
+      "step": 6412
+    },
+    {
+      "epoch": 3.359350445259298,
+      "grad_norm": 1.0283732414245605,
+      "learning_rate": 1.5251226003775222e-06,
+      "loss": 0.1028,
+      "step": 6413
+    },
+    {
+      "epoch": 3.359350445259298,
+      "loss_partial": 0.050537109375,
+      "step": 6413
+    },
+    {
+      "epoch": 3.359350445259298,
+      "loss_partial": 0.04638671875,
+      "step": 6413
+    },
+    {
+      "epoch": 3.359874279727606,
+      "grad_norm": 1.2598834037780762,
+      "learning_rate": 1.5226968207448267e-06,
+      "loss": 0.1022,
+      "step": 6414
+    },
+    {
+      "epoch": 3.359874279727606,
+      "loss_partial": 0.05078125,
+      "step": 6414
+    },
+    {
+      "epoch": 3.359874279727606,
+      "loss_partial": 0.04931640625,
+      "step": 6414
+    },
+    {
+      "epoch": 3.360398114195914,
+      "grad_norm": 1.2780617475509644,
+      "learning_rate": 1.520272812816077e-06,
+      "loss": 0.1004,
+      "step": 6415
+    },
+    {
+      "epoch": 3.360398114195914,
+      "loss_partial": 0.04345703125,
+      "step": 6415
+    },
+    {
+      "epoch": 3.360398114195914,
+      "loss_partial": 0.04541015625,
+      "step": 6415
+    },
+    {
+      "epoch": 3.360921948664222,
+      "grad_norm": 1.265514612197876,
+      "learning_rate": 1.5178505770978714e-06,
+      "loss": 0.0971,
+      "step": 6416
+    },
+    {
+      "epoch": 3.360921948664222,
+      "loss_partial": 0.048583984375,
+      "step": 6416
+    },
+    {
+      "epoch": 3.360921948664222,
+      "loss_partial": 0.045654296875,
+      "step": 6416
+    },
+    {
+      "epoch": 3.36144578313253,
+      "grad_norm": 1.1824263334274292,
+      "learning_rate": 1.5154301140964456e-06,
+      "loss": 0.0994,
+      "step": 6417
+    },
+    {
+      "epoch": 3.36144578313253,
+      "loss_partial": 0.0439453125,
+      "step": 6417
+    },
+    {
+      "epoch": 3.36144578313253,
+      "loss_partial": 0.040771484375,
+      "step": 6417
+    },
+    {
+      "epoch": 3.3619696176008382,
+      "grad_norm": 0.988292396068573,
+      "learning_rate": 1.5130114243176542e-06,
+      "loss": 0.0945,
+      "step": 6418
+    },
+    {
+      "epoch": 3.3619696176008382,
+      "loss_partial": 0.05615234375,
+      "step": 6418
+    },
+    {
+      "epoch": 3.3619696176008382,
+      "loss_partial": 0.04736328125,
+      "step": 6418
+    },
+    {
+      "epoch": 3.3624934520691463,
+      "grad_norm": 1.0897351503372192,
+      "learning_rate": 1.510594508266995e-06,
+      "loss": 0.1006,
+      "step": 6419
+    },
+    {
+      "epoch": 3.3624934520691463,
+      "loss_partial": 0.046875,
+      "step": 6419
+    },
+    {
+      "epoch": 3.3624934520691463,
+      "loss_partial": 0.052001953125,
+      "step": 6419
+    },
+    {
+      "epoch": 3.3630172865374544,
+      "grad_norm": 1.312878131866455,
+      "learning_rate": 1.508179366449587e-06,
+      "loss": 0.0986,
+      "step": 6420
+    },
+    {
+      "epoch": 3.3630172865374544,
+      "loss_partial": 0.051513671875,
+      "step": 6420
+    },
+    {
+      "epoch": 3.3630172865374544,
+      "loss_partial": 0.040283203125,
+      "step": 6420
+    },
+    {
+      "epoch": 3.3635411210057624,
+      "grad_norm": 1.2578788995742798,
+      "learning_rate": 1.5057659993701801e-06,
+      "loss": 0.1014,
+      "step": 6421
+    },
+    {
+      "epoch": 3.3635411210057624,
+      "loss_partial": 0.048828125,
+      "step": 6421
+    },
+    {
+      "epoch": 3.3635411210057624,
+      "loss_partial": 0.0478515625,
+      "step": 6421
+    },
+    {
+      "epoch": 3.36406495547407,
+      "grad_norm": 1.0123885869979858,
+      "learning_rate": 1.5033544075331496e-06,
+      "loss": 0.1005,
+      "step": 6422
+    },
+    {
+      "epoch": 3.36406495547407,
+      "loss_partial": 0.0556640625,
+      "step": 6422
+    },
+    {
+      "epoch": 3.36406495547407,
+      "loss_partial": 0.04296875,
+      "step": 6422
+    },
+    {
+      "epoch": 3.364588789942378,
+      "grad_norm": 1.2782095670700073,
+      "learning_rate": 1.5009445914425093e-06,
+      "loss": 0.1038,
+      "step": 6423
+    },
+    {
+      "epoch": 3.364588789942378,
+      "loss_partial": 0.045654296875,
+      "step": 6423
+    },
+    {
+      "epoch": 3.364588789942378,
+      "loss_partial": 0.046630859375,
+      "step": 6423
+    },
+    {
+      "epoch": 3.365112624410686,
+      "grad_norm": 1.5621134042739868,
+      "learning_rate": 1.4985365516018924e-06,
+      "loss": 0.1014,
+      "step": 6424
+    },
+    {
+      "epoch": 3.365112624410686,
+      "loss_partial": 0.04052734375,
+      "step": 6424
+    },
+    {
+      "epoch": 3.365112624410686,
+      "loss_partial": 0.04931640625,
+      "step": 6424
+    },
+    {
+      "epoch": 3.365636458878994,
+      "grad_norm": 1.2051175832748413,
+      "learning_rate": 1.4961302885145646e-06,
+      "loss": 0.1019,
+      "step": 6425
+    },
+    {
+      "epoch": 3.365636458878994,
+      "loss_partial": 0.044677734375,
+      "step": 6425
+    },
+    {
+      "epoch": 3.365636458878994,
+      "loss_partial": 0.05517578125,
+      "step": 6425
+    },
+    {
+      "epoch": 3.3661602933473023,
+      "grad_norm": 1.221648931503296,
+      "learning_rate": 1.4937258026834177e-06,
+      "loss": 0.0962,
+      "step": 6426
+    },
+    {
+      "epoch": 3.3661602933473023,
+      "loss_partial": 0.056640625,
+      "step": 6426
+    },
+    {
+      "epoch": 3.3661602933473023,
+      "loss_partial": 0.049560546875,
+      "step": 6426
+    },
+    {
+      "epoch": 3.3666841278156103,
+      "grad_norm": 1.3402214050292969,
+      "learning_rate": 1.4913230946109803e-06,
+      "loss": 0.0986,
+      "step": 6427
+    },
+    {
+      "epoch": 3.3666841278156103,
+      "loss_partial": 0.049072265625,
+      "step": 6427
+    },
+    {
+      "epoch": 3.3666841278156103,
+      "loss_partial": 0.0615234375,
+      "step": 6427
+    },
+    {
+      "epoch": 3.3672079622839184,
+      "grad_norm": 1.1677035093307495,
+      "learning_rate": 1.488922164799399e-06,
+      "loss": 0.1059,
+      "step": 6428
+    },
+    {
+      "epoch": 3.3672079622839184,
+      "loss_partial": 0.0419921875,
+      "step": 6428
+    },
+    {
+      "epoch": 3.3672079622839184,
+      "loss_partial": 0.046142578125,
+      "step": 6428
+    },
+    {
+      "epoch": 3.3677317967522264,
+      "grad_norm": 1.2953872680664062,
+      "learning_rate": 1.486523013750456e-06,
+      "loss": 0.0984,
+      "step": 6429
+    },
+    {
+      "epoch": 3.3677317967522264,
+      "loss_partial": 0.044921875,
+      "step": 6429
+    },
+    {
+      "epoch": 3.3677317967522264,
+      "loss_partial": 0.052978515625,
+      "step": 6429
+    },
+    {
+      "epoch": 3.368255631220534,
+      "grad_norm": 1.0318231582641602,
+      "learning_rate": 1.4841256419655537e-06,
+      "loss": 0.1019,
+      "step": 6430
+    },
+    {
+      "epoch": 3.368255631220534,
+      "loss_partial": 0.050048828125,
+      "step": 6430
+    },
+    {
+      "epoch": 3.368255631220534,
+      "loss_partial": 0.047119140625,
+      "step": 6430
+    },
+    {
+      "epoch": 3.368779465688842,
+      "grad_norm": 1.1462652683258057,
+      "learning_rate": 1.4817300499457343e-06,
+      "loss": 0.0977,
+      "step": 6431
+    },
+    {
+      "epoch": 3.368779465688842,
+      "loss_partial": 0.04638671875,
+      "step": 6431
+    },
+    {
+      "epoch": 3.368779465688842,
+      "loss_partial": 0.052978515625,
+      "step": 6431
+    },
+    {
+      "epoch": 3.36930330015715,
+      "grad_norm": 1.4979338645935059,
+      "learning_rate": 1.4793362381916587e-06,
+      "loss": 0.1019,
+      "step": 6432
+    },
+    {
+      "epoch": 3.36930330015715,
+      "loss_partial": 0.051513671875,
+      "step": 6432
+    },
+    {
+      "epoch": 3.36930330015715,
+      "loss_partial": 0.050048828125,
+      "step": 6432
+    },
+    {
+      "epoch": 3.3698271346254582,
+      "grad_norm": 0.9720748066902161,
+      "learning_rate": 1.4769442072036188e-06,
+      "loss": 0.0997,
+      "step": 6433
+    },
+    {
+      "epoch": 3.3698271346254582,
+      "loss_partial": 0.04541015625,
+      "step": 6433
+    },
+    {
+      "epoch": 3.3698271346254582,
+      "loss_partial": 0.044677734375,
+      "step": 6433
+    },
+    {
+      "epoch": 3.3703509690937663,
+      "grad_norm": 1.0968399047851562,
+      "learning_rate": 1.4745539574815327e-06,
+      "loss": 0.1011,
+      "step": 6434
+    },
+    {
+      "epoch": 3.3703509690937663,
+      "loss_partial": 0.049072265625,
+      "step": 6434
+    },
+    {
+      "epoch": 3.3703509690937663,
+      "loss_partial": 0.045166015625,
+      "step": 6434
+    },
+    {
+      "epoch": 3.3708748035620744,
+      "grad_norm": 1.5591484308242798,
+      "learning_rate": 1.4721654895249516e-06,
+      "loss": 0.0977,
+      "step": 6435
+    },
+    {
+      "epoch": 3.3708748035620744,
+      "loss_partial": 0.052001953125,
+      "step": 6435
+    },
+    {
+      "epoch": 3.3708748035620744,
+      "loss_partial": 0.050048828125,
+      "step": 6435
+    },
+    {
+      "epoch": 3.3713986380303824,
+      "grad_norm": 1.1880559921264648,
+      "learning_rate": 1.4697788038330484e-06,
+      "loss": 0.1019,
+      "step": 6436
+    },
+    {
+      "epoch": 3.3713986380303824,
+      "loss_partial": 0.04443359375,
+      "step": 6436
+    },
+    {
+      "epoch": 3.3713986380303824,
+      "loss_partial": 0.04833984375,
+      "step": 6436
+    },
+    {
+      "epoch": 3.3719224724986905,
+      "grad_norm": 1.1241192817687988,
+      "learning_rate": 1.4673939009046268e-06,
+      "loss": 0.0997,
+      "step": 6437
+    },
+    {
+      "epoch": 3.3719224724986905,
+      "loss_partial": 0.0546875,
+      "step": 6437
+    },
+    {
+      "epoch": 3.3719224724986905,
+      "loss_partial": 0.046630859375,
+      "step": 6437
+    },
+    {
+      "epoch": 3.3724463069669985,
+      "grad_norm": 1.2616239786148071,
+      "learning_rate": 1.4650107812381153e-06,
+      "loss": 0.0991,
+      "step": 6438
+    },
+    {
+      "epoch": 3.3724463069669985,
+      "loss_partial": 0.055908203125,
+      "step": 6438
+    },
+    {
+      "epoch": 3.3724463069669985,
+      "loss_partial": 0.04638671875,
+      "step": 6438
+    },
+    {
+      "epoch": 3.3729701414353066,
+      "grad_norm": 1.092430830001831,
+      "learning_rate": 1.4626294453315693e-06,
+      "loss": 0.1012,
+      "step": 6439
+    },
+    {
+      "epoch": 3.3729701414353066,
+      "loss_partial": 0.05859375,
+      "step": 6439
+    },
+    {
+      "epoch": 3.3729701414353066,
+      "loss_partial": 0.047607421875,
+      "step": 6439
+    },
+    {
+      "epoch": 3.3734939759036147,
+      "grad_norm": 1.14517343044281,
+      "learning_rate": 1.4602498936826781e-06,
+      "loss": 0.0997,
+      "step": 6440
+    },
+    {
+      "epoch": 3.3734939759036147,
+      "loss_partial": 0.044677734375,
+      "step": 6440
+    },
+    {
+      "epoch": 3.3734939759036147,
+      "loss_partial": 0.045166015625,
+      "step": 6440
+    },
+    {
+      "epoch": 3.3740178103719227,
+      "grad_norm": 1.1945548057556152,
+      "learning_rate": 1.457872126788752e-06,
+      "loss": 0.0992,
+      "step": 6441
+    },
+    {
+      "epoch": 3.3740178103719227,
+      "loss_partial": 0.05908203125,
+      "step": 6441
+    },
+    {
+      "epoch": 3.3740178103719227,
+      "loss_partial": 0.0458984375,
+      "step": 6441
+    },
+    {
+      "epoch": 3.3745416448402303,
+      "grad_norm": 1.216284155845642,
+      "learning_rate": 1.4554961451467298e-06,
+      "loss": 0.1018,
+      "step": 6442
+    },
+    {
+      "epoch": 3.3745416448402303,
+      "loss_partial": 0.0576171875,
+      "step": 6442
+    },
+    {
+      "epoch": 3.3745416448402303,
+      "loss_partial": 0.05029296875,
+      "step": 6442
+    },
+    {
+      "epoch": 3.3750654793085384,
+      "grad_norm": 0.9619084596633911,
+      "learning_rate": 1.4531219492531712e-06,
+      "loss": 0.099,
+      "step": 6443
+    },
+    {
+      "epoch": 3.3750654793085384,
+      "loss_partial": 0.05615234375,
+      "step": 6443
+    },
+    {
+      "epoch": 3.3750654793085384,
+      "loss_partial": 0.050048828125,
+      "step": 6443
+    },
+    {
+      "epoch": 3.3755893137768465,
+      "grad_norm": 0.941428542137146,
+      "learning_rate": 1.4507495396042747e-06,
+      "loss": 0.104,
+      "step": 6444
+    },
+    {
+      "epoch": 3.3755893137768465,
+      "loss_partial": 0.04296875,
+      "step": 6444
+    },
+    {
+      "epoch": 3.3755893137768465,
+      "loss_partial": 0.048828125,
+      "step": 6444
+    },
+    {
+      "epoch": 3.3761131482451545,
+      "grad_norm": 1.3870673179626465,
+      "learning_rate": 1.4483789166958572e-06,
+      "loss": 0.0956,
+      "step": 6445
+    },
+    {
+      "epoch": 3.3761131482451545,
+      "loss_partial": 0.057373046875,
+      "step": 6445
+    },
+    {
+      "epoch": 3.3761131482451545,
+      "loss_partial": 0.04638671875,
+      "step": 6445
+    },
+    {
+      "epoch": 3.3766369827134626,
+      "grad_norm": 0.9862642884254456,
+      "learning_rate": 1.4460100810233657e-06,
+      "loss": 0.1012,
+      "step": 6446
+    },
+    {
+      "epoch": 3.3766369827134626,
+      "loss_partial": 0.03955078125,
+      "step": 6446
+    },
+    {
+      "epoch": 3.3766369827134626,
+      "loss_partial": 0.05078125,
+      "step": 6446
+    },
+    {
+      "epoch": 3.3771608171817706,
+      "grad_norm": 1.0521141290664673,
+      "learning_rate": 1.4436430330818685e-06,
+      "loss": 0.0984,
+      "step": 6447
+    },
+    {
+      "epoch": 3.3771608171817706,
+      "loss_partial": 0.050048828125,
+      "step": 6447
+    },
+    {
+      "epoch": 3.3771608171817706,
+      "loss_partial": 0.048095703125,
+      "step": 6447
+    },
+    {
+      "epoch": 3.3776846516500787,
+      "grad_norm": 1.028812289237976,
+      "learning_rate": 1.4412777733660678e-06,
+      "loss": 0.0981,
+      "step": 6448
+    },
+    {
+      "epoch": 3.3776846516500787,
+      "loss_partial": 0.05126953125,
+      "step": 6448
+    },
+    {
+      "epoch": 3.3776846516500787,
+      "loss_partial": 0.05029296875,
+      "step": 6448
+    },
+    {
+      "epoch": 3.3782084861183868,
+      "grad_norm": 0.8344314694404602,
+      "learning_rate": 1.4389143023702878e-06,
+      "loss": 0.0988,
+      "step": 6449
+    },
+    {
+      "epoch": 3.3782084861183868,
+      "loss_partial": 0.045654296875,
+      "step": 6449
+    },
+    {
+      "epoch": 3.3782084861183868,
+      "loss_partial": 0.052734375,
+      "step": 6449
+    },
+    {
+      "epoch": 3.3787323205866944,
+      "grad_norm": 0.9969456195831299,
+      "learning_rate": 1.4365526205884783e-06,
+      "loss": 0.099,
+      "step": 6450
+    },
+    {
+      "epoch": 3.3787323205866944,
+      "loss_partial": 0.045166015625,
+      "step": 6450
+    },
+    {
+      "epoch": 3.3787323205866944,
+      "loss_partial": 0.056884765625,
+      "step": 6450
+    },
+    {
+      "epoch": 3.3792561550550024,
+      "grad_norm": 0.9942353963851929,
+      "learning_rate": 1.4341927285142132e-06,
+      "loss": 0.0975,
+      "step": 6451
+    },
+    {
+      "epoch": 3.3792561550550024,
+      "loss_partial": 0.048583984375,
+      "step": 6451
+    },
+    {
+      "epoch": 3.3792561550550024,
+      "loss_partial": 0.05615234375,
+      "step": 6451
+    },
+    {
+      "epoch": 3.3797799895233105,
+      "grad_norm": 0.9536429643630981,
+      "learning_rate": 1.4318346266407023e-06,
+      "loss": 0.0959,
+      "step": 6452
+    },
+    {
+      "epoch": 3.3797799895233105,
+      "loss_partial": 0.049072265625,
+      "step": 6452
+    },
+    {
+      "epoch": 3.3797799895233105,
+      "loss_partial": 0.05029296875,
+      "step": 6452
+    },
+    {
+      "epoch": 3.3803038239916186,
+      "grad_norm": 1.1133290529251099,
+      "learning_rate": 1.4294783154607694e-06,
+      "loss": 0.1017,
+      "step": 6453
+    },
+    {
+      "epoch": 3.3803038239916186,
+      "loss_partial": 0.044677734375,
+      "step": 6453
+    },
+    {
+      "epoch": 3.3803038239916186,
+      "loss_partial": 0.0498046875,
+      "step": 6453
+    },
+    {
+      "epoch": 3.3808276584599266,
+      "grad_norm": 0.9953356385231018,
+      "learning_rate": 1.4271237954668726e-06,
+      "loss": 0.0985,
+      "step": 6454
+    },
+    {
+      "epoch": 3.3808276584599266,
+      "loss_partial": 0.04736328125,
+      "step": 6454
+    },
+    {
+      "epoch": 3.3808276584599266,
+      "loss_partial": 0.0458984375,
+      "step": 6454
+    },
+    {
+      "epoch": 3.3813514929282347,
+      "grad_norm": 0.9352977871894836,
+      "learning_rate": 1.4247710671510851e-06,
+      "loss": 0.0978,
+      "step": 6455
+    },
+    {
+      "epoch": 3.3813514929282347,
+      "loss_partial": 0.04931640625,
+      "step": 6455
+    },
+    {
+      "epoch": 3.3813514929282347,
+      "loss_partial": 0.047607421875,
+      "step": 6455
+    },
+    {
+      "epoch": 3.3818753273965427,
+      "grad_norm": 1.241015911102295,
+      "learning_rate": 1.4224201310051188e-06,
+      "loss": 0.0978,
+      "step": 6456
+    },
+    {
+      "epoch": 3.3818753273965427,
+      "loss_partial": 0.051513671875,
+      "step": 6456
+    },
+    {
+      "epoch": 3.3818753273965427,
+      "loss_partial": 0.046875,
+      "step": 6456
+    },
+    {
+      "epoch": 3.382399161864851,
+      "grad_norm": 1.0519282817840576,
+      "learning_rate": 1.4200709875203034e-06,
+      "loss": 0.1,
+      "step": 6457
+    },
+    {
+      "epoch": 3.382399161864851,
+      "loss_partial": 0.040283203125,
+      "step": 6457
+    },
+    {
+      "epoch": 3.382399161864851,
+      "loss_partial": 0.05224609375,
+      "step": 6457
+    },
+    {
+      "epoch": 3.382922996333159,
+      "grad_norm": 1.0700565576553345,
+      "learning_rate": 1.4177236371875946e-06,
+      "loss": 0.0978,
+      "step": 6458
+    },
+    {
+      "epoch": 3.382922996333159,
+      "loss_partial": 0.052001953125,
+      "step": 6458
+    },
+    {
+      "epoch": 3.382922996333159,
+      "loss_partial": 0.0458984375,
+      "step": 6458
+    },
+    {
+      "epoch": 3.383446830801467,
+      "grad_norm": 1.1104873418807983,
+      "learning_rate": 1.4153780804975714e-06,
+      "loss": 0.0982,
+      "step": 6459
+    },
+    {
+      "epoch": 3.383446830801467,
+      "loss_partial": 0.06103515625,
+      "step": 6459
+    },
+    {
+      "epoch": 3.383446830801467,
+      "loss_partial": 0.057861328125,
+      "step": 6459
+    },
+    {
+      "epoch": 3.383970665269775,
+      "grad_norm": 1.0633652210235596,
+      "learning_rate": 1.413034317940447e-06,
+      "loss": 0.0998,
+      "step": 6460
+    },
+    {
+      "epoch": 3.383970665269775,
+      "loss_partial": 0.055419921875,
+      "step": 6460
+    },
+    {
+      "epoch": 3.383970665269775,
+      "loss_partial": 0.054931640625,
+      "step": 6460
+    },
+    {
+      "epoch": 3.3844944997380826,
+      "grad_norm": 1.0147885084152222,
+      "learning_rate": 1.4106923500060498e-06,
+      "loss": 0.1014,
+      "step": 6461
+    },
+    {
+      "epoch": 3.3844944997380826,
+      "loss_partial": 0.04833984375,
+      "step": 6461
+    },
+    {
+      "epoch": 3.3844944997380826,
+      "loss_partial": 0.041015625,
+      "step": 6461
+    },
+    {
+      "epoch": 3.3850183342063906,
+      "grad_norm": 1.131536602973938,
+      "learning_rate": 1.408352177183837e-06,
+      "loss": 0.0997,
+      "step": 6462
+    },
+    {
+      "epoch": 3.3850183342063906,
+      "loss_partial": 0.047607421875,
+      "step": 6462
+    },
+    {
+      "epoch": 3.3850183342063906,
+      "loss_partial": 0.060791015625,
+      "step": 6462
+    },
+    {
+      "epoch": 3.3855421686746987,
+      "grad_norm": 0.9142684936523438,
+      "learning_rate": 1.4060137999628875e-06,
+      "loss": 0.0992,
+      "step": 6463
+    },
+    {
+      "epoch": 3.3855421686746987,
+      "loss_partial": 0.048583984375,
+      "step": 6463
+    },
+    {
+      "epoch": 3.3855421686746987,
+      "loss_partial": 0.040283203125,
+      "step": 6463
+    },
+    {
+      "epoch": 3.3860660031430068,
+      "grad_norm": 1.1305807828903198,
+      "learning_rate": 1.4036772188319125e-06,
+      "loss": 0.0994,
+      "step": 6464
+    },
+    {
+      "epoch": 3.3860660031430068,
+      "loss_partial": 0.045654296875,
+      "step": 6464
+    },
+    {
+      "epoch": 3.3860660031430068,
+      "loss_partial": 0.0556640625,
+      "step": 6464
+    },
+    {
+      "epoch": 3.386589837611315,
+      "grad_norm": 1.0243998765945435,
+      "learning_rate": 1.4013424342792403e-06,
+      "loss": 0.1011,
+      "step": 6465
+    },
+    {
+      "epoch": 3.386589837611315,
+      "loss_partial": 0.06005859375,
+      "step": 6465
+    },
+    {
+      "epoch": 3.386589837611315,
+      "loss_partial": 0.0439453125,
+      "step": 6465
+    },
+    {
+      "epoch": 3.387113672079623,
+      "grad_norm": 1.006709337234497,
+      "learning_rate": 1.3990094467928273e-06,
+      "loss": 0.0984,
+      "step": 6466
+    },
+    {
+      "epoch": 3.387113672079623,
+      "loss_partial": 0.051513671875,
+      "step": 6466
+    },
+    {
+      "epoch": 3.387113672079623,
+      "loss_partial": 0.047607421875,
+      "step": 6466
+    },
+    {
+      "epoch": 3.387637506547931,
+      "grad_norm": 1.0439304113388062,
+      "learning_rate": 1.3966782568602532e-06,
+      "loss": 0.1023,
+      "step": 6467
+    },
+    {
+      "epoch": 3.387637506547931,
+      "loss_partial": 0.047607421875,
+      "step": 6467
+    },
+    {
+      "epoch": 3.387637506547931,
+      "loss_partial": 0.046630859375,
+      "step": 6467
+    },
+    {
+      "epoch": 3.388161341016239,
+      "grad_norm": 0.9421868920326233,
+      "learning_rate": 1.394348864968722e-06,
+      "loss": 0.1033,
+      "step": 6468
+    },
+    {
+      "epoch": 3.388161341016239,
+      "loss_partial": 0.052978515625,
+      "step": 6468
+    },
+    {
+      "epoch": 3.388161341016239,
+      "loss_partial": 0.052734375,
+      "step": 6468
+    },
+    {
+      "epoch": 3.388685175484547,
+      "grad_norm": 1.2404398918151855,
+      "learning_rate": 1.3920212716050618e-06,
+      "loss": 0.0992,
+      "step": 6469
+    },
+    {
+      "epoch": 3.388685175484547,
+      "loss_partial": 0.0517578125,
+      "step": 6469
+    },
+    {
+      "epoch": 3.388685175484547,
+      "loss_partial": 0.047119140625,
+      "step": 6469
+    },
+    {
+      "epoch": 3.3892090099528547,
+      "grad_norm": 1.1466064453125,
+      "learning_rate": 1.3896954772557258e-06,
+      "loss": 0.1053,
+      "step": 6470
+    },
+    {
+      "epoch": 3.3892090099528547,
+      "loss_partial": 0.049072265625,
+      "step": 6470
+    },
+    {
+      "epoch": 3.3892090099528547,
+      "loss_partial": 0.048828125,
+      "step": 6470
+    },
+    {
+      "epoch": 3.3897328444211627,
+      "grad_norm": 1.0879019498825073,
+      "learning_rate": 1.3873714824067886e-06,
+      "loss": 0.1026,
+      "step": 6471
+    },
+    {
+      "epoch": 3.3897328444211627,
+      "loss_partial": 0.04736328125,
+      "step": 6471
+    },
+    {
+      "epoch": 3.3897328444211627,
+      "loss_partial": 0.060546875,
+      "step": 6471
+    },
+    {
+      "epoch": 3.390256678889471,
+      "grad_norm": 1.013136386871338,
+      "learning_rate": 1.3850492875439546e-06,
+      "loss": 0.0993,
+      "step": 6472
+    },
+    {
+      "epoch": 3.390256678889471,
+      "loss_partial": 0.054931640625,
+      "step": 6472
+    },
+    {
+      "epoch": 3.390256678889471,
+      "loss_partial": 0.0400390625,
+      "step": 6472
+    },
+    {
+      "epoch": 3.390780513357779,
+      "grad_norm": 1.0585724115371704,
+      "learning_rate": 1.3827288931525462e-06,
+      "loss": 0.1026,
+      "step": 6473
+    },
+    {
+      "epoch": 3.390780513357779,
+      "loss_partial": 0.052978515625,
+      "step": 6473
+    },
+    {
+      "epoch": 3.390780513357779,
+      "loss_partial": 0.045654296875,
+      "step": 6473
+    },
+    {
+      "epoch": 3.391304347826087,
+      "grad_norm": 1.0201613903045654,
+      "learning_rate": 1.3804102997175118e-06,
+      "loss": 0.0999,
+      "step": 6474
+    },
+    {
+      "epoch": 3.391304347826087,
+      "loss_partial": 0.04541015625,
+      "step": 6474
+    },
+    {
+      "epoch": 3.391304347826087,
+      "loss_partial": 0.0439453125,
+      "step": 6474
+    },
+    {
+      "epoch": 3.391828182294395,
+      "grad_norm": 1.2199721336364746,
+      "learning_rate": 1.3780935077234204e-06,
+      "loss": 0.0982,
+      "step": 6475
+    },
+    {
+      "epoch": 3.391828182294395,
+      "loss_partial": 0.04248046875,
+      "step": 6475
+    },
+    {
+      "epoch": 3.391828182294395,
+      "loss_partial": 0.042724609375,
+      "step": 6475
+    },
+    {
+      "epoch": 3.392352016762703,
+      "grad_norm": 1.1785438060760498,
+      "learning_rate": 1.3757785176544724e-06,
+      "loss": 0.094,
+      "step": 6476
+    },
+    {
+      "epoch": 3.392352016762703,
+      "loss_partial": 0.048095703125,
+      "step": 6476
+    },
+    {
+      "epoch": 3.392352016762703,
+      "loss_partial": 0.05224609375,
+      "step": 6476
+    },
+    {
+      "epoch": 3.392875851231011,
+      "grad_norm": 1.3254668712615967,
+      "learning_rate": 1.3734653299944834e-06,
+      "loss": 0.099,
+      "step": 6477
+    },
+    {
+      "epoch": 3.392875851231011,
+      "loss_partial": 0.04931640625,
+      "step": 6477
+    },
+    {
+      "epoch": 3.392875851231011,
+      "loss_partial": 0.051513671875,
+      "step": 6477
+    },
+    {
+      "epoch": 3.393399685699319,
+      "grad_norm": 1.0237962007522583,
+      "learning_rate": 1.3711539452268951e-06,
+      "loss": 0.101,
+      "step": 6478
+    },
+    {
+      "epoch": 3.393399685699319,
+      "loss_partial": 0.05810546875,
+      "step": 6478
+    },
+    {
+      "epoch": 3.393399685699319,
+      "loss_partial": 0.054931640625,
+      "step": 6478
+    },
+    {
+      "epoch": 3.393923520167627,
+      "grad_norm": 0.7607654333114624,
+      "learning_rate": 1.3688443638347726e-06,
+      "loss": 0.0988,
+      "step": 6479
+    },
+    {
+      "epoch": 3.393923520167627,
+      "loss_partial": 0.05224609375,
+      "step": 6479
+    },
+    {
+      "epoch": 3.393923520167627,
+      "loss_partial": 0.046630859375,
+      "step": 6479
+    },
+    {
+      "epoch": 3.3944473546359353,
+      "grad_norm": 1.28190279006958,
+      "learning_rate": 1.3665365863008063e-06,
+      "loss": 0.1017,
+      "step": 6480
+    },
+    {
+      "epoch": 3.3944473546359353,
+      "loss_partial": 0.048095703125,
+      "step": 6480
+    },
+    {
+      "epoch": 3.3944473546359353,
+      "loss_partial": 0.057373046875,
+      "step": 6480
+    },
+    {
+      "epoch": 3.394971189104243,
+      "grad_norm": 1.0876152515411377,
+      "learning_rate": 1.3642306131073047e-06,
+      "loss": 0.0996,
+      "step": 6481
+    },
+    {
+      "epoch": 3.394971189104243,
+      "loss_partial": 0.04931640625,
+      "step": 6481
+    },
+    {
+      "epoch": 3.394971189104243,
+      "loss_partial": 0.051513671875,
+      "step": 6481
+    },
+    {
+      "epoch": 3.395495023572551,
+      "grad_norm": 0.8521604537963867,
+      "learning_rate": 1.3619264447362023e-06,
+      "loss": 0.099,
+      "step": 6482
+    },
+    {
+      "epoch": 3.395495023572551,
+      "loss_partial": 0.057373046875,
+      "step": 6482
+    },
+    {
+      "epoch": 3.395495023572551,
+      "loss_partial": 0.051513671875,
+      "step": 6482
+    },
+    {
+      "epoch": 3.396018858040859,
+      "grad_norm": 1.1795262098312378,
+      "learning_rate": 1.3596240816690553e-06,
+      "loss": 0.1006,
+      "step": 6483
+    },
+    {
+      "epoch": 3.396018858040859,
+      "loss_partial": 0.054443359375,
+      "step": 6483
+    },
+    {
+      "epoch": 3.396018858040859,
+      "loss_partial": 0.047119140625,
+      "step": 6483
+    },
+    {
+      "epoch": 3.396542692509167,
+      "grad_norm": 1.129590630531311,
+      "learning_rate": 1.3573235243870476e-06,
+      "loss": 0.1012,
+      "step": 6484
+    },
+    {
+      "epoch": 3.396542692509167,
+      "loss_partial": 0.050537109375,
+      "step": 6484
+    },
+    {
+      "epoch": 3.396542692509167,
+      "loss_partial": 0.051025390625,
+      "step": 6484
+    },
+    {
+      "epoch": 3.397066526977475,
+      "grad_norm": 1.188988447189331,
+      "learning_rate": 1.3550247733709777e-06,
+      "loss": 0.1,
+      "step": 6485
+    },
+    {
+      "epoch": 3.397066526977475,
+      "loss_partial": 0.040771484375,
+      "step": 6485
+    },
+    {
+      "epoch": 3.397066526977475,
+      "loss_partial": 0.05126953125,
+      "step": 6485
+    },
+    {
+      "epoch": 3.397590361445783,
+      "grad_norm": 1.026613712310791,
+      "learning_rate": 1.3527278291012713e-06,
+      "loss": 0.0999,
+      "step": 6486
+    },
+    {
+      "epoch": 3.397590361445783,
+      "loss_partial": 0.04248046875,
+      "step": 6486
+    },
+    {
+      "epoch": 3.397590361445783,
+      "loss_partial": 0.050048828125,
+      "step": 6486
+    },
+    {
+      "epoch": 3.3981141959140913,
+      "grad_norm": 0.9222801327705383,
+      "learning_rate": 1.3504326920579735e-06,
+      "loss": 0.1016,
+      "step": 6487
+    },
+    {
+      "epoch": 3.3981141959140913,
+      "loss_partial": 0.04541015625,
+      "step": 6487
+    },
+    {
+      "epoch": 3.3981141959140913,
+      "loss_partial": 0.0478515625,
+      "step": 6487
+    },
+    {
+      "epoch": 3.3986380303823993,
+      "grad_norm": 1.2398000955581665,
+      "learning_rate": 1.3481393627207574e-06,
+      "loss": 0.094,
+      "step": 6488
+    },
+    {
+      "epoch": 3.3986380303823993,
+      "loss_partial": 0.050537109375,
+      "step": 6488
+    },
+    {
+      "epoch": 3.3986380303823993,
+      "loss_partial": 0.06689453125,
+      "step": 6488
+    },
+    {
+      "epoch": 3.3991618648507074,
+      "grad_norm": 1.3334194421768188,
+      "learning_rate": 1.3458478415689136e-06,
+      "loss": 0.1014,
+      "step": 6489
+    },
+    {
+      "epoch": 3.3991618648507074,
+      "loss_partial": 0.044921875,
+      "step": 6489
+    },
+    {
+      "epoch": 3.3991618648507074,
+      "loss_partial": 0.042236328125,
+      "step": 6489
+    },
+    {
+      "epoch": 3.399685699319015,
+      "grad_norm": 1.079120397567749,
+      "learning_rate": 1.343558129081355e-06,
+      "loss": 0.0973,
+      "step": 6490
+    },
+    {
+      "epoch": 3.399685699319015,
+      "loss_partial": 0.060546875,
+      "step": 6490
+    },
+    {
+      "epoch": 3.399685699319015,
+      "loss_partial": 0.05029296875,
+      "step": 6490
+    },
+    {
+      "epoch": 3.400209533787323,
+      "grad_norm": 1.3999738693237305,
+      "learning_rate": 1.3412702257366172e-06,
+      "loss": 0.0977,
+      "step": 6491
+    },
+    {
+      "epoch": 3.400209533787323,
+      "loss_partial": 0.05126953125,
+      "step": 6491
+    },
+    {
+      "epoch": 3.400209533787323,
+      "loss_partial": 0.044921875,
+      "step": 6491
+    },
+    {
+      "epoch": 3.400733368255631,
+      "grad_norm": 1.328373670578003,
+      "learning_rate": 1.3389841320128593e-06,
+      "loss": 0.1004,
+      "step": 6492
+    },
+    {
+      "epoch": 3.400733368255631,
+      "loss_partial": 0.04736328125,
+      "step": 6492
+    },
+    {
+      "epoch": 3.400733368255631,
+      "loss_partial": 0.04638671875,
+      "step": 6492
+    },
+    {
+      "epoch": 3.401257202723939,
+      "grad_norm": 1.2222033739089966,
+      "learning_rate": 1.336699848387859e-06,
+      "loss": 0.0962,
+      "step": 6493
+    },
+    {
+      "epoch": 3.401257202723939,
+      "loss_partial": 0.0517578125,
+      "step": 6493
+    },
+    {
+      "epoch": 3.401257202723939,
+      "loss_partial": 0.052978515625,
+      "step": 6493
+    },
+    {
+      "epoch": 3.4017810371922472,
+      "grad_norm": 0.9688495993614197,
+      "learning_rate": 1.334417375339019e-06,
+      "loss": 0.101,
+      "step": 6494
+    },
+    {
+      "epoch": 3.4017810371922472,
+      "loss_partial": 0.0517578125,
+      "step": 6494
+    },
+    {
+      "epoch": 3.4017810371922472,
+      "loss_partial": 0.04638671875,
+      "step": 6494
+    },
+    {
+      "epoch": 3.4023048716605553,
+      "grad_norm": 1.0286989212036133,
+      "learning_rate": 1.3321367133433605e-06,
+      "loss": 0.0993,
+      "step": 6495
+    },
+    {
+      "epoch": 3.4023048716605553,
+      "loss_partial": 0.055908203125,
+      "step": 6495
+    },
+    {
+      "epoch": 3.4023048716605553,
+      "loss_partial": 0.05126953125,
+      "step": 6495
+    },
+    {
+      "epoch": 3.4028287061288633,
+      "grad_norm": 1.1184357404708862,
+      "learning_rate": 1.3298578628775261e-06,
+      "loss": 0.1036,
+      "step": 6496
+    },
+    {
+      "epoch": 3.4028287061288633,
+      "loss_partial": 0.04345703125,
+      "step": 6496
+    },
+    {
+      "epoch": 3.4028287061288633,
+      "loss_partial": 0.0400390625,
+      "step": 6496
+    },
+    {
+      "epoch": 3.4033525405971714,
+      "grad_norm": 1.2883731126785278,
+      "learning_rate": 1.3275808244177858e-06,
+      "loss": 0.0996,
+      "step": 6497
+    },
+    {
+      "epoch": 3.4033525405971714,
+      "loss_partial": 0.0458984375,
+      "step": 6497
+    },
+    {
+      "epoch": 3.4033525405971714,
+      "loss_partial": 0.05126953125,
+      "step": 6497
+    },
+    {
+      "epoch": 3.4038763750654795,
+      "grad_norm": 1.0864498615264893,
+      "learning_rate": 1.3253055984400254e-06,
+      "loss": 0.0974,
+      "step": 6498
+    },
+    {
+      "epoch": 3.4038763750654795,
+      "loss_partial": 0.048095703125,
+      "step": 6498
+    },
+    {
+      "epoch": 3.4038763750654795,
+      "loss_partial": 0.04931640625,
+      "step": 6498
+    },
+    {
+      "epoch": 3.4044002095337875,
+      "grad_norm": 1.0221259593963623,
+      "learning_rate": 1.3230321854197515e-06,
+      "loss": 0.0998,
+      "step": 6499
+    },
+    {
+      "epoch": 3.4044002095337875,
+      "loss_partial": 0.043212890625,
+      "step": 6499
+    },
+    {
+      "epoch": 3.4044002095337875,
+      "loss_partial": 0.04736328125,
+      "step": 6499
+    },
+    {
+      "epoch": 3.4049240440020956,
+      "grad_norm": 0.9764578938484192,
+      "learning_rate": 1.3207605858320915e-06,
+      "loss": 0.0997,
+      "step": 6500
+    },
+    {
+      "epoch": 3.4049240440020956,
+      "loss_partial": 0.051025390625,
+      "step": 6500
+    },
+    {
+      "epoch": 3.4049240440020956,
+      "loss_partial": 0.053955078125,
+      "step": 6500
+    },
+    {
+      "epoch": 3.405447878470403,
+      "grad_norm": 1.3125463724136353,
+      "learning_rate": 1.3184908001518015e-06,
+      "loss": 0.1015,
+      "step": 6501
+    },
+    {
+      "epoch": 3.405447878470403,
+      "loss_partial": 0.048583984375,
+      "step": 6501
+    },
+    {
+      "epoch": 3.405447878470403,
+      "loss_partial": 0.0517578125,
+      "step": 6501
+    },
+    {
+      "epoch": 3.4059717129387113,
+      "grad_norm": 0.999309778213501,
+      "learning_rate": 1.3162228288532498e-06,
+      "loss": 0.0996,
+      "step": 6502
+    },
+    {
+      "epoch": 3.4059717129387113,
+      "loss_partial": 0.05126953125,
+      "step": 6502
+    },
+    {
+      "epoch": 3.4059717129387113,
+      "loss_partial": 0.04833984375,
+      "step": 6502
+    },
+    {
+      "epoch": 3.4064955474070193,
+      "grad_norm": 1.098611831665039,
+      "learning_rate": 1.3139566724104268e-06,
+      "loss": 0.103,
+      "step": 6503
+    },
+    {
+      "epoch": 3.4064955474070193,
+      "loss_partial": 0.048583984375,
+      "step": 6503
+    },
+    {
+      "epoch": 3.4064955474070193,
+      "loss_partial": 0.043701171875,
+      "step": 6503
+    },
+    {
+      "epoch": 3.4070193818753274,
+      "grad_norm": 1.6771727800369263,
+      "learning_rate": 1.3116923312969466e-06,
+      "loss": 0.1011,
+      "step": 6504
+    },
+    {
+      "epoch": 3.4070193818753274,
+      "loss_partial": 0.058837890625,
+      "step": 6504
+    },
+    {
+      "epoch": 3.4070193818753274,
+      "loss_partial": 0.044677734375,
+      "step": 6504
+    },
+    {
+      "epoch": 3.4075432163436354,
+      "grad_norm": 1.1759732961654663,
+      "learning_rate": 1.3094298059860422e-06,
+      "loss": 0.0987,
+      "step": 6505
+    },
+    {
+      "epoch": 3.4075432163436354,
+      "loss_partial": 0.053955078125,
+      "step": 6505
+    },
+    {
+      "epoch": 3.4075432163436354,
+      "loss_partial": 0.0419921875,
+      "step": 6505
+    },
+    {
+      "epoch": 3.4080670508119435,
+      "grad_norm": 0.876780092716217,
+      "learning_rate": 1.307169096950568e-06,
+      "loss": 0.0971,
+      "step": 6506
+    },
+    {
+      "epoch": 3.4080670508119435,
+      "loss_partial": 0.0517578125,
+      "step": 6506
+    },
+    {
+      "epoch": 3.4080670508119435,
+      "loss_partial": 0.052001953125,
+      "step": 6506
+    },
+    {
+      "epoch": 3.4085908852802516,
+      "grad_norm": 1.563476324081421,
+      "learning_rate": 1.304910204662998e-06,
+      "loss": 0.1031,
+      "step": 6507
+    },
+    {
+      "epoch": 3.4085908852802516,
+      "loss_partial": 0.04931640625,
+      "step": 6507
+    },
+    {
+      "epoch": 3.4085908852802516,
+      "loss_partial": 0.05322265625,
+      "step": 6507
+    },
+    {
+      "epoch": 3.4091147197485596,
+      "grad_norm": 0.8593183159828186,
+      "learning_rate": 1.3026531295954237e-06,
+      "loss": 0.1006,
+      "step": 6508
+    },
+    {
+      "epoch": 3.4091147197485596,
+      "loss_partial": 0.055908203125,
+      "step": 6508
+    },
+    {
+      "epoch": 3.4091147197485596,
+      "loss_partial": 0.0537109375,
+      "step": 6508
+    },
+    {
+      "epoch": 3.4096385542168672,
+      "grad_norm": 1.3571869134902954,
+      "learning_rate": 1.3003978722195643e-06,
+      "loss": 0.1006,
+      "step": 6509
+    },
+    {
+      "epoch": 3.4096385542168672,
+      "loss_partial": 0.054931640625,
+      "step": 6509
+    },
+    {
+      "epoch": 3.4096385542168672,
+      "loss_partial": 0.047607421875,
+      "step": 6509
+    },
+    {
+      "epoch": 3.4101623886851753,
+      "grad_norm": 0.7809088826179504,
+      "learning_rate": 1.2981444330067539e-06,
+      "loss": 0.0997,
+      "step": 6510
+    },
+    {
+      "epoch": 3.4101623886851753,
+      "loss_partial": 0.051513671875,
+      "step": 6510
+    },
+    {
+      "epoch": 3.4101623886851753,
+      "loss_partial": 0.051513671875,
+      "step": 6510
+    },
+    {
+      "epoch": 3.4106862231534834,
+      "grad_norm": 0.9386767148971558,
+      "learning_rate": 1.2958928124279457e-06,
+      "loss": 0.0991,
+      "step": 6511
+    },
+    {
+      "epoch": 3.4106862231534834,
+      "loss_partial": 0.05712890625,
+      "step": 6511
+    },
+    {
+      "epoch": 3.4106862231534834,
+      "loss_partial": 0.048583984375,
+      "step": 6511
+    },
+    {
+      "epoch": 3.4112100576217914,
+      "grad_norm": 0.923699676990509,
+      "learning_rate": 1.2936430109537125e-06,
+      "loss": 0.1,
+      "step": 6512
+    },
+    {
+      "epoch": 3.4112100576217914,
+      "loss_partial": 0.04638671875,
+      "step": 6512
+    },
+    {
+      "epoch": 3.4112100576217914,
+      "loss_partial": 0.047119140625,
+      "step": 6512
+    },
+    {
+      "epoch": 3.4117338920900995,
+      "grad_norm": 1.0181490182876587,
+      "learning_rate": 1.2913950290542543e-06,
+      "loss": 0.0983,
+      "step": 6513
+    },
+    {
+      "epoch": 3.4117338920900995,
+      "loss_partial": 0.049560546875,
+      "step": 6513
+    },
+    {
+      "epoch": 3.4117338920900995,
+      "loss_partial": 0.05859375,
+      "step": 6513
+    },
+    {
+      "epoch": 3.4122577265584075,
+      "grad_norm": 1.1407597064971924,
+      "learning_rate": 1.2891488671993824e-06,
+      "loss": 0.1023,
+      "step": 6514
+    },
+    {
+      "epoch": 3.4122577265584075,
+      "loss_partial": 0.055908203125,
+      "step": 6514
+    },
+    {
+      "epoch": 3.4122577265584075,
+      "loss_partial": 0.045654296875,
+      "step": 6514
+    },
+    {
+      "epoch": 3.4127815610267156,
+      "grad_norm": 0.9990971684455872,
+      "learning_rate": 1.2869045258585311e-06,
+      "loss": 0.1059,
+      "step": 6515
+    },
+    {
+      "epoch": 3.4127815610267156,
+      "loss_partial": 0.046875,
+      "step": 6515
+    },
+    {
+      "epoch": 3.4127815610267156,
+      "loss_partial": 0.04833984375,
+      "step": 6515
+    },
+    {
+      "epoch": 3.4133053954950237,
+      "grad_norm": 1.052750587463379,
+      "learning_rate": 1.2846620055007541e-06,
+      "loss": 0.0983,
+      "step": 6516
+    },
+    {
+      "epoch": 3.4133053954950237,
+      "loss_partial": 0.048095703125,
+      "step": 6516
+    },
+    {
+      "epoch": 3.4133053954950237,
+      "loss_partial": 0.0576171875,
+      "step": 6516
+    },
+    {
+      "epoch": 3.4138292299633317,
+      "grad_norm": 0.9817789793014526,
+      "learning_rate": 1.2824213065947232e-06,
+      "loss": 0.0976,
+      "step": 6517
+    },
+    {
+      "epoch": 3.4138292299633317,
+      "loss_partial": 0.04833984375,
+      "step": 6517
+    },
+    {
+      "epoch": 3.4138292299633317,
+      "loss_partial": 0.044189453125,
+      "step": 6517
+    },
+    {
+      "epoch": 3.4143530644316398,
+      "grad_norm": 1.0502996444702148,
+      "learning_rate": 1.2801824296087318e-06,
+      "loss": 0.0991,
+      "step": 6518
+    },
+    {
+      "epoch": 3.4143530644316398,
+      "loss_partial": 0.06298828125,
+      "step": 6518
+    },
+    {
+      "epoch": 3.4143530644316398,
+      "loss_partial": 0.0478515625,
+      "step": 6518
+    },
+    {
+      "epoch": 3.414876898899948,
+      "grad_norm": 0.9637386798858643,
+      "learning_rate": 1.2779453750106907e-06,
+      "loss": 0.1013,
+      "step": 6519
+    },
+    {
+      "epoch": 3.414876898899948,
+      "loss_partial": 0.06201171875,
+      "step": 6519
+    },
+    {
+      "epoch": 3.414876898899948,
+      "loss_partial": 0.050048828125,
+      "step": 6519
+    },
+    {
+      "epoch": 3.4154007333682554,
+      "grad_norm": 1.1121485233306885,
+      "learning_rate": 1.2757101432681295e-06,
+      "loss": 0.0991,
+      "step": 6520
+    },
+    {
+      "epoch": 3.4154007333682554,
+      "loss_partial": 0.044921875,
+      "step": 6520
+    },
+    {
+      "epoch": 3.4154007333682554,
+      "loss_partial": 0.041748046875,
+      "step": 6520
+    },
+    {
+      "epoch": 3.4159245678365635,
+      "grad_norm": 1.651775598526001,
+      "learning_rate": 1.2734767348482002e-06,
+      "loss": 0.0975,
+      "step": 6521
+    },
+    {
+      "epoch": 3.4159245678365635,
+      "loss_partial": 0.052734375,
+      "step": 6521
+    },
+    {
+      "epoch": 3.4159245678365635,
+      "loss_partial": 0.051025390625,
+      "step": 6521
+    },
+    {
+      "epoch": 3.4164484023048716,
+      "grad_norm": 1.1550668478012085,
+      "learning_rate": 1.2712451502176704e-06,
+      "loss": 0.103,
+      "step": 6522
+    },
+    {
+      "epoch": 3.4164484023048716,
+      "loss_partial": 0.052490234375,
+      "step": 6522
+    },
+    {
+      "epoch": 3.4164484023048716,
+      "loss_partial": 0.055908203125,
+      "step": 6522
+    },
+    {
+      "epoch": 3.4169722367731796,
+      "grad_norm": 0.8913529515266418,
+      "learning_rate": 1.269015389842927e-06,
+      "loss": 0.103,
+      "step": 6523
+    },
+    {
+      "epoch": 3.4169722367731796,
+      "loss_partial": 0.0537109375,
+      "step": 6523
+    },
+    {
+      "epoch": 3.4169722367731796,
+      "loss_partial": 0.05078125,
+      "step": 6523
+    },
+    {
+      "epoch": 3.4174960712414877,
+      "grad_norm": 1.1345956325531006,
+      "learning_rate": 1.266787454189974e-06,
+      "loss": 0.1014,
+      "step": 6524
+    },
+    {
+      "epoch": 3.4174960712414877,
+      "loss_partial": 0.05859375,
+      "step": 6524
+    },
+    {
+      "epoch": 3.4174960712414877,
+      "loss_partial": 0.045654296875,
+      "step": 6524
+    },
+    {
+      "epoch": 3.4180199057097957,
+      "grad_norm": 1.1188085079193115,
+      "learning_rate": 1.2645613437244408e-06,
+      "loss": 0.0998,
+      "step": 6525
+    },
+    {
+      "epoch": 3.4180199057097957,
+      "loss_partial": 0.048828125,
+      "step": 6525
+    },
+    {
+      "epoch": 3.4180199057097957,
+      "loss_partial": 0.04296875,
+      "step": 6525
+    },
+    {
+      "epoch": 3.418543740178104,
+      "grad_norm": 1.02622389793396,
+      "learning_rate": 1.2623370589115679e-06,
+      "loss": 0.0965,
+      "step": 6526
+    },
+    {
+      "epoch": 3.418543740178104,
+      "loss_partial": 0.0419921875,
+      "step": 6526
+    },
+    {
+      "epoch": 3.418543740178104,
+      "loss_partial": 0.05078125,
+      "step": 6526
+    },
+    {
+      "epoch": 3.419067574646412,
+      "grad_norm": 1.1487371921539307,
+      "learning_rate": 1.260114600216218e-06,
+      "loss": 0.098,
+      "step": 6527
+    },
+    {
+      "epoch": 3.419067574646412,
+      "loss_partial": 0.04052734375,
+      "step": 6527
+    },
+    {
+      "epoch": 3.419067574646412,
+      "loss_partial": 0.046142578125,
+      "step": 6527
+    },
+    {
+      "epoch": 3.41959140911472,
+      "grad_norm": 1.1536980867385864,
+      "learning_rate": 1.2578939681028713e-06,
+      "loss": 0.0981,
+      "step": 6528
+    },
+    {
+      "epoch": 3.41959140911472,
+      "loss_partial": 0.048828125,
+      "step": 6528
+    },
+    {
+      "epoch": 3.41959140911472,
+      "loss_partial": 0.048095703125,
+      "step": 6528
+    },
+    {
+      "epoch": 3.4201152435830275,
+      "grad_norm": 1.0845898389816284,
+      "learning_rate": 1.2556751630356257e-06,
+      "loss": 0.1,
+      "step": 6529
+    },
+    {
+      "epoch": 3.4201152435830275,
+      "loss_partial": 0.04345703125,
+      "step": 6529
+    },
+    {
+      "epoch": 3.4201152435830275,
+      "loss_partial": 0.039794921875,
+      "step": 6529
+    },
+    {
+      "epoch": 3.4206390780513356,
+      "grad_norm": 1.5700348615646362,
+      "learning_rate": 1.2534581854781969e-06,
+      "loss": 0.0974,
+      "step": 6530
+    },
+    {
+      "epoch": 3.4206390780513356,
+      "loss_partial": 0.04443359375,
+      "step": 6530
+    },
+    {
+      "epoch": 3.4206390780513356,
+      "loss_partial": 0.04541015625,
+      "step": 6530
+    },
+    {
+      "epoch": 3.4211629125196437,
+      "grad_norm": 1.0374068021774292,
+      "learning_rate": 1.251243035893921e-06,
+      "loss": 0.102,
+      "step": 6531
+    },
+    {
+      "epoch": 3.4211629125196437,
+      "loss_partial": 0.04248046875,
+      "step": 6531
+    },
+    {
+      "epoch": 3.4211629125196437,
+      "loss_partial": 0.04736328125,
+      "step": 6531
+    },
+    {
+      "epoch": 3.4216867469879517,
+      "grad_norm": 1.1805269718170166,
+      "learning_rate": 1.2490297147457475e-06,
+      "loss": 0.0989,
+      "step": 6532
+    },
+    {
+      "epoch": 3.4216867469879517,
+      "loss_partial": 0.050537109375,
+      "step": 6532
+    },
+    {
+      "epoch": 3.4216867469879517,
+      "loss_partial": 0.051513671875,
+      "step": 6532
+    },
+    {
+      "epoch": 3.42221058145626,
+      "grad_norm": 0.9269155263900757,
+      "learning_rate": 1.2468182224962521e-06,
+      "loss": 0.0999,
+      "step": 6533
+    },
+    {
+      "epoch": 3.42221058145626,
+      "loss_partial": 0.04345703125,
+      "step": 6533
+    },
+    {
+      "epoch": 3.42221058145626,
+      "loss_partial": 0.05126953125,
+      "step": 6533
+    },
+    {
+      "epoch": 3.422734415924568,
+      "grad_norm": 1.0806549787521362,
+      "learning_rate": 1.2446085596076197e-06,
+      "loss": 0.1011,
+      "step": 6534
+    },
+    {
+      "epoch": 3.422734415924568,
+      "loss_partial": 0.059326171875,
+      "step": 6534
+    },
+    {
+      "epoch": 3.422734415924568,
+      "loss_partial": 0.046875,
+      "step": 6534
+    },
+    {
+      "epoch": 3.423258250392876,
+      "grad_norm": 1.20064115524292,
+      "learning_rate": 1.242400726541656e-06,
+      "loss": 0.1006,
+      "step": 6535
+    },
+    {
+      "epoch": 3.423258250392876,
+      "loss_partial": 0.04833984375,
+      "step": 6535
+    },
+    {
+      "epoch": 3.423258250392876,
+      "loss_partial": 0.05126953125,
+      "step": 6535
+    },
+    {
+      "epoch": 3.423782084861184,
+      "grad_norm": 0.8813166618347168,
+      "learning_rate": 1.2401947237597833e-06,
+      "loss": 0.0999,
+      "step": 6536
+    },
+    {
+      "epoch": 3.423782084861184,
+      "loss_partial": 0.046142578125,
+      "step": 6536
+    },
+    {
+      "epoch": 3.423782084861184,
+      "loss_partial": 0.040283203125,
+      "step": 6536
+    },
+    {
+      "epoch": 3.424305919329492,
+      "grad_norm": 1.325202226638794,
+      "learning_rate": 1.2379905517230472e-06,
+      "loss": 0.0998,
+      "step": 6537
+    },
+    {
+      "epoch": 3.424305919329492,
+      "loss_partial": 0.04541015625,
+      "step": 6537
+    },
+    {
+      "epoch": 3.424305919329492,
+      "loss_partial": 0.057373046875,
+      "step": 6537
+    },
+    {
+      "epoch": 3.4248297537978,
+      "grad_norm": 1.4555120468139648,
+      "learning_rate": 1.2357882108921026e-06,
+      "loss": 0.097,
+      "step": 6538
+    },
+    {
+      "epoch": 3.4248297537978,
+      "loss_partial": 0.04833984375,
+      "step": 6538
+    },
+    {
+      "epoch": 3.4248297537978,
+      "loss_partial": 0.0498046875,
+      "step": 6538
+    },
+    {
+      "epoch": 3.425353588266108,
+      "grad_norm": 0.9744271636009216,
+      "learning_rate": 1.2335877017272257e-06,
+      "loss": 0.1008,
+      "step": 6539
+    },
+    {
+      "epoch": 3.425353588266108,
+      "loss_partial": 0.051025390625,
+      "step": 6539
+    },
+    {
+      "epoch": 3.425353588266108,
+      "loss_partial": 0.05029296875,
+      "step": 6539
+    },
+    {
+      "epoch": 3.4258774227344158,
+      "grad_norm": 1.1288347244262695,
+      "learning_rate": 1.2313890246883087e-06,
+      "loss": 0.1007,
+      "step": 6540
+    },
+    {
+      "epoch": 3.4258774227344158,
+      "loss_partial": 0.045166015625,
+      "step": 6540
+    },
+    {
+      "epoch": 3.4258774227344158,
+      "loss_partial": 0.0439453125,
+      "step": 6540
+    },
+    {
+      "epoch": 3.426401257202724,
+      "grad_norm": 1.0883398056030273,
+      "learning_rate": 1.2291921802348628e-06,
+      "loss": 0.1,
+      "step": 6541
+    },
+    {
+      "epoch": 3.426401257202724,
+      "loss_partial": 0.055908203125,
+      "step": 6541
+    },
+    {
+      "epoch": 3.426401257202724,
+      "loss_partial": 0.048583984375,
+      "step": 6541
+    },
+    {
+      "epoch": 3.426925091671032,
+      "grad_norm": 1.3852535486221313,
+      "learning_rate": 1.226997168826014e-06,
+      "loss": 0.0991,
+      "step": 6542
+    },
+    {
+      "epoch": 3.426925091671032,
+      "loss_partial": 0.04736328125,
+      "step": 6542
+    },
+    {
+      "epoch": 3.426925091671032,
+      "loss_partial": 0.044921875,
+      "step": 6542
+    },
+    {
+      "epoch": 3.42744892613934,
+      "grad_norm": 1.6010758876800537,
+      "learning_rate": 1.2248039909205067e-06,
+      "loss": 0.1001,
+      "step": 6543
+    },
+    {
+      "epoch": 3.42744892613934,
+      "loss_partial": 0.05615234375,
+      "step": 6543
+    },
+    {
+      "epoch": 3.42744892613934,
+      "loss_partial": 0.05029296875,
+      "step": 6543
+    },
+    {
+      "epoch": 3.427972760607648,
+      "grad_norm": 0.9161674976348877,
+      "learning_rate": 1.2226126469766985e-06,
+      "loss": 0.1001,
+      "step": 6544
+    },
+    {
+      "epoch": 3.427972760607648,
+      "loss_partial": 0.046142578125,
+      "step": 6544
+    },
+    {
+      "epoch": 3.427972760607648,
+      "loss_partial": 0.041748046875,
+      "step": 6544
+    },
+    {
+      "epoch": 3.428496595075956,
+      "grad_norm": 0.9324565529823303,
+      "learning_rate": 1.2204231374525722e-06,
+      "loss": 0.0986,
+      "step": 6545
+    },
+    {
+      "epoch": 3.428496595075956,
+      "loss_partial": 0.04638671875,
+      "step": 6545
+    },
+    {
+      "epoch": 3.428496595075956,
+      "loss_partial": 0.0478515625,
+      "step": 6545
+    },
+    {
+      "epoch": 3.429020429544264,
+      "grad_norm": 1.2217644453048706,
+      "learning_rate": 1.218235462805718e-06,
+      "loss": 0.0937,
+      "step": 6546
+    },
+    {
+      "epoch": 3.429020429544264,
+      "loss_partial": 0.05908203125,
+      "step": 6546
+    },
+    {
+      "epoch": 3.429020429544264,
+      "loss_partial": 0.06787109375,
+      "step": 6546
+    },
+    {
+      "epoch": 3.429544264012572,
+      "grad_norm": 1.1631593704223633,
+      "learning_rate": 1.2160496234933471e-06,
+      "loss": 0.1017,
+      "step": 6547
+    },
+    {
+      "epoch": 3.429544264012572,
+      "loss_partial": 0.049072265625,
+      "step": 6547
+    },
+    {
+      "epoch": 3.429544264012572,
+      "loss_partial": 0.041015625,
+      "step": 6547
+    },
+    {
+      "epoch": 3.4300680984808802,
+      "grad_norm": 1.0274813175201416,
+      "learning_rate": 1.2138656199722854e-06,
+      "loss": 0.0981,
+      "step": 6548
+    },
+    {
+      "epoch": 3.4300680984808802,
+      "loss_partial": 0.058349609375,
+      "step": 6548
+    },
+    {
+      "epoch": 3.4300680984808802,
+      "loss_partial": 0.051025390625,
+      "step": 6548
+    },
+    {
+      "epoch": 3.430591932949188,
+      "grad_norm": 0.9641110301017761,
+      "learning_rate": 1.2116834526989785e-06,
+      "loss": 0.0982,
+      "step": 6549
+    },
+    {
+      "epoch": 3.430591932949188,
+      "loss_partial": 0.04248046875,
+      "step": 6549
+    },
+    {
+      "epoch": 3.430591932949188,
+      "loss_partial": 0.049072265625,
+      "step": 6549
+    },
+    {
+      "epoch": 3.431115767417496,
+      "grad_norm": 1.4600834846496582,
+      "learning_rate": 1.209503122129485e-06,
+      "loss": 0.0992,
+      "step": 6550
+    },
+    {
+      "epoch": 3.431115767417496,
+      "loss_partial": 0.050048828125,
+      "step": 6550
+    },
+    {
+      "epoch": 3.431115767417496,
+      "loss_partial": 0.042724609375,
+      "step": 6550
+    },
+    {
+      "epoch": 3.431639601885804,
+      "grad_norm": 0.9614516496658325,
+      "learning_rate": 1.2073246287194806e-06,
+      "loss": 0.0961,
+      "step": 6551
+    },
+    {
+      "epoch": 3.431639601885804,
+      "loss_partial": 0.044677734375,
+      "step": 6551
+    },
+    {
+      "epoch": 3.431639601885804,
+      "loss_partial": 0.050537109375,
+      "step": 6551
+    },
+    {
+      "epoch": 3.432163436354112,
+      "grad_norm": 1.1410131454467773,
+      "learning_rate": 1.2051479729242566e-06,
+      "loss": 0.1001,
+      "step": 6552
+    },
+    {
+      "epoch": 3.432163436354112,
+      "loss_partial": 0.048828125,
+      "step": 6552
+    },
+    {
+      "epoch": 3.432163436354112,
+      "loss_partial": 0.046875,
+      "step": 6552
+    },
+    {
+      "epoch": 3.43268727082242,
+      "grad_norm": 0.9886347055435181,
+      "learning_rate": 1.20297315519872e-06,
+      "loss": 0.0971,
+      "step": 6553
+    },
+    {
+      "epoch": 3.43268727082242,
+      "loss_partial": 0.050048828125,
+      "step": 6553
+    },
+    {
+      "epoch": 3.43268727082242,
+      "loss_partial": 0.0517578125,
+      "step": 6553
+    },
+    {
+      "epoch": 3.433211105290728,
+      "grad_norm": 1.1277928352355957,
+      "learning_rate": 1.2008001759973954e-06,
+      "loss": 0.1029,
+      "step": 6554
+    },
+    {
+      "epoch": 3.433211105290728,
+      "loss_partial": 0.042236328125,
+      "step": 6554
+    },
+    {
+      "epoch": 3.433211105290728,
+      "loss_partial": 0.06201171875,
+      "step": 6554
+    },
+    {
+      "epoch": 3.433734939759036,
+      "grad_norm": 0.99554443359375,
+      "learning_rate": 1.1986290357744202e-06,
+      "loss": 0.1004,
+      "step": 6555
+    },
+    {
+      "epoch": 3.433734939759036,
+      "loss_partial": 0.041015625,
+      "step": 6555
+    },
+    {
+      "epoch": 3.433734939759036,
+      "loss_partial": 0.044677734375,
+      "step": 6555
+    },
+    {
+      "epoch": 3.4342587742273443,
+      "grad_norm": 1.3395326137542725,
+      "learning_rate": 1.1964597349835515e-06,
+      "loss": 0.1006,
+      "step": 6556
+    },
+    {
+      "epoch": 3.4342587742273443,
+      "loss_partial": 0.044921875,
+      "step": 6556
+    },
+    {
+      "epoch": 3.4342587742273443,
+      "loss_partial": 0.045166015625,
+      "step": 6556
+    },
+    {
+      "epoch": 3.4347826086956523,
+      "grad_norm": 0.8742262125015259,
+      "learning_rate": 1.194292274078156e-06,
+      "loss": 0.0981,
+      "step": 6557
+    },
+    {
+      "epoch": 3.4347826086956523,
+      "loss_partial": 0.04736328125,
+      "step": 6557
+    },
+    {
+      "epoch": 3.4347826086956523,
+      "loss_partial": 0.05126953125,
+      "step": 6557
+    },
+    {
+      "epoch": 3.4353064431639604,
+      "grad_norm": 1.0171035528182983,
+      "learning_rate": 1.1921266535112253e-06,
+      "loss": 0.1009,
+      "step": 6558
+    },
+    {
+      "epoch": 3.4353064431639604,
+      "loss_partial": 0.0546875,
+      "step": 6558
+    },
+    {
+      "epoch": 3.4353064431639604,
+      "loss_partial": 0.052978515625,
+      "step": 6558
+    },
+    {
+      "epoch": 3.4358302776322684,
+      "grad_norm": 1.042495846748352,
+      "learning_rate": 1.1899628737353575e-06,
+      "loss": 0.0997,
+      "step": 6559
+    },
+    {
+      "epoch": 3.4358302776322684,
+      "loss_partial": 0.05419921875,
+      "step": 6559
+    },
+    {
+      "epoch": 3.4358302776322684,
+      "loss_partial": 0.045654296875,
+      "step": 6559
+    },
+    {
+      "epoch": 3.436354112100576,
+      "grad_norm": 1.006489634513855,
+      "learning_rate": 1.1878009352027687e-06,
+      "loss": 0.1,
+      "step": 6560
+    },
+    {
+      "epoch": 3.436354112100576,
+      "loss_partial": 0.0634765625,
+      "step": 6560
+    },
+    {
+      "epoch": 3.436354112100576,
+      "loss_partial": 0.04736328125,
+      "step": 6560
+    },
+    {
+      "epoch": 3.436877946568884,
+      "grad_norm": 0.929436981678009,
+      "learning_rate": 1.1856408383652906e-06,
+      "loss": 0.1004,
+      "step": 6561
+    },
+    {
+      "epoch": 3.436877946568884,
+      "loss_partial": 0.052001953125,
+      "step": 6561
+    },
+    {
+      "epoch": 3.436877946568884,
+      "loss_partial": 0.04150390625,
+      "step": 6561
+    },
+    {
+      "epoch": 3.437401781037192,
+      "grad_norm": 1.1107146739959717,
+      "learning_rate": 1.1834825836743725e-06,
+      "loss": 0.0957,
+      "step": 6562
+    },
+    {
+      "epoch": 3.437401781037192,
+      "loss_partial": 0.050048828125,
+      "step": 6562
+    },
+    {
+      "epoch": 3.437401781037192,
+      "loss_partial": 0.04638671875,
+      "step": 6562
+    },
+    {
+      "epoch": 3.4379256155055002,
+      "grad_norm": 1.0399123430252075,
+      "learning_rate": 1.1813261715810741e-06,
+      "loss": 0.1018,
+      "step": 6563
+    },
+    {
+      "epoch": 3.4379256155055002,
+      "loss_partial": 0.050048828125,
+      "step": 6563
+    },
+    {
+      "epoch": 3.4379256155055002,
+      "loss_partial": 0.058349609375,
+      "step": 6563
+    },
+    {
+      "epoch": 3.4384494499738083,
+      "grad_norm": 1.0290740728378296,
+      "learning_rate": 1.1791716025360734e-06,
+      "loss": 0.1007,
+      "step": 6564
+    },
+    {
+      "epoch": 3.4384494499738083,
+      "loss_partial": 0.0419921875,
+      "step": 6564
+    },
+    {
+      "epoch": 3.4384494499738083,
+      "loss_partial": 0.047119140625,
+      "step": 6564
+    },
+    {
+      "epoch": 3.4389732844421164,
+      "grad_norm": 0.9612303972244263,
+      "learning_rate": 1.1770188769896606e-06,
+      "loss": 0.1016,
+      "step": 6565
+    },
+    {
+      "epoch": 3.4389732844421164,
+      "loss_partial": 0.03955078125,
+      "step": 6565
+    },
+    {
+      "epoch": 3.4389732844421164,
+      "loss_partial": 0.052734375,
+      "step": 6565
+    },
+    {
+      "epoch": 3.4394971189104244,
+      "grad_norm": 0.8023644089698792,
+      "learning_rate": 1.174867995391744e-06,
+      "loss": 0.0984,
+      "step": 6566
+    },
+    {
+      "epoch": 3.4394971189104244,
+      "loss_partial": 0.05224609375,
+      "step": 6566
+    },
+    {
+      "epoch": 3.4394971189104244,
+      "loss_partial": 0.05322265625,
+      "step": 6566
+    },
+    {
+      "epoch": 3.4400209533787325,
+      "grad_norm": 1.3405057191848755,
+      "learning_rate": 1.172718958191843e-06,
+      "loss": 0.1019,
+      "step": 6567
+    },
+    {
+      "epoch": 3.4400209533787325,
+      "loss_partial": 0.052001953125,
+      "step": 6567
+    },
+    {
+      "epoch": 3.4400209533787325,
+      "loss_partial": 0.04736328125,
+      "step": 6567
+    },
+    {
+      "epoch": 3.44054478784704,
+      "grad_norm": 1.0509672164916992,
+      "learning_rate": 1.1705717658390925e-06,
+      "loss": 0.0957,
+      "step": 6568
+    },
+    {
+      "epoch": 3.44054478784704,
+      "loss_partial": 0.046875,
+      "step": 6568
+    },
+    {
+      "epoch": 3.44054478784704,
+      "loss_partial": 0.0498046875,
+      "step": 6568
+    },
+    {
+      "epoch": 3.441068622315348,
+      "grad_norm": 1.274815559387207,
+      "learning_rate": 1.1684264187822426e-06,
+      "loss": 0.1006,
+      "step": 6569
+    },
+    {
+      "epoch": 3.441068622315348,
+      "loss_partial": 0.04345703125,
+      "step": 6569
+    },
+    {
+      "epoch": 3.441068622315348,
+      "loss_partial": 0.048095703125,
+      "step": 6569
+    },
+    {
+      "epoch": 3.441592456783656,
+      "grad_norm": 0.7874521017074585,
+      "learning_rate": 1.16628291746966e-06,
+      "loss": 0.096,
+      "step": 6570
+    },
+    {
+      "epoch": 3.441592456783656,
+      "loss_partial": 0.051513671875,
+      "step": 6570
+    },
+    {
+      "epoch": 3.441592456783656,
+      "loss_partial": 0.0498046875,
+      "step": 6570
+    },
+    {
+      "epoch": 3.4421162912519643,
+      "grad_norm": 1.06488835811615,
+      "learning_rate": 1.164141262349322e-06,
+      "loss": 0.1014,
+      "step": 6571
+    },
+    {
+      "epoch": 3.4421162912519643,
+      "loss_partial": 0.0517578125,
+      "step": 6571
+    },
+    {
+      "epoch": 3.4421162912519643,
+      "loss_partial": 0.0634765625,
+      "step": 6571
+    },
+    {
+      "epoch": 3.4426401257202723,
+      "grad_norm": 1.3933322429656982,
+      "learning_rate": 1.1620014538688207e-06,
+      "loss": 0.0985,
+      "step": 6572
+    },
+    {
+      "epoch": 3.4426401257202723,
+      "loss_partial": 0.05517578125,
+      "step": 6572
+    },
+    {
+      "epoch": 3.4426401257202723,
+      "loss_partial": 0.057861328125,
+      "step": 6572
+    },
+    {
+      "epoch": 3.4431639601885804,
+      "grad_norm": 0.9471867680549622,
+      "learning_rate": 1.1598634924753604e-06,
+      "loss": 0.1035,
+      "step": 6573
+    },
+    {
+      "epoch": 3.4431639601885804,
+      "loss_partial": 0.0498046875,
+      "step": 6573
+    },
+    {
+      "epoch": 3.4431639601885804,
+      "loss_partial": 0.04541015625,
+      "step": 6573
+    },
+    {
+      "epoch": 3.4436877946568885,
+      "grad_norm": 0.8700320720672607,
+      "learning_rate": 1.1577273786157662e-06,
+      "loss": 0.0981,
+      "step": 6574
+    },
+    {
+      "epoch": 3.4436877946568885,
+      "loss_partial": 0.05419921875,
+      "step": 6574
+    },
+    {
+      "epoch": 3.4436877946568885,
+      "loss_partial": 0.038330078125,
+      "step": 6574
+    },
+    {
+      "epoch": 3.4442116291251965,
+      "grad_norm": 1.2129883766174316,
+      "learning_rate": 1.1555931127364716e-06,
+      "loss": 0.0994,
+      "step": 6575
+    },
+    {
+      "epoch": 3.4442116291251965,
+      "loss_partial": 0.050048828125,
+      "step": 6575
+    },
+    {
+      "epoch": 3.4442116291251965,
+      "loss_partial": 0.04150390625,
+      "step": 6575
+    },
+    {
+      "epoch": 3.4447354635935046,
+      "grad_norm": 0.9956658482551575,
+      "learning_rate": 1.1534606952835247e-06,
+      "loss": 0.0953,
+      "step": 6576
+    },
+    {
+      "epoch": 3.4447354635935046,
+      "loss_partial": 0.04443359375,
+      "step": 6576
+    },
+    {
+      "epoch": 3.4447354635935046,
+      "loss_partial": 0.047119140625,
+      "step": 6576
+    },
+    {
+      "epoch": 3.4452592980618126,
+      "grad_norm": 1.205733299255371,
+      "learning_rate": 1.1513301267025866e-06,
+      "loss": 0.0992,
+      "step": 6577
+    },
+    {
+      "epoch": 3.4452592980618126,
+      "loss_partial": 0.056884765625,
+      "step": 6577
+    },
+    {
+      "epoch": 3.4452592980618126,
+      "loss_partial": 0.041259765625,
+      "step": 6577
+    },
+    {
+      "epoch": 3.4457831325301207,
+      "grad_norm": 0.8515434861183167,
+      "learning_rate": 1.1492014074389334e-06,
+      "loss": 0.0992,
+      "step": 6578
+    },
+    {
+      "epoch": 3.4457831325301207,
+      "loss_partial": 0.057373046875,
+      "step": 6578
+    },
+    {
+      "epoch": 3.4457831325301207,
+      "loss_partial": 0.046630859375,
+      "step": 6578
+    },
+    {
+      "epoch": 3.4463069669984283,
+      "grad_norm": 0.9408016800880432,
+      "learning_rate": 1.1470745379374538e-06,
+      "loss": 0.0995,
+      "step": 6579
+    },
+    {
+      "epoch": 3.4463069669984283,
+      "loss_partial": 0.044189453125,
+      "step": 6579
+    },
+    {
+      "epoch": 3.4463069669984283,
+      "loss_partial": 0.048828125,
+      "step": 6579
+    },
+    {
+      "epoch": 3.4468308014667364,
+      "grad_norm": 1.2072062492370605,
+      "learning_rate": 1.1449495186426518e-06,
+      "loss": 0.0999,
+      "step": 6580
+    },
+    {
+      "epoch": 3.4468308014667364,
+      "loss_partial": 0.045654296875,
+      "step": 6580
+    },
+    {
+      "epoch": 3.4468308014667364,
+      "loss_partial": 0.04833984375,
+      "step": 6580
+    },
+    {
+      "epoch": 3.4473546359350444,
+      "grad_norm": 1.268202304840088,
+      "learning_rate": 1.1428263499986403e-06,
+      "loss": 0.1011,
+      "step": 6581
+    },
+    {
+      "epoch": 3.4473546359350444,
+      "loss_partial": 0.047607421875,
+      "step": 6581
+    },
+    {
+      "epoch": 3.4473546359350444,
+      "loss_partial": 0.05419921875,
+      "step": 6581
+    },
+    {
+      "epoch": 3.4478784704033525,
+      "grad_norm": 1.4356826543807983,
+      "learning_rate": 1.140705032449152e-06,
+      "loss": 0.0985,
+      "step": 6582
+    },
+    {
+      "epoch": 3.4478784704033525,
+      "loss_partial": 0.047607421875,
+      "step": 6582
+    },
+    {
+      "epoch": 3.4478784704033525,
+      "loss_partial": 0.052734375,
+      "step": 6582
+    },
+    {
+      "epoch": 3.4484023048716606,
+      "grad_norm": 0.9103869795799255,
+      "learning_rate": 1.1385855664375289e-06,
+      "loss": 0.0964,
+      "step": 6583
+    },
+    {
+      "epoch": 3.4484023048716606,
+      "loss_partial": 0.055419921875,
+      "step": 6583
+    },
+    {
+      "epoch": 3.4484023048716606,
+      "loss_partial": 0.04833984375,
+      "step": 6583
+    },
+    {
+      "epoch": 3.4489261393399686,
+      "grad_norm": 1.0404558181762695,
+      "learning_rate": 1.1364679524067235e-06,
+      "loss": 0.1029,
+      "step": 6584
+    },
+    {
+      "epoch": 3.4489261393399686,
+      "loss_partial": 0.05322265625,
+      "step": 6584
+    },
+    {
+      "epoch": 3.4489261393399686,
+      "loss_partial": 0.04541015625,
+      "step": 6584
+    },
+    {
+      "epoch": 3.4494499738082767,
+      "grad_norm": 1.1033997535705566,
+      "learning_rate": 1.1343521907993049e-06,
+      "loss": 0.0966,
+      "step": 6585
+    },
+    {
+      "epoch": 3.4494499738082767,
+      "loss_partial": 0.048828125,
+      "step": 6585
+    },
+    {
+      "epoch": 3.4494499738082767,
+      "loss_partial": 0.05224609375,
+      "step": 6585
+    },
+    {
+      "epoch": 3.4499738082765847,
+      "grad_norm": 0.9776666164398193,
+      "learning_rate": 1.1322382820574563e-06,
+      "loss": 0.0985,
+      "step": 6586
+    },
+    {
+      "epoch": 3.4499738082765847,
+      "loss_partial": 0.043701171875,
+      "step": 6586
+    },
+    {
+      "epoch": 3.4499738082765847,
+      "loss_partial": 0.052490234375,
+      "step": 6586
+    },
+    {
+      "epoch": 3.450497642744893,
+      "grad_norm": 0.9917707443237305,
+      "learning_rate": 1.1301262266229695e-06,
+      "loss": 0.0948,
+      "step": 6587
+    },
+    {
+      "epoch": 3.450497642744893,
+      "loss_partial": 0.04443359375,
+      "step": 6587
+    },
+    {
+      "epoch": 3.450497642744893,
+      "loss_partial": 0.049072265625,
+      "step": 6587
+    },
+    {
+      "epoch": 3.4510214772132004,
+      "grad_norm": 1.0949468612670898,
+      "learning_rate": 1.128016024937253e-06,
+      "loss": 0.0952,
+      "step": 6588
+    },
+    {
+      "epoch": 3.4510214772132004,
+      "loss_partial": 0.044189453125,
+      "step": 6588
+    },
+    {
+      "epoch": 3.4510214772132004,
+      "loss_partial": 0.050048828125,
+      "step": 6588
+    },
+    {
+      "epoch": 3.4515453116815085,
+      "grad_norm": 1.011942744255066,
+      "learning_rate": 1.1259076774413235e-06,
+      "loss": 0.0997,
+      "step": 6589
+    },
+    {
+      "epoch": 3.4515453116815085,
+      "loss_partial": 0.04443359375,
+      "step": 6589
+    },
+    {
+      "epoch": 3.4515453116815085,
+      "loss_partial": 0.06396484375,
+      "step": 6589
+    },
+    {
+      "epoch": 3.4520691461498165,
+      "grad_norm": 1.124596118927002,
+      "learning_rate": 1.1238011845758145e-06,
+      "loss": 0.1019,
+      "step": 6590
+    },
+    {
+      "epoch": 3.4520691461498165,
+      "loss_partial": 0.051513671875,
+      "step": 6590
+    },
+    {
+      "epoch": 3.4520691461498165,
+      "loss_partial": 0.04931640625,
+      "step": 6590
+    },
+    {
+      "epoch": 3.4525929806181246,
+      "grad_norm": 1.064733624458313,
+      "learning_rate": 1.1216965467809692e-06,
+      "loss": 0.0972,
+      "step": 6591
+    },
+    {
+      "epoch": 3.4525929806181246,
+      "loss_partial": 0.056640625,
+      "step": 6591
+    },
+    {
+      "epoch": 3.4525929806181246,
+      "loss_partial": 0.046630859375,
+      "step": 6591
+    },
+    {
+      "epoch": 3.4531168150864326,
+      "grad_norm": 1.0816247463226318,
+      "learning_rate": 1.1195937644966437e-06,
+      "loss": 0.1002,
+      "step": 6592
+    },
+    {
+      "epoch": 3.4531168150864326,
+      "loss_partial": 0.047607421875,
+      "step": 6592
+    },
+    {
+      "epoch": 3.4531168150864326,
+      "loss_partial": 0.048583984375,
+      "step": 6592
+    },
+    {
+      "epoch": 3.4536406495547407,
+      "grad_norm": 1.2840672731399536,
+      "learning_rate": 1.117492838162304e-06,
+      "loss": 0.0998,
+      "step": 6593
+    },
+    {
+      "epoch": 3.4536406495547407,
+      "loss_partial": 0.050537109375,
+      "step": 6593
+    },
+    {
+      "epoch": 3.4536406495547407,
+      "loss_partial": 0.05078125,
+      "step": 6593
+    },
+    {
+      "epoch": 3.4541644840230488,
+      "grad_norm": 1.0434184074401855,
+      "learning_rate": 1.1153937682170357e-06,
+      "loss": 0.0956,
+      "step": 6594
+    },
+    {
+      "epoch": 3.4541644840230488,
+      "loss_partial": 0.055419921875,
+      "step": 6594
+    },
+    {
+      "epoch": 3.4541644840230488,
+      "loss_partial": 0.051025390625,
+      "step": 6594
+    },
+    {
+      "epoch": 3.454688318491357,
+      "grad_norm": 1.4457265138626099,
+      "learning_rate": 1.11329655509953e-06,
+      "loss": 0.1036,
+      "step": 6595
+    },
+    {
+      "epoch": 3.454688318491357,
+      "loss_partial": 0.043701171875,
+      "step": 6595
+    },
+    {
+      "epoch": 3.454688318491357,
+      "loss_partial": 0.051025390625,
+      "step": 6595
+    },
+    {
+      "epoch": 3.455212152959665,
+      "grad_norm": 1.04927396774292,
+      "learning_rate": 1.1112011992480899e-06,
+      "loss": 0.101,
+      "step": 6596
+    },
+    {
+      "epoch": 3.455212152959665,
+      "loss_partial": 0.04736328125,
+      "step": 6596
+    },
+    {
+      "epoch": 3.455212152959665,
+      "loss_partial": 0.046630859375,
+      "step": 6596
+    },
+    {
+      "epoch": 3.455735987427973,
+      "grad_norm": 0.9737049341201782,
+      "learning_rate": 1.1091077011006302e-06,
+      "loss": 0.1014,
+      "step": 6597
+    },
+    {
+      "epoch": 3.455735987427973,
+      "loss_partial": 0.055419921875,
+      "step": 6597
+    },
+    {
+      "epoch": 3.455735987427973,
+      "loss_partial": 0.048828125,
+      "step": 6597
+    },
+    {
+      "epoch": 3.456259821896281,
+      "grad_norm": 1.1803101301193237,
+      "learning_rate": 1.1070160610946833e-06,
+      "loss": 0.0983,
+      "step": 6598
+    },
+    {
+      "epoch": 3.456259821896281,
+      "loss_partial": 0.05615234375,
+      "step": 6598
+    },
+    {
+      "epoch": 3.456259821896281,
+      "loss_partial": 0.04736328125,
+      "step": 6598
+    },
+    {
+      "epoch": 3.4567836563645886,
+      "grad_norm": 1.2390943765640259,
+      "learning_rate": 1.1049262796673865e-06,
+      "loss": 0.1006,
+      "step": 6599
+    },
+    {
+      "epoch": 3.4567836563645886,
+      "loss_partial": 0.0478515625,
+      "step": 6599
+    },
+    {
+      "epoch": 3.4567836563645886,
+      "loss_partial": 0.04052734375,
+      "step": 6599
+    },
+    {
+      "epoch": 3.4573074908328967,
+      "grad_norm": 0.9002989530563354,
+      "learning_rate": 1.1028383572554935e-06,
+      "loss": 0.099,
+      "step": 6600
+    },
+    {
+      "epoch": 3.4573074908328967,
+      "loss_partial": 0.053955078125,
+      "step": 6600
+    },
+    {
+      "epoch": 3.4573074908328967,
+      "loss_partial": 0.05224609375,
+      "step": 6600
+    },
+    {
+      "epoch": 3.4578313253012047,
+      "grad_norm": 1.315559983253479,
+      "learning_rate": 1.100752294295364e-06,
+      "loss": 0.0991,
+      "step": 6601
+    },
+    {
+      "epoch": 3.4578313253012047,
+      "loss_partial": 0.048828125,
+      "step": 6601
+    },
+    {
+      "epoch": 3.4578313253012047,
+      "loss_partial": 0.038818359375,
+      "step": 6601
+    },
+    {
+      "epoch": 3.458355159769513,
+      "grad_norm": 1.2620688676834106,
+      "learning_rate": 1.0986680912229742e-06,
+      "loss": 0.0956,
+      "step": 6602
+    },
+    {
+      "epoch": 3.458355159769513,
+      "loss_partial": 0.058837890625,
+      "step": 6602
+    },
+    {
+      "epoch": 3.458355159769513,
+      "loss_partial": 0.04736328125,
+      "step": 6602
+    },
+    {
+      "epoch": 3.458878994237821,
+      "grad_norm": 1.2865017652511597,
+      "learning_rate": 1.0965857484739096e-06,
+      "loss": 0.0987,
+      "step": 6603
+    },
+    {
+      "epoch": 3.458878994237821,
+      "loss_partial": 0.049560546875,
+      "step": 6603
+    },
+    {
+      "epoch": 3.458878994237821,
+      "loss_partial": 0.05712890625,
+      "step": 6603
+    },
+    {
+      "epoch": 3.459402828706129,
+      "grad_norm": 1.0926982164382935,
+      "learning_rate": 1.0945052664833656e-06,
+      "loss": 0.101,
+      "step": 6604
+    },
+    {
+      "epoch": 3.459402828706129,
+      "loss_partial": 0.055908203125,
+      "step": 6604
+    },
+    {
+      "epoch": 3.459402828706129,
+      "loss_partial": 0.046142578125,
+      "step": 6604
+    },
+    {
+      "epoch": 3.459926663174437,
+      "grad_norm": 1.124793291091919,
+      "learning_rate": 1.09242664568615e-06,
+      "loss": 0.101,
+      "step": 6605
+    },
+    {
+      "epoch": 3.459926663174437,
+      "loss_partial": 0.047119140625,
+      "step": 6605
+    },
+    {
+      "epoch": 3.459926663174437,
+      "loss_partial": 0.049560546875,
+      "step": 6605
+    },
+    {
+      "epoch": 3.460450497642745,
+      "grad_norm": 1.112809419631958,
+      "learning_rate": 1.090349886516684e-06,
+      "loss": 0.0972,
+      "step": 6606
+    },
+    {
+      "epoch": 3.460450497642745,
+      "loss_partial": 0.055908203125,
+      "step": 6606
+    },
+    {
+      "epoch": 3.460450497642745,
+      "loss_partial": 0.05810546875,
+      "step": 6606
+    },
+    {
+      "epoch": 3.460974332111053,
+      "grad_norm": 1.1120661497116089,
+      "learning_rate": 1.088274989408996e-06,
+      "loss": 0.1014,
+      "step": 6607
+    },
+    {
+      "epoch": 3.460974332111053,
+      "loss_partial": 0.0419921875,
+      "step": 6607
+    },
+    {
+      "epoch": 3.460974332111053,
+      "loss_partial": 0.044677734375,
+      "step": 6607
+    },
+    {
+      "epoch": 3.4614981665793607,
+      "grad_norm": 1.3310997486114502,
+      "learning_rate": 1.0862019547967261e-06,
+      "loss": 0.0999,
+      "step": 6608
+    },
+    {
+      "epoch": 3.4614981665793607,
+      "loss_partial": 0.044677734375,
+      "step": 6608
+    },
+    {
+      "epoch": 3.4614981665793607,
+      "loss_partial": 0.06103515625,
+      "step": 6608
+    },
+    {
+      "epoch": 3.4620220010476688,
+      "grad_norm": 0.983698844909668,
+      "learning_rate": 1.0841307831131275e-06,
+      "loss": 0.1001,
+      "step": 6609
+    },
+    {
+      "epoch": 3.4620220010476688,
+      "loss_partial": 0.048828125,
+      "step": 6609
+    },
+    {
+      "epoch": 3.4620220010476688,
+      "loss_partial": 0.0634765625,
+      "step": 6609
+    },
+    {
+      "epoch": 3.462545835515977,
+      "grad_norm": 0.8897766470909119,
+      "learning_rate": 1.0820614747910595e-06,
+      "loss": 0.0985,
+      "step": 6610
+    },
+    {
+      "epoch": 3.462545835515977,
+      "loss_partial": 0.0400390625,
+      "step": 6610
+    },
+    {
+      "epoch": 3.462545835515977,
+      "loss_partial": 0.047607421875,
+      "step": 6610
+    },
+    {
+      "epoch": 3.463069669984285,
+      "grad_norm": 1.106707215309143,
+      "learning_rate": 1.0799940302629974e-06,
+      "loss": 0.1023,
+      "step": 6611
+    },
+    {
+      "epoch": 3.463069669984285,
+      "loss_partial": 0.052001953125,
+      "step": 6611
+    },
+    {
+      "epoch": 3.463069669984285,
+      "loss_partial": 0.04833984375,
+      "step": 6611
+    },
+    {
+      "epoch": 3.463593504452593,
+      "grad_norm": 1.2758920192718506,
+      "learning_rate": 1.0779284499610244e-06,
+      "loss": 0.0985,
+      "step": 6612
+    },
+    {
+      "epoch": 3.463593504452593,
+      "loss_partial": 0.048828125,
+      "step": 6612
+    },
+    {
+      "epoch": 3.463593504452593,
+      "loss_partial": 0.056884765625,
+      "step": 6612
+    },
+    {
+      "epoch": 3.464117338920901,
+      "grad_norm": 1.3401157855987549,
+      "learning_rate": 1.0758647343168315e-06,
+      "loss": 0.1006,
+      "step": 6613
+    },
+    {
+      "epoch": 3.464117338920901,
+      "loss_partial": 0.051513671875,
+      "step": 6613
+    },
+    {
+      "epoch": 3.464117338920901,
+      "loss_partial": 0.056640625,
+      "step": 6613
+    },
+    {
+      "epoch": 3.464641173389209,
+      "grad_norm": 1.119368314743042,
+      "learning_rate": 1.0738028837617242e-06,
+      "loss": 0.1003,
+      "step": 6614
+    },
+    {
+      "epoch": 3.464641173389209,
+      "loss_partial": 0.051025390625,
+      "step": 6614
+    },
+    {
+      "epoch": 3.464641173389209,
+      "loss_partial": 0.0478515625,
+      "step": 6614
+    },
+    {
+      "epoch": 3.465165007857517,
+      "grad_norm": 1.589805006980896,
+      "learning_rate": 1.0717428987266177e-06,
+      "loss": 0.1046,
+      "step": 6615
+    },
+    {
+      "epoch": 3.465165007857517,
+      "loss_partial": 0.04248046875,
+      "step": 6615
+    },
+    {
+      "epoch": 3.465165007857517,
+      "loss_partial": 0.050048828125,
+      "step": 6615
+    },
+    {
+      "epoch": 3.465688842325825,
+      "grad_norm": 1.120630145072937,
+      "learning_rate": 1.0696847796420328e-06,
+      "loss": 0.0992,
+      "step": 6616
+    },
+    {
+      "epoch": 3.465688842325825,
+      "loss_partial": 0.054443359375,
+      "step": 6616
+    },
+    {
+      "epoch": 3.465688842325825,
+      "loss_partial": 0.038818359375,
+      "step": 6616
+    },
+    {
+      "epoch": 3.4662126767941333,
+      "grad_norm": 1.0246065855026245,
+      "learning_rate": 1.067628526938107e-06,
+      "loss": 0.0995,
+      "step": 6617
+    },
+    {
+      "epoch": 3.4662126767941333,
+      "loss_partial": 0.054443359375,
+      "step": 6617
+    },
+    {
+      "epoch": 3.4662126767941333,
+      "loss_partial": 0.04443359375,
+      "step": 6617
+    },
+    {
+      "epoch": 3.4667365112624413,
+      "grad_norm": 1.0968207120895386,
+      "learning_rate": 1.0655741410445807e-06,
+      "loss": 0.0948,
+      "step": 6618
+    },
+    {
+      "epoch": 3.4667365112624413,
+      "loss_partial": 0.044189453125,
+      "step": 6618
+    },
+    {
+      "epoch": 3.4667365112624413,
+      "loss_partial": 0.046875,
+      "step": 6618
+    },
+    {
+      "epoch": 3.467260345730749,
+      "grad_norm": 1.0838311910629272,
+      "learning_rate": 1.0635216223908118e-06,
+      "loss": 0.0967,
+      "step": 6619
+    },
+    {
+      "epoch": 3.467260345730749,
+      "loss_partial": 0.0498046875,
+      "step": 6619
+    },
+    {
+      "epoch": 3.467260345730749,
+      "loss_partial": 0.045654296875,
+      "step": 6619
+    },
+    {
+      "epoch": 3.467784180199057,
+      "grad_norm": 1.1535494327545166,
+      "learning_rate": 1.0614709714057624e-06,
+      "loss": 0.1021,
+      "step": 6620
+    },
+    {
+      "epoch": 3.467784180199057,
+      "loss_partial": 0.054931640625,
+      "step": 6620
+    },
+    {
+      "epoch": 3.467784180199057,
+      "loss_partial": 0.040283203125,
+      "step": 6620
+    },
+    {
+      "epoch": 3.468308014667365,
+      "grad_norm": 0.9549482464790344,
+      "learning_rate": 1.0594221885180055e-06,
+      "loss": 0.1009,
+      "step": 6621
+    },
+    {
+      "epoch": 3.468308014667365,
+      "loss_partial": 0.04541015625,
+      "step": 6621
+    },
+    {
+      "epoch": 3.468308014667365,
+      "loss_partial": 0.06005859375,
+      "step": 6621
+    },
+    {
+      "epoch": 3.468831849135673,
+      "grad_norm": 1.1515344381332397,
+      "learning_rate": 1.0573752741557218e-06,
+      "loss": 0.1021,
+      "step": 6622
+    },
+    {
+      "epoch": 3.468831849135673,
+      "loss_partial": 0.060546875,
+      "step": 6622
+    },
+    {
+      "epoch": 3.468831849135673,
+      "loss_partial": 0.053955078125,
+      "step": 6622
+    },
+    {
+      "epoch": 3.469355683603981,
+      "grad_norm": 1.004005789756775,
+      "learning_rate": 1.0553302287467094e-06,
+      "loss": 0.1,
+      "step": 6623
+    },
+    {
+      "epoch": 3.469355683603981,
+      "loss_partial": 0.0654296875,
+      "step": 6623
+    },
+    {
+      "epoch": 3.469355683603981,
+      "loss_partial": 0.04736328125,
+      "step": 6623
+    },
+    {
+      "epoch": 3.4698795180722892,
+      "grad_norm": 1.1401299238204956,
+      "learning_rate": 1.053287052718367e-06,
+      "loss": 0.1,
+      "step": 6624
+    },
+    {
+      "epoch": 3.4698795180722892,
+      "loss_partial": 0.05126953125,
+      "step": 6624
+    },
+    {
+      "epoch": 3.4698795180722892,
+      "loss_partial": 0.048828125,
+      "step": 6624
+    },
+    {
+      "epoch": 3.4704033525405973,
+      "grad_norm": 1.1301445960998535,
+      "learning_rate": 1.0512457464977055e-06,
+      "loss": 0.0984,
+      "step": 6625
+    },
+    {
+      "epoch": 3.4704033525405973,
+      "loss_partial": 0.044189453125,
+      "step": 6625
+    },
+    {
+      "epoch": 3.4704033525405973,
+      "loss_partial": 0.046630859375,
+      "step": 6625
+    },
+    {
+      "epoch": 3.4709271870089053,
+      "grad_norm": 0.8889147043228149,
+      "learning_rate": 1.0492063105113448e-06,
+      "loss": 0.1,
+      "step": 6626
+    },
+    {
+      "epoch": 3.4709271870089053,
+      "loss_partial": 0.045654296875,
+      "step": 6626
+    },
+    {
+      "epoch": 3.4709271870089053,
+      "loss_partial": 0.054443359375,
+      "step": 6626
+    },
+    {
+      "epoch": 3.471451021477213,
+      "grad_norm": 0.9019566178321838,
+      "learning_rate": 1.0471687451855162e-06,
+      "loss": 0.0966,
+      "step": 6627
+    },
+    {
+      "epoch": 3.471451021477213,
+      "loss_partial": 0.05126953125,
+      "step": 6627
+    },
+    {
+      "epoch": 3.471451021477213,
+      "loss_partial": 0.048828125,
+      "step": 6627
+    },
+    {
+      "epoch": 3.471974855945521,
+      "grad_norm": 1.4343818426132202,
+      "learning_rate": 1.0451330509460566e-06,
+      "loss": 0.1002,
+      "step": 6628
+    },
+    {
+      "epoch": 3.471974855945521,
+      "loss_partial": 0.04443359375,
+      "step": 6628
+    },
+    {
+      "epoch": 3.471974855945521,
+      "loss_partial": 0.0419921875,
+      "step": 6628
+    },
+    {
+      "epoch": 3.472498690413829,
+      "grad_norm": 0.9309033751487732,
+      "learning_rate": 1.0430992282184148e-06,
+      "loss": 0.0981,
+      "step": 6629
+    },
+    {
+      "epoch": 3.472498690413829,
+      "loss_partial": 0.04345703125,
+      "step": 6629
+    },
+    {
+      "epoch": 3.472498690413829,
+      "loss_partial": 0.055419921875,
+      "step": 6629
+    },
+    {
+      "epoch": 3.473022524882137,
+      "grad_norm": 1.025378942489624,
+      "learning_rate": 1.041067277427643e-06,
+      "loss": 0.099,
+      "step": 6630
+    },
+    {
+      "epoch": 3.473022524882137,
+      "loss_partial": 0.051025390625,
+      "step": 6630
+    },
+    {
+      "epoch": 3.473022524882137,
+      "loss_partial": 0.0439453125,
+      "step": 6630
+    },
+    {
+      "epoch": 3.473546359350445,
+      "grad_norm": 1.052662968635559,
+      "learning_rate": 1.0390371989984138e-06,
+      "loss": 0.1021,
+      "step": 6631
+    },
+    {
+      "epoch": 3.473546359350445,
+      "loss_partial": 0.042724609375,
+      "step": 6631
+    },
+    {
+      "epoch": 3.473546359350445,
+      "loss_partial": 0.04931640625,
+      "step": 6631
+    },
+    {
+      "epoch": 3.4740701938187533,
+      "grad_norm": 1.3069517612457275,
+      "learning_rate": 1.0370089933549954e-06,
+      "loss": 0.1008,
+      "step": 6632
+    },
+    {
+      "epoch": 3.4740701938187533,
+      "loss_partial": 0.041748046875,
+      "step": 6632
+    },
+    {
+      "epoch": 3.4740701938187533,
+      "loss_partial": 0.0478515625,
+      "step": 6632
+    },
+    {
+      "epoch": 3.4745940282870613,
+      "grad_norm": 0.9702107906341553,
+      "learning_rate": 1.0349826609212733e-06,
+      "loss": 0.0977,
+      "step": 6633
+    },
+    {
+      "epoch": 3.4745940282870613,
+      "loss_partial": 0.046142578125,
+      "step": 6633
+    },
+    {
+      "epoch": 3.4745940282870613,
+      "loss_partial": 0.056884765625,
+      "step": 6633
+    },
+    {
+      "epoch": 3.4751178627553694,
+      "grad_norm": 1.1668428182601929,
+      "learning_rate": 1.032958202120734e-06,
+      "loss": 0.1045,
+      "step": 6634
+    },
+    {
+      "epoch": 3.4751178627553694,
+      "loss_partial": 0.0537109375,
+      "step": 6634
+    },
+    {
+      "epoch": 3.4751178627553694,
+      "loss_partial": 0.05224609375,
+      "step": 6634
+    },
+    {
+      "epoch": 3.4756416972236774,
+      "grad_norm": 1.2465479373931885,
+      "learning_rate": 1.030935617376483e-06,
+      "loss": 0.1023,
+      "step": 6635
+    },
+    {
+      "epoch": 3.4756416972236774,
+      "loss_partial": 0.046875,
+      "step": 6635
+    },
+    {
+      "epoch": 3.4756416972236774,
+      "loss_partial": 0.05224609375,
+      "step": 6635
+    },
+    {
+      "epoch": 3.4761655316919855,
+      "grad_norm": 1.1604554653167725,
+      "learning_rate": 1.0289149071112237e-06,
+      "loss": 0.0952,
+      "step": 6636
+    },
+    {
+      "epoch": 3.4761655316919855,
+      "loss_partial": 0.0498046875,
+      "step": 6636
+    },
+    {
+      "epoch": 3.4761655316919855,
+      "loss_partial": 0.042724609375,
+      "step": 6636
+    },
+    {
+      "epoch": 3.4766893661602936,
+      "grad_norm": 1.0920451879501343,
+      "learning_rate": 1.0268960717472742e-06,
+      "loss": 0.1011,
+      "step": 6637
+    },
+    {
+      "epoch": 3.4766893661602936,
+      "loss_partial": 0.042724609375,
+      "step": 6637
+    },
+    {
+      "epoch": 3.4766893661602936,
+      "loss_partial": 0.0498046875,
+      "step": 6637
+    },
+    {
+      "epoch": 3.477213200628601,
+      "grad_norm": 1.1620676517486572,
+      "learning_rate": 1.024879111706557e-06,
+      "loss": 0.1023,
+      "step": 6638
+    },
+    {
+      "epoch": 3.477213200628601,
+      "loss_partial": 0.04443359375,
+      "step": 6638
+    },
+    {
+      "epoch": 3.477213200628601,
+      "loss_partial": 0.05029296875,
+      "step": 6638
+    },
+    {
+      "epoch": 3.4777370350969092,
+      "grad_norm": 1.1683176755905151,
+      "learning_rate": 1.022864027410606e-06,
+      "loss": 0.0992,
+      "step": 6639
+    },
+    {
+      "epoch": 3.4777370350969092,
+      "loss_partial": 0.050537109375,
+      "step": 6639
+    },
+    {
+      "epoch": 3.4777370350969092,
+      "loss_partial": 0.0517578125,
+      "step": 6639
+    },
+    {
+      "epoch": 3.4782608695652173,
+      "grad_norm": 1.0421253442764282,
+      "learning_rate": 1.02085081928056e-06,
+      "loss": 0.1037,
+      "step": 6640
+    },
+    {
+      "epoch": 3.4782608695652173,
+      "loss_partial": 0.045166015625,
+      "step": 6640
+    },
+    {
+      "epoch": 3.4782608695652173,
+      "loss_partial": 0.052490234375,
+      "step": 6640
+    },
+    {
+      "epoch": 3.4787847040335254,
+      "grad_norm": 1.029639720916748,
+      "learning_rate": 1.0188394877371665e-06,
+      "loss": 0.0972,
+      "step": 6641
+    },
+    {
+      "epoch": 3.4787847040335254,
+      "loss_partial": 0.0439453125,
+      "step": 6641
+    },
+    {
+      "epoch": 3.4787847040335254,
+      "loss_partial": 0.052978515625,
+      "step": 6641
+    },
+    {
+      "epoch": 3.4793085385018334,
+      "grad_norm": 0.9884294867515564,
+      "learning_rate": 1.0168300332007808e-06,
+      "loss": 0.0988,
+      "step": 6642
+    },
+    {
+      "epoch": 3.4793085385018334,
+      "loss_partial": 0.050048828125,
+      "step": 6642
+    },
+    {
+      "epoch": 3.4793085385018334,
+      "loss_partial": 0.050048828125,
+      "step": 6642
+    },
+    {
+      "epoch": 3.4798323729701415,
+      "grad_norm": 0.9297333359718323,
+      "learning_rate": 1.01482245609137e-06,
+      "loss": 0.103,
+      "step": 6643
+    },
+    {
+      "epoch": 3.4798323729701415,
+      "loss_partial": 0.046875,
+      "step": 6643
+    },
+    {
+      "epoch": 3.4798323729701415,
+      "loss_partial": 0.05517578125,
+      "step": 6643
+    },
+    {
+      "epoch": 3.4803562074384495,
+      "grad_norm": 1.035867691040039,
+      "learning_rate": 1.012816756828503e-06,
+      "loss": 0.1,
+      "step": 6644
+    },
+    {
+      "epoch": 3.4803562074384495,
+      "loss_partial": 0.06005859375,
+      "step": 6644
+    },
+    {
+      "epoch": 3.4803562074384495,
+      "loss_partial": 0.046875,
+      "step": 6644
+    },
+    {
+      "epoch": 3.4808800419067576,
+      "grad_norm": 0.946270227432251,
+      "learning_rate": 1.010812935831359e-06,
+      "loss": 0.1012,
+      "step": 6645
+    },
+    {
+      "epoch": 3.4808800419067576,
+      "loss_partial": 0.055908203125,
+      "step": 6645
+    },
+    {
+      "epoch": 3.4808800419067576,
+      "loss_partial": 0.05029296875,
+      "step": 6645
+    },
+    {
+      "epoch": 3.4814038763750657,
+      "grad_norm": 1.32950758934021,
+      "learning_rate": 1.0088109935187219e-06,
+      "loss": 0.1032,
+      "step": 6646
+    },
+    {
+      "epoch": 3.4814038763750657,
+      "loss_partial": 0.0546875,
+      "step": 6646
+    },
+    {
+      "epoch": 3.4814038763750657,
+      "loss_partial": 0.047607421875,
+      "step": 6646
+    },
+    {
+      "epoch": 3.4819277108433733,
+      "grad_norm": 0.8574466705322266,
+      "learning_rate": 1.00681093030899e-06,
+      "loss": 0.0957,
+      "step": 6647
+    },
+    {
+      "epoch": 3.4819277108433733,
+      "loss_partial": 0.046142578125,
+      "step": 6647
+    },
+    {
+      "epoch": 3.4819277108433733,
+      "loss_partial": 0.04443359375,
+      "step": 6647
+    },
+    {
+      "epoch": 3.4824515453116813,
+      "grad_norm": 1.2444380521774292,
+      "learning_rate": 1.0048127466201606e-06,
+      "loss": 0.0982,
+      "step": 6648
+    },
+    {
+      "epoch": 3.4824515453116813,
+      "loss_partial": 0.040283203125,
+      "step": 6648
+    },
+    {
+      "epoch": 3.4824515453116813,
+      "loss_partial": 0.05322265625,
+      "step": 6648
+    },
+    {
+      "epoch": 3.4829753797799894,
+      "grad_norm": 1.0634551048278809,
+      "learning_rate": 1.0028164428698428e-06,
+      "loss": 0.0999,
+      "step": 6649
+    },
+    {
+      "epoch": 3.4829753797799894,
+      "loss_partial": 0.048828125,
+      "step": 6649
+    },
+    {
+      "epoch": 3.4829753797799894,
+      "loss_partial": 0.046630859375,
+      "step": 6649
+    },
+    {
+      "epoch": 3.4834992142482974,
+      "grad_norm": 1.4316964149475098,
+      "learning_rate": 1.0008220194752515e-06,
+      "loss": 0.0984,
+      "step": 6650
+    },
+    {
+      "epoch": 3.4834992142482974,
+      "loss_partial": 0.05224609375,
+      "step": 6650
+    },
+    {
+      "epoch": 3.4834992142482974,
+      "loss_partial": 0.04541015625,
+      "step": 6650
+    },
+    {
+      "epoch": 3.4840230487166055,
+      "grad_norm": 1.0910824537277222,
+      "learning_rate": 9.98829476853208e-07,
+      "loss": 0.0962,
+      "step": 6651
+    },
+    {
+      "epoch": 3.4840230487166055,
+      "loss_partial": 0.0458984375,
+      "step": 6651
+    },
+    {
+      "epoch": 3.4840230487166055,
+      "loss_partial": 0.04248046875,
+      "step": 6651
+    },
+    {
+      "epoch": 3.4845468831849136,
+      "grad_norm": 1.0488595962524414,
+      "learning_rate": 9.968388154201436e-07,
+      "loss": 0.0995,
+      "step": 6652
+    },
+    {
+      "epoch": 3.4845468831849136,
+      "loss_partial": 0.049072265625,
+      "step": 6652
+    },
+    {
+      "epoch": 3.4845468831849136,
+      "loss_partial": 0.05712890625,
+      "step": 6652
+    },
+    {
+      "epoch": 3.4850707176532216,
+      "grad_norm": 1.1208703517913818,
+      "learning_rate": 9.94850035592092e-07,
+      "loss": 0.0948,
+      "step": 6653
+    },
+    {
+      "epoch": 3.4850707176532216,
+      "loss_partial": 0.048583984375,
+      "step": 6653
+    },
+    {
+      "epoch": 3.4850707176532216,
+      "loss_partial": 0.044189453125,
+      "step": 6653
+    },
+    {
+      "epoch": 3.4855945521215297,
+      "grad_norm": 1.1810896396636963,
+      "learning_rate": 9.928631377846954e-07,
+      "loss": 0.0993,
+      "step": 6654
+    },
+    {
+      "epoch": 3.4855945521215297,
+      "loss_partial": 0.051025390625,
+      "step": 6654
+    },
+    {
+      "epoch": 3.4855945521215297,
+      "loss_partial": 0.04345703125,
+      "step": 6654
+    },
+    {
+      "epoch": 3.4861183865898377,
+      "grad_norm": 1.0347825288772583,
+      "learning_rate": 9.90878122413207e-07,
+      "loss": 0.101,
+      "step": 6655
+    },
+    {
+      "epoch": 3.4861183865898377,
+      "loss_partial": 0.04248046875,
+      "step": 6655
+    },
+    {
+      "epoch": 3.4861183865898377,
+      "loss_partial": 0.05517578125,
+      "step": 6655
+    },
+    {
+      "epoch": 3.486642221058146,
+      "grad_norm": 0.8001394867897034,
+      "learning_rate": 9.888949898924793e-07,
+      "loss": 0.1011,
+      "step": 6656
+    },
+    {
+      "epoch": 3.486642221058146,
+      "loss_partial": 0.06103515625,
+      "step": 6656
+    },
+    {
+      "epoch": 3.486642221058146,
+      "loss_partial": 0.04052734375,
+      "step": 6656
+    },
+    {
+      "epoch": 3.487166055526454,
+      "grad_norm": 1.1219724416732788,
+      "learning_rate": 9.869137406369767e-07,
+      "loss": 0.1003,
+      "step": 6657
+    },
+    {
+      "epoch": 3.487166055526454,
+      "loss_partial": 0.053955078125,
+      "step": 6657
+    },
+    {
+      "epoch": 3.487166055526454,
+      "loss_partial": 0.05126953125,
+      "step": 6657
+    },
+    {
+      "epoch": 3.4876898899947615,
+      "grad_norm": 1.1290533542633057,
+      "learning_rate": 9.849343750607642e-07,
+      "loss": 0.1008,
+      "step": 6658
+    },
+    {
+      "epoch": 3.4876898899947615,
+      "loss_partial": 0.04443359375,
+      "step": 6658
+    },
+    {
+      "epoch": 3.4876898899947615,
+      "loss_partial": 0.057861328125,
+      "step": 6658
+    },
+    {
+      "epoch": 3.4882137244630695,
+      "grad_norm": 1.5633844137191772,
+      "learning_rate": 9.829568935775214e-07,
+      "loss": 0.1013,
+      "step": 6659
+    },
+    {
+      "epoch": 3.4882137244630695,
+      "loss_partial": 0.051513671875,
+      "step": 6659
+    },
+    {
+      "epoch": 3.4882137244630695,
+      "loss_partial": 0.056396484375,
+      "step": 6659
+    },
+    {
+      "epoch": 3.4887375589313776,
+      "grad_norm": 0.9674063920974731,
+      "learning_rate": 9.809812966005293e-07,
+      "loss": 0.1026,
+      "step": 6660
+    },
+    {
+      "epoch": 3.4887375589313776,
+      "loss_partial": 0.055908203125,
+      "step": 6660
+    },
+    {
+      "epoch": 3.4887375589313776,
+      "loss_partial": 0.052978515625,
+      "step": 6660
+    },
+    {
+      "epoch": 3.4892613933996857,
+      "grad_norm": 1.0399010181427002,
+      "learning_rate": 9.79007584542674e-07,
+      "loss": 0.1022,
+      "step": 6661
+    },
+    {
+      "epoch": 3.4892613933996857,
+      "loss_partial": 0.047607421875,
+      "step": 6661
+    },
+    {
+      "epoch": 3.4892613933996857,
+      "loss_partial": 0.045654296875,
+      "step": 6661
+    },
+    {
+      "epoch": 3.4897852278679937,
+      "grad_norm": 1.3189581632614136,
+      "learning_rate": 9.77035757816449e-07,
+      "loss": 0.1022,
+      "step": 6662
+    },
+    {
+      "epoch": 3.4897852278679937,
+      "loss_partial": 0.05322265625,
+      "step": 6662
+    },
+    {
+      "epoch": 3.4897852278679937,
+      "loss_partial": 0.0556640625,
+      "step": 6662
+    },
+    {
+      "epoch": 3.490309062336302,
+      "grad_norm": 1.1555781364440918,
+      "learning_rate": 9.750658168339544e-07,
+      "loss": 0.1011,
+      "step": 6663
+    },
+    {
+      "epoch": 3.490309062336302,
+      "loss_partial": 0.0439453125,
+      "step": 6663
+    },
+    {
+      "epoch": 3.490309062336302,
+      "loss_partial": 0.049072265625,
+      "step": 6663
+    },
+    {
+      "epoch": 3.49083289680461,
+      "grad_norm": 1.0958306789398193,
+      "learning_rate": 9.730977620068948e-07,
+      "loss": 0.0984,
+      "step": 6664
+    },
+    {
+      "epoch": 3.49083289680461,
+      "loss_partial": 0.0458984375,
+      "step": 6664
+    },
+    {
+      "epoch": 3.49083289680461,
+      "loss_partial": 0.0517578125,
+      "step": 6664
+    },
+    {
+      "epoch": 3.491356731272918,
+      "grad_norm": 1.1003119945526123,
+      "learning_rate": 9.711315937465838e-07,
+      "loss": 0.1001,
+      "step": 6665
+    },
+    {
+      "epoch": 3.491356731272918,
+      "loss_partial": 0.0634765625,
+      "step": 6665
+    },
+    {
+      "epoch": 3.491356731272918,
+      "loss_partial": 0.0478515625,
+      "step": 6665
+    },
+    {
+      "epoch": 3.491880565741226,
+      "grad_norm": 0.9753367304801941,
+      "learning_rate": 9.691673124639356e-07,
+      "loss": 0.1011,
+      "step": 6666
+    },
+    {
+      "epoch": 3.491880565741226,
+      "loss_partial": 0.044921875,
+      "step": 6666
+    },
+    {
+      "epoch": 3.491880565741226,
+      "loss_partial": 0.0595703125,
+      "step": 6666
+    },
+    {
+      "epoch": 3.4924044002095336,
+      "grad_norm": 1.153364658355713,
+      "learning_rate": 9.672049185694733e-07,
+      "loss": 0.1022,
+      "step": 6667
+    },
+    {
+      "epoch": 3.4924044002095336,
+      "loss_partial": 0.04833984375,
+      "step": 6667
+    },
+    {
+      "epoch": 3.4924044002095336,
+      "loss_partial": 0.051025390625,
+      "step": 6667
+    },
+    {
+      "epoch": 3.4929282346778416,
+      "grad_norm": 0.9489650130271912,
+      "learning_rate": 9.652444124733274e-07,
+      "loss": 0.1009,
+      "step": 6668
+    },
+    {
+      "epoch": 3.4929282346778416,
+      "loss_partial": 0.052001953125,
+      "step": 6668
+    },
+    {
+      "epoch": 3.4929282346778416,
+      "loss_partial": 0.04736328125,
+      "step": 6668
+    },
+    {
+      "epoch": 3.4934520691461497,
+      "grad_norm": 1.2280043363571167,
+      "learning_rate": 9.632857945852291e-07,
+      "loss": 0.0954,
+      "step": 6669
+    },
+    {
+      "epoch": 3.4934520691461497,
+      "loss_partial": 0.056640625,
+      "step": 6669
+    },
+    {
+      "epoch": 3.4934520691461497,
+      "loss_partial": 0.044677734375,
+      "step": 6669
+    },
+    {
+      "epoch": 3.4939759036144578,
+      "grad_norm": 0.7752851247787476,
+      "learning_rate": 9.61329065314519e-07,
+      "loss": 0.1044,
+      "step": 6670
+    },
+    {
+      "epoch": 3.4939759036144578,
+      "loss_partial": 0.039794921875,
+      "step": 6670
+    },
+    {
+      "epoch": 3.4939759036144578,
+      "loss_partial": 0.047607421875,
+      "step": 6670
+    },
+    {
+      "epoch": 3.494499738082766,
+      "grad_norm": 0.90481036901474,
+      "learning_rate": 9.5937422507014e-07,
+      "loss": 0.0952,
+      "step": 6671
+    },
+    {
+      "epoch": 3.494499738082766,
+      "loss_partial": 0.0478515625,
+      "step": 6671
+    },
+    {
+      "epoch": 3.494499738082766,
+      "loss_partial": 0.05517578125,
+      "step": 6671
+    },
+    {
+      "epoch": 3.495023572551074,
+      "grad_norm": 0.8429566621780396,
+      "learning_rate": 9.574212742606426e-07,
+      "loss": 0.0995,
+      "step": 6672
+    },
+    {
+      "epoch": 3.495023572551074,
+      "loss_partial": 0.047119140625,
+      "step": 6672
+    },
+    {
+      "epoch": 3.495023572551074,
+      "loss_partial": 0.053466796875,
+      "step": 6672
+    },
+    {
+      "epoch": 3.495547407019382,
+      "grad_norm": 1.1414539813995361,
+      "learning_rate": 9.554702132941818e-07,
+      "loss": 0.0988,
+      "step": 6673
+    },
+    {
+      "epoch": 3.495547407019382,
+      "loss_partial": 0.0439453125,
+      "step": 6673
+    },
+    {
+      "epoch": 3.495547407019382,
+      "loss_partial": 0.052490234375,
+      "step": 6673
+    },
+    {
+      "epoch": 3.49607124148769,
+      "grad_norm": 0.9086947441101074,
+      "learning_rate": 9.535210425785169e-07,
+      "loss": 0.1017,
+      "step": 6674
+    },
+    {
+      "epoch": 3.49607124148769,
+      "loss_partial": 0.0625,
+      "step": 6674
+    },
+    {
+      "epoch": 3.49607124148769,
+      "loss_partial": 0.053955078125,
+      "step": 6674
+    },
+    {
+      "epoch": 3.496595075955998,
+      "grad_norm": 1.1758536100387573,
+      "learning_rate": 9.515737625210131e-07,
+      "loss": 0.1015,
+      "step": 6675
+    },
+    {
+      "epoch": 3.496595075955998,
+      "loss_partial": 0.045166015625,
+      "step": 6675
+    },
+    {
+      "epoch": 3.496595075955998,
+      "loss_partial": 0.055419921875,
+      "step": 6675
+    },
+    {
+      "epoch": 3.497118910424306,
+      "grad_norm": 0.9843496680259705,
+      "learning_rate": 9.496283735286382e-07,
+      "loss": 0.0999,
+      "step": 6676
+    },
+    {
+      "epoch": 3.497118910424306,
+      "loss_partial": 0.05419921875,
+      "step": 6676
+    },
+    {
+      "epoch": 3.497118910424306,
+      "loss_partial": 0.04931640625,
+      "step": 6676
+    },
+    {
+      "epoch": 3.497642744892614,
+      "grad_norm": 1.0680824518203735,
+      "learning_rate": 9.476848760079671e-07,
+      "loss": 0.1001,
+      "step": 6677
+    },
+    {
+      "epoch": 3.497642744892614,
+      "loss_partial": 0.056884765625,
+      "step": 6677
+    },
+    {
+      "epoch": 3.497642744892614,
+      "loss_partial": 0.058349609375,
+      "step": 6677
+    },
+    {
+      "epoch": 3.498166579360922,
+      "grad_norm": 1.1468327045440674,
+      "learning_rate": 9.457432703651803e-07,
+      "loss": 0.1019,
+      "step": 6678
+    },
+    {
+      "epoch": 3.498166579360922,
+      "loss_partial": 0.0478515625,
+      "step": 6678
+    },
+    {
+      "epoch": 3.498166579360922,
+      "loss_partial": 0.056884765625,
+      "step": 6678
+    },
+    {
+      "epoch": 3.49869041382923,
+      "grad_norm": 0.7854118347167969,
+      "learning_rate": 9.438035570060577e-07,
+      "loss": 0.1006,
+      "step": 6679
+    },
+    {
+      "epoch": 3.49869041382923,
+      "loss_partial": 0.040771484375,
+      "step": 6679
+    },
+    {
+      "epoch": 3.49869041382923,
+      "loss_partial": 0.05224609375,
+      "step": 6679
+    },
+    {
+      "epoch": 3.499214248297538,
+      "grad_norm": 1.2554234266281128,
+      "learning_rate": 9.418657363359918e-07,
+      "loss": 0.0999,
+      "step": 6680
+    },
+    {
+      "epoch": 3.499214248297538,
+      "loss_partial": 0.05322265625,
+      "step": 6680
+    },
+    {
+      "epoch": 3.499214248297538,
+      "loss_partial": 0.048828125,
+      "step": 6680
+    },
+    {
+      "epoch": 3.499738082765846,
+      "grad_norm": 1.0340347290039062,
+      "learning_rate": 9.399298087599729e-07,
+      "loss": 0.0977,
+      "step": 6681
+    },
+    {
+      "epoch": 3.499738082765846,
+      "loss_partial": 0.06103515625,
+      "step": 6681
+    },
+    {
+      "epoch": 3.499738082765846,
+      "loss_partial": 0.06005859375,
+      "step": 6681
+    },
+    {
+      "epoch": 3.500261917234154,
+      "grad_norm": 0.8944905996322632,
+      "learning_rate": 9.379957746825985e-07,
+      "loss": 0.1051,
+      "step": 6682
+    },
+    {
+      "epoch": 3.500261917234154,
+      "loss_partial": 0.046630859375,
+      "step": 6682
+    },
+    {
+      "epoch": 3.500261917234154,
+      "loss_partial": 0.04443359375,
+      "step": 6682
+    },
+    {
+      "epoch": 3.500785751702462,
+      "grad_norm": 1.0328384637832642,
+      "learning_rate": 9.360636345080665e-07,
+      "loss": 0.1037,
+      "step": 6683
+    },
+    {
+      "epoch": 3.500785751702462,
+      "loss_partial": 0.043212890625,
+      "step": 6683
+    },
+    {
+      "epoch": 3.500785751702462,
+      "loss_partial": 0.050537109375,
+      "step": 6683
+    },
+    {
+      "epoch": 3.50130958617077,
+      "grad_norm": 1.2360016107559204,
+      "learning_rate": 9.341333886401882e-07,
+      "loss": 0.0994,
+      "step": 6684
+    },
+    {
+      "epoch": 3.50130958617077,
+      "loss_partial": 0.04638671875,
+      "step": 6684
+    },
+    {
+      "epoch": 3.50130958617077,
+      "loss_partial": 0.04638671875,
+      "step": 6684
+    },
+    {
+      "epoch": 3.501833420639078,
+      "grad_norm": 1.2586876153945923,
+      "learning_rate": 9.322050374823699e-07,
+      "loss": 0.1001,
+      "step": 6685
+    },
+    {
+      "epoch": 3.501833420639078,
+      "loss_partial": 0.043701171875,
+      "step": 6685
+    },
+    {
+      "epoch": 3.501833420639078,
+      "loss_partial": 0.048828125,
+      "step": 6685
+    },
+    {
+      "epoch": 3.502357255107386,
+      "grad_norm": 1.311506986618042,
+      "learning_rate": 9.302785814376247e-07,
+      "loss": 0.099,
+      "step": 6686
+    },
+    {
+      "epoch": 3.502357255107386,
+      "loss_partial": 0.04345703125,
+      "step": 6686
+    },
+    {
+      "epoch": 3.502357255107386,
+      "loss_partial": 0.0400390625,
+      "step": 6686
+    },
+    {
+      "epoch": 3.502881089575694,
+      "grad_norm": 1.2086735963821411,
+      "learning_rate": 9.283540209085717e-07,
+      "loss": 0.1009,
+      "step": 6687
+    },
+    {
+      "epoch": 3.502881089575694,
+      "loss_partial": 0.045166015625,
+      "step": 6687
+    },
+    {
+      "epoch": 3.502881089575694,
+      "loss_partial": 0.044921875,
+      "step": 6687
+    },
+    {
+      "epoch": 3.503404924044002,
+      "grad_norm": 0.897697389125824,
+      "learning_rate": 9.264313562974303e-07,
+      "loss": 0.0989,
+      "step": 6688
+    },
+    {
+      "epoch": 3.503404924044002,
+      "loss_partial": 0.04150390625,
+      "step": 6688
+    },
+    {
+      "epoch": 3.503404924044002,
+      "loss_partial": 0.068359375,
+      "step": 6688
+    },
+    {
+      "epoch": 3.50392875851231,
+      "grad_norm": 0.8923825621604919,
+      "learning_rate": 9.245105880060257e-07,
+      "loss": 0.0966,
+      "step": 6689
+    },
+    {
+      "epoch": 3.50392875851231,
+      "loss_partial": 0.050048828125,
+      "step": 6689
+    },
+    {
+      "epoch": 3.50392875851231,
+      "loss_partial": 0.04443359375,
+      "step": 6689
+    },
+    {
+      "epoch": 3.504452592980618,
+      "grad_norm": 1.2656177282333374,
+      "learning_rate": 9.225917164357867e-07,
+      "loss": 0.0994,
+      "step": 6690
+    },
+    {
+      "epoch": 3.504452592980618,
+      "loss_partial": 0.0478515625,
+      "step": 6690
+    },
+    {
+      "epoch": 3.504452592980618,
+      "loss_partial": 0.0419921875,
+      "step": 6690
+    },
+    {
+      "epoch": 3.504976427448926,
+      "grad_norm": 1.0416102409362793,
+      "learning_rate": 9.206747419877449e-07,
+      "loss": 0.1028,
+      "step": 6691
+    },
+    {
+      "epoch": 3.504976427448926,
+      "loss_partial": 0.046142578125,
+      "step": 6691
+    },
+    {
+      "epoch": 3.504976427448926,
+      "loss_partial": 0.06640625,
+      "step": 6691
+    },
+    {
+      "epoch": 3.505500261917234,
+      "grad_norm": 1.2693239450454712,
+      "learning_rate": 9.187596650625396e-07,
+      "loss": 0.1005,
+      "step": 6692
+    },
+    {
+      "epoch": 3.505500261917234,
+      "loss_partial": 0.045166015625,
+      "step": 6692
+    },
+    {
+      "epoch": 3.505500261917234,
+      "loss_partial": 0.056884765625,
+      "step": 6692
+    },
+    {
+      "epoch": 3.5060240963855422,
+      "grad_norm": 1.3329039812088013,
+      "learning_rate": 9.168464860604065e-07,
+      "loss": 0.0996,
+      "step": 6693
+    },
+    {
+      "epoch": 3.5060240963855422,
+      "loss_partial": 0.055419921875,
+      "step": 6693
+    },
+    {
+      "epoch": 3.5060240963855422,
+      "loss_partial": 0.043701171875,
+      "step": 6693
+    },
+    {
+      "epoch": 3.5065479308538503,
+      "grad_norm": 0.8470185995101929,
+      "learning_rate": 9.1493520538119e-07,
+      "loss": 0.102,
+      "step": 6694
+    },
+    {
+      "epoch": 3.5065479308538503,
+      "loss_partial": 0.057373046875,
+      "step": 6694
+    },
+    {
+      "epoch": 3.5065479308538503,
+      "loss_partial": 0.044677734375,
+      "step": 6694
+    },
+    {
+      "epoch": 3.5070717653221584,
+      "grad_norm": 1.1249175071716309,
+      "learning_rate": 9.130258234243328e-07,
+      "loss": 0.0992,
+      "step": 6695
+    },
+    {
+      "epoch": 3.5070717653221584,
+      "loss_partial": 0.0517578125,
+      "step": 6695
+    },
+    {
+      "epoch": 3.5070717653221584,
+      "loss_partial": 0.0390625,
+      "step": 6695
+    },
+    {
+      "epoch": 3.5075955997904664,
+      "grad_norm": 1.4433248043060303,
+      "learning_rate": 9.11118340588889e-07,
+      "loss": 0.0967,
+      "step": 6696
+    },
+    {
+      "epoch": 3.5075955997904664,
+      "loss_partial": 0.04443359375,
+      "step": 6696
+    },
+    {
+      "epoch": 3.5075955997904664,
+      "loss_partial": 0.040283203125,
+      "step": 6696
+    },
+    {
+      "epoch": 3.5081194342587745,
+      "grad_norm": 1.0339199304580688,
+      "learning_rate": 9.092127572735077e-07,
+      "loss": 0.098,
+      "step": 6697
+    },
+    {
+      "epoch": 3.5081194342587745,
+      "loss_partial": 0.056396484375,
+      "step": 6697
+    },
+    {
+      "epoch": 3.5081194342587745,
+      "loss_partial": 0.04248046875,
+      "step": 6697
+    },
+    {
+      "epoch": 3.508643268727082,
+      "grad_norm": 0.8828780055046082,
+      "learning_rate": 9.073090738764434e-07,
+      "loss": 0.0995,
+      "step": 6698
+    },
+    {
+      "epoch": 3.508643268727082,
+      "loss_partial": 0.0625,
+      "step": 6698
+    },
+    {
+      "epoch": 3.508643268727082,
+      "loss_partial": 0.053955078125,
+      "step": 6698
+    },
+    {
+      "epoch": 3.50916710319539,
+      "grad_norm": 1.0209332704544067,
+      "learning_rate": 9.054072907955569e-07,
+      "loss": 0.0992,
+      "step": 6699
+    },
+    {
+      "epoch": 3.50916710319539,
+      "loss_partial": 0.0419921875,
+      "step": 6699
+    },
+    {
+      "epoch": 3.50916710319539,
+      "loss_partial": 0.05029296875,
+      "step": 6699
+    },
+    {
+      "epoch": 3.509690937663698,
+      "grad_norm": 1.001266360282898,
+      "learning_rate": 9.035074084283058e-07,
+      "loss": 0.0993,
+      "step": 6700
+    },
+    {
+      "epoch": 3.509690937663698,
+      "loss_partial": 0.05615234375,
+      "step": 6700
+    },
+    {
+      "epoch": 3.509690937663698,
+      "loss_partial": 0.049072265625,
+      "step": 6700
+    },
+    {
+      "epoch": 3.5102147721320063,
+      "grad_norm": 1.0217453241348267,
+      "learning_rate": 9.016094271717557e-07,
+      "loss": 0.1027,
+      "step": 6701
+    },
+    {
+      "epoch": 3.5102147721320063,
+      "loss_partial": 0.045166015625,
+      "step": 6701
+    },
+    {
+      "epoch": 3.5102147721320063,
+      "loss_partial": 0.0517578125,
+      "step": 6701
+    },
+    {
+      "epoch": 3.5107386066003143,
+      "grad_norm": 0.8225104808807373,
+      "learning_rate": 8.997133474225739e-07,
+      "loss": 0.0965,
+      "step": 6702
+    },
+    {
+      "epoch": 3.5107386066003143,
+      "loss_partial": 0.053955078125,
+      "step": 6702
+    },
+    {
+      "epoch": 3.5107386066003143,
+      "loss_partial": 0.048095703125,
+      "step": 6702
+    },
+    {
+      "epoch": 3.5112624410686224,
+      "grad_norm": 1.1633222103118896,
+      "learning_rate": 8.978191695770244e-07,
+      "loss": 0.0973,
+      "step": 6703
+    },
+    {
+      "epoch": 3.5112624410686224,
+      "loss_partial": 0.053466796875,
+      "step": 6703
+    },
+    {
+      "epoch": 3.5112624410686224,
+      "loss_partial": 0.045166015625,
+      "step": 6703
+    },
+    {
+      "epoch": 3.5117862755369305,
+      "grad_norm": 1.1730531454086304,
+      "learning_rate": 8.959268940309851e-07,
+      "loss": 0.1017,
+      "step": 6704
+    },
+    {
+      "epoch": 3.5117862755369305,
+      "loss_partial": 0.0546875,
+      "step": 6704
+    },
+    {
+      "epoch": 3.5117862755369305,
+      "loss_partial": 0.04443359375,
+      "step": 6704
+    },
+    {
+      "epoch": 3.512310110005238,
+      "grad_norm": 1.119363784790039,
+      "learning_rate": 8.940365211799263e-07,
+      "loss": 0.0992,
+      "step": 6705
+    },
+    {
+      "epoch": 3.512310110005238,
+      "loss_partial": 0.048828125,
+      "step": 6705
+    },
+    {
+      "epoch": 3.512310110005238,
+      "loss_partial": 0.04736328125,
+      "step": 6705
+    },
+    {
+      "epoch": 3.512833944473546,
+      "grad_norm": 1.2768151760101318,
+      "learning_rate": 8.921480514189262e-07,
+      "loss": 0.0994,
+      "step": 6706
+    },
+    {
+      "epoch": 3.512833944473546,
+      "loss_partial": 0.042236328125,
+      "step": 6706
+    },
+    {
+      "epoch": 3.512833944473546,
+      "loss_partial": 0.055908203125,
+      "step": 6706
+    },
+    {
+      "epoch": 3.513357778941854,
+      "grad_norm": 1.0122497081756592,
+      "learning_rate": 8.902614851426594e-07,
+      "loss": 0.0976,
+      "step": 6707
+    },
+    {
+      "epoch": 3.513357778941854,
+      "loss_partial": 0.036865234375,
+      "step": 6707
+    },
+    {
+      "epoch": 3.513357778941854,
+      "loss_partial": 0.05810546875,
+      "step": 6707
+    },
+    {
+      "epoch": 3.5138816134101623,
+      "grad_norm": 1.045302391052246,
+      "learning_rate": 8.883768227454115e-07,
+      "loss": 0.0971,
+      "step": 6708
+    },
+    {
+      "epoch": 3.5138816134101623,
+      "loss_partial": 0.0439453125,
+      "step": 6708
+    },
+    {
+      "epoch": 3.5138816134101623,
+      "loss_partial": 0.05078125,
+      "step": 6708
+    },
+    {
+      "epoch": 3.5144054478784703,
+      "grad_norm": 1.008175253868103,
+      "learning_rate": 8.864940646210618e-07,
+      "loss": 0.1021,
+      "step": 6709
+    },
+    {
+      "epoch": 3.5144054478784703,
+      "loss_partial": 0.0498046875,
+      "step": 6709
+    },
+    {
+      "epoch": 3.5144054478784703,
+      "loss_partial": 0.04833984375,
+      "step": 6709
+    },
+    {
+      "epoch": 3.5149292823467784,
+      "grad_norm": 1.0004212856292725,
+      "learning_rate": 8.846132111630978e-07,
+      "loss": 0.1,
+      "step": 6710
+    },
+    {
+      "epoch": 3.5149292823467784,
+      "loss_partial": 0.04296875,
+      "step": 6710
+    },
+    {
+      "epoch": 3.5149292823467784,
+      "loss_partial": 0.0498046875,
+      "step": 6710
+    },
+    {
+      "epoch": 3.5154531168150864,
+      "grad_norm": 1.1543198823928833,
+      "learning_rate": 8.827342627646041e-07,
+      "loss": 0.1014,
+      "step": 6711
+    },
+    {
+      "epoch": 3.5154531168150864,
+      "loss_partial": 0.04931640625,
+      "step": 6711
+    },
+    {
+      "epoch": 3.5154531168150864,
+      "loss_partial": 0.049560546875,
+      "step": 6711
+    },
+    {
+      "epoch": 3.5159769512833945,
+      "grad_norm": 1.1701571941375732,
+      "learning_rate": 8.808572198182708e-07,
+      "loss": 0.0964,
+      "step": 6712
+    },
+    {
+      "epoch": 3.5159769512833945,
+      "loss_partial": 0.046875,
+      "step": 6712
+    },
+    {
+      "epoch": 3.5159769512833945,
+      "loss_partial": 0.04541015625,
+      "step": 6712
+    },
+    {
+      "epoch": 3.5165007857517026,
+      "grad_norm": 0.789577305316925,
+      "learning_rate": 8.789820827163865e-07,
+      "loss": 0.0981,
+      "step": 6713
+    },
+    {
+      "epoch": 3.5165007857517026,
+      "loss_partial": 0.04052734375,
+      "step": 6713
+    },
+    {
+      "epoch": 3.5165007857517026,
+      "loss_partial": 0.05322265625,
+      "step": 6713
+    },
+    {
+      "epoch": 3.5170246202200106,
+      "grad_norm": 1.0058391094207764,
+      "learning_rate": 8.771088518508453e-07,
+      "loss": 0.1008,
+      "step": 6714
+    },
+    {
+      "epoch": 3.5170246202200106,
+      "loss_partial": 0.054931640625,
+      "step": 6714
+    },
+    {
+      "epoch": 3.5170246202200106,
+      "loss_partial": 0.04541015625,
+      "step": 6714
+    },
+    {
+      "epoch": 3.5175484546883187,
+      "grad_norm": 0.9990344047546387,
+      "learning_rate": 8.752375276131375e-07,
+      "loss": 0.0987,
+      "step": 6715
+    },
+    {
+      "epoch": 3.5175484546883187,
+      "loss_partial": 0.054443359375,
+      "step": 6715
+    },
+    {
+      "epoch": 3.5175484546883187,
+      "loss_partial": 0.044921875,
+      "step": 6715
+    },
+    {
+      "epoch": 3.5180722891566267,
+      "grad_norm": 0.9413319230079651,
+      "learning_rate": 8.733681103943614e-07,
+      "loss": 0.1003,
+      "step": 6716
+    },
+    {
+      "epoch": 3.5180722891566267,
+      "loss_partial": 0.046630859375,
+      "step": 6716
+    },
+    {
+      "epoch": 3.5180722891566267,
+      "loss_partial": 0.04736328125,
+      "step": 6716
+    },
+    {
+      "epoch": 3.518596123624935,
+      "grad_norm": 1.0628294944763184,
+      "learning_rate": 8.715006005852144e-07,
+      "loss": 0.0968,
+      "step": 6717
+    },
+    {
+      "epoch": 3.518596123624935,
+      "loss_partial": 0.0537109375,
+      "step": 6717
+    },
+    {
+      "epoch": 3.518596123624935,
+      "loss_partial": 0.04931640625,
+      "step": 6717
+    },
+    {
+      "epoch": 3.5191199580932424,
+      "grad_norm": 0.7619753479957581,
+      "learning_rate": 8.696349985759922e-07,
+      "loss": 0.0967,
+      "step": 6718
+    },
+    {
+      "epoch": 3.5191199580932424,
+      "loss_partial": 0.052001953125,
+      "step": 6718
+    },
+    {
+      "epoch": 3.5191199580932424,
+      "loss_partial": 0.056396484375,
+      "step": 6718
+    },
+    {
+      "epoch": 3.5196437925615505,
+      "grad_norm": 1.038830041885376,
+      "learning_rate": 8.677713047565928e-07,
+      "loss": 0.0989,
+      "step": 6719
+    },
+    {
+      "epoch": 3.5196437925615505,
+      "loss_partial": 0.0390625,
+      "step": 6719
+    },
+    {
+      "epoch": 3.5196437925615505,
+      "loss_partial": 0.052978515625,
+      "step": 6719
+    },
+    {
+      "epoch": 3.5201676270298585,
+      "grad_norm": 1.1082897186279297,
+      "learning_rate": 8.659095195165213e-07,
+      "loss": 0.1006,
+      "step": 6720
+    },
+    {
+      "epoch": 3.5201676270298585,
+      "loss_partial": 0.045654296875,
+      "step": 6720
+    },
+    {
+      "epoch": 3.5201676270298585,
+      "loss_partial": 0.054443359375,
+      "step": 6720
+    },
+    {
+      "epoch": 3.5206914614981666,
+      "grad_norm": 1.1709380149841309,
+      "learning_rate": 8.640496432448764e-07,
+      "loss": 0.0996,
+      "step": 6721
+    },
+    {
+      "epoch": 3.5206914614981666,
+      "loss_partial": 0.0458984375,
+      "step": 6721
+    },
+    {
+      "epoch": 3.5206914614981666,
+      "loss_partial": 0.05859375,
+      "step": 6721
+    },
+    {
+      "epoch": 3.5212152959664746,
+      "grad_norm": 0.8356263041496277,
+      "learning_rate": 8.621916763303606e-07,
+      "loss": 0.1004,
+      "step": 6722
+    },
+    {
+      "epoch": 3.5212152959664746,
+      "loss_partial": 0.052978515625,
+      "step": 6722
+    },
+    {
+      "epoch": 3.5212152959664746,
+      "loss_partial": 0.059326171875,
+      "step": 6722
+    },
+    {
+      "epoch": 3.5217391304347827,
+      "grad_norm": 0.7657511234283447,
+      "learning_rate": 8.603356191612777e-07,
+      "loss": 0.1045,
+      "step": 6723
+    },
+    {
+      "epoch": 3.5217391304347827,
+      "loss_partial": 0.048583984375,
+      "step": 6723
+    },
+    {
+      "epoch": 3.5217391304347827,
+      "loss_partial": 0.04638671875,
+      "step": 6723
+    },
+    {
+      "epoch": 3.5222629649030908,
+      "grad_norm": 1.1712802648544312,
+      "learning_rate": 8.58481472125533e-07,
+      "loss": 0.0999,
+      "step": 6724
+    },
+    {
+      "epoch": 3.5222629649030908,
+      "loss_partial": 0.0556640625,
+      "step": 6724
+    },
+    {
+      "epoch": 3.5222629649030908,
+      "loss_partial": 0.051025390625,
+      "step": 6724
+    },
+    {
+      "epoch": 3.5227867993713984,
+      "grad_norm": 0.9499467611312866,
+      "learning_rate": 8.56629235610631e-07,
+      "loss": 0.1018,
+      "step": 6725
+    },
+    {
+      "epoch": 3.5227867993713984,
+      "loss_partial": 0.053466796875,
+      "step": 6725
+    },
+    {
+      "epoch": 3.5227867993713984,
+      "loss_partial": 0.056396484375,
+      "step": 6725
+    },
+    {
+      "epoch": 3.5233106338397064,
+      "grad_norm": 1.3264317512512207,
+      "learning_rate": 8.547789100036786e-07,
+      "loss": 0.0984,
+      "step": 6726
+    },
+    {
+      "epoch": 3.5233106338397064,
+      "loss_partial": 0.04833984375,
+      "step": 6726
+    },
+    {
+      "epoch": 3.5233106338397064,
+      "loss_partial": 0.0498046875,
+      "step": 6726
+    },
+    {
+      "epoch": 3.5238344683080145,
+      "grad_norm": 0.9404909014701843,
+      "learning_rate": 8.529304956913809e-07,
+      "loss": 0.0998,
+      "step": 6727
+    },
+    {
+      "epoch": 3.5238344683080145,
+      "loss_partial": 0.044677734375,
+      "step": 6727
+    },
+    {
+      "epoch": 3.5238344683080145,
+      "loss_partial": 0.045166015625,
+      "step": 6727
+    },
+    {
+      "epoch": 3.5243583027763226,
+      "grad_norm": 1.0161913633346558,
+      "learning_rate": 8.510839930600434e-07,
+      "loss": 0.1,
+      "step": 6728
+    },
+    {
+      "epoch": 3.5243583027763226,
+      "loss_partial": 0.0595703125,
+      "step": 6728
+    },
+    {
+      "epoch": 3.5243583027763226,
+      "loss_partial": 0.051513671875,
+      "step": 6728
+    },
+    {
+      "epoch": 3.5248821372446306,
+      "grad_norm": 0.7371509671211243,
+      "learning_rate": 8.492394024955786e-07,
+      "loss": 0.1031,
+      "step": 6729
+    },
+    {
+      "epoch": 3.5248821372446306,
+      "loss_partial": 0.047607421875,
+      "step": 6729
+    },
+    {
+      "epoch": 3.5248821372446306,
+      "loss_partial": 0.047119140625,
+      "step": 6729
+    },
+    {
+      "epoch": 3.5254059717129387,
+      "grad_norm": 1.1045787334442139,
+      "learning_rate": 8.473967243834924e-07,
+      "loss": 0.097,
+      "step": 6730
+    },
+    {
+      "epoch": 3.5254059717129387,
+      "loss_partial": 0.044921875,
+      "step": 6730
+    },
+    {
+      "epoch": 3.5254059717129387,
+      "loss_partial": 0.046142578125,
+      "step": 6730
+    },
+    {
+      "epoch": 3.5259298061812467,
+      "grad_norm": 0.7701467275619507,
+      "learning_rate": 8.455559591088925e-07,
+      "loss": 0.102,
+      "step": 6731
+    },
+    {
+      "epoch": 3.5259298061812467,
+      "loss_partial": 0.044677734375,
+      "step": 6731
+    },
+    {
+      "epoch": 3.5259298061812467,
+      "loss_partial": 0.048828125,
+      "step": 6731
+    },
+    {
+      "epoch": 3.526453640649555,
+      "grad_norm": 0.8151801228523254,
+      "learning_rate": 8.437171070564865e-07,
+      "loss": 0.0985,
+      "step": 6732
+    },
+    {
+      "epoch": 3.526453640649555,
+      "loss_partial": 0.051513671875,
+      "step": 6732
+    },
+    {
+      "epoch": 3.526453640649555,
+      "loss_partial": 0.07080078125,
+      "step": 6732
+    },
+    {
+      "epoch": 3.526977475117863,
+      "grad_norm": 1.0346862077713013,
+      "learning_rate": 8.418801686105849e-07,
+      "loss": 0.1023,
+      "step": 6733
+    },
+    {
+      "epoch": 3.526977475117863,
+      "loss_partial": 0.046142578125,
+      "step": 6733
+    },
+    {
+      "epoch": 3.526977475117863,
+      "loss_partial": 0.044189453125,
+      "step": 6733
+    },
+    {
+      "epoch": 3.527501309586171,
+      "grad_norm": 0.7967917323112488,
+      "learning_rate": 8.400451441550972e-07,
+      "loss": 0.0997,
+      "step": 6734
+    },
+    {
+      "epoch": 3.527501309586171,
+      "loss_partial": 0.046630859375,
+      "step": 6734
+    },
+    {
+      "epoch": 3.527501309586171,
+      "loss_partial": 0.05224609375,
+      "step": 6734
+    },
+    {
+      "epoch": 3.528025144054479,
+      "grad_norm": 1.2262648344039917,
+      "learning_rate": 8.3821203407353e-07,
+      "loss": 0.0976,
+      "step": 6735
+    },
+    {
+      "epoch": 3.528025144054479,
+      "loss_partial": 0.04833984375,
+      "step": 6735
+    },
+    {
+      "epoch": 3.528025144054479,
+      "loss_partial": 0.04052734375,
+      "step": 6735
+    },
+    {
+      "epoch": 3.528548978522787,
+      "grad_norm": 0.8007270097732544,
+      "learning_rate": 8.363808387489935e-07,
+      "loss": 0.0984,
+      "step": 6736
+    },
+    {
+      "epoch": 3.528548978522787,
+      "loss_partial": 0.06640625,
+      "step": 6736
+    },
+    {
+      "epoch": 3.528548978522787,
+      "loss_partial": 0.046142578125,
+      "step": 6736
+    },
+    {
+      "epoch": 3.5290728129910947,
+      "grad_norm": 0.9027898907661438,
+      "learning_rate": 8.345515585641961e-07,
+      "loss": 0.1053,
+      "step": 6737
+    },
+    {
+      "epoch": 3.5290728129910947,
+      "loss_partial": 0.05078125,
+      "step": 6737
+    },
+    {
+      "epoch": 3.5290728129910947,
+      "loss_partial": 0.044921875,
+      "step": 6737
+    },
+    {
+      "epoch": 3.5295966474594027,
+      "grad_norm": 0.9351586103439331,
+      "learning_rate": 8.327241939014441e-07,
+      "loss": 0.1006,
+      "step": 6738
+    },
+    {
+      "epoch": 3.5295966474594027,
+      "loss_partial": 0.047119140625,
+      "step": 6738
+    },
+    {
+      "epoch": 3.5295966474594027,
+      "loss_partial": 0.056884765625,
+      "step": 6738
+    },
+    {
+      "epoch": 3.5301204819277108,
+      "grad_norm": 0.9065966010093689,
+      "learning_rate": 8.308987451426476e-07,
+      "loss": 0.1017,
+      "step": 6739
+    },
+    {
+      "epoch": 3.5301204819277108,
+      "loss_partial": 0.04931640625,
+      "step": 6739
+    },
+    {
+      "epoch": 3.5301204819277108,
+      "loss_partial": 0.048828125,
+      "step": 6739
+    },
+    {
+      "epoch": 3.530644316396019,
+      "grad_norm": 0.8915482759475708,
+      "learning_rate": 8.290752126693124e-07,
+      "loss": 0.1013,
+      "step": 6740
+    },
+    {
+      "epoch": 3.530644316396019,
+      "loss_partial": 0.052978515625,
+      "step": 6740
+    },
+    {
+      "epoch": 3.530644316396019,
+      "loss_partial": 0.0380859375,
+      "step": 6740
+    },
+    {
+      "epoch": 3.531168150864327,
+      "grad_norm": 0.6974660158157349,
+      "learning_rate": 8.272535968625472e-07,
+      "loss": 0.0968,
+      "step": 6741
+    },
+    {
+      "epoch": 3.531168150864327,
+      "loss_partial": 0.06298828125,
+      "step": 6741
+    },
+    {
+      "epoch": 3.531168150864327,
+      "loss_partial": 0.057373046875,
+      "step": 6741
+    },
+    {
+      "epoch": 3.531691985332635,
+      "grad_norm": 0.9135788679122925,
+      "learning_rate": 8.254338981030574e-07,
+      "loss": 0.1012,
+      "step": 6742
+    },
+    {
+      "epoch": 3.531691985332635,
+      "loss_partial": 0.05322265625,
+      "step": 6742
+    },
+    {
+      "epoch": 3.531691985332635,
+      "loss_partial": 0.051025390625,
+      "step": 6742
+    },
+    {
+      "epoch": 3.532215819800943,
+      "grad_norm": 0.9974222779273987,
+      "learning_rate": 8.236161167711499e-07,
+      "loss": 0.1026,
+      "step": 6743
+    },
+    {
+      "epoch": 3.532215819800943,
+      "loss_partial": 0.057373046875,
+      "step": 6743
+    },
+    {
+      "epoch": 3.532215819800943,
+      "loss_partial": 0.046142578125,
+      "step": 6743
+    },
+    {
+      "epoch": 3.532739654269251,
+      "grad_norm": 0.9359347224235535,
+      "learning_rate": 8.218002532467273e-07,
+      "loss": 0.1009,
+      "step": 6744
+    },
+    {
+      "epoch": 3.532739654269251,
+      "loss_partial": 0.043701171875,
+      "step": 6744
+    },
+    {
+      "epoch": 3.532739654269251,
+      "loss_partial": 0.05322265625,
+      "step": 6744
+    },
+    {
+      "epoch": 3.5332634887375587,
+      "grad_norm": 1.1609212160110474,
+      "learning_rate": 8.199863079092962e-07,
+      "loss": 0.0956,
+      "step": 6745
+    },
+    {
+      "epoch": 3.5332634887375587,
+      "loss_partial": 0.0458984375,
+      "step": 6745
+    },
+    {
+      "epoch": 3.5332634887375587,
+      "loss_partial": 0.04541015625,
+      "step": 6745
+    },
+    {
+      "epoch": 3.5337873232058667,
+      "grad_norm": 0.7135353088378906,
+      "learning_rate": 8.1817428113796e-07,
+      "loss": 0.0948,
+      "step": 6746
+    },
+    {
+      "epoch": 3.5337873232058667,
+      "loss_partial": 0.0390625,
+      "step": 6746
+    },
+    {
+      "epoch": 3.5337873232058667,
+      "loss_partial": 0.05126953125,
+      "step": 6746
+    },
+    {
+      "epoch": 3.534311157674175,
+      "grad_norm": 0.9971004724502563,
+      "learning_rate": 8.163641733114202e-07,
+      "loss": 0.0974,
+      "step": 6747
+    },
+    {
+      "epoch": 3.534311157674175,
+      "loss_partial": 0.0458984375,
+      "step": 6747
+    },
+    {
+      "epoch": 3.534311157674175,
+      "loss_partial": 0.042236328125,
+      "step": 6747
+    },
+    {
+      "epoch": 3.534834992142483,
+      "grad_norm": 0.9771124124526978,
+      "learning_rate": 8.145559848079787e-07,
+      "loss": 0.0969,
+      "step": 6748
+    },
+    {
+      "epoch": 3.534834992142483,
+      "loss_partial": 0.048583984375,
+      "step": 6748
+    },
+    {
+      "epoch": 3.534834992142483,
+      "loss_partial": 0.049072265625,
+      "step": 6748
+    },
+    {
+      "epoch": 3.535358826610791,
+      "grad_norm": 0.909860372543335,
+      "learning_rate": 8.127497160055354e-07,
+      "loss": 0.0995,
+      "step": 6749
+    },
+    {
+      "epoch": 3.535358826610791,
+      "loss_partial": 0.043212890625,
+      "step": 6749
+    },
+    {
+      "epoch": 3.535358826610791,
+      "loss_partial": 0.04541015625,
+      "step": 6749
+    },
+    {
+      "epoch": 3.535882661079099,
+      "grad_norm": 0.899085283279419,
+      "learning_rate": 8.109453672815881e-07,
+      "loss": 0.1006,
+      "step": 6750
+    },
+    {
+      "epoch": 3.535882661079099,
+      "loss_partial": 0.043701171875,
+      "step": 6750
+    },
+    {
+      "epoch": 3.535882661079099,
+      "loss_partial": 0.048583984375,
+      "step": 6750
+    },
+    {
+      "epoch": 3.536406495547407,
+      "grad_norm": 0.8777900338172913,
+      "learning_rate": 8.091429390132366e-07,
+      "loss": 0.0976,
+      "step": 6751
+    },
+    {
+      "epoch": 3.536406495547407,
+      "loss_partial": 0.049560546875,
+      "step": 6751
+    },
+    {
+      "epoch": 3.536406495547407,
+      "loss_partial": 0.05712890625,
+      "step": 6751
+    },
+    {
+      "epoch": 3.536930330015715,
+      "grad_norm": 0.9012749195098877,
+      "learning_rate": 8.073424315771738e-07,
+      "loss": 0.1017,
+      "step": 6752
+    },
+    {
+      "epoch": 3.536930330015715,
+      "loss_partial": 0.04931640625,
+      "step": 6752
+    },
+    {
+      "epoch": 3.536930330015715,
+      "loss_partial": 0.04736328125,
+      "step": 6752
+    },
+    {
+      "epoch": 3.537454164484023,
+      "grad_norm": 0.9664669036865234,
+      "learning_rate": 8.055438453496989e-07,
+      "loss": 0.0988,
+      "step": 6753
+    },
+    {
+      "epoch": 3.537454164484023,
+      "loss_partial": 0.060302734375,
+      "step": 6753
+    },
+    {
+      "epoch": 3.537454164484023,
+      "loss_partial": 0.04443359375,
+      "step": 6753
+    },
+    {
+      "epoch": 3.5379779989523312,
+      "grad_norm": 0.941577672958374,
+      "learning_rate": 8.037471807067043e-07,
+      "loss": 0.104,
+      "step": 6754
+    },
+    {
+      "epoch": 3.5379779989523312,
+      "loss_partial": 0.04833984375,
+      "step": 6754
+    },
+    {
+      "epoch": 3.5379779989523312,
+      "loss_partial": 0.0439453125,
+      "step": 6754
+    },
+    {
+      "epoch": 3.5385018334206393,
+      "grad_norm": 1.0996959209442139,
+      "learning_rate": 8.01952438023681e-07,
+      "loss": 0.1012,
+      "step": 6755
+    },
+    {
+      "epoch": 3.5385018334206393,
+      "loss_partial": 0.053466796875,
+      "step": 6755
+    },
+    {
+      "epoch": 3.5385018334206393,
+      "loss_partial": 0.058349609375,
+      "step": 6755
+    },
+    {
+      "epoch": 3.5390256678889473,
+      "grad_norm": 0.7318110466003418,
+      "learning_rate": 8.001596176757164e-07,
+      "loss": 0.0976,
+      "step": 6756
+    },
+    {
+      "epoch": 3.5390256678889473,
+      "loss_partial": 0.047119140625,
+      "step": 6756
+    },
+    {
+      "epoch": 3.5390256678889473,
+      "loss_partial": 0.047119140625,
+      "step": 6756
+    },
+    {
+      "epoch": 3.539549502357255,
+      "grad_norm": 1.0128209590911865,
+      "learning_rate": 7.983687200375046e-07,
+      "loss": 0.0967,
+      "step": 6757
+    },
+    {
+      "epoch": 3.539549502357255,
+      "loss_partial": 0.052978515625,
+      "step": 6757
+    },
+    {
+      "epoch": 3.539549502357255,
+      "loss_partial": 0.052978515625,
+      "step": 6757
+    },
+    {
+      "epoch": 3.540073336825563,
+      "grad_norm": 0.7750377058982849,
+      "learning_rate": 7.965797454833279e-07,
+      "loss": 0.0993,
+      "step": 6758
+    },
+    {
+      "epoch": 3.540073336825563,
+      "loss_partial": 0.046142578125,
+      "step": 6758
+    },
+    {
+      "epoch": 3.540073336825563,
+      "loss_partial": 0.05029296875,
+      "step": 6758
+    },
+    {
+      "epoch": 3.540597171293871,
+      "grad_norm": 0.8997641205787659,
+      "learning_rate": 7.947926943870732e-07,
+      "loss": 0.1029,
+      "step": 6759
+    },
+    {
+      "epoch": 3.540597171293871,
+      "loss_partial": 0.04736328125,
+      "step": 6759
+    },
+    {
+      "epoch": 3.540597171293871,
+      "loss_partial": 0.048828125,
+      "step": 6759
+    },
+    {
+      "epoch": 3.541121005762179,
+      "grad_norm": 1.0715566873550415,
+      "learning_rate": 7.930075671222214e-07,
+      "loss": 0.1002,
+      "step": 6760
+    },
+    {
+      "epoch": 3.541121005762179,
+      "loss_partial": 0.0546875,
+      "step": 6760
+    },
+    {
+      "epoch": 3.541121005762179,
+      "loss_partial": 0.049072265625,
+      "step": 6760
+    },
+    {
+      "epoch": 3.541644840230487,
+      "grad_norm": 1.3046802282333374,
+      "learning_rate": 7.912243640618544e-07,
+      "loss": 0.0985,
+      "step": 6761
+    },
+    {
+      "epoch": 3.541644840230487,
+      "loss_partial": 0.057373046875,
+      "step": 6761
+    },
+    {
+      "epoch": 3.541644840230487,
+      "loss_partial": 0.049072265625,
+      "step": 6761
+    },
+    {
+      "epoch": 3.5421686746987953,
+      "grad_norm": 1.0201672315597534,
+      "learning_rate": 7.894430855786494e-07,
+      "loss": 0.1001,
+      "step": 6762
+    },
+    {
+      "epoch": 3.5421686746987953,
+      "loss_partial": 0.048583984375,
+      "step": 6762
+    },
+    {
+      "epoch": 3.5421686746987953,
+      "loss_partial": 0.050048828125,
+      "step": 6762
+    },
+    {
+      "epoch": 3.5426925091671033,
+      "grad_norm": 1.2278335094451904,
+      "learning_rate": 7.876637320448833e-07,
+      "loss": 0.099,
+      "step": 6763
+    },
+    {
+      "epoch": 3.5426925091671033,
+      "loss_partial": 0.057373046875,
+      "step": 6763
+    },
+    {
+      "epoch": 3.5426925091671033,
+      "loss_partial": 0.05322265625,
+      "step": 6763
+    },
+    {
+      "epoch": 3.543216343635411,
+      "grad_norm": 0.8617532849311829,
+      "learning_rate": 7.858863038324282e-07,
+      "loss": 0.0967,
+      "step": 6764
+    },
+    {
+      "epoch": 3.543216343635411,
+      "loss_partial": 0.06494140625,
+      "step": 6764
+    },
+    {
+      "epoch": 3.543216343635411,
+      "loss_partial": 0.044921875,
+      "step": 6764
+    },
+    {
+      "epoch": 3.543740178103719,
+      "grad_norm": 1.2586559057235718,
+      "learning_rate": 7.841108013127585e-07,
+      "loss": 0.1033,
+      "step": 6765
+    },
+    {
+      "epoch": 3.543740178103719,
+      "loss_partial": 0.0517578125,
+      "step": 6765
+    },
+    {
+      "epoch": 3.543740178103719,
+      "loss_partial": 0.045654296875,
+      "step": 6765
+    },
+    {
+      "epoch": 3.544264012572027,
+      "grad_norm": 0.9468644261360168,
+      "learning_rate": 7.823372248569427e-07,
+      "loss": 0.0999,
+      "step": 6766
+    },
+    {
+      "epoch": 3.544264012572027,
+      "loss_partial": 0.04931640625,
+      "step": 6766
+    },
+    {
+      "epoch": 3.544264012572027,
+      "loss_partial": 0.047607421875,
+      "step": 6766
+    },
+    {
+      "epoch": 3.544787847040335,
+      "grad_norm": 0.9204089045524597,
+      "learning_rate": 7.805655748356456e-07,
+      "loss": 0.1009,
+      "step": 6767
+    },
+    {
+      "epoch": 3.544787847040335,
+      "loss_partial": 0.04638671875,
+      "step": 6767
+    },
+    {
+      "epoch": 3.544787847040335,
+      "loss_partial": 0.0498046875,
+      "step": 6767
+    },
+    {
+      "epoch": 3.545311681508643,
+      "grad_norm": 1.1012765169143677,
+      "learning_rate": 7.787958516191307e-07,
+      "loss": 0.0986,
+      "step": 6768
+    },
+    {
+      "epoch": 3.545311681508643,
+      "loss_partial": 0.0517578125,
+      "step": 6768
+    },
+    {
+      "epoch": 3.545311681508643,
+      "loss_partial": 0.06689453125,
+      "step": 6768
+    },
+    {
+      "epoch": 3.5458355159769512,
+      "grad_norm": 1.0470318794250488,
+      "learning_rate": 7.770280555772636e-07,
+      "loss": 0.1013,
+      "step": 6769
+    },
+    {
+      "epoch": 3.5458355159769512,
+      "loss_partial": 0.047119140625,
+      "step": 6769
+    },
+    {
+      "epoch": 3.5458355159769512,
+      "loss_partial": 0.048828125,
+      "step": 6769
+    },
+    {
+      "epoch": 3.5463593504452593,
+      "grad_norm": 0.9251560568809509,
+      "learning_rate": 7.752621870794986e-07,
+      "loss": 0.0983,
+      "step": 6770
+    },
+    {
+      "epoch": 3.5463593504452593,
+      "loss_partial": 0.046875,
+      "step": 6770
+    },
+    {
+      "epoch": 3.5463593504452593,
+      "loss_partial": 0.057861328125,
+      "step": 6770
+    },
+    {
+      "epoch": 3.5468831849135674,
+      "grad_norm": 1.0743221044540405,
+      "learning_rate": 7.734982464948925e-07,
+      "loss": 0.1028,
+      "step": 6771
+    },
+    {
+      "epoch": 3.5468831849135674,
+      "loss_partial": 0.0546875,
+      "step": 6771
+    },
+    {
+      "epoch": 3.5468831849135674,
+      "loss_partial": 0.0458984375,
+      "step": 6771
+    },
+    {
+      "epoch": 3.5474070193818754,
+      "grad_norm": 1.3627979755401611,
+      "learning_rate": 7.717362341920975e-07,
+      "loss": 0.1027,
+      "step": 6772
+    },
+    {
+      "epoch": 3.5474070193818754,
+      "loss_partial": 0.05224609375,
+      "step": 6772
+    },
+    {
+      "epoch": 3.5474070193818754,
+      "loss_partial": 0.04541015625,
+      "step": 6772
+    },
+    {
+      "epoch": 3.5479308538501835,
+      "grad_norm": 0.8985278010368347,
+      "learning_rate": 7.699761505393643e-07,
+      "loss": 0.0996,
+      "step": 6773
+    },
+    {
+      "epoch": 3.5479308538501835,
+      "loss_partial": 0.049072265625,
+      "step": 6773
+    },
+    {
+      "epoch": 3.5479308538501835,
+      "loss_partial": 0.04541015625,
+      "step": 6773
+    },
+    {
+      "epoch": 3.5484546883184915,
+      "grad_norm": 0.9026567935943604,
+      "learning_rate": 7.682179959045388e-07,
+      "loss": 0.0974,
+      "step": 6774
+    },
+    {
+      "epoch": 3.5484546883184915,
+      "loss_partial": 0.052001953125,
+      "step": 6774
+    },
+    {
+      "epoch": 3.5484546883184915,
+      "loss_partial": 0.042236328125,
+      "step": 6774
+    },
+    {
+      "epoch": 3.5489785227867996,
+      "grad_norm": 0.9965440034866333,
+      "learning_rate": 7.664617706550648e-07,
+      "loss": 0.1001,
+      "step": 6775
+    },
+    {
+      "epoch": 3.5489785227867996,
+      "loss_partial": 0.04833984375,
+      "step": 6775
+    },
+    {
+      "epoch": 3.5489785227867996,
+      "loss_partial": 0.05224609375,
+      "step": 6775
+    },
+    {
+      "epoch": 3.5495023572551077,
+      "grad_norm": 1.3534291982650757,
+      "learning_rate": 7.647074751579798e-07,
+      "loss": 0.099,
+      "step": 6776
+    },
+    {
+      "epoch": 3.5495023572551077,
+      "loss_partial": 0.051513671875,
+      "step": 6776
+    },
+    {
+      "epoch": 3.5495023572551077,
+      "loss_partial": 0.05029296875,
+      "step": 6776
+    },
+    {
+      "epoch": 3.5500261917234153,
+      "grad_norm": 1.072948694229126,
+      "learning_rate": 7.62955109779926e-07,
+      "loss": 0.1005,
+      "step": 6777
+    },
+    {
+      "epoch": 3.5500261917234153,
+      "loss_partial": 0.043212890625,
+      "step": 6777
+    },
+    {
+      "epoch": 3.5500261917234153,
+      "loss_partial": 0.04931640625,
+      "step": 6777
+    },
+    {
+      "epoch": 3.5505500261917233,
+      "grad_norm": 0.9293152689933777,
+      "learning_rate": 7.612046748871327e-07,
+      "loss": 0.0995,
+      "step": 6778
+    },
+    {
+      "epoch": 3.5505500261917233,
+      "loss_partial": 0.057861328125,
+      "step": 6778
+    },
+    {
+      "epoch": 3.5505500261917233,
+      "loss_partial": 0.0537109375,
+      "step": 6778
+    },
+    {
+      "epoch": 3.5510738606600314,
+      "grad_norm": 1.1205695867538452,
+      "learning_rate": 7.594561708454318e-07,
+      "loss": 0.1036,
+      "step": 6779
+    },
+    {
+      "epoch": 3.5510738606600314,
+      "loss_partial": 0.05078125,
+      "step": 6779
+    },
+    {
+      "epoch": 3.5510738606600314,
+      "loss_partial": 0.06396484375,
+      "step": 6779
+    },
+    {
+      "epoch": 3.5515976951283394,
+      "grad_norm": 0.8804455399513245,
+      "learning_rate": 7.577095980202497e-07,
+      "loss": 0.0995,
+      "step": 6780
+    },
+    {
+      "epoch": 3.5515976951283394,
+      "loss_partial": 0.044921875,
+      "step": 6780
+    },
+    {
+      "epoch": 3.5515976951283394,
+      "loss_partial": 0.051025390625,
+      "step": 6780
+    },
+    {
+      "epoch": 3.5521215295966475,
+      "grad_norm": 0.9789254665374756,
+      "learning_rate": 7.559649567766081e-07,
+      "loss": 0.1004,
+      "step": 6781
+    },
+    {
+      "epoch": 3.5521215295966475,
+      "loss_partial": 0.04931640625,
+      "step": 6781
+    },
+    {
+      "epoch": 3.5521215295966475,
+      "loss_partial": 0.053955078125,
+      "step": 6781
+    },
+    {
+      "epoch": 3.5526453640649556,
+      "grad_norm": 0.9603934288024902,
+      "learning_rate": 7.542222474791272e-07,
+      "loss": 0.0975,
+      "step": 6782
+    },
+    {
+      "epoch": 3.5526453640649556,
+      "loss_partial": 0.058837890625,
+      "step": 6782
+    },
+    {
+      "epoch": 3.5526453640649556,
+      "loss_partial": 0.059326171875,
+      "step": 6782
+    },
+    {
+      "epoch": 3.5531691985332636,
+      "grad_norm": 0.854056715965271,
+      "learning_rate": 7.524814704920236e-07,
+      "loss": 0.101,
+      "step": 6783
+    },
+    {
+      "epoch": 3.5531691985332636,
+      "loss_partial": 0.046630859375,
+      "step": 6783
+    },
+    {
+      "epoch": 3.5531691985332636,
+      "loss_partial": 0.043701171875,
+      "step": 6783
+    },
+    {
+      "epoch": 3.5536930330015712,
+      "grad_norm": 1.0258921384811401,
+      "learning_rate": 7.507426261791073e-07,
+      "loss": 0.0977,
+      "step": 6784
+    },
+    {
+      "epoch": 3.5536930330015712,
+      "loss_partial": 0.055908203125,
+      "step": 6784
+    },
+    {
+      "epoch": 3.5536930330015712,
+      "loss_partial": 0.04052734375,
+      "step": 6784
+    },
+    {
+      "epoch": 3.5542168674698793,
+      "grad_norm": 0.8259691596031189,
+      "learning_rate": 7.490057149037866e-07,
+      "loss": 0.0984,
+      "step": 6785
+    },
+    {
+      "epoch": 3.5542168674698793,
+      "loss_partial": 0.03955078125,
+      "step": 6785
+    },
+    {
+      "epoch": 3.5542168674698793,
+      "loss_partial": 0.048095703125,
+      "step": 6785
+    },
+    {
+      "epoch": 3.5547407019381874,
+      "grad_norm": 1.1767158508300781,
+      "learning_rate": 7.472707370290644e-07,
+      "loss": 0.0956,
+      "step": 6786
+    },
+    {
+      "epoch": 3.5547407019381874,
+      "loss_partial": 0.0517578125,
+      "step": 6786
+    },
+    {
+      "epoch": 3.5547407019381874,
+      "loss_partial": 0.043701171875,
+      "step": 6786
+    },
+    {
+      "epoch": 3.5552645364064954,
+      "grad_norm": 0.9526732563972473,
+      "learning_rate": 7.455376929175417e-07,
+      "loss": 0.0984,
+      "step": 6787
+    },
+    {
+      "epoch": 3.5552645364064954,
+      "loss_partial": 0.05419921875,
+      "step": 6787
+    },
+    {
+      "epoch": 3.5552645364064954,
+      "loss_partial": 0.047119140625,
+      "step": 6787
+    },
+    {
+      "epoch": 3.5557883708748035,
+      "grad_norm": 1.0554531812667847,
+      "learning_rate": 7.438065829314133e-07,
+      "loss": 0.1006,
+      "step": 6788
+    },
+    {
+      "epoch": 3.5557883708748035,
+      "loss_partial": 0.05419921875,
+      "step": 6788
+    },
+    {
+      "epoch": 3.5557883708748035,
+      "loss_partial": 0.04345703125,
+      "step": 6788
+    },
+    {
+      "epoch": 3.5563122053431115,
+      "grad_norm": 1.1311835050582886,
+      "learning_rate": 7.420774074324688e-07,
+      "loss": 0.1007,
+      "step": 6789
+    },
+    {
+      "epoch": 3.5563122053431115,
+      "loss_partial": 0.061279296875,
+      "step": 6789
+    },
+    {
+      "epoch": 3.5563122053431115,
+      "loss_partial": 0.05322265625,
+      "step": 6789
+    },
+    {
+      "epoch": 3.5568360398114196,
+      "grad_norm": 1.021740198135376,
+      "learning_rate": 7.403501667820989e-07,
+      "loss": 0.099,
+      "step": 6790
+    },
+    {
+      "epoch": 3.5568360398114196,
+      "loss_partial": 0.053466796875,
+      "step": 6790
+    },
+    {
+      "epoch": 3.5568360398114196,
+      "loss_partial": 0.05029296875,
+      "step": 6790
+    },
+    {
+      "epoch": 3.5573598742797277,
+      "grad_norm": 1.2112092971801758,
+      "learning_rate": 7.386248613412839e-07,
+      "loss": 0.1013,
+      "step": 6791
+    },
+    {
+      "epoch": 3.5573598742797277,
+      "loss_partial": 0.0537109375,
+      "step": 6791
+    },
+    {
+      "epoch": 3.5573598742797277,
+      "loss_partial": 0.058837890625,
+      "step": 6791
+    },
+    {
+      "epoch": 3.5578837087480357,
+      "grad_norm": 0.8945741057395935,
+      "learning_rate": 7.369014914706019e-07,
+      "loss": 0.0968,
+      "step": 6792
+    },
+    {
+      "epoch": 3.5578837087480357,
+      "loss_partial": 0.04833984375,
+      "step": 6792
+    },
+    {
+      "epoch": 3.5578837087480357,
+      "loss_partial": 0.060546875,
+      "step": 6792
+    },
+    {
+      "epoch": 3.558407543216344,
+      "grad_norm": 0.7882929444313049,
+      "learning_rate": 7.351800575302248e-07,
+      "loss": 0.0974,
+      "step": 6793
+    },
+    {
+      "epoch": 3.558407543216344,
+      "loss_partial": 0.048095703125,
+      "step": 6793
+    },
+    {
+      "epoch": 3.558407543216344,
+      "loss_partial": 0.04541015625,
+      "step": 6793
+    },
+    {
+      "epoch": 3.558931377684652,
+      "grad_norm": 0.8892638087272644,
+      "learning_rate": 7.33460559879926e-07,
+      "loss": 0.1015,
+      "step": 6794
+    },
+    {
+      "epoch": 3.558931377684652,
+      "loss_partial": 0.046875,
+      "step": 6794
+    },
+    {
+      "epoch": 3.558931377684652,
+      "loss_partial": 0.03857421875,
+      "step": 6794
+    },
+    {
+      "epoch": 3.55945521215296,
+      "grad_norm": 1.0048836469650269,
+      "learning_rate": 7.317429988790681e-07,
+      "loss": 0.0969,
+      "step": 6795
+    },
+    {
+      "epoch": 3.55945521215296,
+      "loss_partial": 0.044921875,
+      "step": 6795
+    },
+    {
+      "epoch": 3.55945521215296,
+      "loss_partial": 0.04833984375,
+      "step": 6795
+    },
+    {
+      "epoch": 3.5599790466212675,
+      "grad_norm": 1.1385455131530762,
+      "learning_rate": 7.300273748866094e-07,
+      "loss": 0.1026,
+      "step": 6796
+    },
+    {
+      "epoch": 3.5599790466212675,
+      "loss_partial": 0.04638671875,
+      "step": 6796
+    },
+    {
+      "epoch": 3.5599790466212675,
+      "loss_partial": 0.049560546875,
+      "step": 6796
+    },
+    {
+      "epoch": 3.5605028810895756,
+      "grad_norm": 0.6299471855163574,
+      "learning_rate": 7.283136882611063e-07,
+      "loss": 0.1031,
+      "step": 6797
+    },
+    {
+      "epoch": 3.5605028810895756,
+      "loss_partial": 0.047607421875,
+      "step": 6797
+    },
+    {
+      "epoch": 3.5605028810895756,
+      "loss_partial": 0.05419921875,
+      "step": 6797
+    },
+    {
+      "epoch": 3.5610267155578836,
+      "grad_norm": 0.9592033624649048,
+      "learning_rate": 7.266019393607071e-07,
+      "loss": 0.0974,
+      "step": 6798
+    },
+    {
+      "epoch": 3.5610267155578836,
+      "loss_partial": 0.048095703125,
+      "step": 6798
+    },
+    {
+      "epoch": 3.5610267155578836,
+      "loss_partial": 0.044677734375,
+      "step": 6798
+    },
+    {
+      "epoch": 3.5615505500261917,
+      "grad_norm": 0.9619911909103394,
+      "learning_rate": 7.248921285431587e-07,
+      "loss": 0.1014,
+      "step": 6799
+    },
+    {
+      "epoch": 3.5615505500261917,
+      "loss_partial": 0.04541015625,
+      "step": 6799
+    },
+    {
+      "epoch": 3.5615505500261917,
+      "loss_partial": 0.05029296875,
+      "step": 6799
+    },
+    {
+      "epoch": 3.5620743844944998,
+      "grad_norm": 1.435184121131897,
+      "learning_rate": 7.231842561657998e-07,
+      "loss": 0.0962,
+      "step": 6800
+    },
+    {
+      "epoch": 3.5620743844944998,
+      "loss_partial": 0.05078125,
+      "step": 6800
+    },
+    {
+      "epoch": 3.5620743844944998,
+      "loss_partial": 0.042724609375,
+      "step": 6800
+    },
+    {
+      "epoch": 3.562598218962808,
+      "grad_norm": 0.9030231237411499,
+      "learning_rate": 7.214783225855626e-07,
+      "loss": 0.0966,
+      "step": 6801
+    },
+    {
+      "epoch": 3.562598218962808,
+      "loss_partial": 0.0498046875,
+      "step": 6801
+    },
+    {
+      "epoch": 3.562598218962808,
+      "loss_partial": 0.0556640625,
+      "step": 6801
+    },
+    {
+      "epoch": 3.563122053431116,
+      "grad_norm": 0.9508867263793945,
+      "learning_rate": 7.19774328158982e-07,
+      "loss": 0.1,
+      "step": 6802
+    },
+    {
+      "epoch": 3.563122053431116,
+      "loss_partial": 0.04833984375,
+      "step": 6802
+    },
+    {
+      "epoch": 3.563122053431116,
+      "loss_partial": 0.05419921875,
+      "step": 6802
+    },
+    {
+      "epoch": 3.563645887899424,
+      "grad_norm": 0.739375650882721,
+      "learning_rate": 7.180722732421785e-07,
+      "loss": 0.1002,
+      "step": 6803
+    },
+    {
+      "epoch": 3.563645887899424,
+      "loss_partial": 0.054443359375,
+      "step": 6803
+    },
+    {
+      "epoch": 3.563645887899424,
+      "loss_partial": 0.052490234375,
+      "step": 6803
+    },
+    {
+      "epoch": 3.5641697223677316,
+      "grad_norm": 1.3379935026168823,
+      "learning_rate": 7.163721581908733e-07,
+      "loss": 0.1006,
+      "step": 6804
+    },
+    {
+      "epoch": 3.5641697223677316,
+      "loss_partial": 0.045166015625,
+      "step": 6804
+    },
+    {
+      "epoch": 3.5641697223677316,
+      "loss_partial": 0.045654296875,
+      "step": 6804
+    },
+    {
+      "epoch": 3.5646935568360396,
+      "grad_norm": 0.9961574673652649,
+      "learning_rate": 7.146739833603766e-07,
+      "loss": 0.1003,
+      "step": 6805
+    },
+    {
+      "epoch": 3.5646935568360396,
+      "loss_partial": 0.058349609375,
+      "step": 6805
+    },
+    {
+      "epoch": 3.5646935568360396,
+      "loss_partial": 0.053466796875,
+      "step": 6805
+    },
+    {
+      "epoch": 3.5652173913043477,
+      "grad_norm": 0.8446136116981506,
+      "learning_rate": 7.129777491056e-07,
+      "loss": 0.0966,
+      "step": 6806
+    },
+    {
+      "epoch": 3.5652173913043477,
+      "loss_partial": 0.052734375,
+      "step": 6806
+    },
+    {
+      "epoch": 3.5652173913043477,
+      "loss_partial": 0.04736328125,
+      "step": 6806
+    },
+    {
+      "epoch": 3.5657412257726557,
+      "grad_norm": 0.852781355381012,
+      "learning_rate": 7.112834557810455e-07,
+      "loss": 0.0969,
+      "step": 6807
+    },
+    {
+      "epoch": 3.5657412257726557,
+      "loss_partial": 0.05224609375,
+      "step": 6807
+    },
+    {
+      "epoch": 3.5657412257726557,
+      "loss_partial": 0.052734375,
+      "step": 6807
+    },
+    {
+      "epoch": 3.566265060240964,
+      "grad_norm": 0.9625477194786072,
+      "learning_rate": 7.095911037408087e-07,
+      "loss": 0.0998,
+      "step": 6808
+    },
+    {
+      "epoch": 3.566265060240964,
+      "loss_partial": 0.053955078125,
+      "step": 6808
+    },
+    {
+      "epoch": 3.566265060240964,
+      "loss_partial": 0.046142578125,
+      "step": 6808
+    },
+    {
+      "epoch": 3.566788894709272,
+      "grad_norm": 0.9479097723960876,
+      "learning_rate": 7.079006933385812e-07,
+      "loss": 0.0997,
+      "step": 6809
+    },
+    {
+      "epoch": 3.566788894709272,
+      "loss_partial": 0.06005859375,
+      "step": 6809
+    },
+    {
+      "epoch": 3.566788894709272,
+      "loss_partial": 0.044677734375,
+      "step": 6809
+    },
+    {
+      "epoch": 3.56731272917758,
+      "grad_norm": 0.8275524973869324,
+      "learning_rate": 7.062122249276471e-07,
+      "loss": 0.1032,
+      "step": 6810
+    },
+    {
+      "epoch": 3.56731272917758,
+      "loss_partial": 0.0390625,
+      "step": 6810
+    },
+    {
+      "epoch": 3.56731272917758,
+      "loss_partial": 0.0517578125,
+      "step": 6810
+    },
+    {
+      "epoch": 3.567836563645888,
+      "grad_norm": 0.7482514977455139,
+      "learning_rate": 7.045256988608884e-07,
+      "loss": 0.0995,
+      "step": 6811
+    },
+    {
+      "epoch": 3.567836563645888,
+      "loss_partial": 0.048583984375,
+      "step": 6811
+    },
+    {
+      "epoch": 3.567836563645888,
+      "loss_partial": 0.0478515625,
+      "step": 6811
+    },
+    {
+      "epoch": 3.568360398114196,
+      "grad_norm": 0.8982915282249451,
+      "learning_rate": 7.028411154907766e-07,
+      "loss": 0.0993,
+      "step": 6812
+    },
+    {
+      "epoch": 3.568360398114196,
+      "loss_partial": 0.0478515625,
+      "step": 6812
+    },
+    {
+      "epoch": 3.568360398114196,
+      "loss_partial": 0.05029296875,
+      "step": 6812
+    },
+    {
+      "epoch": 3.568884232582504,
+      "grad_norm": 0.7880661487579346,
+      "learning_rate": 7.011584751693779e-07,
+      "loss": 0.0979,
+      "step": 6813
+    },
+    {
+      "epoch": 3.568884232582504,
+      "loss_partial": 0.048583984375,
+      "step": 6813
+    },
+    {
+      "epoch": 3.568884232582504,
+      "loss_partial": 0.057373046875,
+      "step": 6813
+    },
+    {
+      "epoch": 3.569408067050812,
+      "grad_norm": 1.5956923961639404,
+      "learning_rate": 6.994777782483575e-07,
+      "loss": 0.0988,
+      "step": 6814
+    },
+    {
+      "epoch": 3.569408067050812,
+      "loss_partial": 0.048095703125,
+      "step": 6814
+    },
+    {
+      "epoch": 3.569408067050812,
+      "loss_partial": 0.046630859375,
+      "step": 6814
+    },
+    {
+      "epoch": 3.56993190151912,
+      "grad_norm": 0.9929909706115723,
+      "learning_rate": 6.977990250789679e-07,
+      "loss": 0.0968,
+      "step": 6815
+    },
+    {
+      "epoch": 3.56993190151912,
+      "loss_partial": 0.059326171875,
+      "step": 6815
+    },
+    {
+      "epoch": 3.56993190151912,
+      "loss_partial": 0.045166015625,
+      "step": 6815
+    },
+    {
+      "epoch": 3.570455735987428,
+      "grad_norm": 0.9138315320014954,
+      "learning_rate": 6.961222160120595e-07,
+      "loss": 0.0999,
+      "step": 6816
+    },
+    {
+      "epoch": 3.570455735987428,
+      "loss_partial": 0.05078125,
+      "step": 6816
+    },
+    {
+      "epoch": 3.570455735987428,
+      "loss_partial": 0.049560546875,
+      "step": 6816
+    },
+    {
+      "epoch": 3.570979570455736,
+      "grad_norm": 0.9126499891281128,
+      "learning_rate": 6.944473513980732e-07,
+      "loss": 0.0969,
+      "step": 6817
+    },
+    {
+      "epoch": 3.570979570455736,
+      "loss_partial": 0.06005859375,
+      "step": 6817
+    },
+    {
+      "epoch": 3.570979570455736,
+      "loss_partial": 0.051513671875,
+      "step": 6817
+    },
+    {
+      "epoch": 3.571503404924044,
+      "grad_norm": 1.0475667715072632,
+      "learning_rate": 6.927744315870488e-07,
+      "loss": 0.1013,
+      "step": 6818
+    },
+    {
+      "epoch": 3.571503404924044,
+      "loss_partial": 0.056640625,
+      "step": 6818
+    },
+    {
+      "epoch": 3.571503404924044,
+      "loss_partial": 0.04296875,
+      "step": 6818
+    },
+    {
+      "epoch": 3.572027239392352,
+      "grad_norm": 1.3895145654678345,
+      "learning_rate": 6.911034569286134e-07,
+      "loss": 0.1008,
+      "step": 6819
+    },
+    {
+      "epoch": 3.572027239392352,
+      "loss_partial": 0.05712890625,
+      "step": 6819
+    },
+    {
+      "epoch": 3.572027239392352,
+      "loss_partial": 0.0458984375,
+      "step": 6819
+    },
+    {
+      "epoch": 3.57255107386066,
+      "grad_norm": 1.017366886138916,
+      "learning_rate": 6.894344277719911e-07,
+      "loss": 0.1009,
+      "step": 6820
+    },
+    {
+      "epoch": 3.57255107386066,
+      "loss_partial": 0.04443359375,
+      "step": 6820
+    },
+    {
+      "epoch": 3.57255107386066,
+      "loss_partial": 0.04736328125,
+      "step": 6820
+    },
+    {
+      "epoch": 3.573074908328968,
+      "grad_norm": 1.1311349868774414,
+      "learning_rate": 6.877673444659983e-07,
+      "loss": 0.0994,
+      "step": 6821
+    },
+    {
+      "epoch": 3.573074908328968,
+      "loss_partial": 0.0498046875,
+      "step": 6821
+    },
+    {
+      "epoch": 3.573074908328968,
+      "loss_partial": 0.044189453125,
+      "step": 6821
+    },
+    {
+      "epoch": 3.573598742797276,
+      "grad_norm": 1.146227240562439,
+      "learning_rate": 6.861022073590451e-07,
+      "loss": 0.098,
+      "step": 6822
+    },
+    {
+      "epoch": 3.573598742797276,
+      "loss_partial": 0.04150390625,
+      "step": 6822
+    },
+    {
+      "epoch": 3.573598742797276,
+      "loss_partial": 0.05126953125,
+      "step": 6822
+    },
+    {
+      "epoch": 3.574122577265584,
+      "grad_norm": 1.289410948753357,
+      "learning_rate": 6.844390167991355e-07,
+      "loss": 0.1019,
+      "step": 6823
+    },
+    {
+      "epoch": 3.574122577265584,
+      "loss_partial": 0.061279296875,
+      "step": 6823
+    },
+    {
+      "epoch": 3.574122577265584,
+      "loss_partial": 0.052490234375,
+      "step": 6823
+    },
+    {
+      "epoch": 3.574646411733892,
+      "grad_norm": 1.1269136667251587,
+      "learning_rate": 6.827777731338658e-07,
+      "loss": 0.1005,
+      "step": 6824
+    },
+    {
+      "epoch": 3.574646411733892,
+      "loss_partial": 0.053955078125,
+      "step": 6824
+    },
+    {
+      "epoch": 3.574646411733892,
+      "loss_partial": 0.044921875,
+      "step": 6824
+    },
+    {
+      "epoch": 3.5751702462022,
+      "grad_norm": 1.075740933418274,
+      "learning_rate": 6.811184767104239e-07,
+      "loss": 0.0989,
+      "step": 6825
+    },
+    {
+      "epoch": 3.5751702462022,
+      "loss_partial": 0.046630859375,
+      "step": 6825
+    },
+    {
+      "epoch": 3.5751702462022,
+      "loss_partial": 0.052001953125,
+      "step": 6825
+    },
+    {
+      "epoch": 3.575694080670508,
+      "grad_norm": 0.9338700771331787,
+      "learning_rate": 6.794611278755958e-07,
+      "loss": 0.102,
+      "step": 6826
+    },
+    {
+      "epoch": 3.575694080670508,
+      "loss_partial": 0.050537109375,
+      "step": 6826
+    },
+    {
+      "epoch": 3.575694080670508,
+      "loss_partial": 0.04736328125,
+      "step": 6826
+    },
+    {
+      "epoch": 3.576217915138816,
+      "grad_norm": 0.8577873706817627,
+      "learning_rate": 6.778057269757554e-07,
+      "loss": 0.101,
+      "step": 6827
+    },
+    {
+      "epoch": 3.576217915138816,
+      "loss_partial": 0.050537109375,
+      "step": 6827
+    },
+    {
+      "epoch": 3.576217915138816,
+      "loss_partial": 0.054931640625,
+      "step": 6827
+    },
+    {
+      "epoch": 3.576741749607124,
+      "grad_norm": 1.1407495737075806,
+      "learning_rate": 6.761522743568716e-07,
+      "loss": 0.1009,
+      "step": 6828
+    },
+    {
+      "epoch": 3.576741749607124,
+      "loss_partial": 0.046142578125,
+      "step": 6828
+    },
+    {
+      "epoch": 3.576741749607124,
+      "loss_partial": 0.0478515625,
+      "step": 6828
+    },
+    {
+      "epoch": 3.577265584075432,
+      "grad_norm": 0.9289954900741577,
+      "learning_rate": 6.745007703645045e-07,
+      "loss": 0.0949,
+      "step": 6829
+    },
+    {
+      "epoch": 3.577265584075432,
+      "loss_partial": 0.0537109375,
+      "step": 6829
+    },
+    {
+      "epoch": 3.577265584075432,
+      "loss_partial": 0.04541015625,
+      "step": 6829
+    },
+    {
+      "epoch": 3.57778941854374,
+      "grad_norm": 1.0958659648895264,
+      "learning_rate": 6.728512153438094e-07,
+      "loss": 0.0977,
+      "step": 6830
+    },
+    {
+      "epoch": 3.57778941854374,
+      "loss_partial": 0.048095703125,
+      "step": 6830
+    },
+    {
+      "epoch": 3.57778941854374,
+      "loss_partial": 0.044921875,
+      "step": 6830
+    },
+    {
+      "epoch": 3.5783132530120483,
+      "grad_norm": 1.004984974861145,
+      "learning_rate": 6.712036096395336e-07,
+      "loss": 0.0974,
+      "step": 6831
+    },
+    {
+      "epoch": 3.5783132530120483,
+      "loss_partial": 0.0537109375,
+      "step": 6831
+    },
+    {
+      "epoch": 3.5783132530120483,
+      "loss_partial": 0.048828125,
+      "step": 6831
+    },
+    {
+      "epoch": 3.5788370874803563,
+      "grad_norm": 0.8961698412895203,
+      "learning_rate": 6.695579535960162e-07,
+      "loss": 0.1016,
+      "step": 6832
+    },
+    {
+      "epoch": 3.5788370874803563,
+      "loss_partial": 0.0400390625,
+      "step": 6832
+    },
+    {
+      "epoch": 3.5788370874803563,
+      "loss_partial": 0.04736328125,
+      "step": 6832
+    },
+    {
+      "epoch": 3.5793609219486644,
+      "grad_norm": 0.9150899052619934,
+      "learning_rate": 6.679142475571887e-07,
+      "loss": 0.0965,
+      "step": 6833
+    },
+    {
+      "epoch": 3.5793609219486644,
+      "loss_partial": 0.056396484375,
+      "step": 6833
+    },
+    {
+      "epoch": 3.5793609219486644,
+      "loss_partial": 0.04638671875,
+      "step": 6833
+    },
+    {
+      "epoch": 3.5798847564169725,
+      "grad_norm": 1.0407981872558594,
+      "learning_rate": 6.662724918665753e-07,
+      "loss": 0.106,
+      "step": 6834
+    },
+    {
+      "epoch": 3.5798847564169725,
+      "loss_partial": 0.05419921875,
+      "step": 6834
+    },
+    {
+      "epoch": 3.5798847564169725,
+      "loss_partial": 0.0517578125,
+      "step": 6834
+    },
+    {
+      "epoch": 3.5804085908852805,
+      "grad_norm": 0.9576228260993958,
+      "learning_rate": 6.646326868672937e-07,
+      "loss": 0.1008,
+      "step": 6835
+    },
+    {
+      "epoch": 3.5804085908852805,
+      "loss_partial": 0.042236328125,
+      "step": 6835
+    },
+    {
+      "epoch": 3.5804085908852805,
+      "loss_partial": 0.0439453125,
+      "step": 6835
+    },
+    {
+      "epoch": 3.580932425353588,
+      "grad_norm": 0.8542627096176147,
+      "learning_rate": 6.629948329020508e-07,
+      "loss": 0.0951,
+      "step": 6836
+    },
+    {
+      "epoch": 3.580932425353588,
+      "loss_partial": 0.046875,
+      "step": 6836
+    },
+    {
+      "epoch": 3.580932425353588,
+      "loss_partial": 0.064453125,
+      "step": 6836
+    },
+    {
+      "epoch": 3.581456259821896,
+      "grad_norm": 1.1493959426879883,
+      "learning_rate": 6.613589303131506e-07,
+      "loss": 0.1006,
+      "step": 6837
+    },
+    {
+      "epoch": 3.581456259821896,
+      "loss_partial": 0.048828125,
+      "step": 6837
+    },
+    {
+      "epoch": 3.581456259821896,
+      "loss_partial": 0.052734375,
+      "step": 6837
+    },
+    {
+      "epoch": 3.5819800942902043,
+      "grad_norm": 0.9897215366363525,
+      "learning_rate": 6.597249794424843e-07,
+      "loss": 0.1015,
+      "step": 6838
+    },
+    {
+      "epoch": 3.5819800942902043,
+      "loss_partial": 0.043701171875,
+      "step": 6838
+    },
+    {
+      "epoch": 3.5819800942902043,
+      "loss_partial": 0.045654296875,
+      "step": 6838
+    },
+    {
+      "epoch": 3.5825039287585123,
+      "grad_norm": 0.9479085803031921,
+      "learning_rate": 6.580929806315394e-07,
+      "loss": 0.0977,
+      "step": 6839
+    },
+    {
+      "epoch": 3.5825039287585123,
+      "loss_partial": 0.04541015625,
+      "step": 6839
+    },
+    {
+      "epoch": 3.5825039287585123,
+      "loss_partial": 0.055908203125,
+      "step": 6839
+    },
+    {
+      "epoch": 3.5830277632268204,
+      "grad_norm": 1.003510594367981,
+      "learning_rate": 6.564629342213935e-07,
+      "loss": 0.0993,
+      "step": 6840
+    },
+    {
+      "epoch": 3.5830277632268204,
+      "loss_partial": 0.05322265625,
+      "step": 6840
+    },
+    {
+      "epoch": 3.5830277632268204,
+      "loss_partial": 0.048828125,
+      "step": 6840
+    },
+    {
+      "epoch": 3.5835515976951284,
+      "grad_norm": 1.1200886964797974,
+      "learning_rate": 6.54834840552716e-07,
+      "loss": 0.1,
+      "step": 6841
+    },
+    {
+      "epoch": 3.5835515976951284,
+      "loss_partial": 0.05322265625,
+      "step": 6841
+    },
+    {
+      "epoch": 3.5835515976951284,
+      "loss_partial": 0.049072265625,
+      "step": 6841
+    },
+    {
+      "epoch": 3.5840754321634365,
+      "grad_norm": 0.8837424516677856,
+      "learning_rate": 6.532086999657661e-07,
+      "loss": 0.1062,
+      "step": 6842
+    },
+    {
+      "epoch": 3.5840754321634365,
+      "loss_partial": 0.052490234375,
+      "step": 6842
+    },
+    {
+      "epoch": 3.5840754321634365,
+      "loss_partial": 0.043212890625,
+      "step": 6842
+    },
+    {
+      "epoch": 3.584599266631744,
+      "grad_norm": 1.1852405071258545,
+      "learning_rate": 6.515845128004006e-07,
+      "loss": 0.0977,
+      "step": 6843
+    },
+    {
+      "epoch": 3.584599266631744,
+      "loss_partial": 0.049560546875,
+      "step": 6843
+    },
+    {
+      "epoch": 3.584599266631744,
+      "loss_partial": 0.045654296875,
+      "step": 6843
+    },
+    {
+      "epoch": 3.585123101100052,
+      "grad_norm": 1.069891333580017,
+      "learning_rate": 6.499622793960636e-07,
+      "loss": 0.0957,
+      "step": 6844
+    },
+    {
+      "epoch": 3.585123101100052,
+      "loss_partial": 0.0576171875,
+      "step": 6844
+    },
+    {
+      "epoch": 3.585123101100052,
+      "loss_partial": 0.047119140625,
+      "step": 6844
+    },
+    {
+      "epoch": 3.5856469355683602,
+      "grad_norm": 0.8460632562637329,
+      "learning_rate": 6.483420000917906e-07,
+      "loss": 0.1018,
+      "step": 6845
+    },
+    {
+      "epoch": 3.5856469355683602,
+      "loss_partial": 0.049560546875,
+      "step": 6845
+    },
+    {
+      "epoch": 3.5856469355683602,
+      "loss_partial": 0.048583984375,
+      "step": 6845
+    },
+    {
+      "epoch": 3.5861707700366683,
+      "grad_norm": 1.060423731803894,
+      "learning_rate": 6.467236752262107e-07,
+      "loss": 0.0949,
+      "step": 6846
+    },
+    {
+      "epoch": 3.5861707700366683,
+      "loss_partial": 0.04638671875,
+      "step": 6846
+    },
+    {
+      "epoch": 3.5861707700366683,
+      "loss_partial": 0.049560546875,
+      "step": 6846
+    },
+    {
+      "epoch": 3.5866946045049763,
+      "grad_norm": 1.0419294834136963,
+      "learning_rate": 6.451073051375445e-07,
+      "loss": 0.0968,
+      "step": 6847
+    },
+    {
+      "epoch": 3.5866946045049763,
+      "loss_partial": 0.0615234375,
+      "step": 6847
+    },
+    {
+      "epoch": 3.5866946045049763,
+      "loss_partial": 0.0517578125,
+      "step": 6847
+    },
+    {
+      "epoch": 3.5872184389732844,
+      "grad_norm": 0.9518966674804688,
+      "learning_rate": 6.434928901636018e-07,
+      "loss": 0.1011,
+      "step": 6848
+    },
+    {
+      "epoch": 3.5872184389732844,
+      "loss_partial": 0.0595703125,
+      "step": 6848
+    },
+    {
+      "epoch": 3.5872184389732844,
+      "loss_partial": 0.049072265625,
+      "step": 6848
+    },
+    {
+      "epoch": 3.5877422734415925,
+      "grad_norm": 0.8784857988357544,
+      "learning_rate": 6.418804306417859e-07,
+      "loss": 0.1009,
+      "step": 6849
+    },
+    {
+      "epoch": 3.5877422734415925,
+      "loss_partial": 0.0458984375,
+      "step": 6849
+    },
+    {
+      "epoch": 3.5877422734415925,
+      "loss_partial": 0.044189453125,
+      "step": 6849
+    },
+    {
+      "epoch": 3.5882661079099005,
+      "grad_norm": 1.0252068042755127,
+      "learning_rate": 6.402699269090906e-07,
+      "loss": 0.1027,
+      "step": 6850
+    },
+    {
+      "epoch": 3.5882661079099005,
+      "loss_partial": 0.0712890625,
+      "step": 6850
+    },
+    {
+      "epoch": 3.5882661079099005,
+      "loss_partial": 0.056396484375,
+      "step": 6850
+    },
+    {
+      "epoch": 3.5887899423782086,
+      "grad_norm": 1.0842218399047852,
+      "learning_rate": 6.386613793021023e-07,
+      "loss": 0.1049,
+      "step": 6851
+    },
+    {
+      "epoch": 3.5887899423782086,
+      "loss_partial": 0.04443359375,
+      "step": 6851
+    },
+    {
+      "epoch": 3.5887899423782086,
+      "loss_partial": 0.05859375,
+      "step": 6851
+    },
+    {
+      "epoch": 3.5893137768465166,
+      "grad_norm": 0.8863414525985718,
+      "learning_rate": 6.370547881569978e-07,
+      "loss": 0.1005,
+      "step": 6852
+    },
+    {
+      "epoch": 3.5893137768465166,
+      "loss_partial": 0.0458984375,
+      "step": 6852
+    },
+    {
+      "epoch": 3.5893137768465166,
+      "loss_partial": 0.0537109375,
+      "step": 6852
+    },
+    {
+      "epoch": 3.5898376113148247,
+      "grad_norm": 0.829104483127594,
+      "learning_rate": 6.354501538095437e-07,
+      "loss": 0.1019,
+      "step": 6853
+    },
+    {
+      "epoch": 3.5898376113148247,
+      "loss_partial": 0.0458984375,
+      "step": 6853
+    },
+    {
+      "epoch": 3.5898376113148247,
+      "loss_partial": 0.05078125,
+      "step": 6853
+    },
+    {
+      "epoch": 3.5903614457831328,
+      "grad_norm": 1.0954740047454834,
+      "learning_rate": 6.338474765950975e-07,
+      "loss": 0.1013,
+      "step": 6854
+    },
+    {
+      "epoch": 3.5903614457831328,
+      "loss_partial": 0.05322265625,
+      "step": 6854
+    },
+    {
+      "epoch": 3.5903614457831328,
+      "loss_partial": 0.04638671875,
+      "step": 6854
+    },
+    {
+      "epoch": 3.5908852802514404,
+      "grad_norm": 0.8936727046966553,
+      "learning_rate": 6.322467568486112e-07,
+      "loss": 0.1004,
+      "step": 6855
+    },
+    {
+      "epoch": 3.5908852802514404,
+      "loss_partial": 0.042236328125,
+      "step": 6855
+    },
+    {
+      "epoch": 3.5908852802514404,
+      "loss_partial": 0.04833984375,
+      "step": 6855
+    },
+    {
+      "epoch": 3.5914091147197484,
+      "grad_norm": 1.012110710144043,
+      "learning_rate": 6.306479949046241e-07,
+      "loss": 0.1027,
+      "step": 6856
+    },
+    {
+      "epoch": 3.5914091147197484,
+      "loss_partial": 0.048095703125,
+      "step": 6856
+    },
+    {
+      "epoch": 3.5914091147197484,
+      "loss_partial": 0.05126953125,
+      "step": 6856
+    },
+    {
+      "epoch": 3.5919329491880565,
+      "grad_norm": 1.2262499332427979,
+      "learning_rate": 6.290511910972675e-07,
+      "loss": 0.0991,
+      "step": 6857
+    },
+    {
+      "epoch": 3.5919329491880565,
+      "loss_partial": 0.04296875,
+      "step": 6857
+    },
+    {
+      "epoch": 3.5919329491880565,
+      "loss_partial": 0.05126953125,
+      "step": 6857
+    },
+    {
+      "epoch": 3.5924567836563646,
+      "grad_norm": 1.2528436183929443,
+      "learning_rate": 6.274563457602644e-07,
+      "loss": 0.1015,
+      "step": 6858
+    },
+    {
+      "epoch": 3.5924567836563646,
+      "loss_partial": 0.04345703125,
+      "step": 6858
+    },
+    {
+      "epoch": 3.5924567836563646,
+      "loss_partial": 0.046875,
+      "step": 6858
+    },
+    {
+      "epoch": 3.5929806181246726,
+      "grad_norm": 0.8286058902740479,
+      "learning_rate": 6.258634592269275e-07,
+      "loss": 0.097,
+      "step": 6859
+    },
+    {
+      "epoch": 3.5929806181246726,
+      "loss_partial": 0.049560546875,
+      "step": 6859
+    },
+    {
+      "epoch": 3.5929806181246726,
+      "loss_partial": 0.05859375,
+      "step": 6859
+    },
+    {
+      "epoch": 3.5935044525929807,
+      "grad_norm": 1.0264793634414673,
+      "learning_rate": 6.24272531830159e-07,
+      "loss": 0.0999,
+      "step": 6860
+    },
+    {
+      "epoch": 3.5935044525929807,
+      "loss_partial": 0.05224609375,
+      "step": 6860
+    },
+    {
+      "epoch": 3.5935044525929807,
+      "loss_partial": 0.04248046875,
+      "step": 6860
+    },
+    {
+      "epoch": 3.5940282870612887,
+      "grad_norm": 0.9593798518180847,
+      "learning_rate": 6.226835639024531e-07,
+      "loss": 0.098,
+      "step": 6861
+    },
+    {
+      "epoch": 3.5940282870612887,
+      "loss_partial": 0.05322265625,
+      "step": 6861
+    },
+    {
+      "epoch": 3.5940282870612887,
+      "loss_partial": 0.054931640625,
+      "step": 6861
+    },
+    {
+      "epoch": 3.594552121529597,
+      "grad_norm": 0.910073459148407,
+      "learning_rate": 6.210965557758941e-07,
+      "loss": 0.0999,
+      "step": 6862
+    },
+    {
+      "epoch": 3.594552121529597,
+      "loss_partial": 0.058837890625,
+      "step": 6862
+    },
+    {
+      "epoch": 3.594552121529597,
+      "loss_partial": 0.048583984375,
+      "step": 6862
+    },
+    {
+      "epoch": 3.5950759559979044,
+      "grad_norm": 0.9020259380340576,
+      "learning_rate": 6.19511507782159e-07,
+      "loss": 0.1008,
+      "step": 6863
+    },
+    {
+      "epoch": 3.5950759559979044,
+      "loss_partial": 0.0546875,
+      "step": 6863
+    },
+    {
+      "epoch": 3.5950759559979044,
+      "loss_partial": 0.052734375,
+      "step": 6863
+    },
+    {
+      "epoch": 3.5955997904662125,
+      "grad_norm": 1.104345679283142,
+      "learning_rate": 6.179284202525115e-07,
+      "loss": 0.1024,
+      "step": 6864
+    },
+    {
+      "epoch": 3.5955997904662125,
+      "loss_partial": 0.05126953125,
+      "step": 6864
+    },
+    {
+      "epoch": 3.5955997904662125,
+      "loss_partial": 0.04541015625,
+      "step": 6864
+    },
+    {
+      "epoch": 3.5961236249345205,
+      "grad_norm": 0.7439772486686707,
+      "learning_rate": 6.163472935178061e-07,
+      "loss": 0.1016,
+      "step": 6865
+    },
+    {
+      "epoch": 3.5961236249345205,
+      "loss_partial": 0.052490234375,
+      "step": 6865
+    },
+    {
+      "epoch": 3.5961236249345205,
+      "loss_partial": 0.05615234375,
+      "step": 6865
+    },
+    {
+      "epoch": 3.5966474594028286,
+      "grad_norm": 0.7591316103935242,
+      "learning_rate": 6.147681279084893e-07,
+      "loss": 0.0986,
+      "step": 6866
+    },
+    {
+      "epoch": 3.5966474594028286,
+      "loss_partial": 0.050537109375,
+      "step": 6866
+    },
+    {
+      "epoch": 3.5966474594028286,
+      "loss_partial": 0.05224609375,
+      "step": 6866
+    },
+    {
+      "epoch": 3.5971712938711367,
+      "grad_norm": 1.088936448097229,
+      "learning_rate": 6.131909237545986e-07,
+      "loss": 0.1002,
+      "step": 6867
+    },
+    {
+      "epoch": 3.5971712938711367,
+      "loss_partial": 0.050048828125,
+      "step": 6867
+    },
+    {
+      "epoch": 3.5971712938711367,
+      "loss_partial": 0.05908203125,
+      "step": 6867
+    },
+    {
+      "epoch": 3.5976951283394447,
+      "grad_norm": 0.9651099443435669,
+      "learning_rate": 6.116156813857587e-07,
+      "loss": 0.0957,
+      "step": 6868
+    },
+    {
+      "epoch": 3.5976951283394447,
+      "loss_partial": 0.05615234375,
+      "step": 6868
+    },
+    {
+      "epoch": 3.5976951283394447,
+      "loss_partial": 0.05126953125,
+      "step": 6868
+    },
+    {
+      "epoch": 3.5982189628077528,
+      "grad_norm": 0.8765861392021179,
+      "learning_rate": 6.100424011311845e-07,
+      "loss": 0.1003,
+      "step": 6869
+    },
+    {
+      "epoch": 3.5982189628077528,
+      "loss_partial": 0.040771484375,
+      "step": 6869
+    },
+    {
+      "epoch": 3.5982189628077528,
+      "loss_partial": 0.05029296875,
+      "step": 6869
+    },
+    {
+      "epoch": 3.598742797276061,
+      "grad_norm": 0.8191470503807068,
+      "learning_rate": 6.084710833196839e-07,
+      "loss": 0.0986,
+      "step": 6870
+    },
+    {
+      "epoch": 3.598742797276061,
+      "loss_partial": 0.0498046875,
+      "step": 6870
+    },
+    {
+      "epoch": 3.598742797276061,
+      "loss_partial": 0.0439453125,
+      "step": 6870
+    },
+    {
+      "epoch": 3.599266631744369,
+      "grad_norm": 1.0248239040374756,
+      "learning_rate": 6.069017282796508e-07,
+      "loss": 0.1012,
+      "step": 6871
+    },
+    {
+      "epoch": 3.599266631744369,
+      "loss_partial": 0.05029296875,
+      "step": 6871
+    },
+    {
+      "epoch": 3.599266631744369,
+      "loss_partial": 0.050537109375,
+      "step": 6871
+    },
+    {
+      "epoch": 3.599790466212677,
+      "grad_norm": 1.0230169296264648,
+      "learning_rate": 6.053343363390707e-07,
+      "loss": 0.0991,
+      "step": 6872
+    },
+    {
+      "epoch": 3.599790466212677,
+      "loss_partial": 0.050048828125,
+      "step": 6872
+    },
+    {
+      "epoch": 3.599790466212677,
+      "loss_partial": 0.044921875,
+      "step": 6872
+    },
+    {
+      "epoch": 3.600314300680985,
+      "grad_norm": 0.9123167991638184,
+      "learning_rate": 6.037689078255204e-07,
+      "loss": 0.1032,
+      "step": 6873
+    },
+    {
+      "epoch": 3.600314300680985,
+      "loss_partial": 0.06201171875,
+      "step": 6873
+    },
+    {
+      "epoch": 3.600314300680985,
+      "loss_partial": 0.04296875,
+      "step": 6873
+    },
+    {
+      "epoch": 3.600838135149293,
+      "grad_norm": 0.7328692674636841,
+      "learning_rate": 6.022054430661617e-07,
+      "loss": 0.0992,
+      "step": 6874
+    },
+    {
+      "epoch": 3.600838135149293,
+      "loss_partial": 0.052734375,
+      "step": 6874
+    },
+    {
+      "epoch": 3.600838135149293,
+      "loss_partial": 0.055419921875,
+      "step": 6874
+    },
+    {
+      "epoch": 3.6013619696176007,
+      "grad_norm": 1.1779199838638306,
+      "learning_rate": 6.00643942387753e-07,
+      "loss": 0.1031,
+      "step": 6875
+    },
+    {
+      "epoch": 3.6013619696176007,
+      "loss_partial": 0.049560546875,
+      "step": 6875
+    },
+    {
+      "epoch": 3.6013619696176007,
+      "loss_partial": 0.05615234375,
+      "step": 6875
+    },
+    {
+      "epoch": 3.6018858040859087,
+      "grad_norm": 1.2717244625091553,
+      "learning_rate": 5.990844061166367e-07,
+      "loss": 0.0974,
+      "step": 6876
+    },
+    {
+      "epoch": 3.6018858040859087,
+      "loss_partial": 0.0498046875,
+      "step": 6876
+    },
+    {
+      "epoch": 3.6018858040859087,
+      "loss_partial": 0.042724609375,
+      "step": 6876
+    },
+    {
+      "epoch": 3.602409638554217,
+      "grad_norm": 0.7788599729537964,
+      "learning_rate": 5.975268345787455e-07,
+      "loss": 0.0988,
+      "step": 6877
+    },
+    {
+      "epoch": 3.602409638554217,
+      "loss_partial": 0.044677734375,
+      "step": 6877
+    },
+    {
+      "epoch": 3.602409638554217,
+      "loss_partial": 0.0380859375,
+      "step": 6877
+    },
+    {
+      "epoch": 3.602933473022525,
+      "grad_norm": 0.7724172472953796,
+      "learning_rate": 5.95971228099601e-07,
+      "loss": 0.0962,
+      "step": 6878
+    },
+    {
+      "epoch": 3.602933473022525,
+      "loss_partial": 0.044921875,
+      "step": 6878
+    },
+    {
+      "epoch": 3.602933473022525,
+      "loss_partial": 0.043701171875,
+      "step": 6878
+    },
+    {
+      "epoch": 3.603457307490833,
+      "grad_norm": 1.0275726318359375,
+      "learning_rate": 5.944175870043201e-07,
+      "loss": 0.0995,
+      "step": 6879
+    },
+    {
+      "epoch": 3.603457307490833,
+      "loss_partial": 0.06201171875,
+      "step": 6879
+    },
+    {
+      "epoch": 3.603457307490833,
+      "loss_partial": 0.06005859375,
+      "step": 6879
+    },
+    {
+      "epoch": 3.603981141959141,
+      "grad_norm": 0.8360744714736938,
+      "learning_rate": 5.928659116175994e-07,
+      "loss": 0.1027,
+      "step": 6880
+    },
+    {
+      "epoch": 3.603981141959141,
+      "loss_partial": 0.0654296875,
+      "step": 6880
+    },
+    {
+      "epoch": 3.603981141959141,
+      "loss_partial": 0.053955078125,
+      "step": 6880
+    },
+    {
+      "epoch": 3.604504976427449,
+      "grad_norm": 0.9916101694107056,
+      "learning_rate": 5.91316202263732e-07,
+      "loss": 0.1011,
+      "step": 6881
+    },
+    {
+      "epoch": 3.604504976427449,
+      "loss_partial": 0.05908203125,
+      "step": 6881
+    },
+    {
+      "epoch": 3.604504976427449,
+      "loss_partial": 0.043701171875,
+      "step": 6881
+    },
+    {
+      "epoch": 3.6050288108957567,
+      "grad_norm": 0.6705225706100464,
+      "learning_rate": 5.897684592665964e-07,
+      "loss": 0.0993,
+      "step": 6882
+    },
+    {
+      "epoch": 3.6050288108957567,
+      "loss_partial": 0.040771484375,
+      "step": 6882
+    },
+    {
+      "epoch": 3.6050288108957567,
+      "loss_partial": 0.051025390625,
+      "step": 6882
+    },
+    {
+      "epoch": 3.6055526453640647,
+      "grad_norm": 1.1535588502883911,
+      "learning_rate": 5.882226829496617e-07,
+      "loss": 0.0976,
+      "step": 6883
+    },
+    {
+      "epoch": 3.6055526453640647,
+      "loss_partial": 0.058837890625,
+      "step": 6883
+    },
+    {
+      "epoch": 3.6055526453640647,
+      "loss_partial": 0.053466796875,
+      "step": 6883
+    },
+    {
+      "epoch": 3.606076479832373,
+      "grad_norm": 0.9433351159095764,
+      "learning_rate": 5.866788736359863e-07,
+      "loss": 0.1005,
+      "step": 6884
+    },
+    {
+      "epoch": 3.606076479832373,
+      "loss_partial": 0.045166015625,
+      "step": 6884
+    },
+    {
+      "epoch": 3.606076479832373,
+      "loss_partial": 0.05615234375,
+      "step": 6884
+    },
+    {
+      "epoch": 3.606600314300681,
+      "grad_norm": 0.8969723582267761,
+      "learning_rate": 5.851370316482152e-07,
+      "loss": 0.0966,
+      "step": 6885
+    },
+    {
+      "epoch": 3.606600314300681,
+      "loss_partial": 0.055908203125,
+      "step": 6885
+    },
+    {
+      "epoch": 3.606600314300681,
+      "loss_partial": 0.048828125,
+      "step": 6885
+    },
+    {
+      "epoch": 3.607124148768989,
+      "grad_norm": 0.927071213722229,
+      "learning_rate": 5.835971573085841e-07,
+      "loss": 0.1,
+      "step": 6886
+    },
+    {
+      "epoch": 3.607124148768989,
+      "loss_partial": 0.040283203125,
+      "step": 6886
+    },
+    {
+      "epoch": 3.607124148768989,
+      "loss_partial": 0.049560546875,
+      "step": 6886
+    },
+    {
+      "epoch": 3.607647983237297,
+      "grad_norm": 0.803346574306488,
+      "learning_rate": 5.820592509389189e-07,
+      "loss": 0.0985,
+      "step": 6887
+    },
+    {
+      "epoch": 3.607647983237297,
+      "loss_partial": 0.050048828125,
+      "step": 6887
+    },
+    {
+      "epoch": 3.607647983237297,
+      "loss_partial": 0.051513671875,
+      "step": 6887
+    },
+    {
+      "epoch": 3.608171817705605,
+      "grad_norm": 1.0738346576690674,
+      "learning_rate": 5.805233128606314e-07,
+      "loss": 0.0966,
+      "step": 6888
+    },
+    {
+      "epoch": 3.608171817705605,
+      "loss_partial": 0.0439453125,
+      "step": 6888
+    },
+    {
+      "epoch": 3.608171817705605,
+      "loss_partial": 0.046875,
+      "step": 6888
+    },
+    {
+      "epoch": 3.608695652173913,
+      "grad_norm": 1.1905211210250854,
+      "learning_rate": 5.789893433947236e-07,
+      "loss": 0.0953,
+      "step": 6889
+    },
+    {
+      "epoch": 3.608695652173913,
+      "loss_partial": 0.0537109375,
+      "step": 6889
+    },
+    {
+      "epoch": 3.608695652173913,
+      "loss_partial": 0.058349609375,
+      "step": 6889
+    },
+    {
+      "epoch": 3.609219486642221,
+      "grad_norm": 0.7532727122306824,
+      "learning_rate": 5.774573428617835e-07,
+      "loss": 0.103,
+      "step": 6890
+    },
+    {
+      "epoch": 3.609219486642221,
+      "loss_partial": 0.04833984375,
+      "step": 6890
+    },
+    {
+      "epoch": 3.609219486642221,
+      "loss_partial": 0.046875,
+      "step": 6890
+    },
+    {
+      "epoch": 3.609743321110529,
+      "grad_norm": 0.9168176651000977,
+      "learning_rate": 5.75927311581993e-07,
+      "loss": 0.0975,
+      "step": 6891
+    },
+    {
+      "epoch": 3.609743321110529,
+      "loss_partial": 0.056884765625,
+      "step": 6891
+    },
+    {
+      "epoch": 3.609743321110529,
+      "loss_partial": 0.0615234375,
+      "step": 6891
+    },
+    {
+      "epoch": 3.6102671555788373,
+      "grad_norm": 0.9084877967834473,
+      "learning_rate": 5.743992498751172e-07,
+      "loss": 0.0969,
+      "step": 6892
+    },
+    {
+      "epoch": 3.6102671555788373,
+      "loss_partial": 0.061279296875,
+      "step": 6892
+    },
+    {
+      "epoch": 3.6102671555788373,
+      "loss_partial": 0.059814453125,
+      "step": 6892
+    },
+    {
+      "epoch": 3.6107909900471453,
+      "grad_norm": 0.9646574258804321,
+      "learning_rate": 5.728731580605129e-07,
+      "loss": 0.0983,
+      "step": 6893
+    },
+    {
+      "epoch": 3.6107909900471453,
+      "loss_partial": 0.043212890625,
+      "step": 6893
+    },
+    {
+      "epoch": 3.6107909900471453,
+      "loss_partial": 0.046630859375,
+      "step": 6893
+    },
+    {
+      "epoch": 3.6113148245154534,
+      "grad_norm": 0.9184052348136902,
+      "learning_rate": 5.713490364571217e-07,
+      "loss": 0.1019,
+      "step": 6894
+    },
+    {
+      "epoch": 3.6113148245154534,
+      "loss_partial": 0.05078125,
+      "step": 6894
+    },
+    {
+      "epoch": 3.6113148245154534,
+      "loss_partial": 0.054443359375,
+      "step": 6894
+    },
+    {
+      "epoch": 3.611838658983761,
+      "grad_norm": 1.0274605751037598,
+      "learning_rate": 5.698268853834776e-07,
+      "loss": 0.0993,
+      "step": 6895
+    },
+    {
+      "epoch": 3.611838658983761,
+      "loss_partial": 0.049072265625,
+      "step": 6895
+    },
+    {
+      "epoch": 3.611838658983761,
+      "loss_partial": 0.054443359375,
+      "step": 6895
+    },
+    {
+      "epoch": 3.612362493452069,
+      "grad_norm": 0.8741362690925598,
+      "learning_rate": 5.683067051577007e-07,
+      "loss": 0.1023,
+      "step": 6896
+    },
+    {
+      "epoch": 3.612362493452069,
+      "loss_partial": 0.044677734375,
+      "step": 6896
+    },
+    {
+      "epoch": 3.612362493452069,
+      "loss_partial": 0.03955078125,
+      "step": 6896
+    },
+    {
+      "epoch": 3.612886327920377,
+      "grad_norm": 1.0405843257904053,
+      "learning_rate": 5.667884960974978e-07,
+      "loss": 0.0983,
+      "step": 6897
+    },
+    {
+      "epoch": 3.612886327920377,
+      "loss_partial": 0.046142578125,
+      "step": 6897
+    },
+    {
+      "epoch": 3.612886327920377,
+      "loss_partial": 0.047607421875,
+      "step": 6897
+    },
+    {
+      "epoch": 3.613410162388685,
+      "grad_norm": 0.9614254236221313,
+      "learning_rate": 5.652722585201664e-07,
+      "loss": 0.097,
+      "step": 6898
+    },
+    {
+      "epoch": 3.613410162388685,
+      "loss_partial": 0.05029296875,
+      "step": 6898
+    },
+    {
+      "epoch": 3.613410162388685,
+      "loss_partial": 0.048828125,
+      "step": 6898
+    },
+    {
+      "epoch": 3.6139339968569932,
+      "grad_norm": 0.8307318687438965,
+      "learning_rate": 5.637579927425885e-07,
+      "loss": 0.0982,
+      "step": 6899
+    },
+    {
+      "epoch": 3.6139339968569932,
+      "loss_partial": 0.0498046875,
+      "step": 6899
+    },
+    {
+      "epoch": 3.6139339968569932,
+      "loss_partial": 0.045654296875,
+      "step": 6899
+    },
+    {
+      "epoch": 3.6144578313253013,
+      "grad_norm": 0.9313247799873352,
+      "learning_rate": 5.622456990812397e-07,
+      "loss": 0.0975,
+      "step": 6900
+    },
+    {
+      "epoch": 3.6144578313253013,
+      "loss_partial": 0.062255859375,
+      "step": 6900
+    },
+    {
+      "epoch": 3.6144578313253013,
+      "loss_partial": 0.054931640625,
+      "step": 6900
+    },
+    {
+      "epoch": 3.6149816657936094,
+      "grad_norm": 1.0017938613891602,
+      "learning_rate": 5.607353778521784e-07,
+      "loss": 0.1016,
+      "step": 6901
+    },
+    {
+      "epoch": 3.6149816657936094,
+      "loss_partial": 0.04736328125,
+      "step": 6901
+    },
+    {
+      "epoch": 3.6149816657936094,
+      "loss_partial": 0.04736328125,
+      "step": 6901
+    },
+    {
+      "epoch": 3.615505500261917,
+      "grad_norm": 1.076622486114502,
+      "learning_rate": 5.592270293710522e-07,
+      "loss": 0.1007,
+      "step": 6902
+    },
+    {
+      "epoch": 3.615505500261917,
+      "loss_partial": 0.048828125,
+      "step": 6902
+    },
+    {
+      "epoch": 3.615505500261917,
+      "loss_partial": 0.051025390625,
+      "step": 6902
+    },
+    {
+      "epoch": 3.616029334730225,
+      "grad_norm": 0.9778600335121155,
+      "learning_rate": 5.577206539530966e-07,
+      "loss": 0.0985,
+      "step": 6903
+    },
+    {
+      "epoch": 3.616029334730225,
+      "loss_partial": 0.05859375,
+      "step": 6903
+    },
+    {
+      "epoch": 3.616029334730225,
+      "loss_partial": 0.048828125,
+      "step": 6903
+    },
+    {
+      "epoch": 3.616553169198533,
+      "grad_norm": 0.760021984577179,
+      "learning_rate": 5.562162519131353e-07,
+      "loss": 0.0998,
+      "step": 6904
+    },
+    {
+      "epoch": 3.616553169198533,
+      "loss_partial": 0.056640625,
+      "step": 6904
+    },
+    {
+      "epoch": 3.616553169198533,
+      "loss_partial": 0.055908203125,
+      "step": 6904
+    },
+    {
+      "epoch": 3.617077003666841,
+      "grad_norm": 0.9963350892066956,
+      "learning_rate": 5.547138235655792e-07,
+      "loss": 0.1004,
+      "step": 6905
+    },
+    {
+      "epoch": 3.617077003666841,
+      "loss_partial": 0.042236328125,
+      "step": 6905
+    },
+    {
+      "epoch": 3.617077003666841,
+      "loss_partial": 0.04296875,
+      "step": 6905
+    },
+    {
+      "epoch": 3.617600838135149,
+      "grad_norm": 0.867668092250824,
+      "learning_rate": 5.532133692244257e-07,
+      "loss": 0.1016,
+      "step": 6906
+    },
+    {
+      "epoch": 3.617600838135149,
+      "loss_partial": 0.0478515625,
+      "step": 6906
+    },
+    {
+      "epoch": 3.617600838135149,
+      "loss_partial": 0.043701171875,
+      "step": 6906
+    },
+    {
+      "epoch": 3.6181246726034573,
+      "grad_norm": 0.7841799855232239,
+      "learning_rate": 5.517148892032609e-07,
+      "loss": 0.098,
+      "step": 6907
+    },
+    {
+      "epoch": 3.6181246726034573,
+      "loss_partial": 0.0478515625,
+      "step": 6907
+    },
+    {
+      "epoch": 3.6181246726034573,
+      "loss_partial": 0.03857421875,
+      "step": 6907
+    },
+    {
+      "epoch": 3.6186485070717653,
+      "grad_norm": 1.0196653604507446,
+      "learning_rate": 5.502183838152586e-07,
+      "loss": 0.1006,
+      "step": 6908
+    },
+    {
+      "epoch": 3.6186485070717653,
+      "loss_partial": 0.04833984375,
+      "step": 6908
+    },
+    {
+      "epoch": 3.6186485070717653,
+      "loss_partial": 0.04541015625,
+      "step": 6908
+    },
+    {
+      "epoch": 3.6191723415400734,
+      "grad_norm": 0.8533656001091003,
+      "learning_rate": 5.487238533731776e-07,
+      "loss": 0.0956,
+      "step": 6909
+    },
+    {
+      "epoch": 3.6191723415400734,
+      "loss_partial": 0.047119140625,
+      "step": 6909
+    },
+    {
+      "epoch": 3.6191723415400734,
+      "loss_partial": 0.05322265625,
+      "step": 6909
+    },
+    {
+      "epoch": 3.6196961760083814,
+      "grad_norm": 0.9562242031097412,
+      "learning_rate": 5.472312981893668e-07,
+      "loss": 0.1003,
+      "step": 6910
+    },
+    {
+      "epoch": 3.6196961760083814,
+      "loss_partial": 0.05322265625,
+      "step": 6910
+    },
+    {
+      "epoch": 3.6196961760083814,
+      "loss_partial": 0.04443359375,
+      "step": 6910
+    },
+    {
+      "epoch": 3.6202200104766895,
+      "grad_norm": 0.9675567150115967,
+      "learning_rate": 5.45740718575759e-07,
+      "loss": 0.0985,
+      "step": 6911
+    },
+    {
+      "epoch": 3.6202200104766895,
+      "loss_partial": 0.0546875,
+      "step": 6911
+    },
+    {
+      "epoch": 3.6202200104766895,
+      "loss_partial": 0.049560546875,
+      "step": 6911
+    },
+    {
+      "epoch": 3.6207438449449976,
+      "grad_norm": 1.1357821226119995,
+      "learning_rate": 5.442521148438784e-07,
+      "loss": 0.0987,
+      "step": 6912
+    },
+    {
+      "epoch": 3.6207438449449976,
+      "loss_partial": 0.048583984375,
+      "step": 6912
+    },
+    {
+      "epoch": 3.6207438449449976,
+      "loss_partial": 0.059814453125,
+      "step": 6912
+    },
+    {
+      "epoch": 3.6212676794133056,
+      "grad_norm": 0.847113847732544,
+      "learning_rate": 5.427654873048338e-07,
+      "loss": 0.1036,
+      "step": 6913
+    },
+    {
+      "epoch": 3.6212676794133056,
+      "loss_partial": 0.0537109375,
+      "step": 6913
+    },
+    {
+      "epoch": 3.6212676794133056,
+      "loss_partial": 0.0498046875,
+      "step": 6913
+    },
+    {
+      "epoch": 3.6217915138816132,
+      "grad_norm": 1.164823293685913,
+      "learning_rate": 5.412808362693212e-07,
+      "loss": 0.1013,
+      "step": 6914
+    },
+    {
+      "epoch": 3.6217915138816132,
+      "loss_partial": 0.0517578125,
+      "step": 6914
+    },
+    {
+      "epoch": 3.6217915138816132,
+      "loss_partial": 0.05712890625,
+      "step": 6914
+    },
+    {
+      "epoch": 3.6223153483499213,
+      "grad_norm": 1.003240942955017,
+      "learning_rate": 5.397981620476212e-07,
+      "loss": 0.101,
+      "step": 6915
+    },
+    {
+      "epoch": 3.6223153483499213,
+      "loss_partial": 0.043701171875,
+      "step": 6915
+    },
+    {
+      "epoch": 3.6223153483499213,
+      "loss_partial": 0.04150390625,
+      "step": 6915
+    },
+    {
+      "epoch": 3.6228391828182294,
+      "grad_norm": 1.2893630266189575,
+      "learning_rate": 5.383174649496059e-07,
+      "loss": 0.1046,
+      "step": 6916
+    },
+    {
+      "epoch": 3.6228391828182294,
+      "loss_partial": 0.052001953125,
+      "step": 6916
+    },
+    {
+      "epoch": 3.6228391828182294,
+      "loss_partial": 0.057373046875,
+      "step": 6916
+    },
+    {
+      "epoch": 3.6233630172865374,
+      "grad_norm": 0.919638991355896,
+      "learning_rate": 5.368387452847312e-07,
+      "loss": 0.1014,
+      "step": 6917
+    },
+    {
+      "epoch": 3.6233630172865374,
+      "loss_partial": 0.048095703125,
+      "step": 6917
+    },
+    {
+      "epoch": 3.6233630172865374,
+      "loss_partial": 0.0498046875,
+      "step": 6917
+    },
+    {
+      "epoch": 3.6238868517548455,
+      "grad_norm": 0.8395569920539856,
+      "learning_rate": 5.353620033620411e-07,
+      "loss": 0.0972,
+      "step": 6918
+    },
+    {
+      "epoch": 3.6238868517548455,
+      "loss_partial": 0.047119140625,
+      "step": 6918
+    },
+    {
+      "epoch": 3.6238868517548455,
+      "loss_partial": 0.052490234375,
+      "step": 6918
+    },
+    {
+      "epoch": 3.6244106862231535,
+      "grad_norm": 1.0664716958999634,
+      "learning_rate": 5.33887239490164e-07,
+      "loss": 0.1037,
+      "step": 6919
+    },
+    {
+      "epoch": 3.6244106862231535,
+      "loss_partial": 0.047607421875,
+      "step": 6919
+    },
+    {
+      "epoch": 3.6244106862231535,
+      "loss_partial": 0.046875,
+      "step": 6919
+    },
+    {
+      "epoch": 3.6249345206914616,
+      "grad_norm": 1.270129919052124,
+      "learning_rate": 5.324144539773169e-07,
+      "loss": 0.0985,
+      "step": 6920
+    },
+    {
+      "epoch": 3.6249345206914616,
+      "loss_partial": 0.05712890625,
+      "step": 6920
+    },
+    {
+      "epoch": 3.6249345206914616,
+      "loss_partial": 0.057373046875,
+      "step": 6920
+    },
+    {
+      "epoch": 3.6254583551597697,
+      "grad_norm": 1.309257984161377,
+      "learning_rate": 5.309436471313035e-07,
+      "loss": 0.1042,
+      "step": 6921
+    },
+    {
+      "epoch": 3.6254583551597697,
+      "loss_partial": 0.050537109375,
+      "step": 6921
+    },
+    {
+      "epoch": 3.6254583551597697,
+      "loss_partial": 0.04443359375,
+      "step": 6921
+    },
+    {
+      "epoch": 3.6259821896280773,
+      "grad_norm": 1.0498590469360352,
+      "learning_rate": 5.294748192595134e-07,
+      "loss": 0.1008,
+      "step": 6922
+    },
+    {
+      "epoch": 3.6259821896280773,
+      "loss_partial": 0.046875,
+      "step": 6922
+    },
+    {
+      "epoch": 3.6259821896280773,
+      "loss_partial": 0.047119140625,
+      "step": 6922
+    },
+    {
+      "epoch": 3.6265060240963853,
+      "grad_norm": 1.1138890981674194,
+      "learning_rate": 5.280079706689201e-07,
+      "loss": 0.0992,
+      "step": 6923
+    },
+    {
+      "epoch": 3.6265060240963853,
+      "loss_partial": 0.044189453125,
+      "step": 6923
+    },
+    {
+      "epoch": 3.6265060240963853,
+      "loss_partial": 0.049072265625,
+      "step": 6923
+    },
+    {
+      "epoch": 3.6270298585646934,
+      "grad_norm": 0.8858183026313782,
+      "learning_rate": 5.265431016660894e-07,
+      "loss": 0.1009,
+      "step": 6924
+    },
+    {
+      "epoch": 3.6270298585646934,
+      "loss_partial": 0.0478515625,
+      "step": 6924
+    },
+    {
+      "epoch": 3.6270298585646934,
+      "loss_partial": 0.052490234375,
+      "step": 6924
+    },
+    {
+      "epoch": 3.6275536930330015,
+      "grad_norm": 0.8272022604942322,
+      "learning_rate": 5.250802125571686e-07,
+      "loss": 0.1029,
+      "step": 6925
+    },
+    {
+      "epoch": 3.6275536930330015,
+      "loss_partial": 0.05615234375,
+      "step": 6925
+    },
+    {
+      "epoch": 3.6275536930330015,
+      "loss_partial": 0.04541015625,
+      "step": 6925
+    },
+    {
+      "epoch": 3.6280775275013095,
+      "grad_norm": 1.1645985841751099,
+      "learning_rate": 5.23619303647892e-07,
+      "loss": 0.0974,
+      "step": 6926
+    },
+    {
+      "epoch": 3.6280775275013095,
+      "loss_partial": 0.04736328125,
+      "step": 6926
+    },
+    {
+      "epoch": 3.6280775275013095,
+      "loss_partial": 0.047607421875,
+      "step": 6926
+    },
+    {
+      "epoch": 3.6286013619696176,
+      "grad_norm": 1.2998350858688354,
+      "learning_rate": 5.2216037524358e-07,
+      "loss": 0.0964,
+      "step": 6927
+    },
+    {
+      "epoch": 3.6286013619696176,
+      "loss_partial": 0.048828125,
+      "step": 6927
+    },
+    {
+      "epoch": 3.6286013619696176,
+      "loss_partial": 0.04931640625,
+      "step": 6927
+    },
+    {
+      "epoch": 3.6291251964379256,
+      "grad_norm": 0.9116417765617371,
+      "learning_rate": 5.207034276491407e-07,
+      "loss": 0.1007,
+      "step": 6928
+    },
+    {
+      "epoch": 3.6291251964379256,
+      "loss_partial": 0.04541015625,
+      "step": 6928
+    },
+    {
+      "epoch": 3.6291251964379256,
+      "loss_partial": 0.046142578125,
+      "step": 6928
+    },
+    {
+      "epoch": 3.6296490309062337,
+      "grad_norm": 0.9433156847953796,
+      "learning_rate": 5.192484611690685e-07,
+      "loss": 0.0968,
+      "step": 6929
+    },
+    {
+      "epoch": 3.6296490309062337,
+      "loss_partial": 0.0546875,
+      "step": 6929
+    },
+    {
+      "epoch": 3.6296490309062337,
+      "loss_partial": 0.04931640625,
+      "step": 6929
+    },
+    {
+      "epoch": 3.6301728653745418,
+      "grad_norm": 1.0357568264007568,
+      "learning_rate": 5.177954761074389e-07,
+      "loss": 0.1008,
+      "step": 6930
+    },
+    {
+      "epoch": 3.6301728653745418,
+      "loss_partial": 0.055419921875,
+      "step": 6930
+    },
+    {
+      "epoch": 3.6301728653745418,
+      "loss_partial": 0.0537109375,
+      "step": 6930
+    },
+    {
+      "epoch": 3.63069669984285,
+      "grad_norm": 1.098132848739624,
+      "learning_rate": 5.16344472767919e-07,
+      "loss": 0.0985,
+      "step": 6931
+    },
+    {
+      "epoch": 3.63069669984285,
+      "loss_partial": 0.05078125,
+      "step": 6931
+    },
+    {
+      "epoch": 3.63069669984285,
+      "loss_partial": 0.052001953125,
+      "step": 6931
+    },
+    {
+      "epoch": 3.631220534311158,
+      "grad_norm": 1.0468145608901978,
+      "learning_rate": 5.148954514537596e-07,
+      "loss": 0.1035,
+      "step": 6932
+    },
+    {
+      "epoch": 3.631220534311158,
+      "loss_partial": 0.05029296875,
+      "step": 6932
+    },
+    {
+      "epoch": 3.631220534311158,
+      "loss_partial": 0.047607421875,
+      "step": 6932
+    },
+    {
+      "epoch": 3.631744368779466,
+      "grad_norm": 0.9475061893463135,
+      "learning_rate": 5.134484124677974e-07,
+      "loss": 0.0996,
+      "step": 6933
+    },
+    {
+      "epoch": 3.631744368779466,
+      "loss_partial": 0.050537109375,
+      "step": 6933
+    },
+    {
+      "epoch": 3.631744368779466,
+      "loss_partial": 0.04931640625,
+      "step": 6933
+    },
+    {
+      "epoch": 3.6322682032477736,
+      "grad_norm": 1.0480948686599731,
+      "learning_rate": 5.120033561124527e-07,
+      "loss": 0.1005,
+      "step": 6934
+    },
+    {
+      "epoch": 3.6322682032477736,
+      "loss_partial": 0.048828125,
+      "step": 6934
+    },
+    {
+      "epoch": 3.6322682032477736,
+      "loss_partial": 0.048583984375,
+      "step": 6934
+    },
+    {
+      "epoch": 3.6327920377160816,
+      "grad_norm": 1.013145089149475,
+      "learning_rate": 5.105602826897338e-07,
+      "loss": 0.1004,
+      "step": 6935
+    },
+    {
+      "epoch": 3.6327920377160816,
+      "loss_partial": 0.0625,
+      "step": 6935
+    },
+    {
+      "epoch": 3.6327920377160816,
+      "loss_partial": 0.050537109375,
+      "step": 6935
+    },
+    {
+      "epoch": 3.6333158721843897,
+      "grad_norm": 1.0903825759887695,
+      "learning_rate": 5.091191925012362e-07,
+      "loss": 0.1012,
+      "step": 6936
+    },
+    {
+      "epoch": 3.6333158721843897,
+      "loss_partial": 0.053466796875,
+      "step": 6936
+    },
+    {
+      "epoch": 3.6333158721843897,
+      "loss_partial": 0.05224609375,
+      "step": 6936
+    },
+    {
+      "epoch": 3.6338397066526977,
+      "grad_norm": 1.0872173309326172,
+      "learning_rate": 5.076800858481367e-07,
+      "loss": 0.0999,
+      "step": 6937
+    },
+    {
+      "epoch": 3.6338397066526977,
+      "loss_partial": 0.05517578125,
+      "step": 6937
+    },
+    {
+      "epoch": 3.6338397066526977,
+      "loss_partial": 0.04345703125,
+      "step": 6937
+    },
+    {
+      "epoch": 3.634363541121006,
+      "grad_norm": 0.7724402546882629,
+      "learning_rate": 5.062429630312015e-07,
+      "loss": 0.1026,
+      "step": 6938
+    },
+    {
+      "epoch": 3.634363541121006,
+      "loss_partial": 0.046875,
+      "step": 6938
+    },
+    {
+      "epoch": 3.634363541121006,
+      "loss_partial": 0.049072265625,
+      "step": 6938
+    },
+    {
+      "epoch": 3.634887375589314,
+      "grad_norm": 0.9268259406089783,
+      "learning_rate": 5.048078243507782e-07,
+      "loss": 0.096,
+      "step": 6939
+    },
+    {
+      "epoch": 3.634887375589314,
+      "loss_partial": 0.05712890625,
+      "step": 6939
+    },
+    {
+      "epoch": 3.634887375589314,
+      "loss_partial": 0.05859375,
+      "step": 6939
+    },
+    {
+      "epoch": 3.635411210057622,
+      "grad_norm": 0.925110399723053,
+      "learning_rate": 5.033746701068032e-07,
+      "loss": 0.1007,
+      "step": 6940
+    },
+    {
+      "epoch": 3.635411210057622,
+      "loss_partial": 0.0458984375,
+      "step": 6940
+    },
+    {
+      "epoch": 3.635411210057622,
+      "loss_partial": 0.0654296875,
+      "step": 6940
+    },
+    {
+      "epoch": 3.63593504452593,
+      "grad_norm": 0.932140588760376,
+      "learning_rate": 5.019435005987982e-07,
+      "loss": 0.1022,
+      "step": 6941
+    },
+    {
+      "epoch": 3.63593504452593,
+      "loss_partial": 0.044189453125,
+      "step": 6941
+    },
+    {
+      "epoch": 3.63593504452593,
+      "loss_partial": 0.044921875,
+      "step": 6941
+    },
+    {
+      "epoch": 3.6364588789942376,
+      "grad_norm": 0.9309395551681519,
+      "learning_rate": 5.005143161258663e-07,
+      "loss": 0.0964,
+      "step": 6942
+    },
+    {
+      "epoch": 3.6364588789942376,
+      "loss_partial": 0.06494140625,
+      "step": 6942
+    },
+    {
+      "epoch": 3.6364588789942376,
+      "loss_partial": 0.04638671875,
+      "step": 6942
+    },
+    {
+      "epoch": 3.6369827134625456,
+      "grad_norm": 1.030279517173767,
+      "learning_rate": 4.990871169866995e-07,
+      "loss": 0.0976,
+      "step": 6943
+    },
+    {
+      "epoch": 3.6369827134625456,
+      "loss_partial": 0.054931640625,
+      "step": 6943
+    },
+    {
+      "epoch": 3.6369827134625456,
+      "loss_partial": 0.052001953125,
+      "step": 6943
+    },
+    {
+      "epoch": 3.6375065479308537,
+      "grad_norm": 1.0156840085983276,
+      "learning_rate": 4.976619034795737e-07,
+      "loss": 0.0999,
+      "step": 6944
+    },
+    {
+      "epoch": 3.6375065479308537,
+      "loss_partial": 0.053466796875,
+      "step": 6944
+    },
+    {
+      "epoch": 3.6375065479308537,
+      "loss_partial": 0.053466796875,
+      "step": 6944
+    },
+    {
+      "epoch": 3.6380303823991618,
+      "grad_norm": 0.781561553478241,
+      "learning_rate": 4.962386759023496e-07,
+      "loss": 0.0999,
+      "step": 6945
+    },
+    {
+      "epoch": 3.6380303823991618,
+      "loss_partial": 0.05322265625,
+      "step": 6945
+    },
+    {
+      "epoch": 3.6380303823991618,
+      "loss_partial": 0.042236328125,
+      "step": 6945
+    },
+    {
+      "epoch": 3.63855421686747,
+      "grad_norm": 0.9647900462150574,
+      "learning_rate": 4.948174345524725e-07,
+      "loss": 0.0999,
+      "step": 6946
+    },
+    {
+      "epoch": 3.63855421686747,
+      "loss_partial": 0.056884765625,
+      "step": 6946
+    },
+    {
+      "epoch": 3.63855421686747,
+      "loss_partial": 0.0458984375,
+      "step": 6946
+    },
+    {
+      "epoch": 3.639078051335778,
+      "grad_norm": 0.9929358959197998,
+      "learning_rate": 4.933981797269716e-07,
+      "loss": 0.0964,
+      "step": 6947
+    },
+    {
+      "epoch": 3.639078051335778,
+      "loss_partial": 0.0458984375,
+      "step": 6947
+    },
+    {
+      "epoch": 3.639078051335778,
+      "loss_partial": 0.049072265625,
+      "step": 6947
+    },
+    {
+      "epoch": 3.639601885804086,
+      "grad_norm": 0.9538285732269287,
+      "learning_rate": 4.919809117224661e-07,
+      "loss": 0.0987,
+      "step": 6948
+    },
+    {
+      "epoch": 3.639601885804086,
+      "loss_partial": 0.049072265625,
+      "step": 6948
+    },
+    {
+      "epoch": 3.639601885804086,
+      "loss_partial": 0.0419921875,
+      "step": 6948
+    },
+    {
+      "epoch": 3.640125720272394,
+      "grad_norm": 0.92076575756073,
+      "learning_rate": 4.905656308351536e-07,
+      "loss": 0.0993,
+      "step": 6949
+    },
+    {
+      "epoch": 3.640125720272394,
+      "loss_partial": 0.0400390625,
+      "step": 6949
+    },
+    {
+      "epoch": 3.640125720272394,
+      "loss_partial": 0.04638671875,
+      "step": 6949
+    },
+    {
+      "epoch": 3.640649554740702,
+      "grad_norm": 0.9234386682510376,
+      "learning_rate": 4.891523373608197e-07,
+      "loss": 0.0957,
+      "step": 6950
+    },
+    {
+      "epoch": 3.640649554740702,
+      "loss_partial": 0.062255859375,
+      "step": 6950
+    },
+    {
+      "epoch": 3.640649554740702,
+      "loss_partial": 0.04736328125,
+      "step": 6950
+    },
+    {
+      "epoch": 3.64117338920901,
+      "grad_norm": 1.24509596824646,
+      "learning_rate": 4.87741031594835e-07,
+      "loss": 0.1019,
+      "step": 6951
+    },
+    {
+      "epoch": 3.64117338920901,
+      "loss_partial": 0.04150390625,
+      "step": 6951
+    },
+    {
+      "epoch": 3.64117338920901,
+      "loss_partial": 0.050048828125,
+      "step": 6951
+    },
+    {
+      "epoch": 3.641697223677318,
+      "grad_norm": 0.9802213311195374,
+      "learning_rate": 4.863317138321499e-07,
+      "loss": 0.0976,
+      "step": 6952
+    },
+    {
+      "epoch": 3.641697223677318,
+      "loss_partial": 0.04443359375,
+      "step": 6952
+    },
+    {
+      "epoch": 3.641697223677318,
+      "loss_partial": 0.047119140625,
+      "step": 6952
+    },
+    {
+      "epoch": 3.6422210581456262,
+      "grad_norm": 0.791880190372467,
+      "learning_rate": 4.84924384367308e-07,
+      "loss": 0.0966,
+      "step": 6953
+    },
+    {
+      "epoch": 3.6422210581456262,
+      "loss_partial": 0.0419921875,
+      "step": 6953
+    },
+    {
+      "epoch": 3.6422210581456262,
+      "loss_partial": 0.04296875,
+      "step": 6953
+    },
+    {
+      "epoch": 3.642744892613934,
+      "grad_norm": 0.8598478436470032,
+      "learning_rate": 4.835190434944303e-07,
+      "loss": 0.0938,
+      "step": 6954
+    },
+    {
+      "epoch": 3.642744892613934,
+      "loss_partial": 0.050537109375,
+      "step": 6954
+    },
+    {
+      "epoch": 3.642744892613934,
+      "loss_partial": 0.0458984375,
+      "step": 6954
+    },
+    {
+      "epoch": 3.643268727082242,
+      "grad_norm": 0.9732964038848877,
+      "learning_rate": 4.821156915072245e-07,
+      "loss": 0.0949,
+      "step": 6955
+    },
+    {
+      "epoch": 3.643268727082242,
+      "loss_partial": 0.048828125,
+      "step": 6955
+    },
+    {
+      "epoch": 3.643268727082242,
+      "loss_partial": 0.05517578125,
+      "step": 6955
+    },
+    {
+      "epoch": 3.64379256155055,
+      "grad_norm": 1.2755067348480225,
+      "learning_rate": 4.807143286989812e-07,
+      "loss": 0.0992,
+      "step": 6956
+    },
+    {
+      "epoch": 3.64379256155055,
+      "loss_partial": 0.045166015625,
+      "step": 6956
+    },
+    {
+      "epoch": 3.64379256155055,
+      "loss_partial": 0.05126953125,
+      "step": 6956
+    },
+    {
+      "epoch": 3.644316396018858,
+      "grad_norm": 0.9180019497871399,
+      "learning_rate": 4.793149553625786e-07,
+      "loss": 0.1006,
+      "step": 6957
+    },
+    {
+      "epoch": 3.644316396018858,
+      "loss_partial": 0.052490234375,
+      "step": 6957
+    },
+    {
+      "epoch": 3.644316396018858,
+      "loss_partial": 0.045654296875,
+      "step": 6957
+    },
+    {
+      "epoch": 3.644840230487166,
+      "grad_norm": 0.9153634309768677,
+      "learning_rate": 4.779175717904772e-07,
+      "loss": 0.1042,
+      "step": 6958
+    },
+    {
+      "epoch": 3.644840230487166,
+      "loss_partial": 0.05078125,
+      "step": 6958
+    },
+    {
+      "epoch": 3.644840230487166,
+      "loss_partial": 0.06396484375,
+      "step": 6958
+    },
+    {
+      "epoch": 3.645364064955474,
+      "grad_norm": 0.8723747134208679,
+      "learning_rate": 4.765221782747187e-07,
+      "loss": 0.1002,
+      "step": 6959
+    },
+    {
+      "epoch": 3.645364064955474,
+      "loss_partial": 0.05078125,
+      "step": 6959
+    },
+    {
+      "epoch": 3.645364064955474,
+      "loss_partial": 0.051513671875,
+      "step": 6959
+    },
+    {
+      "epoch": 3.645887899423782,
+      "grad_norm": 0.9001675248146057,
+      "learning_rate": 4.7512877510693335e-07,
+      "loss": 0.1003,
+      "step": 6960
+    },
+    {
+      "epoch": 3.645887899423782,
+      "loss_partial": 0.053955078125,
+      "step": 6960
+    },
+    {
+      "epoch": 3.645887899423782,
+      "loss_partial": 0.04833984375,
+      "step": 6960
+    },
+    {
+      "epoch": 3.64641173389209,
+      "grad_norm": 0.7873924374580383,
+      "learning_rate": 4.737373625783348e-07,
+      "loss": 0.099,
+      "step": 6961
+    },
+    {
+      "epoch": 3.64641173389209,
+      "loss_partial": 0.046875,
+      "step": 6961
+    },
+    {
+      "epoch": 3.64641173389209,
+      "loss_partial": 0.05322265625,
+      "step": 6961
+    },
+    {
+      "epoch": 3.646935568360398,
+      "grad_norm": 0.7559621930122375,
+      "learning_rate": 4.7234794097971934e-07,
+      "loss": 0.0983,
+      "step": 6962
+    },
+    {
+      "epoch": 3.646935568360398,
+      "loss_partial": 0.041748046875,
+      "step": 6962
+    },
+    {
+      "epoch": 3.646935568360398,
+      "loss_partial": 0.054443359375,
+      "step": 6962
+    },
+    {
+      "epoch": 3.647459402828706,
+      "grad_norm": 1.0081373453140259,
+      "learning_rate": 4.7096051060146584e-07,
+      "loss": 0.1026,
+      "step": 6963
+    },
+    {
+      "epoch": 3.647459402828706,
+      "loss_partial": 0.046142578125,
+      "step": 6963
+    },
+    {
+      "epoch": 3.647459402828706,
+      "loss_partial": 0.0498046875,
+      "step": 6963
+    },
+    {
+      "epoch": 3.647983237297014,
+      "grad_norm": 0.7370396256446838,
+      "learning_rate": 4.69575071733539e-07,
+      "loss": 0.0996,
+      "step": 6964
+    },
+    {
+      "epoch": 3.647983237297014,
+      "loss_partial": 0.05078125,
+      "step": 6964
+    },
+    {
+      "epoch": 3.647983237297014,
+      "loss_partial": 0.0517578125,
+      "step": 6964
+    },
+    {
+      "epoch": 3.648507071765322,
+      "grad_norm": 0.8565558791160583,
+      "learning_rate": 4.681916246654894e-07,
+      "loss": 0.1022,
+      "step": 6965
+    },
+    {
+      "epoch": 3.648507071765322,
+      "loss_partial": 0.05712890625,
+      "step": 6965
+    },
+    {
+      "epoch": 3.648507071765322,
+      "loss_partial": 0.0537109375,
+      "step": 6965
+    },
+    {
+      "epoch": 3.64903090623363,
+      "grad_norm": 0.7159857153892517,
+      "learning_rate": 4.66810169686448e-07,
+      "loss": 0.1014,
+      "step": 6966
+    },
+    {
+      "epoch": 3.64903090623363,
+      "loss_partial": 0.05908203125,
+      "step": 6966
+    },
+    {
+      "epoch": 3.64903090623363,
+      "loss_partial": 0.051025390625,
+      "step": 6966
+    },
+    {
+      "epoch": 3.649554740701938,
+      "grad_norm": 1.1417105197906494,
+      "learning_rate": 4.6543070708512827e-07,
+      "loss": 0.0984,
+      "step": 6967
+    },
+    {
+      "epoch": 3.649554740701938,
+      "loss_partial": 0.056640625,
+      "step": 6967
+    },
+    {
+      "epoch": 3.649554740701938,
+      "loss_partial": 0.047119140625,
+      "step": 6967
+    },
+    {
+      "epoch": 3.6500785751702463,
+      "grad_norm": 0.8989397287368774,
+      "learning_rate": 4.6405323714983184e-07,
+      "loss": 0.1014,
+      "step": 6968
+    },
+    {
+      "epoch": 3.6500785751702463,
+      "loss_partial": 0.043212890625,
+      "step": 6968
+    },
+    {
+      "epoch": 3.6500785751702463,
+      "loss_partial": 0.04150390625,
+      "step": 6968
+    },
+    {
+      "epoch": 3.6506024096385543,
+      "grad_norm": 1.026750087738037,
+      "learning_rate": 4.626777601684396e-07,
+      "loss": 0.0945,
+      "step": 6969
+    },
+    {
+      "epoch": 3.6506024096385543,
+      "loss_partial": 0.044921875,
+      "step": 6969
+    },
+    {
+      "epoch": 3.6506024096385543,
+      "loss_partial": 0.046875,
+      "step": 6969
+    },
+    {
+      "epoch": 3.6511262441068624,
+      "grad_norm": 1.1375234127044678,
+      "learning_rate": 4.6130427642841924e-07,
+      "loss": 0.0966,
+      "step": 6970
+    },
+    {
+      "epoch": 3.6511262441068624,
+      "loss_partial": 0.05078125,
+      "step": 6970
+    },
+    {
+      "epoch": 3.6511262441068624,
+      "loss_partial": 0.048583984375,
+      "step": 6970
+    },
+    {
+      "epoch": 3.6516500785751704,
+      "grad_norm": 1.014006495475769,
+      "learning_rate": 4.59932786216819e-07,
+      "loss": 0.0988,
+      "step": 6971
+    },
+    {
+      "epoch": 3.6516500785751704,
+      "loss_partial": 0.0458984375,
+      "step": 6971
+    },
+    {
+      "epoch": 3.6516500785751704,
+      "loss_partial": 0.047119140625,
+      "step": 6971
+    },
+    {
+      "epoch": 3.6521739130434785,
+      "grad_norm": 1.187825322151184,
+      "learning_rate": 4.5856328982027075e-07,
+      "loss": 0.0994,
+      "step": 6972
+    },
+    {
+      "epoch": 3.6521739130434785,
+      "loss_partial": 0.05224609375,
+      "step": 6972
+    },
+    {
+      "epoch": 3.6521739130434785,
+      "loss_partial": 0.054443359375,
+      "step": 6972
+    },
+    {
+      "epoch": 3.6526977475117866,
+      "grad_norm": 0.8875767588615417,
+      "learning_rate": 4.5719578752499436e-07,
+      "loss": 0.0998,
+      "step": 6973
+    },
+    {
+      "epoch": 3.6526977475117866,
+      "loss_partial": 0.053466796875,
+      "step": 6973
+    },
+    {
+      "epoch": 3.6526977475117866,
+      "loss_partial": 0.055419921875,
+      "step": 6973
+    },
+    {
+      "epoch": 3.653221581980094,
+      "grad_norm": 0.8363103270530701,
+      "learning_rate": 4.5583027961678793e-07,
+      "loss": 0.1013,
+      "step": 6974
+    },
+    {
+      "epoch": 3.653221581980094,
+      "loss_partial": 0.044189453125,
+      "step": 6974
+    },
+    {
+      "epoch": 3.653221581980094,
+      "loss_partial": 0.05517578125,
+      "step": 6974
+    },
+    {
+      "epoch": 3.6537454164484022,
+      "grad_norm": 0.9672171473503113,
+      "learning_rate": 4.5446676638103315e-07,
+      "loss": 0.0979,
+      "step": 6975
+    },
+    {
+      "epoch": 3.6537454164484022,
+      "loss_partial": 0.054443359375,
+      "step": 6975
+    },
+    {
+      "epoch": 3.6537454164484022,
+      "loss_partial": 0.05029296875,
+      "step": 6975
+    },
+    {
+      "epoch": 3.6542692509167103,
+      "grad_norm": 0.6929641366004944,
+      "learning_rate": 4.531052481026943e-07,
+      "loss": 0.0997,
+      "step": 6976
+    },
+    {
+      "epoch": 3.6542692509167103,
+      "loss_partial": 0.045654296875,
+      "step": 6976
+    },
+    {
+      "epoch": 3.6542692509167103,
+      "loss_partial": 0.0458984375,
+      "step": 6976
+    },
+    {
+      "epoch": 3.6547930853850183,
+      "grad_norm": 1.195488691329956,
+      "learning_rate": 4.5174572506632486e-07,
+      "loss": 0.1012,
+      "step": 6977
+    },
+    {
+      "epoch": 3.6547930853850183,
+      "loss_partial": 0.0498046875,
+      "step": 6977
+    },
+    {
+      "epoch": 3.6547930853850183,
+      "loss_partial": 0.05078125,
+      "step": 6977
+    },
+    {
+      "epoch": 3.6553169198533264,
+      "grad_norm": 0.90199875831604,
+      "learning_rate": 4.5038819755605423e-07,
+      "loss": 0.0994,
+      "step": 6978
+    },
+    {
+      "epoch": 3.6553169198533264,
+      "loss_partial": 0.04296875,
+      "step": 6978
+    },
+    {
+      "epoch": 3.6553169198533264,
+      "loss_partial": 0.043701171875,
+      "step": 6978
+    },
+    {
+      "epoch": 3.6558407543216345,
+      "grad_norm": 0.8512787222862244,
+      "learning_rate": 4.490326658555966e-07,
+      "loss": 0.0975,
+      "step": 6979
+    },
+    {
+      "epoch": 3.6558407543216345,
+      "loss_partial": 0.052734375,
+      "step": 6979
+    },
+    {
+      "epoch": 3.6558407543216345,
+      "loss_partial": 0.0419921875,
+      "step": 6979
+    },
+    {
+      "epoch": 3.6563645887899425,
+      "grad_norm": 0.8800835013389587,
+      "learning_rate": 4.476791302482508e-07,
+      "loss": 0.1007,
+      "step": 6980
+    },
+    {
+      "epoch": 3.6563645887899425,
+      "loss_partial": 0.0498046875,
+      "step": 6980
+    },
+    {
+      "epoch": 3.6563645887899425,
+      "loss_partial": 0.057373046875,
+      "step": 6980
+    },
+    {
+      "epoch": 3.65688842325825,
+      "grad_norm": 0.867641806602478,
+      "learning_rate": 4.463275910168963e-07,
+      "loss": 0.0998,
+      "step": 6981
+    },
+    {
+      "epoch": 3.65688842325825,
+      "loss_partial": 0.044677734375,
+      "step": 6981
+    },
+    {
+      "epoch": 3.65688842325825,
+      "loss_partial": 0.05712890625,
+      "step": 6981
+    },
+    {
+      "epoch": 3.657412257726558,
+      "grad_norm": 0.8030350208282471,
+      "learning_rate": 4.4497804844399807e-07,
+      "loss": 0.0981,
+      "step": 6982
+    },
+    {
+      "epoch": 3.657412257726558,
+      "loss_partial": 0.05078125,
+      "step": 6982
+    },
+    {
+      "epoch": 3.657412257726558,
+      "loss_partial": 0.0537109375,
+      "step": 6982
+    },
+    {
+      "epoch": 3.6579360921948663,
+      "grad_norm": 0.8234720826148987,
+      "learning_rate": 4.436305028116006e-07,
+      "loss": 0.0999,
+      "step": 6983
+    },
+    {
+      "epoch": 3.6579360921948663,
+      "loss_partial": 0.05126953125,
+      "step": 6983
+    },
+    {
+      "epoch": 3.6579360921948663,
+      "loss_partial": 0.048828125,
+      "step": 6983
+    },
+    {
+      "epoch": 3.6584599266631743,
+      "grad_norm": 1.0051378011703491,
+      "learning_rate": 4.4228495440133304e-07,
+      "loss": 0.0966,
+      "step": 6984
+    },
+    {
+      "epoch": 3.6584599266631743,
+      "loss_partial": 0.050048828125,
+      "step": 6984
+    },
+    {
+      "epoch": 3.6584599266631743,
+      "loss_partial": 0.052001953125,
+      "step": 6984
+    },
+    {
+      "epoch": 3.6589837611314824,
+      "grad_norm": 0.8061665296554565,
+      "learning_rate": 4.409414034944071e-07,
+      "loss": 0.0979,
+      "step": 6985
+    },
+    {
+      "epoch": 3.6589837611314824,
+      "loss_partial": 0.052001953125,
+      "step": 6985
+    },
+    {
+      "epoch": 3.6589837611314824,
+      "loss_partial": 0.050048828125,
+      "step": 6985
+    },
+    {
+      "epoch": 3.6595075955997904,
+      "grad_norm": 1.0145708322525024,
+      "learning_rate": 4.39599850371617e-07,
+      "loss": 0.1009,
+      "step": 6986
+    },
+    {
+      "epoch": 3.6595075955997904,
+      "loss_partial": 0.04736328125,
+      "step": 6986
+    },
+    {
+      "epoch": 3.6595075955997904,
+      "loss_partial": 0.048828125,
+      "step": 6986
+    },
+    {
+      "epoch": 3.6600314300680985,
+      "grad_norm": 0.8691889643669128,
+      "learning_rate": 4.3826029531333746e-07,
+      "loss": 0.1032,
+      "step": 6987
+    },
+    {
+      "epoch": 3.6600314300680985,
+      "loss_partial": 0.04736328125,
+      "step": 6987
+    },
+    {
+      "epoch": 3.6600314300680985,
+      "loss_partial": 0.04736328125,
+      "step": 6987
+    },
+    {
+      "epoch": 3.6605552645364066,
+      "grad_norm": 0.7640069723129272,
+      "learning_rate": 4.369227385995278e-07,
+      "loss": 0.0963,
+      "step": 6988
+    },
+    {
+      "epoch": 3.6605552645364066,
+      "loss_partial": 0.054931640625,
+      "step": 6988
+    },
+    {
+      "epoch": 3.6605552645364066,
+      "loss_partial": 0.04736328125,
+      "step": 6988
+    },
+    {
+      "epoch": 3.6610790990047146,
+      "grad_norm": 0.8818397521972656,
+      "learning_rate": 4.3558718050973e-07,
+      "loss": 0.0988,
+      "step": 6989
+    },
+    {
+      "epoch": 3.6610790990047146,
+      "loss_partial": 0.046630859375,
+      "step": 6989
+    },
+    {
+      "epoch": 3.6610790990047146,
+      "loss_partial": 0.046630859375,
+      "step": 6989
+    },
+    {
+      "epoch": 3.6616029334730227,
+      "grad_norm": 1.0476740598678589,
+      "learning_rate": 4.342536213230675e-07,
+      "loss": 0.0952,
+      "step": 6990
+    },
+    {
+      "epoch": 3.6616029334730227,
+      "loss_partial": 0.04443359375,
+      "step": 6990
+    },
+    {
+      "epoch": 3.6616029334730227,
+      "loss_partial": 0.04833984375,
+      "step": 6990
+    },
+    {
+      "epoch": 3.6621267679413307,
+      "grad_norm": 0.8392003774642944,
+      "learning_rate": 4.329220613182461e-07,
+      "loss": 0.1027,
+      "step": 6991
+    },
+    {
+      "epoch": 3.6621267679413307,
+      "loss_partial": 0.051513671875,
+      "step": 6991
+    },
+    {
+      "epoch": 3.6621267679413307,
+      "loss_partial": 0.046875,
+      "step": 6991
+    },
+    {
+      "epoch": 3.662650602409639,
+      "grad_norm": 1.170973539352417,
+      "learning_rate": 4.3159250077355107e-07,
+      "loss": 0.0978,
+      "step": 6992
+    },
+    {
+      "epoch": 3.662650602409639,
+      "loss_partial": 0.05712890625,
+      "step": 6992
+    },
+    {
+      "epoch": 3.662650602409639,
+      "loss_partial": 0.052978515625,
+      "step": 6992
+    },
+    {
+      "epoch": 3.6631744368779464,
+      "grad_norm": 0.7489129304885864,
+      "learning_rate": 4.302649399668557e-07,
+      "loss": 0.0974,
+      "step": 6993
+    },
+    {
+      "epoch": 3.6631744368779464,
+      "loss_partial": 0.04248046875,
+      "step": 6993
+    },
+    {
+      "epoch": 3.6631744368779464,
+      "loss_partial": 0.056640625,
+      "step": 6993
+    },
+    {
+      "epoch": 3.6636982713462545,
+      "grad_norm": 1.0039618015289307,
+      "learning_rate": 4.2893937917561026e-07,
+      "loss": 0.1003,
+      "step": 6994
+    },
+    {
+      "epoch": 3.6636982713462545,
+      "loss_partial": 0.045166015625,
+      "step": 6994
+    },
+    {
+      "epoch": 3.6636982713462545,
+      "loss_partial": 0.056396484375,
+      "step": 6994
+    },
+    {
+      "epoch": 3.6642221058145625,
+      "grad_norm": 0.9261347055435181,
+      "learning_rate": 4.2761581867684755e-07,
+      "loss": 0.1003,
+      "step": 6995
+    },
+    {
+      "epoch": 3.6642221058145625,
+      "loss_partial": 0.05810546875,
+      "step": 6995
+    },
+    {
+      "epoch": 3.6642221058145625,
+      "loss_partial": 0.045166015625,
+      "step": 6995
+    },
+    {
+      "epoch": 3.6647459402828706,
+      "grad_norm": 0.9506723284721375,
+      "learning_rate": 4.262942587471852e-07,
+      "loss": 0.0984,
+      "step": 6996
+    },
+    {
+      "epoch": 3.6647459402828706,
+      "loss_partial": 0.045166015625,
+      "step": 6996
+    },
+    {
+      "epoch": 3.6647459402828706,
+      "loss_partial": 0.06005859375,
+      "step": 6996
+    },
+    {
+      "epoch": 3.6652697747511787,
+      "grad_norm": 0.9592220187187195,
+      "learning_rate": 4.2497469966282125e-07,
+      "loss": 0.0987,
+      "step": 6997
+    },
+    {
+      "epoch": 3.6652697747511787,
+      "loss_partial": 0.046630859375,
+      "step": 6997
+    },
+    {
+      "epoch": 3.6652697747511787,
+      "loss_partial": 0.059326171875,
+      "step": 6997
+    },
+    {
+      "epoch": 3.6657936092194867,
+      "grad_norm": 1.029296875,
+      "learning_rate": 4.23657141699535e-07,
+      "loss": 0.1023,
+      "step": 6998
+    },
+    {
+      "epoch": 3.6657936092194867,
+      "loss_partial": 0.048095703125,
+      "step": 6998
+    },
+    {
+      "epoch": 3.6657936092194867,
+      "loss_partial": 0.045654296875,
+      "step": 6998
+    },
+    {
+      "epoch": 3.6663174436877948,
+      "grad_norm": 0.7102771401405334,
+      "learning_rate": 4.2234158513268733e-07,
+      "loss": 0.0995,
+      "step": 6999
+    },
+    {
+      "epoch": 3.6663174436877948,
+      "loss_partial": 0.044677734375,
+      "step": 6999
+    },
+    {
+      "epoch": 3.6663174436877948,
+      "loss_partial": 0.040283203125,
+      "step": 6999
+    },
+    {
+      "epoch": 3.666841278156103,
+      "grad_norm": 0.8308125734329224,
+      "learning_rate": 4.210280302372216e-07,
+      "loss": 0.0979,
+      "step": 7000
+    },
+    {
+      "epoch": 3.666841278156103,
+      "loss_partial": 0.05419921875,
+      "step": 7000
+    },
+    {
+      "epoch": 3.666841278156103,
+      "loss_partial": 0.04931640625,
+      "step": 7000
+    },
+    {
+      "epoch": 3.6673651126244105,
+      "grad_norm": 0.8045811057090759,
+      "learning_rate": 4.1971647728766386e-07,
+      "loss": 0.0982,
+      "step": 7001
+    },
+    {
+      "epoch": 3.6673651126244105,
+      "loss_partial": 0.046630859375,
+      "step": 7001
+    },
+    {
+      "epoch": 3.6673651126244105,
+      "loss_partial": 0.04296875,
+      "step": 7001
+    },
+    {
+      "epoch": 3.6678889470927185,
+      "grad_norm": 0.8190149664878845,
+      "learning_rate": 4.184069265581192e-07,
+      "loss": 0.0973,
+      "step": 7002
+    },
+    {
+      "epoch": 3.6678889470927185,
+      "loss_partial": 0.056396484375,
+      "step": 7002
+    },
+    {
+      "epoch": 3.6678889470927185,
+      "loss_partial": 0.05322265625,
+      "step": 7002
+    },
+    {
+      "epoch": 3.6684127815610266,
+      "grad_norm": 0.7964325547218323,
+      "learning_rate": 4.170993783222765e-07,
+      "loss": 0.0987,
+      "step": 7003
+    },
+    {
+      "epoch": 3.6684127815610266,
+      "loss_partial": 0.052490234375,
+      "step": 7003
+    },
+    {
+      "epoch": 3.6684127815610266,
+      "loss_partial": 0.054931640625,
+      "step": 7003
+    },
+    {
+      "epoch": 3.6689366160293346,
+      "grad_norm": 0.9092559218406677,
+      "learning_rate": 4.1579383285340504e-07,
+      "loss": 0.0989,
+      "step": 7004
+    },
+    {
+      "epoch": 3.6689366160293346,
+      "loss_partial": 0.044189453125,
+      "step": 7004
+    },
+    {
+      "epoch": 3.6689366160293346,
+      "loss_partial": 0.054931640625,
+      "step": 7004
+    },
+    {
+      "epoch": 3.6694604504976427,
+      "grad_norm": 1.081549048423767,
+      "learning_rate": 4.144902904243553e-07,
+      "loss": 0.1008,
+      "step": 7005
+    },
+    {
+      "epoch": 3.6694604504976427,
+      "loss_partial": 0.046630859375,
+      "step": 7005
+    },
+    {
+      "epoch": 3.6694604504976427,
+      "loss_partial": 0.046875,
+      "step": 7005
+    },
+    {
+      "epoch": 3.6699842849659507,
+      "grad_norm": 0.9575638175010681,
+      "learning_rate": 4.1318875130756055e-07,
+      "loss": 0.0978,
+      "step": 7006
+    },
+    {
+      "epoch": 3.6699842849659507,
+      "loss_partial": 0.05810546875,
+      "step": 7006
+    },
+    {
+      "epoch": 3.6699842849659507,
+      "loss_partial": 0.050048828125,
+      "step": 7006
+    },
+    {
+      "epoch": 3.670508119434259,
+      "grad_norm": 0.9125267267227173,
+      "learning_rate": 4.1188921577503205e-07,
+      "loss": 0.0991,
+      "step": 7007
+    },
+    {
+      "epoch": 3.670508119434259,
+      "loss_partial": 0.044921875,
+      "step": 7007
+    },
+    {
+      "epoch": 3.670508119434259,
+      "loss_partial": 0.044189453125,
+      "step": 7007
+    },
+    {
+      "epoch": 3.671031953902567,
+      "grad_norm": 0.9099379181861877,
+      "learning_rate": 4.1059168409836815e-07,
+      "loss": 0.0963,
+      "step": 7008
+    },
+    {
+      "epoch": 3.671031953902567,
+      "loss_partial": 0.0537109375,
+      "step": 7008
+    },
+    {
+      "epoch": 3.671031953902567,
+      "loss_partial": 0.05322265625,
+      "step": 7008
+    },
+    {
+      "epoch": 3.671555788370875,
+      "grad_norm": 0.9910556674003601,
+      "learning_rate": 4.092961565487408e-07,
+      "loss": 0.0967,
+      "step": 7009
+    },
+    {
+      "epoch": 3.671555788370875,
+      "loss_partial": 0.06298828125,
+      "step": 7009
+    },
+    {
+      "epoch": 3.671555788370875,
+      "loss_partial": 0.04833984375,
+      "step": 7009
+    },
+    {
+      "epoch": 3.672079622839183,
+      "grad_norm": 1.0452613830566406,
+      "learning_rate": 4.0800263339691114e-07,
+      "loss": 0.1022,
+      "step": 7010
+    },
+    {
+      "epoch": 3.672079622839183,
+      "loss_partial": 0.046142578125,
+      "step": 7010
+    },
+    {
+      "epoch": 3.672079622839183,
+      "loss_partial": 0.05029296875,
+      "step": 7010
+    },
+    {
+      "epoch": 3.672603457307491,
+      "grad_norm": 0.8442836999893188,
+      "learning_rate": 4.0671111491321633e-07,
+      "loss": 0.0994,
+      "step": 7011
+    },
+    {
+      "epoch": 3.672603457307491,
+      "loss_partial": 0.051025390625,
+      "step": 7011
+    },
+    {
+      "epoch": 3.672603457307491,
+      "loss_partial": 0.046875,
+      "step": 7011
+    },
+    {
+      "epoch": 3.673127291775799,
+      "grad_norm": 0.6505481004714966,
+      "learning_rate": 4.054216013675749e-07,
+      "loss": 0.0987,
+      "step": 7012
+    },
+    {
+      "epoch": 3.673127291775799,
+      "loss_partial": 0.04052734375,
+      "step": 7012
+    },
+    {
+      "epoch": 3.673127291775799,
+      "loss_partial": 0.05029296875,
+      "step": 7012
+    },
+    {
+      "epoch": 3.6736511262441067,
+      "grad_norm": 0.9362701177597046,
+      "learning_rate": 4.041340930294868e-07,
+      "loss": 0.0977,
+      "step": 7013
+    },
+    {
+      "epoch": 3.6736511262441067,
+      "loss_partial": 0.05078125,
+      "step": 7013
+    },
+    {
+      "epoch": 3.6736511262441067,
+      "loss_partial": 0.06201171875,
+      "step": 7013
+    },
+    {
+      "epoch": 3.674174960712415,
+      "grad_norm": 1.0643036365509033,
+      "learning_rate": 4.0284859016803567e-07,
+      "loss": 0.1004,
+      "step": 7014
+    },
+    {
+      "epoch": 3.674174960712415,
+      "loss_partial": 0.05322265625,
+      "step": 7014
+    },
+    {
+      "epoch": 3.674174960712415,
+      "loss_partial": 0.052490234375,
+      "step": 7014
+    },
+    {
+      "epoch": 3.674698795180723,
+      "grad_norm": 0.9071156978607178,
+      "learning_rate": 4.015650930518811e-07,
+      "loss": 0.1032,
+      "step": 7015
+    },
+    {
+      "epoch": 3.674698795180723,
+      "loss_partial": 0.05810546875,
+      "step": 7015
+    },
+    {
+      "epoch": 3.674698795180723,
+      "loss_partial": 0.039794921875,
+      "step": 7015
+    },
+    {
+      "epoch": 3.675222629649031,
+      "grad_norm": 1.0311230421066284,
+      "learning_rate": 4.002836019492684e-07,
+      "loss": 0.1023,
+      "step": 7016
+    },
+    {
+      "epoch": 3.675222629649031,
+      "loss_partial": 0.044921875,
+      "step": 7016
+    },
+    {
+      "epoch": 3.675222629649031,
+      "loss_partial": 0.04248046875,
+      "step": 7016
+    },
+    {
+      "epoch": 3.675746464117339,
+      "grad_norm": 1.0556442737579346,
+      "learning_rate": 3.99004117128019e-07,
+      "loss": 0.0953,
+      "step": 7017
+    },
+    {
+      "epoch": 3.675746464117339,
+      "loss_partial": 0.050048828125,
+      "step": 7017
+    },
+    {
+      "epoch": 3.675746464117339,
+      "loss_partial": 0.057373046875,
+      "step": 7017
+    },
+    {
+      "epoch": 3.676270298585647,
+      "grad_norm": 0.8476938605308533,
+      "learning_rate": 3.97726638855539e-07,
+      "loss": 0.0988,
+      "step": 7018
+    },
+    {
+      "epoch": 3.676270298585647,
+      "loss_partial": 0.055908203125,
+      "step": 7018
+    },
+    {
+      "epoch": 3.676270298585647,
+      "loss_partial": 0.047119140625,
+      "step": 7018
+    },
+    {
+      "epoch": 3.676794133053955,
+      "grad_norm": 0.9599350690841675,
+      "learning_rate": 3.964511673988125e-07,
+      "loss": 0.1008,
+      "step": 7019
+    },
+    {
+      "epoch": 3.676794133053955,
+      "loss_partial": 0.05029296875,
+      "step": 7019
+    },
+    {
+      "epoch": 3.676794133053955,
+      "loss_partial": 0.04931640625,
+      "step": 7019
+    },
+    {
+      "epoch": 3.6773179675222627,
+      "grad_norm": 1.0290316343307495,
+      "learning_rate": 3.951777030244064e-07,
+      "loss": 0.101,
+      "step": 7020
+    },
+    {
+      "epoch": 3.6773179675222627,
+      "loss_partial": 0.05322265625,
+      "step": 7020
+    },
+    {
+      "epoch": 3.6773179675222627,
+      "loss_partial": 0.049560546875,
+      "step": 7020
+    },
+    {
+      "epoch": 3.6778418019905708,
+      "grad_norm": 0.8971187472343445,
+      "learning_rate": 3.9390624599846327e-07,
+      "loss": 0.1017,
+      "step": 7021
+    },
+    {
+      "epoch": 3.6778418019905708,
+      "loss_partial": 0.052734375,
+      "step": 7021
+    },
+    {
+      "epoch": 3.6778418019905708,
+      "loss_partial": 0.05224609375,
+      "step": 7021
+    },
+    {
+      "epoch": 3.678365636458879,
+      "grad_norm": 0.9947978258132935,
+      "learning_rate": 3.926367965867151e-07,
+      "loss": 0.1001,
+      "step": 7022
+    },
+    {
+      "epoch": 3.678365636458879,
+      "loss_partial": 0.0498046875,
+      "step": 7022
+    },
+    {
+      "epoch": 3.678365636458879,
+      "loss_partial": 0.05224609375,
+      "step": 7022
+    },
+    {
+      "epoch": 3.678889470927187,
+      "grad_norm": 0.7300689220428467,
+      "learning_rate": 3.9136935505446505e-07,
+      "loss": 0.098,
+      "step": 7023
+    },
+    {
+      "epoch": 3.678889470927187,
+      "loss_partial": 0.05517578125,
+      "step": 7023
+    },
+    {
+      "epoch": 3.678889470927187,
+      "loss_partial": 0.0556640625,
+      "step": 7023
+    },
+    {
+      "epoch": 3.679413305395495,
+      "grad_norm": 0.9493797421455383,
+      "learning_rate": 3.901039216666025e-07,
+      "loss": 0.0984,
+      "step": 7024
+    },
+    {
+      "epoch": 3.679413305395495,
+      "loss_partial": 0.04150390625,
+      "step": 7024
+    },
+    {
+      "epoch": 3.679413305395495,
+      "loss_partial": 0.049560546875,
+      "step": 7024
+    },
+    {
+      "epoch": 3.679937139863803,
+      "grad_norm": 0.9585814476013184,
+      "learning_rate": 3.888404966875925e-07,
+      "loss": 0.0954,
+      "step": 7025
+    },
+    {
+      "epoch": 3.679937139863803,
+      "loss_partial": 0.04931640625,
+      "step": 7025
+    },
+    {
+      "epoch": 3.679937139863803,
+      "loss_partial": 0.05859375,
+      "step": 7025
+    },
+    {
+      "epoch": 3.680460974332111,
+      "grad_norm": 0.8448718786239624,
+      "learning_rate": 3.8757908038148715e-07,
+      "loss": 0.1037,
+      "step": 7026
+    },
+    {
+      "epoch": 3.680460974332111,
+      "loss_partial": 0.04296875,
+      "step": 7026
+    },
+    {
+      "epoch": 3.680460974332111,
+      "loss_partial": 0.054931640625,
+      "step": 7026
+    },
+    {
+      "epoch": 3.680984808800419,
+      "grad_norm": 1.0890240669250488,
+      "learning_rate": 3.863196730119123e-07,
+      "loss": 0.1043,
+      "step": 7027
+    },
+    {
+      "epoch": 3.680984808800419,
+      "loss_partial": 0.051513671875,
+      "step": 7027
+    },
+    {
+      "epoch": 3.680984808800419,
+      "loss_partial": 0.06298828125,
+      "step": 7027
+    },
+    {
+      "epoch": 3.681508643268727,
+      "grad_norm": 0.8327417969703674,
+      "learning_rate": 3.8506227484207624e-07,
+      "loss": 0.0992,
+      "step": 7028
+    },
+    {
+      "epoch": 3.681508643268727,
+      "loss_partial": 0.04736328125,
+      "step": 7028
+    },
+    {
+      "epoch": 3.681508643268727,
+      "loss_partial": 0.052734375,
+      "step": 7028
+    },
+    {
+      "epoch": 3.6820324777370352,
+      "grad_norm": 1.0689846277236938,
+      "learning_rate": 3.838068861347688e-07,
+      "loss": 0.0987,
+      "step": 7029
+    },
+    {
+      "epoch": 3.6820324777370352,
+      "loss_partial": 0.048583984375,
+      "step": 7029
+    },
+    {
+      "epoch": 3.6820324777370352,
+      "loss_partial": 0.045166015625,
+      "step": 7029
+    },
+    {
+      "epoch": 3.6825563122053433,
+      "grad_norm": 0.7608569860458374,
+      "learning_rate": 3.8255350715235673e-07,
+      "loss": 0.0994,
+      "step": 7030
+    },
+    {
+      "epoch": 3.6825563122053433,
+      "loss_partial": 0.052978515625,
+      "step": 7030
+    },
+    {
+      "epoch": 3.6825563122053433,
+      "loss_partial": 0.05517578125,
+      "step": 7030
+    },
+    {
+      "epoch": 3.6830801466736514,
+      "grad_norm": 0.7352872490882874,
+      "learning_rate": 3.8130213815678827e-07,
+      "loss": 0.0961,
+      "step": 7031
+    },
+    {
+      "epoch": 3.6830801466736514,
+      "loss_partial": 0.045654296875,
+      "step": 7031
+    },
+    {
+      "epoch": 3.6830801466736514,
+      "loss_partial": 0.052490234375,
+      "step": 7031
+    },
+    {
+      "epoch": 3.6836039811419594,
+      "grad_norm": 1.2094796895980835,
+      "learning_rate": 3.800527794095943e-07,
+      "loss": 0.0987,
+      "step": 7032
+    },
+    {
+      "epoch": 3.6836039811419594,
+      "loss_partial": 0.05078125,
+      "step": 7032
+    },
+    {
+      "epoch": 3.6836039811419594,
+      "loss_partial": 0.05859375,
+      "step": 7032
+    },
+    {
+      "epoch": 3.684127815610267,
+      "grad_norm": 1.0531474351882935,
+      "learning_rate": 3.7880543117187917e-07,
+      "loss": 0.0999,
+      "step": 7033
+    },
+    {
+      "epoch": 3.684127815610267,
+      "loss_partial": 0.047607421875,
+      "step": 7033
+    },
+    {
+      "epoch": 3.684127815610267,
+      "loss_partial": 0.047119140625,
+      "step": 7033
+    },
+    {
+      "epoch": 3.684651650078575,
+      "grad_norm": 1.0063157081604004,
+      "learning_rate": 3.7756009370433444e-07,
+      "loss": 0.0991,
+      "step": 7034
+    },
+    {
+      "epoch": 3.684651650078575,
+      "loss_partial": 0.056396484375,
+      "step": 7034
+    },
+    {
+      "epoch": 3.684651650078575,
+      "loss_partial": 0.0458984375,
+      "step": 7034
+    },
+    {
+      "epoch": 3.685175484546883,
+      "grad_norm": 0.7185530066490173,
+      "learning_rate": 3.7631676726722746e-07,
+      "loss": 0.0983,
+      "step": 7035
+    },
+    {
+      "epoch": 3.685175484546883,
+      "loss_partial": 0.044921875,
+      "step": 7035
+    },
+    {
+      "epoch": 3.685175484546883,
+      "loss_partial": 0.060791015625,
+      "step": 7035
+    },
+    {
+      "epoch": 3.685699319015191,
+      "grad_norm": 1.0967986583709717,
+      "learning_rate": 3.7507545212040474e-07,
+      "loss": 0.1002,
+      "step": 7036
+    },
+    {
+      "epoch": 3.685699319015191,
+      "loss_partial": 0.04443359375,
+      "step": 7036
+    },
+    {
+      "epoch": 3.685699319015191,
+      "loss_partial": 0.048095703125,
+      "step": 7036
+    },
+    {
+      "epoch": 3.6862231534834993,
+      "grad_norm": 0.8785573840141296,
+      "learning_rate": 3.738361485232922e-07,
+      "loss": 0.1022,
+      "step": 7037
+    },
+    {
+      "epoch": 3.6862231534834993,
+      "loss_partial": 0.05322265625,
+      "step": 7037
+    },
+    {
+      "epoch": 3.6862231534834993,
+      "loss_partial": 0.056884765625,
+      "step": 7037
+    },
+    {
+      "epoch": 3.6867469879518073,
+      "grad_norm": 0.7648268342018127,
+      "learning_rate": 3.7259885673490036e-07,
+      "loss": 0.0988,
+      "step": 7038
+    },
+    {
+      "epoch": 3.6867469879518073,
+      "loss_partial": 0.048095703125,
+      "step": 7038
+    },
+    {
+      "epoch": 3.6867469879518073,
+      "loss_partial": 0.052490234375,
+      "step": 7038
+    },
+    {
+      "epoch": 3.6872708224201154,
+      "grad_norm": 1.1178176403045654,
+      "learning_rate": 3.713635770138124e-07,
+      "loss": 0.1015,
+      "step": 7039
+    },
+    {
+      "epoch": 3.6872708224201154,
+      "loss_partial": 0.0517578125,
+      "step": 7039
+    },
+    {
+      "epoch": 3.6872708224201154,
+      "loss_partial": 0.052978515625,
+      "step": 7039
+    },
+    {
+      "epoch": 3.687794656888423,
+      "grad_norm": 0.8499972224235535,
+      "learning_rate": 3.701303096181963e-07,
+      "loss": 0.0969,
+      "step": 7040
+    },
+    {
+      "epoch": 3.687794656888423,
+      "loss_partial": 0.04833984375,
+      "step": 7040
+    },
+    {
+      "epoch": 3.687794656888423,
+      "loss_partial": 0.055419921875,
+      "step": 7040
+    },
+    {
+      "epoch": 3.688318491356731,
+      "grad_norm": 0.8860012888908386,
+      "learning_rate": 3.688990548057947e-07,
+      "loss": 0.0984,
+      "step": 7041
+    },
+    {
+      "epoch": 3.688318491356731,
+      "loss_partial": 0.0546875,
+      "step": 7041
+    },
+    {
+      "epoch": 3.688318491356731,
+      "loss_partial": 0.0556640625,
+      "step": 7041
+    },
+    {
+      "epoch": 3.688842325825039,
+      "grad_norm": 0.8909804821014404,
+      "learning_rate": 3.676698128339351e-07,
+      "loss": 0.103,
+      "step": 7042
+    },
+    {
+      "epoch": 3.688842325825039,
+      "loss_partial": 0.051513671875,
+      "step": 7042
+    },
+    {
+      "epoch": 3.688842325825039,
+      "loss_partial": 0.051025390625,
+      "step": 7042
+    },
+    {
+      "epoch": 3.689366160293347,
+      "grad_norm": 0.9006908535957336,
+      "learning_rate": 3.664425839595187e-07,
+      "loss": 0.1003,
+      "step": 7043
+    },
+    {
+      "epoch": 3.689366160293347,
+      "loss_partial": 0.045166015625,
+      "step": 7043
+    },
+    {
+      "epoch": 3.689366160293347,
+      "loss_partial": 0.04833984375,
+      "step": 7043
+    },
+    {
+      "epoch": 3.6898899947616552,
+      "grad_norm": 0.8037259578704834,
+      "learning_rate": 3.6521736843903144e-07,
+      "loss": 0.0975,
+      "step": 7044
+    },
+    {
+      "epoch": 3.6898899947616552,
+      "loss_partial": 0.049072265625,
+      "step": 7044
+    },
+    {
+      "epoch": 3.6898899947616552,
+      "loss_partial": 0.0478515625,
+      "step": 7044
+    },
+    {
+      "epoch": 3.6904138292299633,
+      "grad_norm": 0.992772102355957,
+      "learning_rate": 3.6399416652853293e-07,
+      "loss": 0.0961,
+      "step": 7045
+    },
+    {
+      "epoch": 3.6904138292299633,
+      "loss_partial": 0.050537109375,
+      "step": 7045
+    },
+    {
+      "epoch": 3.6904138292299633,
+      "loss_partial": 0.051513671875,
+      "step": 7045
+    },
+    {
+      "epoch": 3.6909376636982714,
+      "grad_norm": 0.7631925940513611,
+      "learning_rate": 3.627729784836664e-07,
+      "loss": 0.0992,
+      "step": 7046
+    },
+    {
+      "epoch": 3.6909376636982714,
+      "loss_partial": 0.048583984375,
+      "step": 7046
+    },
+    {
+      "epoch": 3.6909376636982714,
+      "loss_partial": 0.05322265625,
+      "step": 7046
+    },
+    {
+      "epoch": 3.6914614981665794,
+      "grad_norm": 1.009413242340088,
+      "learning_rate": 3.615538045596534e-07,
+      "loss": 0.1016,
+      "step": 7047
+    },
+    {
+      "epoch": 3.6914614981665794,
+      "loss_partial": 0.049560546875,
+      "step": 7047
+    },
+    {
+      "epoch": 3.6914614981665794,
+      "loss_partial": 0.060546875,
+      "step": 7047
+    },
+    {
+      "epoch": 3.6919853326348875,
+      "grad_norm": 0.8372737765312195,
+      "learning_rate": 3.603366450112933e-07,
+      "loss": 0.1021,
+      "step": 7048
+    },
+    {
+      "epoch": 3.6919853326348875,
+      "loss_partial": 0.05322265625,
+      "step": 7048
+    },
+    {
+      "epoch": 3.6919853326348875,
+      "loss_partial": 0.04638671875,
+      "step": 7048
+    },
+    {
+      "epoch": 3.6925091671031955,
+      "grad_norm": 0.8925758004188538,
+      "learning_rate": 3.5912150009296265e-07,
+      "loss": 0.0988,
+      "step": 7049
+    },
+    {
+      "epoch": 3.6925091671031955,
+      "loss_partial": 0.043701171875,
+      "step": 7049
+    },
+    {
+      "epoch": 3.6925091671031955,
+      "loss_partial": 0.03759765625,
+      "step": 7049
+    },
+    {
+      "epoch": 3.6930330015715036,
+      "grad_norm": 0.8749141097068787,
+      "learning_rate": 3.579083700586228e-07,
+      "loss": 0.0993,
+      "step": 7050
+    },
+    {
+      "epoch": 3.6930330015715036,
+      "loss_partial": 0.04736328125,
+      "step": 7050
+    },
+    {
+      "epoch": 3.6930330015715036,
+      "loss_partial": 0.05029296875,
+      "step": 7050
+    },
+    {
+      "epoch": 3.6935568360398117,
+      "grad_norm": 0.9127289056777954,
+      "learning_rate": 3.5669725516180866e-07,
+      "loss": 0.1002,
+      "step": 7051
+    },
+    {
+      "epoch": 3.6935568360398117,
+      "loss_partial": 0.044921875,
+      "step": 7051
+    },
+    {
+      "epoch": 3.6935568360398117,
+      "loss_partial": 0.0498046875,
+      "step": 7051
+    },
+    {
+      "epoch": 3.6940806705081193,
+      "grad_norm": 1.3551580905914307,
+      "learning_rate": 3.554881556556355e-07,
+      "loss": 0.1037,
+      "step": 7052
+    },
+    {
+      "epoch": 3.6940806705081193,
+      "loss_partial": 0.052490234375,
+      "step": 7052
+    },
+    {
+      "epoch": 3.6940806705081193,
+      "loss_partial": 0.046875,
+      "step": 7052
+    },
+    {
+      "epoch": 3.6946045049764273,
+      "grad_norm": 1.0567266941070557,
+      "learning_rate": 3.5428107179279805e-07,
+      "loss": 0.0959,
+      "step": 7053
+    },
+    {
+      "epoch": 3.6946045049764273,
+      "loss_partial": 0.05224609375,
+      "step": 7053
+    },
+    {
+      "epoch": 3.6946045049764273,
+      "loss_partial": 0.041259765625,
+      "step": 7053
+    },
+    {
+      "epoch": 3.6951283394447354,
+      "grad_norm": 0.9645189642906189,
+      "learning_rate": 3.530760038255698e-07,
+      "loss": 0.1009,
+      "step": 7054
+    },
+    {
+      "epoch": 3.6951283394447354,
+      "loss_partial": 0.05322265625,
+      "step": 7054
+    },
+    {
+      "epoch": 3.6951283394447354,
+      "loss_partial": 0.046142578125,
+      "step": 7054
+    },
+    {
+      "epoch": 3.6956521739130435,
+      "grad_norm": 1.0131651163101196,
+      "learning_rate": 3.5187295200580064e-07,
+      "loss": 0.1006,
+      "step": 7055
+    },
+    {
+      "epoch": 3.6956521739130435,
+      "loss_partial": 0.055908203125,
+      "step": 7055
+    },
+    {
+      "epoch": 3.6956521739130435,
+      "loss_partial": 0.04931640625,
+      "step": 7055
+    },
+    {
+      "epoch": 3.6961760083813515,
+      "grad_norm": 1.0603179931640625,
+      "learning_rate": 3.506719165849237e-07,
+      "loss": 0.1011,
+      "step": 7056
+    },
+    {
+      "epoch": 3.6961760083813515,
+      "loss_partial": 0.055419921875,
+      "step": 7056
+    },
+    {
+      "epoch": 3.6961760083813515,
+      "loss_partial": 0.04052734375,
+      "step": 7056
+    },
+    {
+      "epoch": 3.6966998428496596,
+      "grad_norm": 1.001496434211731,
+      "learning_rate": 3.49472897813945e-07,
+      "loss": 0.1012,
+      "step": 7057
+    },
+    {
+      "epoch": 3.6966998428496596,
+      "loss_partial": 0.0458984375,
+      "step": 7057
+    },
+    {
+      "epoch": 3.6966998428496596,
+      "loss_partial": 0.058837890625,
+      "step": 7057
+    },
+    {
+      "epoch": 3.6972236773179676,
+      "grad_norm": 1.0895674228668213,
+      "learning_rate": 3.48275895943454e-07,
+      "loss": 0.0981,
+      "step": 7058
+    },
+    {
+      "epoch": 3.6972236773179676,
+      "loss_partial": 0.054931640625,
+      "step": 7058
+    },
+    {
+      "epoch": 3.6972236773179676,
+      "loss_partial": 0.047119140625,
+      "step": 7058
+    },
+    {
+      "epoch": 3.6977475117862757,
+      "grad_norm": 1.1824145317077637,
+      "learning_rate": 3.470809112236173e-07,
+      "loss": 0.0994,
+      "step": 7059
+    },
+    {
+      "epoch": 3.6977475117862757,
+      "loss_partial": 0.04443359375,
+      "step": 7059
+    },
+    {
+      "epoch": 3.6977475117862757,
+      "loss_partial": 0.05517578125,
+      "step": 7059
+    },
+    {
+      "epoch": 3.6982713462545833,
+      "grad_norm": 1.252869725227356,
+      "learning_rate": 3.458879439041773e-07,
+      "loss": 0.094,
+      "step": 7060
+    },
+    {
+      "epoch": 3.6982713462545833,
+      "loss_partial": 0.05712890625,
+      "step": 7060
+    },
+    {
+      "epoch": 3.6982713462545833,
+      "loss_partial": 0.044189453125,
+      "step": 7060
+    },
+    {
+      "epoch": 3.6987951807228914,
+      "grad_norm": 1.5228301286697388,
+      "learning_rate": 3.4469699423445693e-07,
+      "loss": 0.1024,
+      "step": 7061
+    },
+    {
+      "epoch": 3.6987951807228914,
+      "loss_partial": 0.046142578125,
+      "step": 7061
+    },
+    {
+      "epoch": 3.6987951807228914,
+      "loss_partial": 0.0458984375,
+      "step": 7061
+    },
+    {
+      "epoch": 3.6993190151911994,
+      "grad_norm": 1.0018982887268066,
+      "learning_rate": 3.435080624633602e-07,
+      "loss": 0.0998,
+      "step": 7062
+    },
+    {
+      "epoch": 3.6993190151911994,
+      "loss_partial": 0.05029296875,
+      "step": 7062
+    },
+    {
+      "epoch": 3.6993190151911994,
+      "loss_partial": 0.051025390625,
+      "step": 7062
+    },
+    {
+      "epoch": 3.6998428496595075,
+      "grad_norm": 0.759630560874939,
+      "learning_rate": 3.4232114883936406e-07,
+      "loss": 0.0941,
+      "step": 7063
+    },
+    {
+      "epoch": 3.6998428496595075,
+      "loss_partial": 0.051025390625,
+      "step": 7063
+    },
+    {
+      "epoch": 3.6998428496595075,
+      "loss_partial": 0.06298828125,
+      "step": 7063
+    },
+    {
+      "epoch": 3.7003666841278156,
+      "grad_norm": 0.8340038061141968,
+      "learning_rate": 3.411362536105256e-07,
+      "loss": 0.1024,
+      "step": 7064
+    },
+    {
+      "epoch": 3.7003666841278156,
+      "loss_partial": 0.05322265625,
+      "step": 7064
+    },
+    {
+      "epoch": 3.7003666841278156,
+      "loss_partial": 0.0458984375,
+      "step": 7064
+    },
+    {
+      "epoch": 3.7008905185961236,
+      "grad_norm": 1.0935779809951782,
+      "learning_rate": 3.3995337702448226e-07,
+      "loss": 0.1006,
+      "step": 7065
+    },
+    {
+      "epoch": 3.7008905185961236,
+      "loss_partial": 0.048828125,
+      "step": 7065
+    },
+    {
+      "epoch": 3.7008905185961236,
+      "loss_partial": 0.045166015625,
+      "step": 7065
+    },
+    {
+      "epoch": 3.7014143530644317,
+      "grad_norm": 0.8765901923179626,
+      "learning_rate": 3.3877251932844744e-07,
+      "loss": 0.0994,
+      "step": 7066
+    },
+    {
+      "epoch": 3.7014143530644317,
+      "loss_partial": 0.044921875,
+      "step": 7066
+    },
+    {
+      "epoch": 3.7014143530644317,
+      "loss_partial": 0.043701171875,
+      "step": 7066
+    },
+    {
+      "epoch": 3.7019381875327397,
+      "grad_norm": 0.8998542428016663,
+      "learning_rate": 3.375936807692126e-07,
+      "loss": 0.0997,
+      "step": 7067
+    },
+    {
+      "epoch": 3.7019381875327397,
+      "loss_partial": 0.052490234375,
+      "step": 7067
+    },
+    {
+      "epoch": 3.7019381875327397,
+      "loss_partial": 0.06494140625,
+      "step": 7067
+    },
+    {
+      "epoch": 3.702462022001048,
+      "grad_norm": 0.9322919249534607,
+      "learning_rate": 3.3641686159314735e-07,
+      "loss": 0.1011,
+      "step": 7068
+    },
+    {
+      "epoch": 3.702462022001048,
+      "loss_partial": 0.0478515625,
+      "step": 7068
+    },
+    {
+      "epoch": 3.702462022001048,
+      "loss_partial": 0.05810546875,
+      "step": 7068
+    },
+    {
+      "epoch": 3.702985856469356,
+      "grad_norm": 0.9383575916290283,
+      "learning_rate": 3.3524206204620156e-07,
+      "loss": 0.0968,
+      "step": 7069
+    },
+    {
+      "epoch": 3.702985856469356,
+      "loss_partial": 0.04833984375,
+      "step": 7069
+    },
+    {
+      "epoch": 3.702985856469356,
+      "loss_partial": 0.045654296875,
+      "step": 7069
+    },
+    {
+      "epoch": 3.703509690937664,
+      "grad_norm": 0.82320237159729,
+      "learning_rate": 3.340692823738978e-07,
+      "loss": 0.0963,
+      "step": 7070
+    },
+    {
+      "epoch": 3.703509690937664,
+      "loss_partial": 0.051025390625,
+      "step": 7070
+    },
+    {
+      "epoch": 3.703509690937664,
+      "loss_partial": 0.05078125,
+      "step": 7070
+    },
+    {
+      "epoch": 3.704033525405972,
+      "grad_norm": 0.9946268796920776,
+      "learning_rate": 3.328985228213422e-07,
+      "loss": 0.0977,
+      "step": 7071
+    },
+    {
+      "epoch": 3.704033525405972,
+      "loss_partial": 0.053955078125,
+      "step": 7071
+    },
+    {
+      "epoch": 3.704033525405972,
+      "loss_partial": 0.04541015625,
+      "step": 7071
+    },
+    {
+      "epoch": 3.7045573598742796,
+      "grad_norm": 1.1896928548812866,
+      "learning_rate": 3.317297836332167e-07,
+      "loss": 0.0977,
+      "step": 7072
+    },
+    {
+      "epoch": 3.7045573598742796,
+      "loss_partial": 0.048583984375,
+      "step": 7072
+    },
+    {
+      "epoch": 3.7045573598742796,
+      "loss_partial": 0.0478515625,
+      "step": 7072
+    },
+    {
+      "epoch": 3.7050811943425876,
+      "grad_norm": 0.8188559412956238,
+      "learning_rate": 3.3056306505377946e-07,
+      "loss": 0.1016,
+      "step": 7073
+    },
+    {
+      "epoch": 3.7050811943425876,
+      "loss_partial": 0.044677734375,
+      "step": 7073
+    },
+    {
+      "epoch": 3.7050811943425876,
+      "loss_partial": 0.052978515625,
+      "step": 7073
+    },
+    {
+      "epoch": 3.7056050288108957,
+      "grad_norm": 0.9102766513824463,
+      "learning_rate": 3.2939836732686634e-07,
+      "loss": 0.1008,
+      "step": 7074
+    },
+    {
+      "epoch": 3.7056050288108957,
+      "loss_partial": 0.0498046875,
+      "step": 7074
+    },
+    {
+      "epoch": 3.7056050288108957,
+      "loss_partial": 0.0478515625,
+      "step": 7074
+    },
+    {
+      "epoch": 3.7061288632792038,
+      "grad_norm": 0.8371280431747437,
+      "learning_rate": 3.282356906958939e-07,
+      "loss": 0.0995,
+      "step": 7075
+    },
+    {
+      "epoch": 3.7061288632792038,
+      "loss_partial": 0.055908203125,
+      "step": 7075
+    },
+    {
+      "epoch": 3.7061288632792038,
+      "loss_partial": 0.046142578125,
+      "step": 7075
+    },
+    {
+      "epoch": 3.706652697747512,
+      "grad_norm": 0.8745279908180237,
+      "learning_rate": 3.2707503540385433e-07,
+      "loss": 0.0982,
+      "step": 7076
+    },
+    {
+      "epoch": 3.706652697747512,
+      "loss_partial": 0.056640625,
+      "step": 7076
+    },
+    {
+      "epoch": 3.706652697747512,
+      "loss_partial": 0.052001953125,
+      "step": 7076
+    },
+    {
+      "epoch": 3.70717653221582,
+      "grad_norm": 0.9332588315010071,
+      "learning_rate": 3.2591640169331697e-07,
+      "loss": 0.1025,
+      "step": 7077
+    },
+    {
+      "epoch": 3.70717653221582,
+      "loss_partial": 0.052978515625,
+      "step": 7077
+    },
+    {
+      "epoch": 3.70717653221582,
+      "loss_partial": 0.05078125,
+      "step": 7077
+    },
+    {
+      "epoch": 3.707700366684128,
+      "grad_norm": 0.7797408699989319,
+      "learning_rate": 3.247597898064292e-07,
+      "loss": 0.0993,
+      "step": 7078
+    },
+    {
+      "epoch": 3.707700366684128,
+      "loss_partial": 0.04443359375,
+      "step": 7078
+    },
+    {
+      "epoch": 3.707700366684128,
+      "loss_partial": 0.0517578125,
+      "step": 7078
+    },
+    {
+      "epoch": 3.7082242011524356,
+      "grad_norm": 0.826316773891449,
+      "learning_rate": 3.2360519998491437e-07,
+      "loss": 0.1003,
+      "step": 7079
+    },
+    {
+      "epoch": 3.7082242011524356,
+      "loss_partial": 0.048095703125,
+      "step": 7079
+    },
+    {
+      "epoch": 3.7082242011524356,
+      "loss_partial": 0.048095703125,
+      "step": 7079
+    },
+    {
+      "epoch": 3.7087480356207436,
+      "grad_norm": 0.8121337890625,
+      "learning_rate": 3.2245263247007607e-07,
+      "loss": 0.1033,
+      "step": 7080
+    },
+    {
+      "epoch": 3.7087480356207436,
+      "loss_partial": 0.053466796875,
+      "step": 7080
+    },
+    {
+      "epoch": 3.7087480356207436,
+      "loss_partial": 0.04541015625,
+      "step": 7080
+    },
+    {
+      "epoch": 3.7092718700890517,
+      "grad_norm": 0.7945500612258911,
+      "learning_rate": 3.213020875027939e-07,
+      "loss": 0.0992,
+      "step": 7081
+    },
+    {
+      "epoch": 3.7092718700890517,
+      "loss_partial": 0.04443359375,
+      "step": 7081
+    },
+    {
+      "epoch": 3.7092718700890517,
+      "loss_partial": 0.053466796875,
+      "step": 7081
+    },
+    {
+      "epoch": 3.7097957045573597,
+      "grad_norm": 0.8085187673568726,
+      "learning_rate": 3.2015356532352326e-07,
+      "loss": 0.0978,
+      "step": 7082
+    },
+    {
+      "epoch": 3.7097957045573597,
+      "loss_partial": 0.064453125,
+      "step": 7082
+    },
+    {
+      "epoch": 3.7097957045573597,
+      "loss_partial": 0.053955078125,
+      "step": 7082
+    },
+    {
+      "epoch": 3.710319539025668,
+      "grad_norm": 0.6944863796234131,
+      "learning_rate": 3.1900706617229994e-07,
+      "loss": 0.0974,
+      "step": 7083
+    },
+    {
+      "epoch": 3.710319539025668,
+      "loss_partial": 0.04931640625,
+      "step": 7083
+    },
+    {
+      "epoch": 3.710319539025668,
+      "loss_partial": 0.048583984375,
+      "step": 7083
+    },
+    {
+      "epoch": 3.710843373493976,
+      "grad_norm": 1.195313811302185,
+      "learning_rate": 3.1786259028873335e-07,
+      "loss": 0.0992,
+      "step": 7084
+    },
+    {
+      "epoch": 3.710843373493976,
+      "loss_partial": 0.0478515625,
+      "step": 7084
+    },
+    {
+      "epoch": 3.710843373493976,
+      "loss_partial": 0.046875,
+      "step": 7084
+    },
+    {
+      "epoch": 3.711367207962284,
+      "grad_norm": 1.283443570137024,
+      "learning_rate": 3.167201379120144e-07,
+      "loss": 0.1021,
+      "step": 7085
+    },
+    {
+      "epoch": 3.711367207962284,
+      "loss_partial": 0.052734375,
+      "step": 7085
+    },
+    {
+      "epoch": 3.711367207962284,
+      "loss_partial": 0.04833984375,
+      "step": 7085
+    },
+    {
+      "epoch": 3.711891042430592,
+      "grad_norm": 0.9150662422180176,
+      "learning_rate": 3.155797092809043e-07,
+      "loss": 0.0994,
+      "step": 7086
+    },
+    {
+      "epoch": 3.711891042430592,
+      "loss_partial": 0.0400390625,
+      "step": 7086
+    },
+    {
+      "epoch": 3.711891042430592,
+      "loss_partial": 0.049072265625,
+      "step": 7086
+    },
+    {
+      "epoch": 3.7124148768989,
+      "grad_norm": 0.6999467015266418,
+      "learning_rate": 3.1444130463375023e-07,
+      "loss": 0.1038,
+      "step": 7087
+    },
+    {
+      "epoch": 3.7124148768989,
+      "loss_partial": 0.057373046875,
+      "step": 7087
+    },
+    {
+      "epoch": 3.7124148768989,
+      "loss_partial": 0.041259765625,
+      "step": 7087
+    },
+    {
+      "epoch": 3.712938711367208,
+      "grad_norm": 0.7546278238296509,
+      "learning_rate": 3.133049242084707e-07,
+      "loss": 0.0995,
+      "step": 7088
+    },
+    {
+      "epoch": 3.712938711367208,
+      "loss_partial": 0.064453125,
+      "step": 7088
+    },
+    {
+      "epoch": 3.712938711367208,
+      "loss_partial": 0.06640625,
+      "step": 7088
+    },
+    {
+      "epoch": 3.713462545835516,
+      "grad_norm": 0.7548868656158447,
+      "learning_rate": 3.121705682425602e-07,
+      "loss": 0.1034,
+      "step": 7089
+    },
+    {
+      "epoch": 3.713462545835516,
+      "loss_partial": 0.041015625,
+      "step": 7089
+    },
+    {
+      "epoch": 3.713462545835516,
+      "loss_partial": 0.05908203125,
+      "step": 7089
+    },
+    {
+      "epoch": 3.713986380303824,
+      "grad_norm": 0.8037089109420776,
+      "learning_rate": 3.1103823697309354e-07,
+      "loss": 0.1037,
+      "step": 7090
+    },
+    {
+      "epoch": 3.713986380303824,
+      "loss_partial": 0.048583984375,
+      "step": 7090
+    },
+    {
+      "epoch": 3.713986380303824,
+      "loss_partial": 0.04736328125,
+      "step": 7090
+    },
+    {
+      "epoch": 3.7145102147721323,
+      "grad_norm": 0.8070679306983948,
+      "learning_rate": 3.099079306367203e-07,
+      "loss": 0.0989,
+      "step": 7091
+    },
+    {
+      "epoch": 3.7145102147721323,
+      "loss_partial": 0.0458984375,
+      "step": 7091
+    },
+    {
+      "epoch": 3.7145102147721323,
+      "loss_partial": 0.059326171875,
+      "step": 7091
+    },
+    {
+      "epoch": 3.71503404924044,
+      "grad_norm": 0.8684533834457397,
+      "learning_rate": 3.087796494696682e-07,
+      "loss": 0.1025,
+      "step": 7092
+    },
+    {
+      "epoch": 3.71503404924044,
+      "loss_partial": 0.05712890625,
+      "step": 7092
+    },
+    {
+      "epoch": 3.71503404924044,
+      "loss_partial": 0.04833984375,
+      "step": 7092
+    },
+    {
+      "epoch": 3.715557883708748,
+      "grad_norm": 0.7791542410850525,
+      "learning_rate": 3.076533937077408e-07,
+      "loss": 0.0998,
+      "step": 7093
+    },
+    {
+      "epoch": 3.715557883708748,
+      "loss_partial": 0.051025390625,
+      "step": 7093
+    },
+    {
+      "epoch": 3.715557883708748,
+      "loss_partial": 0.047607421875,
+      "step": 7093
+    },
+    {
+      "epoch": 3.716081718177056,
+      "grad_norm": 0.8508850932121277,
+      "learning_rate": 3.065291635863166e-07,
+      "loss": 0.0992,
+      "step": 7094
+    },
+    {
+      "epoch": 3.716081718177056,
+      "loss_partial": 0.04052734375,
+      "step": 7094
+    },
+    {
+      "epoch": 3.716081718177056,
+      "loss_partial": 0.045166015625,
+      "step": 7094
+    },
+    {
+      "epoch": 3.716605552645364,
+      "grad_norm": 0.9178784489631653,
+      "learning_rate": 3.054069593403564e-07,
+      "loss": 0.0995,
+      "step": 7095
+    },
+    {
+      "epoch": 3.716605552645364,
+      "loss_partial": 0.046630859375,
+      "step": 7095
+    },
+    {
+      "epoch": 3.716605552645364,
+      "loss_partial": 0.046630859375,
+      "step": 7095
+    },
+    {
+      "epoch": 3.717129387113672,
+      "grad_norm": 0.7876203656196594,
+      "learning_rate": 3.042867812043915e-07,
+      "loss": 0.0995,
+      "step": 7096
+    },
+    {
+      "epoch": 3.717129387113672,
+      "loss_partial": 0.044677734375,
+      "step": 7096
+    },
+    {
+      "epoch": 3.717129387113672,
+      "loss_partial": 0.062255859375,
+      "step": 7096
+    },
+    {
+      "epoch": 3.71765322158198,
+      "grad_norm": 1.3240597248077393,
+      "learning_rate": 3.0316862941253356e-07,
+      "loss": 0.101,
+      "step": 7097
+    },
+    {
+      "epoch": 3.71765322158198,
+      "loss_partial": 0.052734375,
+      "step": 7097
+    },
+    {
+      "epoch": 3.71765322158198,
+      "loss_partial": 0.05712890625,
+      "step": 7097
+    },
+    {
+      "epoch": 3.7181770560502883,
+      "grad_norm": 0.730465829372406,
+      "learning_rate": 3.0205250419846677e-07,
+      "loss": 0.098,
+      "step": 7098
+    },
+    {
+      "epoch": 3.7181770560502883,
+      "loss_partial": 0.042236328125,
+      "step": 7098
+    },
+    {
+      "epoch": 3.7181770560502883,
+      "loss_partial": 0.043701171875,
+      "step": 7098
+    },
+    {
+      "epoch": 3.718700890518596,
+      "grad_norm": 0.9869339466094971,
+      "learning_rate": 3.0093840579545783e-07,
+      "loss": 0.1013,
+      "step": 7099
+    },
+    {
+      "epoch": 3.718700890518596,
+      "loss_partial": 0.05810546875,
+      "step": 7099
+    },
+    {
+      "epoch": 3.718700890518596,
+      "loss_partial": 0.047607421875,
+      "step": 7099
+    },
+    {
+      "epoch": 3.719224724986904,
+      "grad_norm": 0.8349632620811462,
+      "learning_rate": 2.99826334436345e-07,
+      "loss": 0.0993,
+      "step": 7100
+    },
+    {
+      "epoch": 3.719224724986904,
+      "loss_partial": 0.059814453125,
+      "step": 7100
+    },
+    {
+      "epoch": 3.719224724986904,
+      "loss_partial": 0.060302734375,
+      "step": 7100
+    },
+    {
+      "epoch": 3.719748559455212,
+      "grad_norm": 1.0406023263931274,
+      "learning_rate": 2.9871629035354457e-07,
+      "loss": 0.1007,
+      "step": 7101
+    },
+    {
+      "epoch": 3.719748559455212,
+      "loss_partial": 0.051025390625,
+      "step": 7101
+    },
+    {
+      "epoch": 3.719748559455212,
+      "loss_partial": 0.04248046875,
+      "step": 7101
+    },
+    {
+      "epoch": 3.72027239392352,
+      "grad_norm": 1.7565070390701294,
+      "learning_rate": 2.9760827377904644e-07,
+      "loss": 0.0958,
+      "step": 7102
+    },
+    {
+      "epoch": 3.72027239392352,
+      "loss_partial": 0.05029296875,
+      "step": 7102
+    },
+    {
+      "epoch": 3.72027239392352,
+      "loss_partial": 0.052490234375,
+      "step": 7102
+    },
+    {
+      "epoch": 3.720796228391828,
+      "grad_norm": 0.907753050327301,
+      "learning_rate": 2.9650228494442433e-07,
+      "loss": 0.1011,
+      "step": 7103
+    },
+    {
+      "epoch": 3.720796228391828,
+      "loss_partial": 0.0615234375,
+      "step": 7103
+    },
+    {
+      "epoch": 3.720796228391828,
+      "loss_partial": 0.052001953125,
+      "step": 7103
+    },
+    {
+      "epoch": 3.721320062860136,
+      "grad_norm": 1.333847165107727,
+      "learning_rate": 2.9539832408081893e-07,
+      "loss": 0.0985,
+      "step": 7104
+    },
+    {
+      "epoch": 3.721320062860136,
+      "loss_partial": 0.047119140625,
+      "step": 7104
+    },
+    {
+      "epoch": 3.721320062860136,
+      "loss_partial": 0.049560546875,
+      "step": 7104
+    },
+    {
+      "epoch": 3.7218438973284442,
+      "grad_norm": 1.1889126300811768,
+      "learning_rate": 2.9429639141895226e-07,
+      "loss": 0.0999,
+      "step": 7105
+    },
+    {
+      "epoch": 3.7218438973284442,
+      "loss_partial": 0.05419921875,
+      "step": 7105
+    },
+    {
+      "epoch": 3.7218438973284442,
+      "loss_partial": 0.04541015625,
+      "step": 7105
+    },
+    {
+      "epoch": 3.7223677317967523,
+      "grad_norm": 0.8880273699760437,
+      "learning_rate": 2.931964871891202e-07,
+      "loss": 0.102,
+      "step": 7106
+    },
+    {
+      "epoch": 3.7223677317967523,
+      "loss_partial": 0.0478515625,
+      "step": 7106
+    },
+    {
+      "epoch": 3.7223677317967523,
+      "loss_partial": 0.04833984375,
+      "step": 7106
+    },
+    {
+      "epoch": 3.7228915662650603,
+      "grad_norm": 0.70538729429245,
+      "learning_rate": 2.920986116211988e-07,
+      "loss": 0.1037,
+      "step": 7107
+    },
+    {
+      "epoch": 3.7228915662650603,
+      "loss_partial": 0.0478515625,
+      "step": 7107
+    },
+    {
+      "epoch": 3.7228915662650603,
+      "loss_partial": 0.0458984375,
+      "step": 7107
+    },
+    {
+      "epoch": 3.7234154007333684,
+      "grad_norm": 1.1551069021224976,
+      "learning_rate": 2.910027649446345e-07,
+      "loss": 0.099,
+      "step": 7108
+    },
+    {
+      "epoch": 3.7234154007333684,
+      "loss_partial": 0.0546875,
+      "step": 7108
+    },
+    {
+      "epoch": 3.7234154007333684,
+      "loss_partial": 0.0517578125,
+      "step": 7108
+    },
+    {
+      "epoch": 3.7239392352016765,
+      "grad_norm": 1.0674763917922974,
+      "learning_rate": 2.899089473884553e-07,
+      "loss": 0.1012,
+      "step": 7109
+    },
+    {
+      "epoch": 3.7239392352016765,
+      "loss_partial": 0.05078125,
+      "step": 7109
+    },
+    {
+      "epoch": 3.7239392352016765,
+      "loss_partial": 0.048583984375,
+      "step": 7109
+    },
+    {
+      "epoch": 3.7244630696699845,
+      "grad_norm": 0.9531827569007874,
+      "learning_rate": 2.888171591812583e-07,
+      "loss": 0.1011,
+      "step": 7110
+    },
+    {
+      "epoch": 3.7244630696699845,
+      "loss_partial": 0.059814453125,
+      "step": 7110
+    },
+    {
+      "epoch": 3.7244630696699845,
+      "loss_partial": 0.0556640625,
+      "step": 7110
+    },
+    {
+      "epoch": 3.724986904138292,
+      "grad_norm": 0.7513861060142517,
+      "learning_rate": 2.8772740055122425e-07,
+      "loss": 0.1024,
+      "step": 7111
+    },
+    {
+      "epoch": 3.724986904138292,
+      "loss_partial": 0.047119140625,
+      "step": 7111
+    },
+    {
+      "epoch": 3.724986904138292,
+      "loss_partial": 0.046630859375,
+      "step": 7111
+    },
+    {
+      "epoch": 3.7255107386066,
+      "grad_norm": 1.2837390899658203,
+      "learning_rate": 2.866396717261055e-07,
+      "loss": 0.0991,
+      "step": 7112
+    },
+    {
+      "epoch": 3.7255107386066,
+      "loss_partial": 0.046630859375,
+      "step": 7112
+    },
+    {
+      "epoch": 3.7255107386066,
+      "loss_partial": 0.050048828125,
+      "step": 7112
+    },
+    {
+      "epoch": 3.7260345730749083,
+      "grad_norm": 0.8866897225379944,
+      "learning_rate": 2.8555397293322905e-07,
+      "loss": 0.1017,
+      "step": 7113
+    },
+    {
+      "epoch": 3.7260345730749083,
+      "loss_partial": 0.050537109375,
+      "step": 7113
+    },
+    {
+      "epoch": 3.7260345730749083,
+      "loss_partial": 0.061279296875,
+      "step": 7113
+    },
+    {
+      "epoch": 3.7265584075432163,
+      "grad_norm": 1.0499013662338257,
+      "learning_rate": 2.8447030439949896e-07,
+      "loss": 0.1019,
+      "step": 7114
+    },
+    {
+      "epoch": 3.7265584075432163,
+      "loss_partial": 0.046142578125,
+      "step": 7114
+    },
+    {
+      "epoch": 3.7265584075432163,
+      "loss_partial": 0.0546875,
+      "step": 7114
+    },
+    {
+      "epoch": 3.7270822420115244,
+      "grad_norm": 1.0193711519241333,
+      "learning_rate": 2.8338866635139737e-07,
+      "loss": 0.0983,
+      "step": 7115
+    },
+    {
+      "epoch": 3.7270822420115244,
+      "loss_partial": 0.06005859375,
+      "step": 7115
+    },
+    {
+      "epoch": 3.7270822420115244,
+      "loss_partial": 0.05029296875,
+      "step": 7115
+    },
+    {
+      "epoch": 3.7276060764798324,
+      "grad_norm": 1.0535567998886108,
+      "learning_rate": 2.8230905901497906e-07,
+      "loss": 0.1024,
+      "step": 7116
+    },
+    {
+      "epoch": 3.7276060764798324,
+      "loss_partial": 0.0498046875,
+      "step": 7116
+    },
+    {
+      "epoch": 3.7276060764798324,
+      "loss_partial": 0.044921875,
+      "step": 7116
+    },
+    {
+      "epoch": 3.7281299109481405,
+      "grad_norm": 1.284761667251587,
+      "learning_rate": 2.8123148261587465e-07,
+      "loss": 0.1013,
+      "step": 7117
+    },
+    {
+      "epoch": 3.7281299109481405,
+      "loss_partial": 0.05615234375,
+      "step": 7117
+    },
+    {
+      "epoch": 3.7281299109481405,
+      "loss_partial": 0.046142578125,
+      "step": 7117
+    },
+    {
+      "epoch": 3.7286537454164486,
+      "grad_norm": 0.9634237289428711,
+      "learning_rate": 2.801559373792906e-07,
+      "loss": 0.1016,
+      "step": 7118
+    },
+    {
+      "epoch": 3.7286537454164486,
+      "loss_partial": 0.05078125,
+      "step": 7118
+    },
+    {
+      "epoch": 3.7286537454164486,
+      "loss_partial": 0.0546875,
+      "step": 7118
+    },
+    {
+      "epoch": 3.729177579884756,
+      "grad_norm": 0.9219918847084045,
+      "learning_rate": 2.7908242353001047e-07,
+      "loss": 0.1003,
+      "step": 7119
+    },
+    {
+      "epoch": 3.729177579884756,
+      "loss_partial": 0.06103515625,
+      "step": 7119
+    },
+    {
+      "epoch": 3.729177579884756,
+      "loss_partial": 0.05029296875,
+      "step": 7119
+    },
+    {
+      "epoch": 3.7297014143530642,
+      "grad_norm": 0.8492686748504639,
+      "learning_rate": 2.780109412923915e-07,
+      "loss": 0.1005,
+      "step": 7120
+    },
+    {
+      "epoch": 3.7297014143530642,
+      "loss_partial": 0.047119140625,
+      "step": 7120
+    },
+    {
+      "epoch": 3.7297014143530642,
+      "loss_partial": 0.0458984375,
+      "step": 7120
+    },
+    {
+      "epoch": 3.7302252488213723,
+      "grad_norm": 1.198464274406433,
+      "learning_rate": 2.769414908903667e-07,
+      "loss": 0.1006,
+      "step": 7121
+    },
+    {
+      "epoch": 3.7302252488213723,
+      "loss_partial": 0.043701171875,
+      "step": 7121
+    },
+    {
+      "epoch": 3.7302252488213723,
+      "loss_partial": 0.04345703125,
+      "step": 7121
+    },
+    {
+      "epoch": 3.7307490832896804,
+      "grad_norm": 0.8039165139198303,
+      "learning_rate": 2.758740725474451e-07,
+      "loss": 0.1,
+      "step": 7122
+    },
+    {
+      "epoch": 3.7307490832896804,
+      "loss_partial": 0.042724609375,
+      "step": 7122
+    },
+    {
+      "epoch": 3.7307490832896804,
+      "loss_partial": 0.04931640625,
+      "step": 7122
+    },
+    {
+      "epoch": 3.7312729177579884,
+      "grad_norm": 1.1720327138900757,
+      "learning_rate": 2.7480868648670943e-07,
+      "loss": 0.098,
+      "step": 7123
+    },
+    {
+      "epoch": 3.7312729177579884,
+      "loss_partial": 0.0478515625,
+      "step": 7123
+    },
+    {
+      "epoch": 3.7312729177579884,
+      "loss_partial": 0.048828125,
+      "step": 7123
+    },
+    {
+      "epoch": 3.7317967522262965,
+      "grad_norm": 0.9429211616516113,
+      "learning_rate": 2.737453329308193e-07,
+      "loss": 0.0997,
+      "step": 7124
+    },
+    {
+      "epoch": 3.7317967522262965,
+      "loss_partial": 0.05078125,
+      "step": 7124
+    },
+    {
+      "epoch": 3.7317967522262965,
+      "loss_partial": 0.049560546875,
+      "step": 7124
+    },
+    {
+      "epoch": 3.7323205866946045,
+      "grad_norm": 0.8099144101142883,
+      "learning_rate": 2.726840121020102e-07,
+      "loss": 0.1027,
+      "step": 7125
+    },
+    {
+      "epoch": 3.7323205866946045,
+      "loss_partial": 0.05029296875,
+      "step": 7125
+    },
+    {
+      "epoch": 3.7323205866946045,
+      "loss_partial": 0.051513671875,
+      "step": 7125
+    },
+    {
+      "epoch": 3.7328444211629126,
+      "grad_norm": 0.7815111875534058,
+      "learning_rate": 2.7162472422209043e-07,
+      "loss": 0.0997,
+      "step": 7126
+    },
+    {
+      "epoch": 3.7328444211629126,
+      "loss_partial": 0.06298828125,
+      "step": 7126
+    },
+    {
+      "epoch": 3.7328444211629126,
+      "loss_partial": 0.05517578125,
+      "step": 7126
+    },
+    {
+      "epoch": 3.7333682556312207,
+      "grad_norm": 0.8915308117866516,
+      "learning_rate": 2.705674695124438e-07,
+      "loss": 0.1002,
+      "step": 7127
+    },
+    {
+      "epoch": 3.7333682556312207,
+      "loss_partial": 0.04443359375,
+      "step": 7127
+    },
+    {
+      "epoch": 3.7333682556312207,
+      "loss_partial": 0.052734375,
+      "step": 7127
+    },
+    {
+      "epoch": 3.7338920900995287,
+      "grad_norm": 1.051234483718872,
+      "learning_rate": 2.6951224819403374e-07,
+      "loss": 0.0972,
+      "step": 7128
+    },
+    {
+      "epoch": 3.7338920900995287,
+      "loss_partial": 0.0478515625,
+      "step": 7128
+    },
+    {
+      "epoch": 3.7338920900995287,
+      "loss_partial": 0.046142578125,
+      "step": 7128
+    },
+    {
+      "epoch": 3.7344159245678368,
+      "grad_norm": 0.7752313613891602,
+      "learning_rate": 2.684590604873905e-07,
+      "loss": 0.099,
+      "step": 7129
+    },
+    {
+      "epoch": 3.7344159245678368,
+      "loss_partial": 0.039794921875,
+      "step": 7129
+    },
+    {
+      "epoch": 3.7344159245678368,
+      "loss_partial": 0.047607421875,
+      "step": 7129
+    },
+    {
+      "epoch": 3.734939759036145,
+      "grad_norm": 0.7881588935852051,
+      "learning_rate": 2.674079066126256e-07,
+      "loss": 0.1013,
+      "step": 7130
+    },
+    {
+      "epoch": 3.734939759036145,
+      "loss_partial": 0.041015625,
+      "step": 7130
+    },
+    {
+      "epoch": 3.734939759036145,
+      "loss_partial": 0.0458984375,
+      "step": 7130
+    },
+    {
+      "epoch": 3.7354635935044525,
+      "grad_norm": 0.831866443157196,
+      "learning_rate": 2.663587867894235e-07,
+      "loss": 0.0926,
+      "step": 7131
+    },
+    {
+      "epoch": 3.7354635935044525,
+      "loss_partial": 0.045654296875,
+      "step": 7131
+    },
+    {
+      "epoch": 3.7354635935044525,
+      "loss_partial": 0.0458984375,
+      "step": 7131
+    },
+    {
+      "epoch": 3.7359874279727605,
+      "grad_norm": 1.0025383234024048,
+      "learning_rate": 2.653117012370443e-07,
+      "loss": 0.0989,
+      "step": 7132
+    },
+    {
+      "epoch": 3.7359874279727605,
+      "loss_partial": 0.052734375,
+      "step": 7132
+    },
+    {
+      "epoch": 3.7359874279727605,
+      "loss_partial": 0.0595703125,
+      "step": 7132
+    },
+    {
+      "epoch": 3.7365112624410686,
+      "grad_norm": 0.7972069978713989,
+      "learning_rate": 2.6426665017432186e-07,
+      "loss": 0.0986,
+      "step": 7133
+    },
+    {
+      "epoch": 3.7365112624410686,
+      "loss_partial": 0.048095703125,
+      "step": 7133
+    },
+    {
+      "epoch": 3.7365112624410686,
+      "loss_partial": 0.058349609375,
+      "step": 7133
+    },
+    {
+      "epoch": 3.7370350969093766,
+      "grad_norm": 1.0954275131225586,
+      "learning_rate": 2.632236338196659e-07,
+      "loss": 0.0988,
+      "step": 7134
+    },
+    {
+      "epoch": 3.7370350969093766,
+      "loss_partial": 0.052490234375,
+      "step": 7134
+    },
+    {
+      "epoch": 3.7370350969093766,
+      "loss_partial": 0.048828125,
+      "step": 7134
+    },
+    {
+      "epoch": 3.7375589313776847,
+      "grad_norm": 0.9502957463264465,
+      "learning_rate": 2.6218265239105867e-07,
+      "loss": 0.0985,
+      "step": 7135
+    },
+    {
+      "epoch": 3.7375589313776847,
+      "loss_partial": 0.04248046875,
+      "step": 7135
+    },
+    {
+      "epoch": 3.7375589313776847,
+      "loss_partial": 0.06591796875,
+      "step": 7135
+    },
+    {
+      "epoch": 3.7380827658459927,
+      "grad_norm": 0.9089735746383667,
+      "learning_rate": 2.611437061060618e-07,
+      "loss": 0.0998,
+      "step": 7136
+    },
+    {
+      "epoch": 3.7380827658459927,
+      "loss_partial": 0.05517578125,
+      "step": 7136
+    },
+    {
+      "epoch": 3.7380827658459927,
+      "loss_partial": 0.05859375,
+      "step": 7136
+    },
+    {
+      "epoch": 3.738606600314301,
+      "grad_norm": 0.7428134679794312,
+      "learning_rate": 2.601067951818059e-07,
+      "loss": 0.0994,
+      "step": 7137
+    },
+    {
+      "epoch": 3.738606600314301,
+      "loss_partial": 0.044921875,
+      "step": 7137
+    },
+    {
+      "epoch": 3.738606600314301,
+      "loss_partial": 0.04638671875,
+      "step": 7137
+    },
+    {
+      "epoch": 3.7391304347826084,
+      "grad_norm": 1.3583110570907593,
+      "learning_rate": 2.5907191983500114e-07,
+      "loss": 0.1006,
+      "step": 7138
+    },
+    {
+      "epoch": 3.7391304347826084,
+      "loss_partial": 0.0458984375,
+      "step": 7138
+    },
+    {
+      "epoch": 3.7391304347826084,
+      "loss_partial": 0.055419921875,
+      "step": 7138
+    },
+    {
+      "epoch": 3.7396542692509165,
+      "grad_norm": 0.9580907821655273,
+      "learning_rate": 2.580390802819277e-07,
+      "loss": 0.0995,
+      "step": 7139
+    },
+    {
+      "epoch": 3.7396542692509165,
+      "loss_partial": 0.049560546875,
+      "step": 7139
+    },
+    {
+      "epoch": 3.7396542692509165,
+      "loss_partial": 0.047119140625,
+      "step": 7139
+    },
+    {
+      "epoch": 3.7401781037192245,
+      "grad_norm": 1.1115446090698242,
+      "learning_rate": 2.5700827673844633e-07,
+      "loss": 0.0998,
+      "step": 7140
+    },
+    {
+      "epoch": 3.7401781037192245,
+      "loss_partial": 0.048583984375,
+      "step": 7140
+    },
+    {
+      "epoch": 3.7401781037192245,
+      "loss_partial": 0.056396484375,
+      "step": 7140
+    },
+    {
+      "epoch": 3.7407019381875326,
+      "grad_norm": 0.8841076493263245,
+      "learning_rate": 2.5597950941998573e-07,
+      "loss": 0.0972,
+      "step": 7141
+    },
+    {
+      "epoch": 3.7407019381875326,
+      "loss_partial": 0.053955078125,
+      "step": 7141
+    },
+    {
+      "epoch": 3.7407019381875326,
+      "loss_partial": 0.0546875,
+      "step": 7141
+    },
+    {
+      "epoch": 3.7412257726558407,
+      "grad_norm": 1.3082438707351685,
+      "learning_rate": 2.5495277854155174e-07,
+      "loss": 0.1007,
+      "step": 7142
+    },
+    {
+      "epoch": 3.7412257726558407,
+      "loss_partial": 0.052734375,
+      "step": 7142
+    },
+    {
+      "epoch": 3.7412257726558407,
+      "loss_partial": 0.06494140625,
+      "step": 7142
+    },
+    {
+      "epoch": 3.7417496071241487,
+      "grad_norm": 1.0772154331207275,
+      "learning_rate": 2.5392808431772384e-07,
+      "loss": 0.1018,
+      "step": 7143
+    },
+    {
+      "epoch": 3.7417496071241487,
+      "loss_partial": 0.050048828125,
+      "step": 7143
+    },
+    {
+      "epoch": 3.7417496071241487,
+      "loss_partial": 0.043212890625,
+      "step": 7143
+    },
+    {
+      "epoch": 3.742273441592457,
+      "grad_norm": 1.0135540962219238,
+      "learning_rate": 2.5290542696266076e-07,
+      "loss": 0.1009,
+      "step": 7144
+    },
+    {
+      "epoch": 3.742273441592457,
+      "loss_partial": 0.046875,
+      "step": 7144
+    },
+    {
+      "epoch": 3.742273441592457,
+      "loss_partial": 0.049560546875,
+      "step": 7144
+    },
+    {
+      "epoch": 3.742797276060765,
+      "grad_norm": 0.8892979025840759,
+      "learning_rate": 2.518848066900881e-07,
+      "loss": 0.0984,
+      "step": 7145
+    },
+    {
+      "epoch": 3.742797276060765,
+      "loss_partial": 0.049072265625,
+      "step": 7145
+    },
+    {
+      "epoch": 3.742797276060765,
+      "loss_partial": 0.04541015625,
+      "step": 7145
+    },
+    {
+      "epoch": 3.743321110529073,
+      "grad_norm": 0.9693982005119324,
+      "learning_rate": 2.5086622371331083e-07,
+      "loss": 0.0984,
+      "step": 7146
+    },
+    {
+      "epoch": 3.743321110529073,
+      "loss_partial": 0.044189453125,
+      "step": 7146
+    },
+    {
+      "epoch": 3.743321110529073,
+      "loss_partial": 0.0517578125,
+      "step": 7146
+    },
+    {
+      "epoch": 3.743844944997381,
+      "grad_norm": 1.1030834913253784,
+      "learning_rate": 2.498496782452053e-07,
+      "loss": 0.0984,
+      "step": 7147
+    },
+    {
+      "epoch": 3.743844944997381,
+      "loss_partial": 0.050537109375,
+      "step": 7147
+    },
+    {
+      "epoch": 3.743844944997381,
+      "loss_partial": 0.041259765625,
+      "step": 7147
+    },
+    {
+      "epoch": 3.744368779465689,
+      "grad_norm": 0.788975715637207,
+      "learning_rate": 2.4883517049822504e-07,
+      "loss": 0.099,
+      "step": 7148
+    },
+    {
+      "epoch": 3.744368779465689,
+      "loss_partial": 0.0439453125,
+      "step": 7148
+    },
+    {
+      "epoch": 3.744368779465689,
+      "loss_partial": 0.048828125,
+      "step": 7148
+    },
+    {
+      "epoch": 3.744892613933997,
+      "grad_norm": 0.9468040466308594,
+      "learning_rate": 2.4782270068439363e-07,
+      "loss": 0.1003,
+      "step": 7149
+    },
+    {
+      "epoch": 3.744892613933997,
+      "loss_partial": 0.05029296875,
+      "step": 7149
+    },
+    {
+      "epoch": 3.744892613933997,
+      "loss_partial": 0.04541015625,
+      "step": 7149
+    },
+    {
+      "epoch": 3.745416448402305,
+      "grad_norm": 1.1177403926849365,
+      "learning_rate": 2.46812269015313e-07,
+      "loss": 0.0989,
+      "step": 7150
+    },
+    {
+      "epoch": 3.745416448402305,
+      "loss_partial": 0.04736328125,
+      "step": 7150
+    },
+    {
+      "epoch": 3.745416448402305,
+      "loss_partial": 0.045166015625,
+      "step": 7150
+    },
+    {
+      "epoch": 3.7459402828706128,
+      "grad_norm": 0.987156867980957,
+      "learning_rate": 2.4580387570215523e-07,
+      "loss": 0.0946,
+      "step": 7151
+    },
+    {
+      "epoch": 3.7459402828706128,
+      "loss_partial": 0.04150390625,
+      "step": 7151
+    },
+    {
+      "epoch": 3.7459402828706128,
+      "loss_partial": 0.056396484375,
+      "step": 7151
+    },
+    {
+      "epoch": 3.746464117338921,
+      "grad_norm": 0.8651735782623291,
+      "learning_rate": 2.4479752095567077e-07,
+      "loss": 0.1013,
+      "step": 7152
+    },
+    {
+      "epoch": 3.746464117338921,
+      "loss_partial": 0.05126953125,
+      "step": 7152
+    },
+    {
+      "epoch": 3.746464117338921,
+      "loss_partial": 0.06298828125,
+      "step": 7152
+    },
+    {
+      "epoch": 3.746987951807229,
+      "grad_norm": 0.9423000812530518,
+      "learning_rate": 2.437932049861791e-07,
+      "loss": 0.1004,
+      "step": 7153
+    },
+    {
+      "epoch": 3.746987951807229,
+      "loss_partial": 0.06005859375,
+      "step": 7153
+    },
+    {
+      "epoch": 3.746987951807229,
+      "loss_partial": 0.0478515625,
+      "step": 7153
+    },
+    {
+      "epoch": 3.747511786275537,
+      "grad_norm": 1.2001559734344482,
+      "learning_rate": 2.4279092800357786e-07,
+      "loss": 0.1006,
+      "step": 7154
+    },
+    {
+      "epoch": 3.747511786275537,
+      "loss_partial": 0.06298828125,
+      "step": 7154
+    },
+    {
+      "epoch": 3.747511786275537,
+      "loss_partial": 0.058837890625,
+      "step": 7154
+    },
+    {
+      "epoch": 3.748035620743845,
+      "grad_norm": 1.240252137184143,
+      "learning_rate": 2.4179069021733395e-07,
+      "loss": 0.0978,
+      "step": 7155
+    },
+    {
+      "epoch": 3.748035620743845,
+      "loss_partial": 0.044921875,
+      "step": 7155
+    },
+    {
+      "epoch": 3.748035620743845,
+      "loss_partial": 0.052734375,
+      "step": 7155
+    },
+    {
+      "epoch": 3.748559455212153,
+      "grad_norm": 0.9651560187339783,
+      "learning_rate": 2.4079249183649454e-07,
+      "loss": 0.1029,
+      "step": 7156
+    },
+    {
+      "epoch": 3.748559455212153,
+      "loss_partial": 0.0595703125,
+      "step": 7156
+    },
+    {
+      "epoch": 3.748559455212153,
+      "loss_partial": 0.053955078125,
+      "step": 7156
+    },
+    {
+      "epoch": 3.749083289680461,
+      "grad_norm": 0.7411653995513916,
+      "learning_rate": 2.397963330696751e-07,
+      "loss": 0.0996,
+      "step": 7157
+    },
+    {
+      "epoch": 3.749083289680461,
+      "loss_partial": 0.043701171875,
+      "step": 7157
+    },
+    {
+      "epoch": 3.749083289680461,
+      "loss_partial": 0.0498046875,
+      "step": 7157
+    },
+    {
+      "epoch": 3.7496071241487687,
+      "grad_norm": 0.9720988869667053,
+      "learning_rate": 2.388022141250668e-07,
+      "loss": 0.0994,
+      "step": 7158
+    },
+    {
+      "epoch": 3.7496071241487687,
+      "loss_partial": 0.04736328125,
+      "step": 7158
+    },
+    {
+      "epoch": 3.7496071241487687,
+      "loss_partial": 0.04541015625,
+      "step": 7158
+    },
+    {
+      "epoch": 3.750130958617077,
+      "grad_norm": 0.8509067893028259,
+      "learning_rate": 2.378101352104334e-07,
+      "loss": 0.1012,
+      "step": 7159
+    },
+    {
+      "epoch": 3.750130958617077,
+      "loss_partial": 0.04443359375,
+      "step": 7159
+    },
+    {
+      "epoch": 3.750130958617077,
+      "loss_partial": 0.052978515625,
+      "step": 7159
+    },
+    {
+      "epoch": 3.750654793085385,
+      "grad_norm": 0.8176167607307434,
+      "learning_rate": 2.3682009653311465e-07,
+      "loss": 0.0993,
+      "step": 7160
+    },
+    {
+      "epoch": 3.750654793085385,
+      "loss_partial": 0.048583984375,
+      "step": 7160
+    },
+    {
+      "epoch": 3.750654793085385,
+      "loss_partial": 0.043212890625,
+      "step": 7160
+    },
+    {
+      "epoch": 3.751178627553693,
+      "grad_norm": 0.9764745831489563,
+      "learning_rate": 2.3583209830002175e-07,
+      "loss": 0.1015,
+      "step": 7161
+    },
+    {
+      "epoch": 3.751178627553693,
+      "loss_partial": 0.05859375,
+      "step": 7161
+    },
+    {
+      "epoch": 3.751178627553693,
+      "loss_partial": 0.04150390625,
+      "step": 7161
+    },
+    {
+      "epoch": 3.751702462022001,
+      "grad_norm": 1.0653645992279053,
+      "learning_rate": 2.3484614071764167e-07,
+      "loss": 0.1042,
+      "step": 7162
+    },
+    {
+      "epoch": 3.751702462022001,
+      "loss_partial": 0.05517578125,
+      "step": 7162
+    },
+    {
+      "epoch": 3.751702462022001,
+      "loss_partial": 0.041748046875,
+      "step": 7162
+    },
+    {
+      "epoch": 3.752226296490309,
+      "grad_norm": 0.8047468662261963,
+      "learning_rate": 2.3386222399203074e-07,
+      "loss": 0.1005,
+      "step": 7163
+    },
+    {
+      "epoch": 3.752226296490309,
+      "loss_partial": 0.0673828125,
+      "step": 7163
+    },
+    {
+      "epoch": 3.752226296490309,
+      "loss_partial": 0.048095703125,
+      "step": 7163
+    },
+    {
+      "epoch": 3.752750130958617,
+      "grad_norm": 0.8608628511428833,
+      "learning_rate": 2.3288034832882335e-07,
+      "loss": 0.0984,
+      "step": 7164
+    },
+    {
+      "epoch": 3.752750130958617,
+      "loss_partial": 0.048095703125,
+      "step": 7164
+    },
+    {
+      "epoch": 3.752750130958617,
+      "loss_partial": 0.048828125,
+      "step": 7164
+    },
+    {
+      "epoch": 3.753273965426925,
+      "grad_norm": 1.1236233711242676,
+      "learning_rate": 2.3190051393322754e-07,
+      "loss": 0.0993,
+      "step": 7165
+    },
+    {
+      "epoch": 3.753273965426925,
+      "loss_partial": 0.0498046875,
+      "step": 7165
+    },
+    {
+      "epoch": 3.753273965426925,
+      "loss_partial": 0.04833984375,
+      "step": 7165
+    },
+    {
+      "epoch": 3.753797799895233,
+      "grad_norm": 0.8856029510498047,
+      "learning_rate": 2.3092272101001733e-07,
+      "loss": 0.101,
+      "step": 7166
+    },
+    {
+      "epoch": 3.753797799895233,
+      "loss_partial": 0.038330078125,
+      "step": 7166
+    },
+    {
+      "epoch": 3.753797799895233,
+      "loss_partial": 0.04296875,
+      "step": 7166
+    },
+    {
+      "epoch": 3.7543216343635413,
+      "grad_norm": 0.9196661114692688,
+      "learning_rate": 2.2994696976354926e-07,
+      "loss": 0.0986,
+      "step": 7167
+    },
+    {
+      "epoch": 3.7543216343635413,
+      "loss_partial": 0.0400390625,
+      "step": 7167
+    },
+    {
+      "epoch": 3.7543216343635413,
+      "loss_partial": 0.05810546875,
+      "step": 7167
+    },
+    {
+      "epoch": 3.7548454688318493,
+      "grad_norm": 0.8480843901634216,
+      "learning_rate": 2.2897326039774792e-07,
+      "loss": 0.0991,
+      "step": 7168
+    },
+    {
+      "epoch": 3.7548454688318493,
+      "loss_partial": 0.05615234375,
+      "step": 7168
+    },
+    {
+      "epoch": 3.7548454688318493,
+      "loss_partial": 0.061279296875,
+      "step": 7168
+    },
+    {
+      "epoch": 3.7553693033001574,
+      "grad_norm": 0.8956894278526306,
+      "learning_rate": 2.280015931161139e-07,
+      "loss": 0.099,
+      "step": 7169
+    },
+    {
+      "epoch": 3.7553693033001574,
+      "loss_partial": 0.050537109375,
+      "step": 7169
+    },
+    {
+      "epoch": 3.7553693033001574,
+      "loss_partial": 0.0498046875,
+      "step": 7169
+    },
+    {
+      "epoch": 3.755893137768465,
+      "grad_norm": 0.9493269920349121,
+      "learning_rate": 2.2703196812171814e-07,
+      "loss": 0.0992,
+      "step": 7170
+    },
+    {
+      "epoch": 3.755893137768465,
+      "loss_partial": 0.0498046875,
+      "step": 7170
+    },
+    {
+      "epoch": 3.755893137768465,
+      "loss_partial": 0.044921875,
+      "step": 7170
+    },
+    {
+      "epoch": 3.756416972236773,
+      "grad_norm": 0.8920361399650574,
+      "learning_rate": 2.2606438561720511e-07,
+      "loss": 0.1019,
+      "step": 7171
+    },
+    {
+      "epoch": 3.756416972236773,
+      "loss_partial": 0.060302734375,
+      "step": 7171
+    },
+    {
+      "epoch": 3.756416972236773,
+      "loss_partial": 0.04345703125,
+      "step": 7171
+    },
+    {
+      "epoch": 3.756940806705081,
+      "grad_norm": 0.8731986880302429,
+      "learning_rate": 2.2509884580479757e-07,
+      "loss": 0.1021,
+      "step": 7172
+    },
+    {
+      "epoch": 3.756940806705081,
+      "loss_partial": 0.051025390625,
+      "step": 7172
+    },
+    {
+      "epoch": 3.756940806705081,
+      "loss_partial": 0.0537109375,
+      "step": 7172
+    },
+    {
+      "epoch": 3.757464641173389,
+      "grad_norm": 1.3483023643493652,
+      "learning_rate": 2.2413534888628406e-07,
+      "loss": 0.0988,
+      "step": 7173
+    },
+    {
+      "epoch": 3.757464641173389,
+      "loss_partial": 0.05712890625,
+      "step": 7173
+    },
+    {
+      "epoch": 3.757464641173389,
+      "loss_partial": 0.05615234375,
+      "step": 7173
+    },
+    {
+      "epoch": 3.7579884756416972,
+      "grad_norm": 0.9358108639717102,
+      "learning_rate": 2.2317389506303132e-07,
+      "loss": 0.0995,
+      "step": 7174
+    },
+    {
+      "epoch": 3.7579884756416972,
+      "loss_partial": 0.048583984375,
+      "step": 7174
+    },
+    {
+      "epoch": 3.7579884756416972,
+      "loss_partial": 0.051025390625,
+      "step": 7174
+    },
+    {
+      "epoch": 3.7585123101100053,
+      "grad_norm": 0.7129615545272827,
+      "learning_rate": 2.2221448453597528e-07,
+      "loss": 0.0997,
+      "step": 7175
+    },
+    {
+      "epoch": 3.7585123101100053,
+      "loss_partial": 0.055419921875,
+      "step": 7175
+    },
+    {
+      "epoch": 3.7585123101100053,
+      "loss_partial": 0.059326171875,
+      "step": 7175
+    },
+    {
+      "epoch": 3.7590361445783134,
+      "grad_norm": 0.7492048740386963,
+      "learning_rate": 2.2125711750562884e-07,
+      "loss": 0.1021,
+      "step": 7176
+    },
+    {
+      "epoch": 3.7590361445783134,
+      "loss_partial": 0.0498046875,
+      "step": 7176
+    },
+    {
+      "epoch": 3.7590361445783134,
+      "loss_partial": 0.0537109375,
+      "step": 7176
+    },
+    {
+      "epoch": 3.7595599790466214,
+      "grad_norm": 0.9866016507148743,
+      "learning_rate": 2.2030179417207641e-07,
+      "loss": 0.1019,
+      "step": 7177
+    },
+    {
+      "epoch": 3.7595599790466214,
+      "loss_partial": 0.046875,
+      "step": 7177
+    },
+    {
+      "epoch": 3.7595599790466214,
+      "loss_partial": 0.04931640625,
+      "step": 7177
+    },
+    {
+      "epoch": 3.760083813514929,
+      "grad_norm": 0.8041554689407349,
+      "learning_rate": 2.1934851473497165e-07,
+      "loss": 0.0979,
+      "step": 7178
+    },
+    {
+      "epoch": 3.760083813514929,
+      "loss_partial": 0.046875,
+      "step": 7178
+    },
+    {
+      "epoch": 3.760083813514929,
+      "loss_partial": 0.05224609375,
+      "step": 7178
+    },
+    {
+      "epoch": 3.760607647983237,
+      "grad_norm": 0.7759349346160889,
+      "learning_rate": 2.1839727939354517e-07,
+      "loss": 0.1027,
+      "step": 7179
+    },
+    {
+      "epoch": 3.760607647983237,
+      "loss_partial": 0.042236328125,
+      "step": 7179
+    },
+    {
+      "epoch": 3.760607647983237,
+      "loss_partial": 0.047607421875,
+      "step": 7179
+    },
+    {
+      "epoch": 3.761131482451545,
+      "grad_norm": 0.8366078734397888,
+      "learning_rate": 2.1744808834659903e-07,
+      "loss": 0.0958,
+      "step": 7180
+    },
+    {
+      "epoch": 3.761131482451545,
+      "loss_partial": 0.047607421875,
+      "step": 7180
+    },
+    {
+      "epoch": 3.761131482451545,
+      "loss_partial": 0.050048828125,
+      "step": 7180
+    },
+    {
+      "epoch": 3.761655316919853,
+      "grad_norm": 0.7687367796897888,
+      "learning_rate": 2.1650094179251014e-07,
+      "loss": 0.0951,
+      "step": 7181
+    },
+    {
+      "epoch": 3.761655316919853,
+      "loss_partial": 0.047119140625,
+      "step": 7181
+    },
+    {
+      "epoch": 3.761655316919853,
+      "loss_partial": 0.044189453125,
+      "step": 7181
+    },
+    {
+      "epoch": 3.7621791513881613,
+      "grad_norm": 0.9283986687660217,
+      "learning_rate": 2.1555583992922458e-07,
+      "loss": 0.1012,
+      "step": 7182
+    },
+    {
+      "epoch": 3.7621791513881613,
+      "loss_partial": 0.037353515625,
+      "step": 7182
+    },
+    {
+      "epoch": 3.7621791513881613,
+      "loss_partial": 0.039794921875,
+      "step": 7182
+    },
+    {
+      "epoch": 3.7627029858564693,
+      "grad_norm": 0.8405429720878601,
+      "learning_rate": 2.1461278295426435e-07,
+      "loss": 0.093,
+      "step": 7183
+    },
+    {
+      "epoch": 3.7627029858564693,
+      "loss_partial": 0.0732421875,
+      "step": 7183
+    },
+    {
+      "epoch": 3.7627029858564693,
+      "loss_partial": 0.052734375,
+      "step": 7183
+    },
+    {
+      "epoch": 3.7632268203247774,
+      "grad_norm": 0.7732101082801819,
+      "learning_rate": 2.136717710647196e-07,
+      "loss": 0.1002,
+      "step": 7184
+    },
+    {
+      "epoch": 3.7632268203247774,
+      "loss_partial": 0.04638671875,
+      "step": 7184
+    },
+    {
+      "epoch": 3.7632268203247774,
+      "loss_partial": 0.04345703125,
+      "step": 7184
+    },
+    {
+      "epoch": 3.7637506547930855,
+      "grad_norm": 0.9788275957107544,
+      "learning_rate": 2.1273280445725851e-07,
+      "loss": 0.0988,
+      "step": 7185
+    },
+    {
+      "epoch": 3.7637506547930855,
+      "loss_partial": 0.06103515625,
+      "step": 7185
+    },
+    {
+      "epoch": 3.7637506547930855,
+      "loss_partial": 0.053955078125,
+      "step": 7185
+    },
+    {
+      "epoch": 3.7642744892613935,
+      "grad_norm": 0.9822759032249451,
+      "learning_rate": 2.1179588332811973e-07,
+      "loss": 0.1034,
+      "step": 7186
+    },
+    {
+      "epoch": 3.7642744892613935,
+      "loss_partial": 0.044189453125,
+      "step": 7186
+    },
+    {
+      "epoch": 3.7642744892613935,
+      "loss_partial": 0.052734375,
+      "step": 7186
+    },
+    {
+      "epoch": 3.7647983237297016,
+      "grad_norm": 1.0957845449447632,
+      "learning_rate": 2.1086100787311214e-07,
+      "loss": 0.1006,
+      "step": 7187
+    },
+    {
+      "epoch": 3.7647983237297016,
+      "loss_partial": 0.051513671875,
+      "step": 7187
+    },
+    {
+      "epoch": 3.7647983237297016,
+      "loss_partial": 0.04541015625,
+      "step": 7187
+    },
+    {
+      "epoch": 3.7653221581980096,
+      "grad_norm": 0.8254935145378113,
+      "learning_rate": 2.0992817828761947e-07,
+      "loss": 0.1014,
+      "step": 7188
+    },
+    {
+      "epoch": 3.7653221581980096,
+      "loss_partial": 0.052978515625,
+      "step": 7188
+    },
+    {
+      "epoch": 3.7653221581980096,
+      "loss_partial": 0.03955078125,
+      "step": 7188
+    },
+    {
+      "epoch": 3.7658459926663177,
+      "grad_norm": 0.7938843965530396,
+      "learning_rate": 2.0899739476659907e-07,
+      "loss": 0.1,
+      "step": 7189
+    },
+    {
+      "epoch": 3.7658459926663177,
+      "loss_partial": 0.0498046875,
+      "step": 7189
+    },
+    {
+      "epoch": 3.7658459926663177,
+      "loss_partial": 0.054931640625,
+      "step": 7189
+    },
+    {
+      "epoch": 3.7663698271346253,
+      "grad_norm": 0.865682065486908,
+      "learning_rate": 2.080686575045754e-07,
+      "loss": 0.102,
+      "step": 7190
+    },
+    {
+      "epoch": 3.7663698271346253,
+      "loss_partial": 0.059814453125,
+      "step": 7190
+    },
+    {
+      "epoch": 3.7663698271346253,
+      "loss_partial": 0.04931640625,
+      "step": 7190
+    },
+    {
+      "epoch": 3.7668936616029334,
+      "grad_norm": 0.8974693417549133,
+      "learning_rate": 2.07141966695652e-07,
+      "loss": 0.1034,
+      "step": 7191
+    },
+    {
+      "epoch": 3.7668936616029334,
+      "loss_partial": 0.04541015625,
+      "step": 7191
+    },
+    {
+      "epoch": 3.7668936616029334,
+      "loss_partial": 0.046630859375,
+      "step": 7191
+    },
+    {
+      "epoch": 3.7674174960712414,
+      "grad_norm": 1.0123120546340942,
+      "learning_rate": 2.0621732253349958e-07,
+      "loss": 0.1028,
+      "step": 7192
+    },
+    {
+      "epoch": 3.7674174960712414,
+      "loss_partial": 0.0556640625,
+      "step": 7192
+    },
+    {
+      "epoch": 3.7674174960712414,
+      "loss_partial": 0.046142578125,
+      "step": 7192
+    },
+    {
+      "epoch": 3.7679413305395495,
+      "grad_norm": 0.8858882784843445,
+      "learning_rate": 2.0529472521136463e-07,
+      "loss": 0.1012,
+      "step": 7193
+    },
+    {
+      "epoch": 3.7679413305395495,
+      "loss_partial": 0.04296875,
+      "step": 7193
+    },
+    {
+      "epoch": 3.7679413305395495,
+      "loss_partial": 0.056396484375,
+      "step": 7193
+    },
+    {
+      "epoch": 3.7684651650078576,
+      "grad_norm": 0.8145493268966675,
+      "learning_rate": 2.043741749220629e-07,
+      "loss": 0.1032,
+      "step": 7194
+    },
+    {
+      "epoch": 3.7684651650078576,
+      "loss_partial": 0.05615234375,
+      "step": 7194
+    },
+    {
+      "epoch": 3.7684651650078576,
+      "loss_partial": 0.041748046875,
+      "step": 7194
+    },
+    {
+      "epoch": 3.7689889994761656,
+      "grad_norm": 1.0227088928222656,
+      "learning_rate": 2.03455671857985e-07,
+      "loss": 0.0987,
+      "step": 7195
+    },
+    {
+      "epoch": 3.7689889994761656,
+      "loss_partial": 0.042724609375,
+      "step": 7195
+    },
+    {
+      "epoch": 3.7689889994761656,
+      "loss_partial": 0.052001953125,
+      "step": 7195
+    },
+    {
+      "epoch": 3.7695128339444737,
+      "grad_norm": 0.8775200247764587,
+      "learning_rate": 2.025392162110895e-07,
+      "loss": 0.0983,
+      "step": 7196
+    },
+    {
+      "epoch": 3.7695128339444737,
+      "loss_partial": 0.05224609375,
+      "step": 7196
+    },
+    {
+      "epoch": 3.7695128339444737,
+      "loss_partial": 0.052490234375,
+      "step": 7196
+    },
+    {
+      "epoch": 3.7700366684127813,
+      "grad_norm": 0.7259947657585144,
+      "learning_rate": 2.0162480817291442e-07,
+      "loss": 0.0938,
+      "step": 7197
+    },
+    {
+      "epoch": 3.7700366684127813,
+      "loss_partial": 0.046875,
+      "step": 7197
+    },
+    {
+      "epoch": 3.7700366684127813,
+      "loss_partial": 0.061279296875,
+      "step": 7197
+    },
+    {
+      "epoch": 3.7705605028810893,
+      "grad_norm": 0.833050012588501,
+      "learning_rate": 2.0071244793456235e-07,
+      "loss": 0.0974,
+      "step": 7198
+    },
+    {
+      "epoch": 3.7705605028810893,
+      "loss_partial": 0.046142578125,
+      "step": 7198
+    },
+    {
+      "epoch": 3.7705605028810893,
+      "loss_partial": 0.045654296875,
+      "step": 7198
+    },
+    {
+      "epoch": 3.7710843373493974,
+      "grad_norm": 0.871094286441803,
+      "learning_rate": 1.9980213568671192e-07,
+      "loss": 0.0961,
+      "step": 7199
+    },
+    {
+      "epoch": 3.7710843373493974,
+      "loss_partial": 0.0634765625,
+      "step": 7199
+    },
+    {
+      "epoch": 3.7710843373493974,
+      "loss_partial": 0.042724609375,
+      "step": 7199
+    },
+    {
+      "epoch": 3.7716081718177055,
+      "grad_norm": 0.9263244271278381,
+      "learning_rate": 1.9889387161961205e-07,
+      "loss": 0.1041,
+      "step": 7200
+    },
+    {
+      "epoch": 3.7716081718177055,
+      "loss_partial": 0.05615234375,
+      "step": 7200
+    },
+    {
+      "epoch": 3.7716081718177055,
+      "loss_partial": 0.056396484375,
+      "step": 7200
+    },
+    {
+      "epoch": 3.7721320062860135,
+      "grad_norm": 0.774478018283844,
+      "learning_rate": 1.9798765592308534e-07,
+      "loss": 0.0995,
+      "step": 7201
+    },
+    {
+      "epoch": 3.7721320062860135,
+      "loss_partial": 0.046142578125,
+      "step": 7201
+    },
+    {
+      "epoch": 3.7721320062860135,
+      "loss_partial": 0.0400390625,
+      "step": 7201
+    },
+    {
+      "epoch": 3.7726558407543216,
+      "grad_norm": 0.9585740566253662,
+      "learning_rate": 1.9708348878652583e-07,
+      "loss": 0.0981,
+      "step": 7202
+    },
+    {
+      "epoch": 3.7726558407543216,
+      "loss_partial": 0.04541015625,
+      "step": 7202
+    },
+    {
+      "epoch": 3.7726558407543216,
+      "loss_partial": 0.0546875,
+      "step": 7202
+    },
+    {
+      "epoch": 3.7731796752226296,
+      "grad_norm": 0.6821768283843994,
+      "learning_rate": 1.9618137039889573e-07,
+      "loss": 0.0997,
+      "step": 7203
+    },
+    {
+      "epoch": 3.7731796752226296,
+      "loss_partial": 0.048583984375,
+      "step": 7203
+    },
+    {
+      "epoch": 3.7731796752226296,
+      "loss_partial": 0.0517578125,
+      "step": 7203
+    },
+    {
+      "epoch": 3.7737035096909377,
+      "grad_norm": 0.9416640400886536,
+      "learning_rate": 1.952813009487342e-07,
+      "loss": 0.0984,
+      "step": 7204
+    },
+    {
+      "epoch": 3.7737035096909377,
+      "loss_partial": 0.05859375,
+      "step": 7204
+    },
+    {
+      "epoch": 3.7737035096909377,
+      "loss_partial": 0.052734375,
+      "step": 7204
+    },
+    {
+      "epoch": 3.7742273441592458,
+      "grad_norm": 0.6537460088729858,
+      "learning_rate": 1.943832806241519e-07,
+      "loss": 0.1016,
+      "step": 7205
+    },
+    {
+      "epoch": 3.7742273441592458,
+      "loss_partial": 0.0439453125,
+      "step": 7205
+    },
+    {
+      "epoch": 3.7742273441592458,
+      "loss_partial": 0.04541015625,
+      "step": 7205
+    },
+    {
+      "epoch": 3.774751178627554,
+      "grad_norm": 0.8197981119155884,
+      "learning_rate": 1.9348730961282646e-07,
+      "loss": 0.0954,
+      "step": 7206
+    },
+    {
+      "epoch": 3.774751178627554,
+      "loss_partial": 0.05224609375,
+      "step": 7206
+    },
+    {
+      "epoch": 3.774751178627554,
+      "loss_partial": 0.047607421875,
+      "step": 7206
+    },
+    {
+      "epoch": 3.775275013095862,
+      "grad_norm": 0.9333531856536865,
+      "learning_rate": 1.9259338810201256e-07,
+      "loss": 0.0991,
+      "step": 7207
+    },
+    {
+      "epoch": 3.775275013095862,
+      "loss_partial": 0.045654296875,
+      "step": 7207
+    },
+    {
+      "epoch": 3.775275013095862,
+      "loss_partial": 0.043212890625,
+      "step": 7207
+    },
+    {
+      "epoch": 3.77579884756417,
+      "grad_norm": 1.0101877450942993,
+      "learning_rate": 1.9170151627853185e-07,
+      "loss": 0.0974,
+      "step": 7208
+    },
+    {
+      "epoch": 3.77579884756417,
+      "loss_partial": 0.04541015625,
+      "step": 7208
+    },
+    {
+      "epoch": 3.77579884756417,
+      "loss_partial": 0.049560546875,
+      "step": 7208
+    },
+    {
+      "epoch": 3.776322682032478,
+      "grad_norm": 0.7525274157524109,
+      "learning_rate": 1.90811694328783e-07,
+      "loss": 0.0963,
+      "step": 7209
+    },
+    {
+      "epoch": 3.776322682032478,
+      "loss_partial": 0.0537109375,
+      "step": 7209
+    },
+    {
+      "epoch": 3.776322682032478,
+      "loss_partial": 0.050048828125,
+      "step": 7209
+    },
+    {
+      "epoch": 3.7768465165007856,
+      "grad_norm": 0.8412090539932251,
+      "learning_rate": 1.899239224387317e-07,
+      "loss": 0.1039,
+      "step": 7210
+    },
+    {
+      "epoch": 3.7768465165007856,
+      "loss_partial": 0.051025390625,
+      "step": 7210
+    },
+    {
+      "epoch": 3.7768465165007856,
+      "loss_partial": 0.0537109375,
+      "step": 7210
+    },
+    {
+      "epoch": 3.7773703509690937,
+      "grad_norm": 0.8312450051307678,
+      "learning_rate": 1.8903820079391732e-07,
+      "loss": 0.0952,
+      "step": 7211
+    },
+    {
+      "epoch": 3.7773703509690937,
+      "loss_partial": 0.046142578125,
+      "step": 7211
+    },
+    {
+      "epoch": 3.7773703509690937,
+      "loss_partial": 0.04736328125,
+      "step": 7211
+    },
+    {
+      "epoch": 3.7778941854374017,
+      "grad_norm": 0.673801600933075,
+      "learning_rate": 1.8815452957944845e-07,
+      "loss": 0.0987,
+      "step": 7212
+    },
+    {
+      "epoch": 3.7778941854374017,
+      "loss_partial": 0.05908203125,
+      "step": 7212
+    },
+    {
+      "epoch": 3.7778941854374017,
+      "loss_partial": 0.050537109375,
+      "step": 7212
+    },
+    {
+      "epoch": 3.77841801990571,
+      "grad_norm": 0.8905302882194519,
+      "learning_rate": 1.872729089800085e-07,
+      "loss": 0.0972,
+      "step": 7213
+    },
+    {
+      "epoch": 3.77841801990571,
+      "loss_partial": 0.037841796875,
+      "step": 7213
+    },
+    {
+      "epoch": 3.77841801990571,
+      "loss_partial": 0.05615234375,
+      "step": 7213
+    },
+    {
+      "epoch": 3.778941854374018,
+      "grad_norm": 0.9159709811210632,
+      "learning_rate": 1.8639333917985226e-07,
+      "loss": 0.096,
+      "step": 7214
+    },
+    {
+      "epoch": 3.778941854374018,
+      "loss_partial": 0.046875,
+      "step": 7214
+    },
+    {
+      "epoch": 3.778941854374018,
+      "loss_partial": 0.04736328125,
+      "step": 7214
+    },
+    {
+      "epoch": 3.779465688842326,
+      "grad_norm": 1.1397589445114136,
+      "learning_rate": 1.855158203628016e-07,
+      "loss": 0.0979,
+      "step": 7215
+    },
+    {
+      "epoch": 3.779465688842326,
+      "loss_partial": 0.05322265625,
+      "step": 7215
+    },
+    {
+      "epoch": 3.779465688842326,
+      "loss_partial": 0.04248046875,
+      "step": 7215
+    },
+    {
+      "epoch": 3.779989523310634,
+      "grad_norm": 0.9275434017181396,
+      "learning_rate": 1.84640352712252e-07,
+      "loss": 0.0968,
+      "step": 7216
+    },
+    {
+      "epoch": 3.779989523310634,
+      "loss_partial": 0.038818359375,
+      "step": 7216
+    },
+    {
+      "epoch": 3.779989523310634,
+      "loss_partial": 0.049560546875,
+      "step": 7216
+    },
+    {
+      "epoch": 3.7805133577789416,
+      "grad_norm": 0.8104423880577087,
+      "learning_rate": 1.8376693641117381e-07,
+      "loss": 0.098,
+      "step": 7217
+    },
+    {
+      "epoch": 3.7805133577789416,
+      "loss_partial": 0.05517578125,
+      "step": 7217
+    },
+    {
+      "epoch": 3.7805133577789416,
+      "loss_partial": 0.048828125,
+      "step": 7217
+    },
+    {
+      "epoch": 3.7810371922472497,
+      "grad_norm": 0.8323699831962585,
+      "learning_rate": 1.8289557164210437e-07,
+      "loss": 0.1002,
+      "step": 7218
+    },
+    {
+      "epoch": 3.7810371922472497,
+      "loss_partial": 0.04541015625,
+      "step": 7218
+    },
+    {
+      "epoch": 3.7810371922472497,
+      "loss_partial": 0.039306640625,
+      "step": 7218
+    },
+    {
+      "epoch": 3.7815610267155577,
+      "grad_norm": 0.7476119995117188,
+      "learning_rate": 1.820262585871535e-07,
+      "loss": 0.0962,
+      "step": 7219
+    },
+    {
+      "epoch": 3.7815610267155577,
+      "loss_partial": 0.0478515625,
+      "step": 7219
+    },
+    {
+      "epoch": 3.7815610267155577,
+      "loss_partial": 0.053466796875,
+      "step": 7219
+    },
+    {
+      "epoch": 3.7820848611838658,
+      "grad_norm": 0.8842277526855469,
+      "learning_rate": 1.8115899742800035e-07,
+      "loss": 0.101,
+      "step": 7220
+    },
+    {
+      "epoch": 3.7820848611838658,
+      "loss_partial": 0.052978515625,
+      "step": 7220
+    },
+    {
+      "epoch": 3.7820848611838658,
+      "loss_partial": 0.042724609375,
+      "step": 7220
+    },
+    {
+      "epoch": 3.782608695652174,
+      "grad_norm": 0.7543587684631348,
+      "learning_rate": 1.8029378834589994e-07,
+      "loss": 0.1039,
+      "step": 7221
+    },
+    {
+      "epoch": 3.782608695652174,
+      "loss_partial": 0.04443359375,
+      "step": 7221
+    },
+    {
+      "epoch": 3.782608695652174,
+      "loss_partial": 0.051025390625,
+      "step": 7221
+    },
+    {
+      "epoch": 3.783132530120482,
+      "grad_norm": 0.9740478992462158,
+      "learning_rate": 1.7943063152167428e-07,
+      "loss": 0.1016,
+      "step": 7222
+    },
+    {
+      "epoch": 3.783132530120482,
+      "loss_partial": 0.048095703125,
+      "step": 7222
+    },
+    {
+      "epoch": 3.783132530120482,
+      "loss_partial": 0.054443359375,
+      "step": 7222
+    },
+    {
+      "epoch": 3.78365636458879,
+      "grad_norm": 0.7428008317947388,
+      "learning_rate": 1.7856952713571908e-07,
+      "loss": 0.0994,
+      "step": 7223
+    },
+    {
+      "epoch": 3.78365636458879,
+      "loss_partial": 0.05078125,
+      "step": 7223
+    },
+    {
+      "epoch": 3.78365636458879,
+      "loss_partial": 0.048828125,
+      "step": 7223
+    },
+    {
+      "epoch": 3.784180199057098,
+      "grad_norm": 0.9615840911865234,
+      "learning_rate": 1.77710475367997e-07,
+      "loss": 0.0966,
+      "step": 7224
+    },
+    {
+      "epoch": 3.784180199057098,
+      "loss_partial": 0.052490234375,
+      "step": 7224
+    },
+    {
+      "epoch": 3.784180199057098,
+      "loss_partial": 0.056640625,
+      "step": 7224
+    },
+    {
+      "epoch": 3.784704033525406,
+      "grad_norm": 0.8966307044029236,
+      "learning_rate": 1.7685347639804784e-07,
+      "loss": 0.0988,
+      "step": 7225
+    },
+    {
+      "epoch": 3.784704033525406,
+      "loss_partial": 0.053955078125,
+      "step": 7225
+    },
+    {
+      "epoch": 3.784704033525406,
+      "loss_partial": 0.053955078125,
+      "step": 7225
+    },
+    {
+      "epoch": 3.785227867993714,
+      "grad_norm": 0.7340103983879089,
+      "learning_rate": 1.7599853040497717e-07,
+      "loss": 0.0984,
+      "step": 7226
+    },
+    {
+      "epoch": 3.785227867993714,
+      "loss_partial": 0.046875,
+      "step": 7226
+    },
+    {
+      "epoch": 3.785227867993714,
+      "loss_partial": 0.052490234375,
+      "step": 7226
+    },
+    {
+      "epoch": 3.785751702462022,
+      "grad_norm": 0.7788717150688171,
+      "learning_rate": 1.751456375674654e-07,
+      "loss": 0.0996,
+      "step": 7227
+    },
+    {
+      "epoch": 3.785751702462022,
+      "loss_partial": 0.040283203125,
+      "step": 7227
+    },
+    {
+      "epoch": 3.785751702462022,
+      "loss_partial": 0.0439453125,
+      "step": 7227
+    },
+    {
+      "epoch": 3.7862755369303303,
+      "grad_norm": 0.8042730689048767,
+      "learning_rate": 1.742947980637577e-07,
+      "loss": 0.0971,
+      "step": 7228
+    },
+    {
+      "epoch": 3.7862755369303303,
+      "loss_partial": 0.038330078125,
+      "step": 7228
+    },
+    {
+      "epoch": 3.7862755369303303,
+      "loss_partial": 0.05078125,
+      "step": 7228
+    },
+    {
+      "epoch": 3.786799371398638,
+      "grad_norm": 0.8033611178398132,
+      "learning_rate": 1.7344601207167854e-07,
+      "loss": 0.1007,
+      "step": 7229
+    },
+    {
+      "epoch": 3.786799371398638,
+      "loss_partial": 0.05419921875,
+      "step": 7229
+    },
+    {
+      "epoch": 3.786799371398638,
+      "loss_partial": 0.052978515625,
+      "step": 7229
+    },
+    {
+      "epoch": 3.787323205866946,
+      "grad_norm": 0.8040266036987305,
+      "learning_rate": 1.7259927976861602e-07,
+      "loss": 0.0981,
+      "step": 7230
+    },
+    {
+      "epoch": 3.787323205866946,
+      "loss_partial": 0.05078125,
+      "step": 7230
+    },
+    {
+      "epoch": 3.787323205866946,
+      "loss_partial": 0.05908203125,
+      "step": 7230
+    },
+    {
+      "epoch": 3.787847040335254,
+      "grad_norm": 0.9284880757331848,
+      "learning_rate": 1.717546013315341e-07,
+      "loss": 0.1004,
+      "step": 7231
+    },
+    {
+      "epoch": 3.787847040335254,
+      "loss_partial": 0.0439453125,
+      "step": 7231
+    },
+    {
+      "epoch": 3.787847040335254,
+      "loss_partial": 0.05419921875,
+      "step": 7231
+    },
+    {
+      "epoch": 3.788370874803562,
+      "grad_norm": 1.3058319091796875,
+      "learning_rate": 1.7091197693696382e-07,
+      "loss": 0.1029,
+      "step": 7232
+    },
+    {
+      "epoch": 3.788370874803562,
+      "loss_partial": 0.05712890625,
+      "step": 7232
+    },
+    {
+      "epoch": 3.788370874803562,
+      "loss_partial": 0.048583984375,
+      "step": 7232
+    },
+    {
+      "epoch": 3.78889470927187,
+      "grad_norm": 1.039591670036316,
+      "learning_rate": 1.7007140676100985e-07,
+      "loss": 0.101,
+      "step": 7233
+    },
+    {
+      "epoch": 3.78889470927187,
+      "loss_partial": 0.04443359375,
+      "step": 7233
+    },
+    {
+      "epoch": 3.78889470927187,
+      "loss_partial": 0.040771484375,
+      "step": 7233
+    },
+    {
+      "epoch": 3.789418543740178,
+      "grad_norm": 0.8222671747207642,
+      "learning_rate": 1.6923289097934504e-07,
+      "loss": 0.0956,
+      "step": 7234
+    },
+    {
+      "epoch": 3.789418543740178,
+      "loss_partial": 0.056396484375,
+      "step": 7234
+    },
+    {
+      "epoch": 3.789418543740178,
+      "loss_partial": 0.046142578125,
+      "step": 7234
+    },
+    {
+      "epoch": 3.7899423782084862,
+      "grad_norm": 0.784389078617096,
+      "learning_rate": 1.6839642976721472e-07,
+      "loss": 0.1023,
+      "step": 7235
+    },
+    {
+      "epoch": 3.7899423782084862,
+      "loss_partial": 0.041015625,
+      "step": 7235
+    },
+    {
+      "epoch": 3.7899423782084862,
+      "loss_partial": 0.046875,
+      "step": 7235
+    },
+    {
+      "epoch": 3.7904662126767943,
+      "grad_norm": 0.9421697854995728,
+      "learning_rate": 1.6756202329943237e-07,
+      "loss": 0.0996,
+      "step": 7236
+    },
+    {
+      "epoch": 3.7904662126767943,
+      "loss_partial": 0.048583984375,
+      "step": 7236
+    },
+    {
+      "epoch": 3.7904662126767943,
+      "loss_partial": 0.05322265625,
+      "step": 7236
+    },
+    {
+      "epoch": 3.790990047145102,
+      "grad_norm": 1.0577630996704102,
+      "learning_rate": 1.6672967175038634e-07,
+      "loss": 0.0998,
+      "step": 7237
+    },
+    {
+      "epoch": 3.790990047145102,
+      "loss_partial": 0.04638671875,
+      "step": 7237
+    },
+    {
+      "epoch": 3.790990047145102,
+      "loss_partial": 0.049560546875,
+      "step": 7237
+    },
+    {
+      "epoch": 3.79151388161341,
+      "grad_norm": 0.9063138365745544,
+      "learning_rate": 1.6589937529403188e-07,
+      "loss": 0.1035,
+      "step": 7238
+    },
+    {
+      "epoch": 3.79151388161341,
+      "loss_partial": 0.043212890625,
+      "step": 7238
+    },
+    {
+      "epoch": 3.79151388161341,
+      "loss_partial": 0.053466796875,
+      "step": 7238
+    },
+    {
+      "epoch": 3.792037716081718,
+      "grad_norm": 0.859327495098114,
+      "learning_rate": 1.6507113410389687e-07,
+      "loss": 0.1024,
+      "step": 7239
+    },
+    {
+      "epoch": 3.792037716081718,
+      "loss_partial": 0.04833984375,
+      "step": 7239
+    },
+    {
+      "epoch": 3.792037716081718,
+      "loss_partial": 0.059326171875,
+      "step": 7239
+    },
+    {
+      "epoch": 3.792561550550026,
+      "grad_norm": 1.1248475313186646,
+      "learning_rate": 1.6424494835307615e-07,
+      "loss": 0.1005,
+      "step": 7240
+    },
+    {
+      "epoch": 3.792561550550026,
+      "loss_partial": 0.05517578125,
+      "step": 7240
+    },
+    {
+      "epoch": 3.792561550550026,
+      "loss_partial": 0.052001953125,
+      "step": 7240
+    },
+    {
+      "epoch": 3.793085385018334,
+      "grad_norm": 1.2567957639694214,
+      "learning_rate": 1.6342081821423826e-07,
+      "loss": 0.1013,
+      "step": 7241
+    },
+    {
+      "epoch": 3.793085385018334,
+      "loss_partial": 0.04931640625,
+      "step": 7241
+    },
+    {
+      "epoch": 3.793085385018334,
+      "loss_partial": 0.047607421875,
+      "step": 7241
+    },
+    {
+      "epoch": 3.793609219486642,
+      "grad_norm": 0.9492605924606323,
+      "learning_rate": 1.6259874385962217e-07,
+      "loss": 0.1,
+      "step": 7242
+    },
+    {
+      "epoch": 3.793609219486642,
+      "loss_partial": 0.045166015625,
+      "step": 7242
+    },
+    {
+      "epoch": 3.793609219486642,
+      "loss_partial": 0.048095703125,
+      "step": 7242
+    },
+    {
+      "epoch": 3.7941330539549503,
+      "grad_norm": 0.9487074613571167,
+      "learning_rate": 1.61778725461037e-07,
+      "loss": 0.1018,
+      "step": 7243
+    },
+    {
+      "epoch": 3.7941330539549503,
+      "loss_partial": 0.052001953125,
+      "step": 7243
+    },
+    {
+      "epoch": 3.7941330539549503,
+      "loss_partial": 0.056640625,
+      "step": 7243
+    },
+    {
+      "epoch": 3.7946568884232583,
+      "grad_norm": 0.9357080459594727,
+      "learning_rate": 1.609607631898591e-07,
+      "loss": 0.1014,
+      "step": 7244
+    },
+    {
+      "epoch": 3.7946568884232583,
+      "loss_partial": 0.044677734375,
+      "step": 7244
+    },
+    {
+      "epoch": 3.7946568884232583,
+      "loss_partial": 0.051025390625,
+      "step": 7244
+    },
+    {
+      "epoch": 3.7951807228915664,
+      "grad_norm": 0.9135075211524963,
+      "learning_rate": 1.6014485721703943e-07,
+      "loss": 0.1002,
+      "step": 7245
+    },
+    {
+      "epoch": 3.7951807228915664,
+      "loss_partial": 0.046875,
+      "step": 7245
+    },
+    {
+      "epoch": 3.7951807228915664,
+      "loss_partial": 0.04736328125,
+      "step": 7245
+    },
+    {
+      "epoch": 3.7957045573598744,
+      "grad_norm": 0.7253185510635376,
+      "learning_rate": 1.5933100771309606e-07,
+      "loss": 0.0996,
+      "step": 7246
+    },
+    {
+      "epoch": 3.7957045573598744,
+      "loss_partial": 0.052001953125,
+      "step": 7246
+    },
+    {
+      "epoch": 3.7957045573598744,
+      "loss_partial": 0.054931640625,
+      "step": 7246
+    },
+    {
+      "epoch": 3.7962283918281825,
+      "grad_norm": 0.8644667863845825,
+      "learning_rate": 1.5851921484812072e-07,
+      "loss": 0.0976,
+      "step": 7247
+    },
+    {
+      "epoch": 3.7962283918281825,
+      "loss_partial": 0.05908203125,
+      "step": 7247
+    },
+    {
+      "epoch": 3.7962283918281825,
+      "loss_partial": 0.056396484375,
+      "step": 7247
+    },
+    {
+      "epoch": 3.7967522262964906,
+      "grad_norm": 0.7731298208236694,
+      "learning_rate": 1.5770947879177102e-07,
+      "loss": 0.1026,
+      "step": 7248
+    },
+    {
+      "epoch": 3.7967522262964906,
+      "loss_partial": 0.05078125,
+      "step": 7248
+    },
+    {
+      "epoch": 3.7967522262964906,
+      "loss_partial": 0.045166015625,
+      "step": 7248
+    },
+    {
+      "epoch": 3.797276060764798,
+      "grad_norm": 0.8559470176696777,
+      "learning_rate": 1.56901799713276e-07,
+      "loss": 0.1025,
+      "step": 7249
+    },
+    {
+      "epoch": 3.797276060764798,
+      "loss_partial": 0.05126953125,
+      "step": 7249
+    },
+    {
+      "epoch": 3.797276060764798,
+      "loss_partial": 0.053955078125,
+      "step": 7249
+    },
+    {
+      "epoch": 3.7977998952331062,
+      "grad_norm": 1.0173673629760742,
+      "learning_rate": 1.5609617778143848e-07,
+      "loss": 0.0997,
+      "step": 7250
+    },
+    {
+      "epoch": 3.7977998952331062,
+      "loss_partial": 0.0673828125,
+      "step": 7250
+    },
+    {
+      "epoch": 3.7977998952331062,
+      "loss_partial": 0.0517578125,
+      "step": 7250
+    },
+    {
+      "epoch": 3.7983237297014143,
+      "grad_norm": 0.7665934562683105,
+      "learning_rate": 1.5529261316462706e-07,
+      "loss": 0.1006,
+      "step": 7251
+    },
+    {
+      "epoch": 3.7983237297014143,
+      "loss_partial": 0.04833984375,
+      "step": 7251
+    },
+    {
+      "epoch": 3.7983237297014143,
+      "loss_partial": 0.04833984375,
+      "step": 7251
+    },
+    {
+      "epoch": 3.7988475641697224,
+      "grad_norm": 0.8651757836341858,
+      "learning_rate": 1.5449110603078077e-07,
+      "loss": 0.0992,
+      "step": 7252
+    },
+    {
+      "epoch": 3.7988475641697224,
+      "loss_partial": 0.048583984375,
+      "step": 7252
+    },
+    {
+      "epoch": 3.7988475641697224,
+      "loss_partial": 0.051025390625,
+      "step": 7252
+    },
+    {
+      "epoch": 3.7993713986380304,
+      "grad_norm": 0.7949160933494568,
+      "learning_rate": 1.5369165654741113e-07,
+      "loss": 0.0979,
+      "step": 7253
+    },
+    {
+      "epoch": 3.7993713986380304,
+      "loss_partial": 0.04736328125,
+      "step": 7253
+    },
+    {
+      "epoch": 3.7993713986380304,
+      "loss_partial": 0.052001953125,
+      "step": 7253
+    },
+    {
+      "epoch": 3.7998952331063385,
+      "grad_norm": 0.9550251364707947,
+      "learning_rate": 1.5289426488159676e-07,
+      "loss": 0.1029,
+      "step": 7254
+    },
+    {
+      "epoch": 3.7998952331063385,
+      "loss_partial": 0.0498046875,
+      "step": 7254
+    },
+    {
+      "epoch": 3.7998952331063385,
+      "loss_partial": 0.04296875,
+      "step": 7254
+    },
+    {
+      "epoch": 3.8004190675746465,
+      "grad_norm": 0.7944902181625366,
+      "learning_rate": 1.5209893119998875e-07,
+      "loss": 0.0983,
+      "step": 7255
+    },
+    {
+      "epoch": 3.8004190675746465,
+      "loss_partial": 0.05224609375,
+      "step": 7255
+    },
+    {
+      "epoch": 3.8004190675746465,
+      "loss_partial": 0.05322265625,
+      "step": 7255
+    },
+    {
+      "epoch": 3.800942902042954,
+      "grad_norm": 1.0781021118164062,
+      "learning_rate": 1.513056556688075e-07,
+      "loss": 0.0992,
+      "step": 7256
+    },
+    {
+      "epoch": 3.800942902042954,
+      "loss_partial": 0.052490234375,
+      "step": 7256
+    },
+    {
+      "epoch": 3.800942902042954,
+      "loss_partial": 0.057373046875,
+      "step": 7256
+    },
+    {
+      "epoch": 3.801466736511262,
+      "grad_norm": 0.905755877494812,
+      "learning_rate": 1.505144384538404e-07,
+      "loss": 0.098,
+      "step": 7257
+    },
+    {
+      "epoch": 3.801466736511262,
+      "loss_partial": 0.0439453125,
+      "step": 7257
+    },
+    {
+      "epoch": 3.801466736511262,
+      "loss_partial": 0.047119140625,
+      "step": 7257
+    },
+    {
+      "epoch": 3.8019905709795703,
+      "grad_norm": 0.8981906771659851,
+      "learning_rate": 1.4972527972044848e-07,
+      "loss": 0.0983,
+      "step": 7258
+    },
+    {
+      "epoch": 3.8019905709795703,
+      "loss_partial": 0.055419921875,
+      "step": 7258
+    },
+    {
+      "epoch": 3.8019905709795703,
+      "loss_partial": 0.05419921875,
+      "step": 7258
+    },
+    {
+      "epoch": 3.8025144054478783,
+      "grad_norm": 1.0976026058197021,
+      "learning_rate": 1.4893817963356095e-07,
+      "loss": 0.1052,
+      "step": 7259
+    },
+    {
+      "epoch": 3.8025144054478783,
+      "loss_partial": 0.04150390625,
+      "step": 7259
+    },
+    {
+      "epoch": 3.8025144054478783,
+      "loss_partial": 0.05908203125,
+      "step": 7259
+    },
+    {
+      "epoch": 3.8030382399161864,
+      "grad_norm": 0.9004393219947815,
+      "learning_rate": 1.481531383576773e-07,
+      "loss": 0.1007,
+      "step": 7260
+    },
+    {
+      "epoch": 3.8030382399161864,
+      "loss_partial": 0.050537109375,
+      "step": 7260
+    },
+    {
+      "epoch": 3.8030382399161864,
+      "loss_partial": 0.04736328125,
+      "step": 7260
+    },
+    {
+      "epoch": 3.8035620743844945,
+      "grad_norm": 0.9876416325569153,
+      "learning_rate": 1.473701560568641e-07,
+      "loss": 0.1006,
+      "step": 7261
+    },
+    {
+      "epoch": 3.8035620743844945,
+      "loss_partial": 0.047607421875,
+      "step": 7261
+    },
+    {
+      "epoch": 3.8035620743844945,
+      "loss_partial": 0.05322265625,
+      "step": 7261
+    },
+    {
+      "epoch": 3.8040859088528025,
+      "grad_norm": 0.8047674894332886,
+      "learning_rate": 1.465892328947638e-07,
+      "loss": 0.1003,
+      "step": 7262
+    },
+    {
+      "epoch": 3.8040859088528025,
+      "loss_partial": 0.052490234375,
+      "step": 7262
+    },
+    {
+      "epoch": 3.8040859088528025,
+      "loss_partial": 0.0419921875,
+      "step": 7262
+    },
+    {
+      "epoch": 3.8046097433211106,
+      "grad_norm": 0.9507255554199219,
+      "learning_rate": 1.4581036903458025e-07,
+      "loss": 0.1013,
+      "step": 7263
+    },
+    {
+      "epoch": 3.8046097433211106,
+      "loss_partial": 0.050537109375,
+      "step": 7263
+    },
+    {
+      "epoch": 3.8046097433211106,
+      "loss_partial": 0.052978515625,
+      "step": 7263
+    },
+    {
+      "epoch": 3.8051335777894186,
+      "grad_norm": 0.9177258610725403,
+      "learning_rate": 1.4503356463909547e-07,
+      "loss": 0.1019,
+      "step": 7264
+    },
+    {
+      "epoch": 3.8051335777894186,
+      "loss_partial": 0.06103515625,
+      "step": 7264
+    },
+    {
+      "epoch": 3.8051335777894186,
+      "loss_partial": 0.0517578125,
+      "step": 7264
+    },
+    {
+      "epoch": 3.8056574122577267,
+      "grad_norm": 0.7873740196228027,
+      "learning_rate": 1.4425881987065183e-07,
+      "loss": 0.1021,
+      "step": 7265
+    },
+    {
+      "epoch": 3.8056574122577267,
+      "loss_partial": 0.060546875,
+      "step": 7265
+    },
+    {
+      "epoch": 3.8056574122577267,
+      "loss_partial": 0.0517578125,
+      "step": 7265
+    },
+    {
+      "epoch": 3.8061812467260348,
+      "grad_norm": 0.7599384784698486,
+      "learning_rate": 1.434861348911698e-07,
+      "loss": 0.1054,
+      "step": 7266
+    },
+    {
+      "epoch": 3.8061812467260348,
+      "loss_partial": 0.0615234375,
+      "step": 7266
+    },
+    {
+      "epoch": 3.8061812467260348,
+      "loss_partial": 0.05908203125,
+      "step": 7266
+    },
+    {
+      "epoch": 3.806705081194343,
+      "grad_norm": 1.1903008222579956,
+      "learning_rate": 1.4271550986213468e-07,
+      "loss": 0.0998,
+      "step": 7267
+    },
+    {
+      "epoch": 3.806705081194343,
+      "loss_partial": 0.050537109375,
+      "step": 7267
+    },
+    {
+      "epoch": 3.806705081194343,
+      "loss_partial": 0.0556640625,
+      "step": 7267
+    },
+    {
+      "epoch": 3.807228915662651,
+      "grad_norm": 0.8760203719139099,
+      "learning_rate": 1.4194694494460315e-07,
+      "loss": 0.1013,
+      "step": 7268
+    },
+    {
+      "epoch": 3.807228915662651,
+      "loss_partial": 0.049560546875,
+      "step": 7268
+    },
+    {
+      "epoch": 3.807228915662651,
+      "loss_partial": 0.0537109375,
+      "step": 7268
+    },
+    {
+      "epoch": 3.8077527501309585,
+      "grad_norm": 0.8339706063270569,
+      "learning_rate": 1.4118044029919785e-07,
+      "loss": 0.0961,
+      "step": 7269
+    },
+    {
+      "epoch": 3.8077527501309585,
+      "loss_partial": 0.056884765625,
+      "step": 7269
+    },
+    {
+      "epoch": 3.8077527501309585,
+      "loss_partial": 0.03955078125,
+      "step": 7269
+    },
+    {
+      "epoch": 3.8082765845992665,
+      "grad_norm": 0.7062340378761292,
+      "learning_rate": 1.4041599608611733e-07,
+      "loss": 0.0965,
+      "step": 7270
+    },
+    {
+      "epoch": 3.8082765845992665,
+      "loss_partial": 0.046142578125,
+      "step": 7270
+    },
+    {
+      "epoch": 3.8082765845992665,
+      "loss_partial": 0.05908203125,
+      "step": 7270
+    },
+    {
+      "epoch": 3.8088004190675746,
+      "grad_norm": 0.8797844052314758,
+      "learning_rate": 1.3965361246512266e-07,
+      "loss": 0.0988,
+      "step": 7271
+    },
+    {
+      "epoch": 3.8088004190675746,
+      "loss_partial": 0.049560546875,
+      "step": 7271
+    },
+    {
+      "epoch": 3.8088004190675746,
+      "loss_partial": 0.043212890625,
+      "step": 7271
+    },
+    {
+      "epoch": 3.8093242535358827,
+      "grad_norm": 0.9563020467758179,
+      "learning_rate": 1.388932895955486e-07,
+      "loss": 0.1013,
+      "step": 7272
+    },
+    {
+      "epoch": 3.8093242535358827,
+      "loss_partial": 0.05126953125,
+      "step": 7272
+    },
+    {
+      "epoch": 3.8093242535358827,
+      "loss_partial": 0.04345703125,
+      "step": 7272
+    },
+    {
+      "epoch": 3.8098480880041907,
+      "grad_norm": 0.8701832294464111,
+      "learning_rate": 1.3813502763629804e-07,
+      "loss": 0.0972,
+      "step": 7273
+    },
+    {
+      "epoch": 3.8098480880041907,
+      "loss_partial": 0.044921875,
+      "step": 7273
+    },
+    {
+      "epoch": 3.8098480880041907,
+      "loss_partial": 0.040283203125,
+      "step": 7273
+    },
+    {
+      "epoch": 3.810371922472499,
+      "grad_norm": 1.0007028579711914,
+      "learning_rate": 1.3737882674584314e-07,
+      "loss": 0.0995,
+      "step": 7274
+    },
+    {
+      "epoch": 3.810371922472499,
+      "loss_partial": 0.0498046875,
+      "step": 7274
+    },
+    {
+      "epoch": 3.810371922472499,
+      "loss_partial": 0.05078125,
+      "step": 7274
+    },
+    {
+      "epoch": 3.810895756940807,
+      "grad_norm": 0.7906286716461182,
+      "learning_rate": 1.366246870822241e-07,
+      "loss": 0.097,
+      "step": 7275
+    },
+    {
+      "epoch": 3.810895756940807,
+      "loss_partial": 0.05517578125,
+      "step": 7275
+    },
+    {
+      "epoch": 3.810895756940807,
+      "loss_partial": 0.047607421875,
+      "step": 7275
+    },
+    {
+      "epoch": 3.8114195914091145,
+      "grad_norm": 1.215386152267456,
+      "learning_rate": 1.3587260880305375e-07,
+      "loss": 0.1018,
+      "step": 7276
+    },
+    {
+      "epoch": 3.8114195914091145,
+      "loss_partial": 0.046142578125,
+      "step": 7276
+    },
+    {
+      "epoch": 3.8114195914091145,
+      "loss_partial": 0.0478515625,
+      "step": 7276
+    },
+    {
+      "epoch": 3.8119434258774225,
+      "grad_norm": 0.9266985654830933,
+      "learning_rate": 1.3512259206550748e-07,
+      "loss": 0.0946,
+      "step": 7277
+    },
+    {
+      "epoch": 3.8119434258774225,
+      "loss_partial": 0.04150390625,
+      "step": 7277
+    },
+    {
+      "epoch": 3.8119434258774225,
+      "loss_partial": 0.050537109375,
+      "step": 7277
+    },
+    {
+      "epoch": 3.8124672603457306,
+      "grad_norm": 0.7176968455314636,
+      "learning_rate": 1.343746370263388e-07,
+      "loss": 0.0947,
+      "step": 7278
+    },
+    {
+      "epoch": 3.8124672603457306,
+      "loss_partial": 0.05419921875,
+      "step": 7278
+    },
+    {
+      "epoch": 3.8124672603457306,
+      "loss_partial": 0.054931640625,
+      "step": 7278
+    },
+    {
+      "epoch": 3.8129910948140386,
+      "grad_norm": 0.9569528698921204,
+      "learning_rate": 1.3362874384186375e-07,
+      "loss": 0.0996,
+      "step": 7279
+    },
+    {
+      "epoch": 3.8129910948140386,
+      "loss_partial": 0.04736328125,
+      "step": 7279
+    },
+    {
+      "epoch": 3.8129910948140386,
+      "loss_partial": 0.05419921875,
+      "step": 7279
+    },
+    {
+      "epoch": 3.8135149292823467,
+      "grad_norm": 0.8352741003036499,
+      "learning_rate": 1.3288491266796877e-07,
+      "loss": 0.0938,
+      "step": 7280
+    },
+    {
+      "epoch": 3.8135149292823467,
+      "loss_partial": 0.053955078125,
+      "step": 7280
+    },
+    {
+      "epoch": 3.8135149292823467,
+      "loss_partial": 0.0517578125,
+      "step": 7280
+    },
+    {
+      "epoch": 3.8140387637506548,
+      "grad_norm": 0.9874464869499207,
+      "learning_rate": 1.321431436601106e-07,
+      "loss": 0.0964,
+      "step": 7281
+    },
+    {
+      "epoch": 3.8140387637506548,
+      "loss_partial": 0.054931640625,
+      "step": 7281
+    },
+    {
+      "epoch": 3.8140387637506548,
+      "loss_partial": 0.060546875,
+      "step": 7281
+    },
+    {
+      "epoch": 3.814562598218963,
+      "grad_norm": 0.8641014099121094,
+      "learning_rate": 1.3140343697331526e-07,
+      "loss": 0.0977,
+      "step": 7282
+    },
+    {
+      "epoch": 3.814562598218963,
+      "loss_partial": 0.051025390625,
+      "step": 7282
+    },
+    {
+      "epoch": 3.814562598218963,
+      "loss_partial": 0.0478515625,
+      "step": 7282
+    },
+    {
+      "epoch": 3.815086432687271,
+      "grad_norm": 0.8845450282096863,
+      "learning_rate": 1.3066579276217573e-07,
+      "loss": 0.0985,
+      "step": 7283
+    },
+    {
+      "epoch": 3.815086432687271,
+      "loss_partial": 0.048095703125,
+      "step": 7283
+    },
+    {
+      "epoch": 3.815086432687271,
+      "loss_partial": 0.0517578125,
+      "step": 7283
+    },
+    {
+      "epoch": 3.815610267155579,
+      "grad_norm": 0.8011734485626221,
+      "learning_rate": 1.2993021118085535e-07,
+      "loss": 0.0977,
+      "step": 7284
+    },
+    {
+      "epoch": 3.815610267155579,
+      "loss_partial": 0.05322265625,
+      "step": 7284
+    },
+    {
+      "epoch": 3.815610267155579,
+      "loss_partial": 0.064453125,
+      "step": 7284
+    },
+    {
+      "epoch": 3.816134101623887,
+      "grad_norm": 0.7794197201728821,
+      "learning_rate": 1.2919669238308564e-07,
+      "loss": 0.1038,
+      "step": 7285
+    },
+    {
+      "epoch": 3.816134101623887,
+      "loss_partial": 0.044189453125,
+      "step": 7285
+    },
+    {
+      "epoch": 3.816134101623887,
+      "loss_partial": 0.044921875,
+      "step": 7285
+    },
+    {
+      "epoch": 3.816657936092195,
+      "grad_norm": 0.7540302276611328,
+      "learning_rate": 1.2846523652216835e-07,
+      "loss": 0.0972,
+      "step": 7286
+    },
+    {
+      "epoch": 3.816657936092195,
+      "loss_partial": 0.05419921875,
+      "step": 7286
+    },
+    {
+      "epoch": 3.816657936092195,
+      "loss_partial": 0.055908203125,
+      "step": 7286
+    },
+    {
+      "epoch": 3.817181770560503,
+      "grad_norm": 0.7353787422180176,
+      "learning_rate": 1.2773584375097236e-07,
+      "loss": 0.1019,
+      "step": 7287
+    },
+    {
+      "epoch": 3.817181770560503,
+      "loss_partial": 0.048828125,
+      "step": 7287
+    },
+    {
+      "epoch": 3.817181770560503,
+      "loss_partial": 0.058349609375,
+      "step": 7287
+    },
+    {
+      "epoch": 3.8177056050288107,
+      "grad_norm": 0.9126858711242676,
+      "learning_rate": 1.2700851422193794e-07,
+      "loss": 0.1,
+      "step": 7288
+    },
+    {
+      "epoch": 3.8177056050288107,
+      "loss_partial": 0.053955078125,
+      "step": 7288
+    },
+    {
+      "epoch": 3.8177056050288107,
+      "loss_partial": 0.041748046875,
+      "step": 7288
+    },
+    {
+      "epoch": 3.818229439497119,
+      "grad_norm": 0.8635374903678894,
+      "learning_rate": 1.2628324808707016e-07,
+      "loss": 0.1008,
+      "step": 7289
+    },
+    {
+      "epoch": 3.818229439497119,
+      "loss_partial": 0.046630859375,
+      "step": 7289
+    },
+    {
+      "epoch": 3.818229439497119,
+      "loss_partial": 0.052978515625,
+      "step": 7289
+    },
+    {
+      "epoch": 3.818753273965427,
+      "grad_norm": 0.9325918555259705,
+      "learning_rate": 1.2556004549794775e-07,
+      "loss": 0.0998,
+      "step": 7290
+    },
+    {
+      "epoch": 3.818753273965427,
+      "loss_partial": 0.048828125,
+      "step": 7290
+    },
+    {
+      "epoch": 3.818753273965427,
+      "loss_partial": 0.041015625,
+      "step": 7290
+    },
+    {
+      "epoch": 3.819277108433735,
+      "grad_norm": 0.9368718862533569,
+      "learning_rate": 1.2483890660571318e-07,
+      "loss": 0.1033,
+      "step": 7291
+    },
+    {
+      "epoch": 3.819277108433735,
+      "loss_partial": 0.044921875,
+      "step": 7291
+    },
+    {
+      "epoch": 3.819277108433735,
+      "loss_partial": 0.043701171875,
+      "step": 7291
+    },
+    {
+      "epoch": 3.819800942902043,
+      "grad_norm": 0.8344730734825134,
+      "learning_rate": 1.2411983156108143e-07,
+      "loss": 0.0958,
+      "step": 7292
+    },
+    {
+      "epoch": 3.819800942902043,
+      "loss_partial": 0.0517578125,
+      "step": 7292
+    },
+    {
+      "epoch": 3.819800942902043,
+      "loss_partial": 0.041015625,
+      "step": 7292
+    },
+    {
+      "epoch": 3.820324777370351,
+      "grad_norm": 0.80898118019104,
+      "learning_rate": 1.2340282051433338e-07,
+      "loss": 0.0988,
+      "step": 7293
+    },
+    {
+      "epoch": 3.820324777370351,
+      "loss_partial": 0.04736328125,
+      "step": 7293
+    },
+    {
+      "epoch": 3.820324777370351,
+      "loss_partial": 0.046630859375,
+      "step": 7293
+    },
+    {
+      "epoch": 3.820848611838659,
+      "grad_norm": 0.7553834915161133,
+      "learning_rate": 1.2268787361532143e-07,
+      "loss": 0.0998,
+      "step": 7294
+    },
+    {
+      "epoch": 3.820848611838659,
+      "loss_partial": 0.052001953125,
+      "step": 7294
+    },
+    {
+      "epoch": 3.820848611838659,
+      "loss_partial": 0.055908203125,
+      "step": 7294
+    },
+    {
+      "epoch": 3.821372446306967,
+      "grad_norm": 0.7955374717712402,
+      "learning_rate": 1.2197499101346377e-07,
+      "loss": 0.0989,
+      "step": 7295
+    },
+    {
+      "epoch": 3.821372446306967,
+      "loss_partial": 0.049560546875,
+      "step": 7295
+    },
+    {
+      "epoch": 3.821372446306967,
+      "loss_partial": 0.0517578125,
+      "step": 7295
+    },
+    {
+      "epoch": 3.8218962807752748,
+      "grad_norm": 0.8111139535903931,
+      "learning_rate": 1.2126417285775017e-07,
+      "loss": 0.0965,
+      "step": 7296
+    },
+    {
+      "epoch": 3.8218962807752748,
+      "loss_partial": 0.049072265625,
+      "step": 7296
+    },
+    {
+      "epoch": 3.8218962807752748,
+      "loss_partial": 0.041259765625,
+      "step": 7296
+    },
+    {
+      "epoch": 3.822420115243583,
+      "grad_norm": 0.742584228515625,
+      "learning_rate": 1.205554192967351e-07,
+      "loss": 0.0993,
+      "step": 7297
+    },
+    {
+      "epoch": 3.822420115243583,
+      "loss_partial": 0.03662109375,
+      "step": 7297
+    },
+    {
+      "epoch": 3.822420115243583,
+      "loss_partial": 0.04296875,
+      "step": 7297
+    },
+    {
+      "epoch": 3.822943949711891,
+      "grad_norm": 0.8428556323051453,
+      "learning_rate": 1.1984873047854561e-07,
+      "loss": 0.097,
+      "step": 7298
+    },
+    {
+      "epoch": 3.822943949711891,
+      "loss_partial": 0.044677734375,
+      "step": 7298
+    },
+    {
+      "epoch": 3.822943949711891,
+      "loss_partial": 0.041748046875,
+      "step": 7298
+    },
+    {
+      "epoch": 3.823467784180199,
+      "grad_norm": 0.6970475912094116,
+      "learning_rate": 1.1914410655087361e-07,
+      "loss": 0.0966,
+      "step": 7299
+    },
+    {
+      "epoch": 3.823467784180199,
+      "loss_partial": 0.047607421875,
+      "step": 7299
+    },
+    {
+      "epoch": 3.823467784180199,
+      "loss_partial": 0.046142578125,
+      "step": 7299
+    },
+    {
+      "epoch": 3.823991618648507,
+      "grad_norm": 0.953976035118103,
+      "learning_rate": 1.1844154766098237e-07,
+      "loss": 0.0995,
+      "step": 7300
+    },
+    {
+      "epoch": 3.823991618648507,
+      "loss_partial": 0.042236328125,
+      "step": 7300
+    },
+    {
+      "epoch": 3.823991618648507,
+      "loss_partial": 0.050048828125,
+      "step": 7300
+    },
+    {
+      "epoch": 3.824515453116815,
+      "grad_norm": 0.8593195676803589,
+      "learning_rate": 1.1774105395570335e-07,
+      "loss": 0.0967,
+      "step": 7301
+    },
+    {
+      "epoch": 3.824515453116815,
+      "loss_partial": 0.052490234375,
+      "step": 7301
+    },
+    {
+      "epoch": 3.824515453116815,
+      "loss_partial": 0.04736328125,
+      "step": 7301
+    },
+    {
+      "epoch": 3.825039287585123,
+      "grad_norm": 0.7130874395370483,
+      "learning_rate": 1.1704262558143165e-07,
+      "loss": 0.0996,
+      "step": 7302
+    },
+    {
+      "epoch": 3.825039287585123,
+      "loss_partial": 0.0478515625,
+      "step": 7302
+    },
+    {
+      "epoch": 3.825039287585123,
+      "loss_partial": 0.042236328125,
+      "step": 7302
+    },
+    {
+      "epoch": 3.825563122053431,
+      "grad_norm": 0.7171645164489746,
+      "learning_rate": 1.1634626268413829e-07,
+      "loss": 0.0979,
+      "step": 7303
+    },
+    {
+      "epoch": 3.825563122053431,
+      "loss_partial": 0.0498046875,
+      "step": 7303
+    },
+    {
+      "epoch": 3.825563122053431,
+      "loss_partial": 0.04248046875,
+      "step": 7303
+    },
+    {
+      "epoch": 3.8260869565217392,
+      "grad_norm": 0.9411013126373291,
+      "learning_rate": 1.1565196540935686e-07,
+      "loss": 0.0995,
+      "step": 7304
+    },
+    {
+      "epoch": 3.8260869565217392,
+      "loss_partial": 0.052001953125,
+      "step": 7304
+    },
+    {
+      "epoch": 3.8260869565217392,
+      "loss_partial": 0.04931640625,
+      "step": 7304
+    },
+    {
+      "epoch": 3.8266107909900473,
+      "grad_norm": 0.9792027473449707,
+      "learning_rate": 1.149597339021924e-07,
+      "loss": 0.0953,
+      "step": 7305
+    },
+    {
+      "epoch": 3.8266107909900473,
+      "loss_partial": 0.0439453125,
+      "step": 7305
+    },
+    {
+      "epoch": 3.8266107909900473,
+      "loss_partial": 0.0673828125,
+      "step": 7305
+    },
+    {
+      "epoch": 3.8271346254583554,
+      "grad_norm": 0.688230037689209,
+      "learning_rate": 1.1426956830731472e-07,
+      "loss": 0.0985,
+      "step": 7306
+    },
+    {
+      "epoch": 3.8271346254583554,
+      "loss_partial": 0.05078125,
+      "step": 7306
+    },
+    {
+      "epoch": 3.8271346254583554,
+      "loss_partial": 0.04443359375,
+      "step": 7306
+    },
+    {
+      "epoch": 3.8276584599266634,
+      "grad_norm": 0.7570824027061462,
+      "learning_rate": 1.1358146876896737e-07,
+      "loss": 0.099,
+      "step": 7307
+    },
+    {
+      "epoch": 3.8276584599266634,
+      "loss_partial": 0.04833984375,
+      "step": 7307
+    },
+    {
+      "epoch": 3.8276584599266634,
+      "loss_partial": 0.05322265625,
+      "step": 7307
+    },
+    {
+      "epoch": 3.828182294394971,
+      "grad_norm": 0.8589209914207458,
+      "learning_rate": 1.1289543543095638e-07,
+      "loss": 0.0997,
+      "step": 7308
+    },
+    {
+      "epoch": 3.828182294394971,
+      "loss_partial": 0.046875,
+      "step": 7308
+    },
+    {
+      "epoch": 3.828182294394971,
+      "loss_partial": 0.041259765625,
+      "step": 7308
+    },
+    {
+      "epoch": 3.828706128863279,
+      "grad_norm": 0.6337153315544128,
+      "learning_rate": 1.1221146843665932e-07,
+      "loss": 0.0981,
+      "step": 7309
+    },
+    {
+      "epoch": 3.828706128863279,
+      "loss_partial": 0.048828125,
+      "step": 7309
+    },
+    {
+      "epoch": 3.828706128863279,
+      "loss_partial": 0.05224609375,
+      "step": 7309
+    },
+    {
+      "epoch": 3.829229963331587,
+      "grad_norm": 0.8420665860176086,
+      "learning_rate": 1.115295679290218e-07,
+      "loss": 0.1001,
+      "step": 7310
+    },
+    {
+      "epoch": 3.829229963331587,
+      "loss_partial": 0.05712890625,
+      "step": 7310
+    },
+    {
+      "epoch": 3.829229963331587,
+      "loss_partial": 0.049560546875,
+      "step": 7310
+    },
+    {
+      "epoch": 3.829753797799895,
+      "grad_norm": 0.7784604430198669,
+      "learning_rate": 1.1084973405055543e-07,
+      "loss": 0.1013,
+      "step": 7311
+    },
+    {
+      "epoch": 3.829753797799895,
+      "loss_partial": 0.0625,
+      "step": 7311
+    },
+    {
+      "epoch": 3.829753797799895,
+      "loss_partial": 0.050537109375,
+      "step": 7311
+    },
+    {
+      "epoch": 3.8302776322682033,
+      "grad_norm": 1.0021781921386719,
+      "learning_rate": 1.1017196694334209e-07,
+      "loss": 0.0993,
+      "step": 7312
+    },
+    {
+      "epoch": 3.8302776322682033,
+      "loss_partial": 0.0498046875,
+      "step": 7312
+    },
+    {
+      "epoch": 3.8302776322682033,
+      "loss_partial": 0.04931640625,
+      "step": 7312
+    },
+    {
+      "epoch": 3.8308014667365113,
+      "grad_norm": 0.9793874025344849,
+      "learning_rate": 1.0949626674903179e-07,
+      "loss": 0.0977,
+      "step": 7313
+    },
+    {
+      "epoch": 3.8308014667365113,
+      "loss_partial": 0.0478515625,
+      "step": 7313
+    },
+    {
+      "epoch": 3.8308014667365113,
+      "loss_partial": 0.04931640625,
+      "step": 7313
+    },
+    {
+      "epoch": 3.8313253012048194,
+      "grad_norm": 0.9061664342880249,
+      "learning_rate": 1.0882263360883938e-07,
+      "loss": 0.104,
+      "step": 7314
+    },
+    {
+      "epoch": 3.8313253012048194,
+      "loss_partial": 0.044189453125,
+      "step": 7314
+    },
+    {
+      "epoch": 3.8313253012048194,
+      "loss_partial": 0.045654296875,
+      "step": 7314
+    },
+    {
+      "epoch": 3.831849135673127,
+      "grad_norm": 0.8576682806015015,
+      "learning_rate": 1.0815106766355221e-07,
+      "loss": 0.0978,
+      "step": 7315
+    },
+    {
+      "epoch": 3.831849135673127,
+      "loss_partial": 0.055419921875,
+      "step": 7315
+    },
+    {
+      "epoch": 3.831849135673127,
+      "loss_partial": 0.042236328125,
+      "step": 7315
+    },
+    {
+      "epoch": 3.832372970141435,
+      "grad_norm": 0.8116436004638672,
+      "learning_rate": 1.0748156905352247e-07,
+      "loss": 0.097,
+      "step": 7316
+    },
+    {
+      "epoch": 3.832372970141435,
+      "loss_partial": 0.07373046875,
+      "step": 7316
+    },
+    {
+      "epoch": 3.832372970141435,
+      "loss_partial": 0.04638671875,
+      "step": 7316
+    },
+    {
+      "epoch": 3.832896804609743,
+      "grad_norm": 0.9789495468139648,
+      "learning_rate": 1.0681413791867157e-07,
+      "loss": 0.102,
+      "step": 7317
+    },
+    {
+      "epoch": 3.832896804609743,
+      "loss_partial": 0.047119140625,
+      "step": 7317
+    },
+    {
+      "epoch": 3.832896804609743,
+      "loss_partial": 0.043701171875,
+      "step": 7317
+    },
+    {
+      "epoch": 3.833420639078051,
+      "grad_norm": 0.7253962755203247,
+      "learning_rate": 1.061487743984868e-07,
+      "loss": 0.098,
+      "step": 7318
+    },
+    {
+      "epoch": 3.833420639078051,
+      "loss_partial": 0.050537109375,
+      "step": 7318
+    },
+    {
+      "epoch": 3.833420639078051,
+      "loss_partial": 0.051513671875,
+      "step": 7318
+    },
+    {
+      "epoch": 3.8339444735463593,
+      "grad_norm": 0.8685640692710876,
+      "learning_rate": 1.0548547863202696e-07,
+      "loss": 0.1011,
+      "step": 7319
+    },
+    {
+      "epoch": 3.8339444735463593,
+      "loss_partial": 0.0419921875,
+      "step": 7319
+    },
+    {
+      "epoch": 3.8339444735463593,
+      "loss_partial": 0.05224609375,
+      "step": 7319
+    },
+    {
+      "epoch": 3.8344683080146673,
+      "grad_norm": 0.8026111125946045,
+      "learning_rate": 1.048242507579167e-07,
+      "loss": 0.0957,
+      "step": 7320
+    },
+    {
+      "epoch": 3.8344683080146673,
+      "loss_partial": 0.050537109375,
+      "step": 7320
+    },
+    {
+      "epoch": 3.8344683080146673,
+      "loss_partial": 0.04931640625,
+      "step": 7320
+    },
+    {
+      "epoch": 3.8349921424829754,
+      "grad_norm": 0.8045693635940552,
+      "learning_rate": 1.0416509091434879e-07,
+      "loss": 0.0998,
+      "step": 7321
+    },
+    {
+      "epoch": 3.8349921424829754,
+      "loss_partial": 0.064453125,
+      "step": 7321
+    },
+    {
+      "epoch": 3.8349921424829754,
+      "loss_partial": 0.05078125,
+      "step": 7321
+    },
+    {
+      "epoch": 3.8355159769512834,
+      "grad_norm": 0.8476414680480957,
+      "learning_rate": 1.0350799923908194e-07,
+      "loss": 0.101,
+      "step": 7322
+    },
+    {
+      "epoch": 3.8355159769512834,
+      "loss_partial": 0.050048828125,
+      "step": 7322
+    },
+    {
+      "epoch": 3.8355159769512834,
+      "loss_partial": 0.046630859375,
+      "step": 7322
+    },
+    {
+      "epoch": 3.8360398114195915,
+      "grad_norm": 0.6961175799369812,
+      "learning_rate": 1.028529758694452e-07,
+      "loss": 0.0982,
+      "step": 7323
+    },
+    {
+      "epoch": 3.8360398114195915,
+      "loss_partial": 0.04931640625,
+      "step": 7323
+    },
+    {
+      "epoch": 3.8360398114195915,
+      "loss_partial": 0.0537109375,
+      "step": 7323
+    },
+    {
+      "epoch": 3.8365636458878996,
+      "grad_norm": 0.7691652774810791,
+      "learning_rate": 1.022000209423346e-07,
+      "loss": 0.1021,
+      "step": 7324
+    },
+    {
+      "epoch": 3.8365636458878996,
+      "loss_partial": 0.045166015625,
+      "step": 7324
+    },
+    {
+      "epoch": 3.8365636458878996,
+      "loss_partial": 0.051025390625,
+      "step": 7324
+    },
+    {
+      "epoch": 3.8370874803562076,
+      "grad_norm": 0.9861047267913818,
+      "learning_rate": 1.0154913459421323e-07,
+      "loss": 0.0988,
+      "step": 7325
+    },
+    {
+      "epoch": 3.8370874803562076,
+      "loss_partial": 0.053466796875,
+      "step": 7325
+    },
+    {
+      "epoch": 3.8370874803562076,
+      "loss_partial": 0.049560546875,
+      "step": 7325
+    },
+    {
+      "epoch": 3.8376113148245157,
+      "grad_norm": 1.0337406396865845,
+      "learning_rate": 1.0090031696111335e-07,
+      "loss": 0.1006,
+      "step": 7326
+    },
+    {
+      "epoch": 3.8376113148245157,
+      "loss_partial": 0.048095703125,
+      "step": 7326
+    },
+    {
+      "epoch": 3.8376113148245157,
+      "loss_partial": 0.0419921875,
+      "step": 7326
+    },
+    {
+      "epoch": 3.8381351492928237,
+      "grad_norm": 0.722740888595581,
+      "learning_rate": 1.0025356817863208e-07,
+      "loss": 0.1018,
+      "step": 7327
+    },
+    {
+      "epoch": 3.8381351492928237,
+      "loss_partial": 0.050048828125,
+      "step": 7327
+    },
+    {
+      "epoch": 3.8381351492928237,
+      "loss_partial": 0.055908203125,
+      "step": 7327
+    },
+    {
+      "epoch": 3.8386589837611313,
+      "grad_norm": 0.7252707481384277,
+      "learning_rate": 9.960888838193794e-08,
+      "loss": 0.1018,
+      "step": 7328
+    },
+    {
+      "epoch": 3.8386589837611313,
+      "loss_partial": 0.04736328125,
+      "step": 7328
+    },
+    {
+      "epoch": 3.8386589837611313,
+      "loss_partial": 0.04833984375,
+      "step": 7328
+    },
+    {
+      "epoch": 3.8391828182294394,
+      "grad_norm": 0.665025532245636,
+      "learning_rate": 9.896627770576316e-08,
+      "loss": 0.1004,
+      "step": 7329
+    },
+    {
+      "epoch": 3.8391828182294394,
+      "loss_partial": 0.045654296875,
+      "step": 7329
+    },
+    {
+      "epoch": 3.8391828182294394,
+      "loss_partial": 0.038330078125,
+      "step": 7329
+    },
+    {
+      "epoch": 3.8397066526977475,
+      "grad_norm": 0.9269627332687378,
+      "learning_rate": 9.832573628440922e-08,
+      "loss": 0.0979,
+      "step": 7330
+    },
+    {
+      "epoch": 3.8397066526977475,
+      "loss_partial": 0.04931640625,
+      "step": 7330
+    },
+    {
+      "epoch": 3.8397066526977475,
+      "loss_partial": 0.04931640625,
+      "step": 7330
+    },
+    {
+      "epoch": 3.8402304871660555,
+      "grad_norm": 0.7570080757141113,
+      "learning_rate": 9.76872642517479e-08,
+      "loss": 0.0959,
+      "step": 7331
+    },
+    {
+      "epoch": 3.8402304871660555,
+      "loss_partial": 0.049072265625,
+      "step": 7331
+    },
+    {
+      "epoch": 3.8402304871660555,
+      "loss_partial": 0.059326171875,
+      "step": 7331
+    },
+    {
+      "epoch": 3.8407543216343636,
+      "grad_norm": 0.8611403703689575,
+      "learning_rate": 9.705086174121358e-08,
+      "loss": 0.1008,
+      "step": 7332
+    },
+    {
+      "epoch": 3.8407543216343636,
+      "loss_partial": 0.053466796875,
+      "step": 7332
+    },
+    {
+      "epoch": 3.8407543216343636,
+      "loss_partial": 0.051513671875,
+      "step": 7332
+    },
+    {
+      "epoch": 3.8412781561026716,
+      "grad_norm": 0.7928159832954407,
+      "learning_rate": 9.64165288858121e-08,
+      "loss": 0.0952,
+      "step": 7333
+    },
+    {
+      "epoch": 3.8412781561026716,
+      "loss_partial": 0.053955078125,
+      "step": 7333
+    },
+    {
+      "epoch": 3.8412781561026716,
+      "loss_partial": 0.050048828125,
+      "step": 7333
+    },
+    {
+      "epoch": 3.8418019905709797,
+      "grad_norm": 0.9601020812988281,
+      "learning_rate": 9.578426581811295e-08,
+      "loss": 0.1009,
+      "step": 7334
+    },
+    {
+      "epoch": 3.8418019905709797,
+      "loss_partial": 0.03955078125,
+      "step": 7334
+    },
+    {
+      "epoch": 3.8418019905709797,
+      "loss_partial": 0.041748046875,
+      "step": 7334
+    },
+    {
+      "epoch": 3.8423258250392873,
+      "grad_norm": 0.9141706824302673,
+      "learning_rate": 9.515407267025601e-08,
+      "loss": 0.0958,
+      "step": 7335
+    },
+    {
+      "epoch": 3.8423258250392873,
+      "loss_partial": 0.0478515625,
+      "step": 7335
+    },
+    {
+      "epoch": 3.8423258250392873,
+      "loss_partial": 0.0498046875,
+      "step": 7335
+    },
+    {
+      "epoch": 3.8428496595075954,
+      "grad_norm": 0.9691590070724487,
+      "learning_rate": 9.452594957394923e-08,
+      "loss": 0.0981,
+      "step": 7336
+    },
+    {
+      "epoch": 3.8428496595075954,
+      "loss_partial": 0.046630859375,
+      "step": 7336
+    },
+    {
+      "epoch": 3.8428496595075954,
+      "loss_partial": 0.046875,
+      "step": 7336
+    },
+    {
+      "epoch": 3.8433734939759034,
+      "grad_norm": 0.7228115200996399,
+      "learning_rate": 9.38998966604654e-08,
+      "loss": 0.0985,
+      "step": 7337
+    },
+    {
+      "epoch": 3.8433734939759034,
+      "loss_partial": 0.04296875,
+      "step": 7337
+    },
+    {
+      "epoch": 3.8433734939759034,
+      "loss_partial": 0.05224609375,
+      "step": 7337
+    },
+    {
+      "epoch": 3.8438973284442115,
+      "grad_norm": 0.7446838021278381,
+      "learning_rate": 9.32759140606454e-08,
+      "loss": 0.0981,
+      "step": 7338
+    },
+    {
+      "epoch": 3.8438973284442115,
+      "loss_partial": 0.06103515625,
+      "step": 7338
+    },
+    {
+      "epoch": 3.8438973284442115,
+      "loss_partial": 0.044921875,
+      "step": 7338
+    },
+    {
+      "epoch": 3.8444211629125196,
+      "grad_norm": 0.9821571707725525,
+      "learning_rate": 9.265400190489826e-08,
+      "loss": 0.0995,
+      "step": 7339
+    },
+    {
+      "epoch": 3.8444211629125196,
+      "loss_partial": 0.04541015625,
+      "step": 7339
+    },
+    {
+      "epoch": 3.8444211629125196,
+      "loss_partial": 0.043701171875,
+      "step": 7339
+    },
+    {
+      "epoch": 3.8449449973808276,
+      "grad_norm": 0.716852068901062,
+      "learning_rate": 9.20341603232e-08,
+      "loss": 0.1001,
+      "step": 7340
+    },
+    {
+      "epoch": 3.8449449973808276,
+      "loss_partial": 0.0458984375,
+      "step": 7340
+    },
+    {
+      "epoch": 3.8449449973808276,
+      "loss_partial": 0.0537109375,
+      "step": 7340
+    },
+    {
+      "epoch": 3.8454688318491357,
+      "grad_norm": 0.830723762512207,
+      "learning_rate": 9.141638944509257e-08,
+      "loss": 0.0977,
+      "step": 7341
+    },
+    {
+      "epoch": 3.8454688318491357,
+      "loss_partial": 0.058349609375,
+      "step": 7341
+    },
+    {
+      "epoch": 3.8454688318491357,
+      "loss_partial": 0.0673828125,
+      "step": 7341
+    },
+    {
+      "epoch": 3.8459926663174437,
+      "grad_norm": 0.8561007380485535,
+      "learning_rate": 9.0800689399686e-08,
+      "loss": 0.1009,
+      "step": 7342
+    },
+    {
+      "epoch": 3.8459926663174437,
+      "loss_partial": 0.041015625,
+      "step": 7342
+    },
+    {
+      "epoch": 3.8459926663174437,
+      "loss_partial": 0.048828125,
+      "step": 7342
+    },
+    {
+      "epoch": 3.846516500785752,
+      "grad_norm": 0.6863334774971008,
+      "learning_rate": 9.01870603156596e-08,
+      "loss": 0.0966,
+      "step": 7343
+    },
+    {
+      "epoch": 3.846516500785752,
+      "loss_partial": 0.046142578125,
+      "step": 7343
+    },
+    {
+      "epoch": 3.846516500785752,
+      "loss_partial": 0.04345703125,
+      "step": 7343
+    },
+    {
+      "epoch": 3.84704033525406,
+      "grad_norm": 0.8942787051200867,
+      "learning_rate": 8.957550232125744e-08,
+      "loss": 0.0972,
+      "step": 7344
+    },
+    {
+      "epoch": 3.84704033525406,
+      "loss_partial": 0.06787109375,
+      "step": 7344
+    },
+    {
+      "epoch": 3.84704033525406,
+      "loss_partial": 0.047607421875,
+      "step": 7344
+    },
+    {
+      "epoch": 3.847564169722368,
+      "grad_norm": 0.8105770349502563,
+      "learning_rate": 8.896601554429063e-08,
+      "loss": 0.102,
+      "step": 7345
+    },
+    {
+      "epoch": 3.847564169722368,
+      "loss_partial": 0.05322265625,
+      "step": 7345
+    },
+    {
+      "epoch": 3.847564169722368,
+      "loss_partial": 0.0458984375,
+      "step": 7345
+    },
+    {
+      "epoch": 3.848088004190676,
+      "grad_norm": 0.9721204042434692,
+      "learning_rate": 8.835860011213837e-08,
+      "loss": 0.1015,
+      "step": 7346
+    },
+    {
+      "epoch": 3.848088004190676,
+      "loss_partial": 0.047119140625,
+      "step": 7346
+    },
+    {
+      "epoch": 3.848088004190676,
+      "loss_partial": 0.050537109375,
+      "step": 7346
+    },
+    {
+      "epoch": 3.8486118386589836,
+      "grad_norm": 0.7822608947753906,
+      "learning_rate": 8.775325615174801e-08,
+      "loss": 0.1,
+      "step": 7347
+    },
+    {
+      "epoch": 3.8486118386589836,
+      "loss_partial": 0.040283203125,
+      "step": 7347
+    },
+    {
+      "epoch": 3.8486118386589836,
+      "loss_partial": 0.052001953125,
+      "step": 7347
+    },
+    {
+      "epoch": 3.8491356731272917,
+      "grad_norm": 0.8084084391593933,
+      "learning_rate": 8.714998378963057e-08,
+      "loss": 0.0968,
+      "step": 7348
+    },
+    {
+      "epoch": 3.8491356731272917,
+      "loss_partial": 0.05126953125,
+      "step": 7348
+    },
+    {
+      "epoch": 3.8491356731272917,
+      "loss_partial": 0.052978515625,
+      "step": 7348
+    },
+    {
+      "epoch": 3.8496595075955997,
+      "grad_norm": 0.8961449861526489,
+      "learning_rate": 8.654878315186743e-08,
+      "loss": 0.1024,
+      "step": 7349
+    },
+    {
+      "epoch": 3.8496595075955997,
+      "loss_partial": 0.04638671875,
+      "step": 7349
+    },
+    {
+      "epoch": 3.8496595075955997,
+      "loss_partial": 0.047119140625,
+      "step": 7349
+    },
+    {
+      "epoch": 3.8501833420639078,
+      "grad_norm": 0.8428197503089905,
+      "learning_rate": 8.594965436410696e-08,
+      "loss": 0.0967,
+      "step": 7350
+    },
+    {
+      "epoch": 3.8501833420639078,
+      "loss_partial": 0.046875,
+      "step": 7350
+    },
+    {
+      "epoch": 3.8501833420639078,
+      "loss_partial": 0.04833984375,
+      "step": 7350
+    },
+    {
+      "epoch": 3.850707176532216,
+      "grad_norm": 0.7106029391288757,
+      "learning_rate": 8.535259755156012e-08,
+      "loss": 0.1005,
+      "step": 7351
+    },
+    {
+      "epoch": 3.850707176532216,
+      "loss_partial": 0.05615234375,
+      "step": 7351
+    },
+    {
+      "epoch": 3.850707176532216,
+      "loss_partial": 0.042236328125,
+      "step": 7351
+    },
+    {
+      "epoch": 3.851231011000524,
+      "grad_norm": 0.6839548945426941,
+      "learning_rate": 8.475761283901151e-08,
+      "loss": 0.099,
+      "step": 7352
+    },
+    {
+      "epoch": 3.851231011000524,
+      "loss_partial": 0.05810546875,
+      "step": 7352
+    },
+    {
+      "epoch": 3.851231011000524,
+      "loss_partial": 0.051513671875,
+      "step": 7352
+    },
+    {
+      "epoch": 3.851754845468832,
+      "grad_norm": 0.8430762887001038,
+      "learning_rate": 8.416470035080725e-08,
+      "loss": 0.1001,
+      "step": 7353
+    },
+    {
+      "epoch": 3.851754845468832,
+      "loss_partial": 0.05126953125,
+      "step": 7353
+    },
+    {
+      "epoch": 3.851754845468832,
+      "loss_partial": 0.048583984375,
+      "step": 7353
+    },
+    {
+      "epoch": 3.85227867993714,
+      "grad_norm": 0.9940699934959412,
+      "learning_rate": 8.357386021086267e-08,
+      "loss": 0.1023,
+      "step": 7354
+    },
+    {
+      "epoch": 3.85227867993714,
+      "loss_partial": 0.04833984375,
+      "step": 7354
+    },
+    {
+      "epoch": 3.85227867993714,
+      "loss_partial": 0.052978515625,
+      "step": 7354
+    },
+    {
+      "epoch": 3.8528025144054476,
+      "grad_norm": 0.7766056060791016,
+      "learning_rate": 8.298509254266008e-08,
+      "loss": 0.0973,
+      "step": 7355
+    },
+    {
+      "epoch": 3.8528025144054476,
+      "loss_partial": 0.04833984375,
+      "step": 7355
+    },
+    {
+      "epoch": 3.8528025144054476,
+      "loss_partial": 0.0439453125,
+      "step": 7355
+    },
+    {
+      "epoch": 3.8533263488737557,
+      "grad_norm": 0.6547693610191345,
+      "learning_rate": 8.239839746924882e-08,
+      "loss": 0.0999,
+      "step": 7356
+    },
+    {
+      "epoch": 3.8533263488737557,
+      "loss_partial": 0.05419921875,
+      "step": 7356
+    },
+    {
+      "epoch": 3.8533263488737557,
+      "loss_partial": 0.0556640625,
+      "step": 7356
+    },
+    {
+      "epoch": 3.8538501833420638,
+      "grad_norm": 1.0350075960159302,
+      "learning_rate": 8.181377511324306e-08,
+      "loss": 0.1,
+      "step": 7357
+    },
+    {
+      "epoch": 3.8538501833420638,
+      "loss_partial": 0.057373046875,
+      "step": 7357
+    },
+    {
+      "epoch": 3.8538501833420638,
+      "loss_partial": 0.04931640625,
+      "step": 7357
+    },
+    {
+      "epoch": 3.854374017810372,
+      "grad_norm": 0.81016606092453,
+      "learning_rate": 8.123122559682728e-08,
+      "loss": 0.0993,
+      "step": 7358
+    },
+    {
+      "epoch": 3.854374017810372,
+      "loss_partial": 0.04541015625,
+      "step": 7358
+    },
+    {
+      "epoch": 3.854374017810372,
+      "loss_partial": 0.05126953125,
+      "step": 7358
+    },
+    {
+      "epoch": 3.85489785227868,
+      "grad_norm": 0.7116239070892334,
+      "learning_rate": 8.065074904174741e-08,
+      "loss": 0.0986,
+      "step": 7359
+    },
+    {
+      "epoch": 3.85489785227868,
+      "loss_partial": 0.04443359375,
+      "step": 7359
+    },
+    {
+      "epoch": 3.85489785227868,
+      "loss_partial": 0.04638671875,
+      "step": 7359
+    },
+    {
+      "epoch": 3.855421686746988,
+      "grad_norm": 0.7237023711204529,
+      "learning_rate": 8.007234556932308e-08,
+      "loss": 0.1018,
+      "step": 7360
+    },
+    {
+      "epoch": 3.855421686746988,
+      "loss_partial": 0.04638671875,
+      "step": 7360
+    },
+    {
+      "epoch": 3.855421686746988,
+      "loss_partial": 0.0556640625,
+      "step": 7360
+    },
+    {
+      "epoch": 3.855945521215296,
+      "grad_norm": 0.9269258379936218,
+      "learning_rate": 7.949601530043537e-08,
+      "loss": 0.0973,
+      "step": 7361
+    },
+    {
+      "epoch": 3.855945521215296,
+      "loss_partial": 0.062255859375,
+      "step": 7361
+    },
+    {
+      "epoch": 3.855945521215296,
+      "loss_partial": 0.046875,
+      "step": 7361
+    },
+    {
+      "epoch": 3.856469355683604,
+      "grad_norm": 0.8060240149497986,
+      "learning_rate": 7.892175835553462e-08,
+      "loss": 0.1029,
+      "step": 7362
+    },
+    {
+      "epoch": 3.856469355683604,
+      "loss_partial": 0.05322265625,
+      "step": 7362
+    },
+    {
+      "epoch": 3.856469355683604,
+      "loss_partial": 0.04248046875,
+      "step": 7362
+    },
+    {
+      "epoch": 3.856993190151912,
+      "grad_norm": 0.7748936414718628,
+      "learning_rate": 7.834957485463479e-08,
+      "loss": 0.0936,
+      "step": 7363
+    },
+    {
+      "epoch": 3.856993190151912,
+      "loss_partial": 0.05419921875,
+      "step": 7363
+    },
+    {
+      "epoch": 3.856993190151912,
+      "loss_partial": 0.05029296875,
+      "step": 7363
+    },
+    {
+      "epoch": 3.85751702462022,
+      "grad_norm": 0.600139319896698,
+      "learning_rate": 7.777946491732136e-08,
+      "loss": 0.1027,
+      "step": 7364
+    },
+    {
+      "epoch": 3.85751702462022,
+      "loss_partial": 0.051513671875,
+      "step": 7364
+    },
+    {
+      "epoch": 3.85751702462022,
+      "loss_partial": 0.047607421875,
+      "step": 7364
+    },
+    {
+      "epoch": 3.8580408590885282,
+      "grad_norm": 0.7828904390335083,
+      "learning_rate": 7.721142866274234e-08,
+      "loss": 0.0989,
+      "step": 7365
+    },
+    {
+      "epoch": 3.8580408590885282,
+      "loss_partial": 0.046630859375,
+      "step": 7365
+    },
+    {
+      "epoch": 3.8580408590885282,
+      "loss_partial": 0.0634765625,
+      "step": 7365
+    },
+    {
+      "epoch": 3.8585646935568363,
+      "grad_norm": 0.8279246687889099,
+      "learning_rate": 7.664546620961389e-08,
+      "loss": 0.099,
+      "step": 7366
+    },
+    {
+      "epoch": 3.8585646935568363,
+      "loss_partial": 0.044921875,
+      "step": 7366
+    },
+    {
+      "epoch": 3.8585646935568363,
+      "loss_partial": 0.044677734375,
+      "step": 7366
+    },
+    {
+      "epoch": 3.859088528025144,
+      "grad_norm": 0.9139337539672852,
+      "learning_rate": 7.608157767621804e-08,
+      "loss": 0.1014,
+      "step": 7367
+    },
+    {
+      "epoch": 3.859088528025144,
+      "loss_partial": 0.045166015625,
+      "step": 7367
+    },
+    {
+      "epoch": 3.859088528025144,
+      "loss_partial": 0.048583984375,
+      "step": 7367
+    },
+    {
+      "epoch": 3.859612362493452,
+      "grad_norm": 0.7608875632286072,
+      "learning_rate": 7.55197631804061e-08,
+      "loss": 0.0982,
+      "step": 7368
+    },
+    {
+      "epoch": 3.859612362493452,
+      "loss_partial": 0.05126953125,
+      "step": 7368
+    },
+    {
+      "epoch": 3.859612362493452,
+      "loss_partial": 0.0546875,
+      "step": 7368
+    },
+    {
+      "epoch": 3.86013619696176,
+      "grad_norm": 0.7813320159912109,
+      "learning_rate": 7.496002283959191e-08,
+      "loss": 0.1012,
+      "step": 7369
+    },
+    {
+      "epoch": 3.86013619696176,
+      "loss_partial": 0.048095703125,
+      "step": 7369
+    },
+    {
+      "epoch": 3.86013619696176,
+      "loss_partial": 0.050537109375,
+      "step": 7369
+    },
+    {
+      "epoch": 3.860660031430068,
+      "grad_norm": 1.016677975654602,
+      "learning_rate": 7.440235677075857e-08,
+      "loss": 0.0995,
+      "step": 7370
+    },
+    {
+      "epoch": 3.860660031430068,
+      "loss_partial": 0.057861328125,
+      "step": 7370
+    },
+    {
+      "epoch": 3.860660031430068,
+      "loss_partial": 0.0478515625,
+      "step": 7370
+    },
+    {
+      "epoch": 3.861183865898376,
+      "grad_norm": 0.9563276767730713,
+      "learning_rate": 7.384676509045397e-08,
+      "loss": 0.1001,
+      "step": 7371
+    },
+    {
+      "epoch": 3.861183865898376,
+      "loss_partial": 0.046875,
+      "step": 7371
+    },
+    {
+      "epoch": 3.861183865898376,
+      "loss_partial": 0.043212890625,
+      "step": 7371
+    },
+    {
+      "epoch": 3.861707700366684,
+      "grad_norm": 0.7947534322738647,
+      "learning_rate": 7.329324791479409e-08,
+      "loss": 0.0987,
+      "step": 7372
+    },
+    {
+      "epoch": 3.861707700366684,
+      "loss_partial": 0.0537109375,
+      "step": 7372
+    },
+    {
+      "epoch": 3.861707700366684,
+      "loss_partial": 0.056640625,
+      "step": 7372
+    },
+    {
+      "epoch": 3.8622315348349923,
+      "grad_norm": 0.8053327202796936,
+      "learning_rate": 7.274180535946085e-08,
+      "loss": 0.1016,
+      "step": 7373
+    },
+    {
+      "epoch": 3.8622315348349923,
+      "loss_partial": 0.0458984375,
+      "step": 7373
+    },
+    {
+      "epoch": 3.8622315348349923,
+      "loss_partial": 0.049560546875,
+      "step": 7373
+    },
+    {
+      "epoch": 3.8627553693033003,
+      "grad_norm": 0.8652783036231995,
+      "learning_rate": 7.219243753970206e-08,
+      "loss": 0.0985,
+      "step": 7374
+    },
+    {
+      "epoch": 3.8627553693033003,
+      "loss_partial": 0.05078125,
+      "step": 7374
+    },
+    {
+      "epoch": 3.8627553693033003,
+      "loss_partial": 0.047119140625,
+      "step": 7374
+    },
+    {
+      "epoch": 3.863279203771608,
+      "grad_norm": 0.9737511873245239,
+      "learning_rate": 7.164514457033256e-08,
+      "loss": 0.1003,
+      "step": 7375
+    },
+    {
+      "epoch": 3.863279203771608,
+      "loss_partial": 0.044189453125,
+      "step": 7375
+    },
+    {
+      "epoch": 3.863279203771608,
+      "loss_partial": 0.05322265625,
+      "step": 7375
+    },
+    {
+      "epoch": 3.863803038239916,
+      "grad_norm": 0.9715057015419006,
+      "learning_rate": 7.109992656573195e-08,
+      "loss": 0.0959,
+      "step": 7376
+    },
+    {
+      "epoch": 3.863803038239916,
+      "loss_partial": 0.0498046875,
+      "step": 7376
+    },
+    {
+      "epoch": 3.863803038239916,
+      "loss_partial": 0.052001953125,
+      "step": 7376
+    },
+    {
+      "epoch": 3.864326872708224,
+      "grad_norm": 0.7412678003311157,
+      "learning_rate": 7.055678363984908e-08,
+      "loss": 0.0988,
+      "step": 7377
+    },
+    {
+      "epoch": 3.864326872708224,
+      "loss_partial": 0.047607421875,
+      "step": 7377
+    },
+    {
+      "epoch": 3.864326872708224,
+      "loss_partial": 0.057373046875,
+      "step": 7377
+    },
+    {
+      "epoch": 3.864850707176532,
+      "grad_norm": 0.752336859703064,
+      "learning_rate": 7.001571590619537e-08,
+      "loss": 0.0977,
+      "step": 7378
+    },
+    {
+      "epoch": 3.864850707176532,
+      "loss_partial": 0.052734375,
+      "step": 7378
+    },
+    {
+      "epoch": 3.864850707176532,
+      "loss_partial": 0.056640625,
+      "step": 7378
+    },
+    {
+      "epoch": 3.86537454164484,
+      "grad_norm": 0.8070597052574158,
+      "learning_rate": 6.947672347785261e-08,
+      "loss": 0.101,
+      "step": 7379
+    },
+    {
+      "epoch": 3.86537454164484,
+      "loss_partial": 0.04736328125,
+      "step": 7379
+    },
+    {
+      "epoch": 3.86537454164484,
+      "loss_partial": 0.06689453125,
+      "step": 7379
+    },
+    {
+      "epoch": 3.8658983761131482,
+      "grad_norm": 0.7085438966751099,
+      "learning_rate": 6.893980646746623e-08,
+      "loss": 0.0981,
+      "step": 7380
+    },
+    {
+      "epoch": 3.8658983761131482,
+      "loss_partial": 0.048095703125,
+      "step": 7380
+    },
+    {
+      "epoch": 3.8658983761131482,
+      "loss_partial": 0.0537109375,
+      "step": 7380
+    },
+    {
+      "epoch": 3.8664222105814563,
+      "grad_norm": 0.8483445644378662,
+      "learning_rate": 6.840496498724758e-08,
+      "loss": 0.1012,
+      "step": 7381
+    },
+    {
+      "epoch": 3.8664222105814563,
+      "loss_partial": 0.04248046875,
+      "step": 7381
+    },
+    {
+      "epoch": 3.8664222105814563,
+      "loss_partial": 0.050048828125,
+      "step": 7381
+    },
+    {
+      "epoch": 3.8669460450497644,
+      "grad_norm": 0.9146727323532104,
+      "learning_rate": 6.787219914897725e-08,
+      "loss": 0.1001,
+      "step": 7382
+    },
+    {
+      "epoch": 3.8669460450497644,
+      "loss_partial": 0.05029296875,
+      "step": 7382
+    },
+    {
+      "epoch": 3.8669460450497644,
+      "loss_partial": 0.05615234375,
+      "step": 7382
+    },
+    {
+      "epoch": 3.8674698795180724,
+      "grad_norm": 0.9597032070159912,
+      "learning_rate": 6.73415090639984e-08,
+      "loss": 0.0984,
+      "step": 7383
+    },
+    {
+      "epoch": 3.8674698795180724,
+      "loss_partial": 0.06103515625,
+      "step": 7383
+    },
+    {
+      "epoch": 3.8674698795180724,
+      "loss_partial": 0.05126953125,
+      "step": 7383
+    },
+    {
+      "epoch": 3.8679937139863805,
+      "grad_norm": 1.1790015697479248,
+      "learning_rate": 6.681289484322228e-08,
+      "loss": 0.1027,
+      "step": 7384
+    },
+    {
+      "epoch": 3.8679937139863805,
+      "loss_partial": 0.0478515625,
+      "step": 7384
+    },
+    {
+      "epoch": 3.8679937139863805,
+      "loss_partial": 0.054443359375,
+      "step": 7384
+    },
+    {
+      "epoch": 3.8685175484546885,
+      "grad_norm": 0.8694273829460144,
+      "learning_rate": 6.628635659712612e-08,
+      "loss": 0.0958,
+      "step": 7385
+    },
+    {
+      "epoch": 3.8685175484546885,
+      "loss_partial": 0.055908203125,
+      "step": 7385
+    },
+    {
+      "epoch": 3.8685175484546885,
+      "loss_partial": 0.0498046875,
+      "step": 7385
+    },
+    {
+      "epoch": 3.8690413829229966,
+      "grad_norm": 0.8800280094146729,
+      "learning_rate": 6.576189443575409e-08,
+      "loss": 0.0982,
+      "step": 7386
+    },
+    {
+      "epoch": 3.8690413829229966,
+      "loss_partial": 0.046875,
+      "step": 7386
+    },
+    {
+      "epoch": 3.8690413829229966,
+      "loss_partial": 0.04150390625,
+      "step": 7386
+    },
+    {
+      "epoch": 3.869565217391304,
+      "grad_norm": 0.8862608671188354,
+      "learning_rate": 6.523950846871407e-08,
+      "loss": 0.0992,
+      "step": 7387
+    },
+    {
+      "epoch": 3.869565217391304,
+      "loss_partial": 0.049072265625,
+      "step": 7387
+    },
+    {
+      "epoch": 3.869565217391304,
+      "loss_partial": 0.04833984375,
+      "step": 7387
+    },
+    {
+      "epoch": 3.8700890518596123,
+      "grad_norm": 0.7242209911346436,
+      "learning_rate": 6.471919880518207e-08,
+      "loss": 0.0993,
+      "step": 7388
+    },
+    {
+      "epoch": 3.8700890518596123,
+      "loss_partial": 0.052734375,
+      "step": 7388
+    },
+    {
+      "epoch": 3.8700890518596123,
+      "loss_partial": 0.050537109375,
+      "step": 7388
+    },
+    {
+      "epoch": 3.8706128863279203,
+      "grad_norm": 0.8997414112091064,
+      "learning_rate": 6.420096555389888e-08,
+      "loss": 0.0991,
+      "step": 7389
+    },
+    {
+      "epoch": 3.8706128863279203,
+      "loss_partial": 0.0517578125,
+      "step": 7389
+    },
+    {
+      "epoch": 3.8706128863279203,
+      "loss_partial": 0.044189453125,
+      "step": 7389
+    },
+    {
+      "epoch": 3.8711367207962284,
+      "grad_norm": 0.6797974109649658,
+      "learning_rate": 6.36848088231734e-08,
+      "loss": 0.1024,
+      "step": 7390
+    },
+    {
+      "epoch": 3.8711367207962284,
+      "loss_partial": 0.0498046875,
+      "step": 7390
+    },
+    {
+      "epoch": 3.8711367207962284,
+      "loss_partial": 0.04296875,
+      "step": 7390
+    },
+    {
+      "epoch": 3.8716605552645365,
+      "grad_norm": 0.7733314037322998,
+      "learning_rate": 6.317072872087827e-08,
+      "loss": 0.0979,
+      "step": 7391
+    },
+    {
+      "epoch": 3.8716605552645365,
+      "loss_partial": 0.06787109375,
+      "step": 7391
+    },
+    {
+      "epoch": 3.8716605552645365,
+      "loss_partial": 0.056396484375,
+      "step": 7391
+    },
+    {
+      "epoch": 3.8721843897328445,
+      "grad_norm": 0.7699337601661682,
+      "learning_rate": 6.265872535445416e-08,
+      "loss": 0.1041,
+      "step": 7392
+    },
+    {
+      "epoch": 3.8721843897328445,
+      "loss_partial": 0.045166015625,
+      "step": 7392
+    },
+    {
+      "epoch": 3.8721843897328445,
+      "loss_partial": 0.052001953125,
+      "step": 7392
+    },
+    {
+      "epoch": 3.8727082242011526,
+      "grad_norm": 1.0287021398544312,
+      "learning_rate": 6.214879883090441e-08,
+      "loss": 0.103,
+      "step": 7393
+    },
+    {
+      "epoch": 3.8727082242011526,
+      "loss_partial": 0.044189453125,
+      "step": 7393
+    },
+    {
+      "epoch": 3.8727082242011526,
+      "loss_partial": 0.04443359375,
+      "step": 7393
+    },
+    {
+      "epoch": 3.87323205866946,
+      "grad_norm": 1.047179937362671,
+      "learning_rate": 6.164094925680264e-08,
+      "loss": 0.0998,
+      "step": 7394
+    },
+    {
+      "epoch": 3.87323205866946,
+      "loss_partial": 0.0498046875,
+      "step": 7394
+    },
+    {
+      "epoch": 3.87323205866946,
+      "loss_partial": 0.052001953125,
+      "step": 7394
+    },
+    {
+      "epoch": 3.8737558931377682,
+      "grad_norm": 0.7936915755271912,
+      "learning_rate": 6.113517673828506e-08,
+      "loss": 0.1028,
+      "step": 7395
+    },
+    {
+      "epoch": 3.8737558931377682,
+      "loss_partial": 0.0517578125,
+      "step": 7395
+    },
+    {
+      "epoch": 3.8737558931377682,
+      "loss_partial": 0.052490234375,
+      "step": 7395
+    },
+    {
+      "epoch": 3.8742797276060763,
+      "grad_norm": 1.0484509468078613,
+      "learning_rate": 6.063148138105602e-08,
+      "loss": 0.101,
+      "step": 7396
+    },
+    {
+      "epoch": 3.8742797276060763,
+      "loss_partial": 0.041748046875,
+      "step": 7396
+    },
+    {
+      "epoch": 3.8742797276060763,
+      "loss_partial": 0.048095703125,
+      "step": 7396
+    },
+    {
+      "epoch": 3.8748035620743844,
+      "grad_norm": 0.8636603951454163,
+      "learning_rate": 6.012986329038462e-08,
+      "loss": 0.1003,
+      "step": 7397
+    },
+    {
+      "epoch": 3.8748035620743844,
+      "loss_partial": 0.046875,
+      "step": 7397
+    },
+    {
+      "epoch": 3.8748035620743844,
+      "loss_partial": 0.05078125,
+      "step": 7397
+    },
+    {
+      "epoch": 3.8753273965426924,
+      "grad_norm": 0.7971204519271851,
+      "learning_rate": 5.963032257110479e-08,
+      "loss": 0.0976,
+      "step": 7398
+    },
+    {
+      "epoch": 3.8753273965426924,
+      "loss_partial": 0.048583984375,
+      "step": 7398
+    },
+    {
+      "epoch": 3.8753273965426924,
+      "loss_partial": 0.048095703125,
+      "step": 7398
+    },
+    {
+      "epoch": 3.8758512310110005,
+      "grad_norm": 0.8101974725723267,
+      "learning_rate": 5.913285932761859e-08,
+      "loss": 0.0991,
+      "step": 7399
+    },
+    {
+      "epoch": 3.8758512310110005,
+      "loss_partial": 0.05859375,
+      "step": 7399
+    },
+    {
+      "epoch": 3.8758512310110005,
+      "loss_partial": 0.057861328125,
+      "step": 7399
+    },
+    {
+      "epoch": 3.8763750654793085,
+      "grad_norm": 0.7768456935882568,
+      "learning_rate": 5.863747366389172e-08,
+      "loss": 0.1022,
+      "step": 7400
+    },
+    {
+      "epoch": 3.8763750654793085,
+      "loss_partial": 0.045166015625,
+      "step": 7400
+    },
+    {
+      "epoch": 3.8763750654793085,
+      "loss_partial": 0.05078125,
+      "step": 7400
+    },
+    {
+      "epoch": 3.8768988999476166,
+      "grad_norm": 0.7343011498451233,
+      "learning_rate": 5.8144165683458045e-08,
+      "loss": 0.1001,
+      "step": 7401
+    },
+    {
+      "epoch": 3.8768988999476166,
+      "loss_partial": 0.045166015625,
+      "step": 7401
+    },
+    {
+      "epoch": 3.8768988999476166,
+      "loss_partial": 0.05078125,
+      "step": 7401
+    },
+    {
+      "epoch": 3.8774227344159247,
+      "grad_norm": 0.8554125428199768,
+      "learning_rate": 5.7652935489415084e-08,
+      "loss": 0.097,
+      "step": 7402
+    },
+    {
+      "epoch": 3.8774227344159247,
+      "loss_partial": 0.050537109375,
+      "step": 7402
+    },
+    {
+      "epoch": 3.8774227344159247,
+      "loss_partial": 0.052734375,
+      "step": 7402
+    },
+    {
+      "epoch": 3.8779465688842327,
+      "grad_norm": 0.812255322933197,
+      "learning_rate": 5.716378318442628e-08,
+      "loss": 0.0997,
+      "step": 7403
+    },
+    {
+      "epoch": 3.8779465688842327,
+      "loss_partial": 0.05224609375,
+      "step": 7403
+    },
+    {
+      "epoch": 3.8779465688842327,
+      "loss_partial": 0.044189453125,
+      "step": 7403
+    },
+    {
+      "epoch": 3.878470403352541,
+      "grad_norm": 0.8001856207847595,
+      "learning_rate": 5.6676708870722075e-08,
+      "loss": 0.0977,
+      "step": 7404
+    },
+    {
+      "epoch": 3.878470403352541,
+      "loss_partial": 0.054443359375,
+      "step": 7404
+    },
+    {
+      "epoch": 3.878470403352541,
+      "loss_partial": 0.046875,
+      "step": 7404
+    },
+    {
+      "epoch": 3.878994237820849,
+      "grad_norm": 0.7574948072433472,
+      "learning_rate": 5.619171265009882e-08,
+      "loss": 0.0995,
+      "step": 7405
+    },
+    {
+      "epoch": 3.878994237820849,
+      "loss_partial": 0.049072265625,
+      "step": 7405
+    },
+    {
+      "epoch": 3.878994237820849,
+      "loss_partial": 0.055419921875,
+      "step": 7405
+    },
+    {
+      "epoch": 3.8795180722891565,
+      "grad_norm": 0.7271586656570435,
+      "learning_rate": 5.570879462391654e-08,
+      "loss": 0.1022,
+      "step": 7406
+    },
+    {
+      "epoch": 3.8795180722891565,
+      "loss_partial": 0.058349609375,
+      "step": 7406
+    },
+    {
+      "epoch": 3.8795180722891565,
+      "loss_partial": 0.04736328125,
+      "step": 7406
+    },
+    {
+      "epoch": 3.8800419067574645,
+      "grad_norm": 0.6899347901344299,
+      "learning_rate": 5.5227954893102284e-08,
+      "loss": 0.0977,
+      "step": 7407
+    },
+    {
+      "epoch": 3.8800419067574645,
+      "loss_partial": 0.048828125,
+      "step": 7407
+    },
+    {
+      "epoch": 3.8800419067574645,
+      "loss_partial": 0.0478515625,
+      "step": 7407
+    },
+    {
+      "epoch": 3.8805657412257726,
+      "grad_norm": 0.8641816973686218,
+      "learning_rate": 5.474919355814901e-08,
+      "loss": 0.0984,
+      "step": 7408
+    },
+    {
+      "epoch": 3.8805657412257726,
+      "loss_partial": 0.053466796875,
+      "step": 7408
+    },
+    {
+      "epoch": 3.8805657412257726,
+      "loss_partial": 0.0439453125,
+      "step": 7408
+    },
+    {
+      "epoch": 3.8810895756940806,
+      "grad_norm": 0.7497325539588928,
+      "learning_rate": 5.427251071911555e-08,
+      "loss": 0.0962,
+      "step": 7409
+    },
+    {
+      "epoch": 3.8810895756940806,
+      "loss_partial": 0.04296875,
+      "step": 7409
+    },
+    {
+      "epoch": 3.8810895756940806,
+      "loss_partial": 0.043212890625,
+      "step": 7409
+    },
+    {
+      "epoch": 3.8816134101623887,
+      "grad_norm": 0.7772818803787231,
+      "learning_rate": 5.3797906475623326e-08,
+      "loss": 0.0988,
+      "step": 7410
+    },
+    {
+      "epoch": 3.8816134101623887,
+      "loss_partial": 0.04931640625,
+      "step": 7410
+    },
+    {
+      "epoch": 3.8816134101623887,
+      "loss_partial": 0.0537109375,
+      "step": 7410
+    },
+    {
+      "epoch": 3.8821372446306968,
+      "grad_norm": 0.6374032497406006,
+      "learning_rate": 5.3325380926864126e-08,
+      "loss": 0.1017,
+      "step": 7411
+    },
+    {
+      "epoch": 3.8821372446306968,
+      "loss_partial": 0.05615234375,
+      "step": 7411
+    },
+    {
+      "epoch": 3.8821372446306968,
+      "loss_partial": 0.04931640625,
+      "step": 7411
+    },
+    {
+      "epoch": 3.882661079099005,
+      "grad_norm": 0.7355490326881409,
+      "learning_rate": 5.285493417159226e-08,
+      "loss": 0.1,
+      "step": 7412
+    },
+    {
+      "epoch": 3.882661079099005,
+      "loss_partial": 0.04296875,
+      "step": 7412
+    },
+    {
+      "epoch": 3.882661079099005,
+      "loss_partial": 0.053466796875,
+      "step": 7412
+    },
+    {
+      "epoch": 3.883184913567313,
+      "grad_norm": 0.6920931339263916,
+      "learning_rate": 5.238656630812799e-08,
+      "loss": 0.0971,
+      "step": 7413
+    },
+    {
+      "epoch": 3.883184913567313,
+      "loss_partial": 0.0654296875,
+      "step": 7413
+    },
+    {
+      "epoch": 3.883184913567313,
+      "loss_partial": 0.0419921875,
+      "step": 7413
+    },
+    {
+      "epoch": 3.8837087480356205,
+      "grad_norm": 0.7030735611915588,
+      "learning_rate": 5.192027743435635e-08,
+      "loss": 0.1001,
+      "step": 7414
+    },
+    {
+      "epoch": 3.8837087480356205,
+      "loss_partial": 0.05517578125,
+      "step": 7414
+    },
+    {
+      "epoch": 3.8837087480356205,
+      "loss_partial": 0.046142578125,
+      "step": 7414
+    },
+    {
+      "epoch": 3.8842325825039286,
+      "grad_norm": 0.9355078339576721,
+      "learning_rate": 5.145606764772937e-08,
+      "loss": 0.1034,
+      "step": 7415
+    },
+    {
+      "epoch": 3.8842325825039286,
+      "loss_partial": 0.05615234375,
+      "step": 7415
+    },
+    {
+      "epoch": 3.8842325825039286,
+      "loss_partial": 0.042236328125,
+      "step": 7415
+    },
+    {
+      "epoch": 3.8847564169722366,
+      "grad_norm": 0.982075035572052,
+      "learning_rate": 5.099393704526501e-08,
+      "loss": 0.0995,
+      "step": 7416
+    },
+    {
+      "epoch": 3.8847564169722366,
+      "loss_partial": 0.05712890625,
+      "step": 7416
+    },
+    {
+      "epoch": 3.8847564169722366,
+      "loss_partial": 0.05517578125,
+      "step": 7416
+    },
+    {
+      "epoch": 3.8852802514405447,
+      "grad_norm": 0.8094208240509033,
+      "learning_rate": 5.0533885723544895e-08,
+      "loss": 0.1021,
+      "step": 7417
+    },
+    {
+      "epoch": 3.8852802514405447,
+      "loss_partial": 0.06787109375,
+      "step": 7417
+    },
+    {
+      "epoch": 3.8852802514405447,
+      "loss_partial": 0.047119140625,
+      "step": 7417
+    },
+    {
+      "epoch": 3.8858040859088527,
+      "grad_norm": 0.6555085778236389,
+      "learning_rate": 5.007591377871657e-08,
+      "loss": 0.0967,
+      "step": 7418
+    },
+    {
+      "epoch": 3.8858040859088527,
+      "loss_partial": 0.05322265625,
+      "step": 7418
+    },
+    {
+      "epoch": 3.8858040859088527,
+      "loss_partial": 0.050537109375,
+      "step": 7418
+    },
+    {
+      "epoch": 3.886327920377161,
+      "grad_norm": 0.8038281202316284,
+      "learning_rate": 4.962002130649457e-08,
+      "loss": 0.1045,
+      "step": 7419
+    },
+    {
+      "epoch": 3.886327920377161,
+      "loss_partial": 0.04736328125,
+      "step": 7419
+    },
+    {
+      "epoch": 3.886327920377161,
+      "loss_partial": 0.0576171875,
+      "step": 7419
+    },
+    {
+      "epoch": 3.886851754845469,
+      "grad_norm": 1.1238198280334473,
+      "learning_rate": 4.916620840215602e-08,
+      "loss": 0.0984,
+      "step": 7420
+    },
+    {
+      "epoch": 3.886851754845469,
+      "loss_partial": 0.05712890625,
+      "step": 7420
+    },
+    {
+      "epoch": 3.886851754845469,
+      "loss_partial": 0.039794921875,
+      "step": 7420
+    },
+    {
+      "epoch": 3.887375589313777,
+      "grad_norm": 0.7634682059288025,
+      "learning_rate": 4.871447516054617e-08,
+      "loss": 0.102,
+      "step": 7421
+    },
+    {
+      "epoch": 3.887375589313777,
+      "loss_partial": 0.058349609375,
+      "step": 7421
+    },
+    {
+      "epoch": 3.887375589313777,
+      "loss_partial": 0.044921875,
+      "step": 7421
+    },
+    {
+      "epoch": 3.887899423782085,
+      "grad_norm": 0.7863073348999023,
+      "learning_rate": 4.826482167607393e-08,
+      "loss": 0.1017,
+      "step": 7422
+    },
+    {
+      "epoch": 3.887899423782085,
+      "loss_partial": 0.046875,
+      "step": 7422
+    },
+    {
+      "epoch": 3.887899423782085,
+      "loss_partial": 0.0478515625,
+      "step": 7422
+    },
+    {
+      "epoch": 3.888423258250393,
+      "grad_norm": 0.9614227414131165,
+      "learning_rate": 4.781724804271526e-08,
+      "loss": 0.0992,
+      "step": 7423
+    },
+    {
+      "epoch": 3.888423258250393,
+      "loss_partial": 0.04248046875,
+      "step": 7423
+    },
+    {
+      "epoch": 3.888423258250393,
+      "loss_partial": 0.055908203125,
+      "step": 7423
+    },
+    {
+      "epoch": 3.888947092718701,
+      "grad_norm": 0.7850943803787231,
+      "learning_rate": 4.737175435400754e-08,
+      "loss": 0.1007,
+      "step": 7424
+    },
+    {
+      "epoch": 3.888947092718701,
+      "loss_partial": 0.046875,
+      "step": 7424
+    },
+    {
+      "epoch": 3.888947092718701,
+      "loss_partial": 0.045166015625,
+      "step": 7424
+    },
+    {
+      "epoch": 3.889470927187009,
+      "grad_norm": 0.9101271033287048,
+      "learning_rate": 4.6928340703059625e-08,
+      "loss": 0.0998,
+      "step": 7425
+    },
+    {
+      "epoch": 3.889470927187009,
+      "loss_partial": 0.042236328125,
+      "step": 7425
+    },
+    {
+      "epoch": 3.889470927187009,
+      "loss_partial": 0.059326171875,
+      "step": 7425
+    },
+    {
+      "epoch": 3.8899947616553168,
+      "grad_norm": 0.7546268105506897,
+      "learning_rate": 4.648700718253962e-08,
+      "loss": 0.1018,
+      "step": 7426
+    },
+    {
+      "epoch": 3.8899947616553168,
+      "loss_partial": 0.047119140625,
+      "step": 7426
+    },
+    {
+      "epoch": 3.8899947616553168,
+      "loss_partial": 0.06103515625,
+      "step": 7426
+    },
+    {
+      "epoch": 3.890518596123625,
+      "grad_norm": 0.7814974188804626,
+      "learning_rate": 4.604775388468485e-08,
+      "loss": 0.1003,
+      "step": 7427
+    },
+    {
+      "epoch": 3.890518596123625,
+      "loss_partial": 0.043212890625,
+      "step": 7427
+    },
+    {
+      "epoch": 3.890518596123625,
+      "loss_partial": 0.05908203125,
+      "step": 7427
+    },
+    {
+      "epoch": 3.891042430591933,
+      "grad_norm": 0.8696131110191345,
+      "learning_rate": 4.561058090129633e-08,
+      "loss": 0.0977,
+      "step": 7428
+    },
+    {
+      "epoch": 3.891042430591933,
+      "loss_partial": 0.047119140625,
+      "step": 7428
+    },
+    {
+      "epoch": 3.891042430591933,
+      "loss_partial": 0.04931640625,
+      "step": 7428
+    },
+    {
+      "epoch": 3.891566265060241,
+      "grad_norm": 0.9193167686462402,
+      "learning_rate": 4.517548832373986e-08,
+      "loss": 0.0995,
+      "step": 7429
+    },
+    {
+      "epoch": 3.891566265060241,
+      "loss_partial": 0.051513671875,
+      "step": 7429
+    },
+    {
+      "epoch": 3.891566265060241,
+      "loss_partial": 0.06005859375,
+      "step": 7429
+    },
+    {
+      "epoch": 3.892090099528549,
+      "grad_norm": 0.8891531229019165,
+      "learning_rate": 4.474247624294714e-08,
+      "loss": 0.1003,
+      "step": 7430
+    },
+    {
+      "epoch": 3.892090099528549,
+      "loss_partial": 0.046630859375,
+      "step": 7430
+    },
+    {
+      "epoch": 3.892090099528549,
+      "loss_partial": 0.036865234375,
+      "step": 7430
+    },
+    {
+      "epoch": 3.892613933996857,
+      "grad_norm": 0.9356659054756165,
+      "learning_rate": 4.431154474941579e-08,
+      "loss": 0.0987,
+      "step": 7431
+    },
+    {
+      "epoch": 3.892613933996857,
+      "loss_partial": 0.047607421875,
+      "step": 7431
+    },
+    {
+      "epoch": 3.892613933996857,
+      "loss_partial": 0.052490234375,
+      "step": 7431
+    },
+    {
+      "epoch": 3.893137768465165,
+      "grad_norm": 0.9246692061424255,
+      "learning_rate": 4.388269393320821e-08,
+      "loss": 0.1028,
+      "step": 7432
+    },
+    {
+      "epoch": 3.893137768465165,
+      "loss_partial": 0.05126953125,
+      "step": 7432
+    },
+    {
+      "epoch": 3.893137768465165,
+      "loss_partial": 0.05029296875,
+      "step": 7432
+    },
+    {
+      "epoch": 3.893661602933473,
+      "grad_norm": 0.8418059945106506,
+      "learning_rate": 4.345592388394937e-08,
+      "loss": 0.1011,
+      "step": 7433
+    },
+    {
+      "epoch": 3.893661602933473,
+      "loss_partial": 0.048828125,
+      "step": 7433
+    },
+    {
+      "epoch": 3.893661602933473,
+      "loss_partial": 0.052978515625,
+      "step": 7433
+    },
+    {
+      "epoch": 3.894185437401781,
+      "grad_norm": 0.7478862404823303,
+      "learning_rate": 4.303123469083459e-08,
+      "loss": 0.1026,
+      "step": 7434
+    },
+    {
+      "epoch": 3.894185437401781,
+      "loss_partial": 0.050537109375,
+      "step": 7434
+    },
+    {
+      "epoch": 3.894185437401781,
+      "loss_partial": 0.04541015625,
+      "step": 7434
+    },
+    {
+      "epoch": 3.894709271870089,
+      "grad_norm": 0.7113259434700012,
+      "learning_rate": 4.260862644261843e-08,
+      "loss": 0.0989,
+      "step": 7435
+    },
+    {
+      "epoch": 3.894709271870089,
+      "loss_partial": 0.052001953125,
+      "step": 7435
+    },
+    {
+      "epoch": 3.894709271870089,
+      "loss_partial": 0.052734375,
+      "step": 7435
+    },
+    {
+      "epoch": 3.895233106338397,
+      "grad_norm": 0.7945064306259155,
+      "learning_rate": 4.2188099227625787e-08,
+      "loss": 0.1024,
+      "step": 7436
+    },
+    {
+      "epoch": 3.895233106338397,
+      "loss_partial": 0.046630859375,
+      "step": 7436
+    },
+    {
+      "epoch": 3.895233106338397,
+      "loss_partial": 0.048828125,
+      "step": 7436
+    },
+    {
+      "epoch": 3.895756940806705,
+      "grad_norm": 0.659666121006012,
+      "learning_rate": 4.1769653133743036e-08,
+      "loss": 0.0992,
+      "step": 7437
+    },
+    {
+      "epoch": 3.895756940806705,
+      "loss_partial": 0.04541015625,
+      "step": 7437
+    },
+    {
+      "epoch": 3.895756940806705,
+      "loss_partial": 0.0498046875,
+      "step": 7437
+    },
+    {
+      "epoch": 3.896280775275013,
+      "grad_norm": 0.9133692383766174,
+      "learning_rate": 4.135328824842244e-08,
+      "loss": 0.0974,
+      "step": 7438
+    },
+    {
+      "epoch": 3.896280775275013,
+      "loss_partial": 0.04296875,
+      "step": 7438
+    },
+    {
+      "epoch": 3.896280775275013,
+      "loss_partial": 0.054931640625,
+      "step": 7438
+    },
+    {
+      "epoch": 3.896804609743321,
+      "grad_norm": 0.7764058709144592,
+      "learning_rate": 4.093900465868217e-08,
+      "loss": 0.0998,
+      "step": 7439
+    },
+    {
+      "epoch": 3.896804609743321,
+      "loss_partial": 0.054443359375,
+      "step": 7439
+    },
+    {
+      "epoch": 3.896804609743321,
+      "loss_partial": 0.0498046875,
+      "step": 7439
+    },
+    {
+      "epoch": 3.897328444211629,
+      "grad_norm": 0.7195848226547241,
+      "learning_rate": 4.052680245110408e-08,
+      "loss": 0.1016,
+      "step": 7440
+    },
+    {
+      "epoch": 3.897328444211629,
+      "loss_partial": 0.042236328125,
+      "step": 7440
+    },
+    {
+      "epoch": 3.897328444211629,
+      "loss_partial": 0.046875,
+      "step": 7440
+    },
+    {
+      "epoch": 3.897852278679937,
+      "grad_norm": 1.0015285015106201,
+      "learning_rate": 4.011668171183813e-08,
+      "loss": 0.0968,
+      "step": 7441
+    },
+    {
+      "epoch": 3.897852278679937,
+      "loss_partial": 0.058837890625,
+      "step": 7441
+    },
+    {
+      "epoch": 3.897852278679937,
+      "loss_partial": 0.056396484375,
+      "step": 7441
+    },
+    {
+      "epoch": 3.8983761131482453,
+      "grad_norm": 0.684630274772644,
+      "learning_rate": 3.9708642526595784e-08,
+      "loss": 0.0997,
+      "step": 7442
+    },
+    {
+      "epoch": 3.8983761131482453,
+      "loss_partial": 0.0517578125,
+      "step": 7442
+    },
+    {
+      "epoch": 3.8983761131482453,
+      "loss_partial": 0.043701171875,
+      "step": 7442
+    },
+    {
+      "epoch": 3.8988999476165533,
+      "grad_norm": 0.9452307224273682,
+      "learning_rate": 3.930268498065326e-08,
+      "loss": 0.0954,
+      "step": 7443
+    },
+    {
+      "epoch": 3.8988999476165533,
+      "loss_partial": 0.049072265625,
+      "step": 7443
+    },
+    {
+      "epoch": 3.8988999476165533,
+      "loss_partial": 0.052734375,
+      "step": 7443
+    },
+    {
+      "epoch": 3.8994237820848614,
+      "grad_norm": 0.7229102253913879,
+      "learning_rate": 3.889880915885491e-08,
+      "loss": 0.1003,
+      "step": 7444
+    },
+    {
+      "epoch": 3.8994237820848614,
+      "loss_partial": 0.043701171875,
+      "step": 7444
+    },
+    {
+      "epoch": 3.8994237820848614,
+      "loss_partial": 0.05712890625,
+      "step": 7444
+    },
+    {
+      "epoch": 3.8999476165531695,
+      "grad_norm": 0.7915159463882446,
+      "learning_rate": 3.8497015145608774e-08,
+      "loss": 0.099,
+      "step": 7445
+    },
+    {
+      "epoch": 3.8999476165531695,
+      "loss_partial": 0.0537109375,
+      "step": 7445
+    },
+    {
+      "epoch": 3.8999476165531695,
+      "loss_partial": 0.05029296875,
+      "step": 7445
+    },
+    {
+      "epoch": 3.900471451021477,
+      "grad_norm": 0.7874122858047485,
+      "learning_rate": 3.809730302488546e-08,
+      "loss": 0.0957,
+      "step": 7446
+    },
+    {
+      "epoch": 3.900471451021477,
+      "loss_partial": 0.052734375,
+      "step": 7446
+    },
+    {
+      "epoch": 3.900471451021477,
+      "loss_partial": 0.0458984375,
+      "step": 7446
+    },
+    {
+      "epoch": 3.900995285489785,
+      "grad_norm": 1.1257028579711914,
+      "learning_rate": 3.769967288022369e-08,
+      "loss": 0.1021,
+      "step": 7447
+    },
+    {
+      "epoch": 3.900995285489785,
+      "loss_partial": 0.05224609375,
+      "step": 7447
+    },
+    {
+      "epoch": 3.900995285489785,
+      "loss_partial": 0.052001953125,
+      "step": 7447
+    },
+    {
+      "epoch": 3.901519119958093,
+      "grad_norm": 0.8264292478561401,
+      "learning_rate": 3.730412479472589e-08,
+      "loss": 0.0999,
+      "step": 7448
+    },
+    {
+      "epoch": 3.901519119958093,
+      "loss_partial": 0.048583984375,
+      "step": 7448
+    },
+    {
+      "epoch": 3.901519119958093,
+      "loss_partial": 0.039306640625,
+      "step": 7448
+    },
+    {
+      "epoch": 3.9020429544264013,
+      "grad_norm": 0.8999307155609131,
+      "learning_rate": 3.6910658851057044e-08,
+      "loss": 0.0949,
+      "step": 7449
+    },
+    {
+      "epoch": 3.9020429544264013,
+      "loss_partial": 0.0478515625,
+      "step": 7449
+    },
+    {
+      "epoch": 3.9020429544264013,
+      "loss_partial": 0.042236328125,
+      "step": 7449
+    },
+    {
+      "epoch": 3.9025667888947093,
+      "grad_norm": 1.2485287189483643,
+      "learning_rate": 3.651927513145137e-08,
+      "loss": 0.0984,
+      "step": 7450
+    },
+    {
+      "epoch": 3.9025667888947093,
+      "loss_partial": 0.059814453125,
+      "step": 7450
+    },
+    {
+      "epoch": 3.9025667888947093,
+      "loss_partial": 0.0458984375,
+      "step": 7450
+    },
+    {
+      "epoch": 3.9030906233630174,
+      "grad_norm": 0.9591355323791504,
+      "learning_rate": 3.6129973717705655e-08,
+      "loss": 0.0981,
+      "step": 7451
+    },
+    {
+      "epoch": 3.9030906233630174,
+      "loss_partial": 0.047119140625,
+      "step": 7451
+    },
+    {
+      "epoch": 3.9030906233630174,
+      "loss_partial": 0.05078125,
+      "step": 7451
+    },
+    {
+      "epoch": 3.9036144578313254,
+      "grad_norm": 0.6951770186424255,
+      "learning_rate": 3.5742754691179274e-08,
+      "loss": 0.1005,
+      "step": 7452
+    },
+    {
+      "epoch": 3.9036144578313254,
+      "loss_partial": 0.050048828125,
+      "step": 7452
+    },
+    {
+      "epoch": 3.9036144578313254,
+      "loss_partial": 0.0419921875,
+      "step": 7452
+    },
+    {
+      "epoch": 3.904138292299633,
+      "grad_norm": 0.8991752862930298,
+      "learning_rate": 3.5357618132800805e-08,
+      "loss": 0.0953,
+      "step": 7453
+    },
+    {
+      "epoch": 3.904138292299633,
+      "loss_partial": 0.041015625,
+      "step": 7453
+    },
+    {
+      "epoch": 3.904138292299633,
+      "loss_partial": 0.04638671875,
+      "step": 7453
+    },
+    {
+      "epoch": 3.904662126767941,
+      "grad_norm": 0.8238940238952637,
+      "learning_rate": 3.497456412306033e-08,
+      "loss": 0.093,
+      "step": 7454
+    },
+    {
+      "epoch": 3.904662126767941,
+      "loss_partial": 0.043701171875,
+      "step": 7454
+    },
+    {
+      "epoch": 3.904662126767941,
+      "loss_partial": 0.0439453125,
+      "step": 7454
+    },
+    {
+      "epoch": 3.905185961236249,
+      "grad_norm": 0.9373859763145447,
+      "learning_rate": 3.4593592742014905e-08,
+      "loss": 0.0969,
+      "step": 7455
+    },
+    {
+      "epoch": 3.905185961236249,
+      "loss_partial": 0.04931640625,
+      "step": 7455
+    },
+    {
+      "epoch": 3.905185961236249,
+      "loss_partial": 0.052978515625,
+      "step": 7455
+    },
+    {
+      "epoch": 3.9057097957045572,
+      "grad_norm": 0.8809216022491455,
+      "learning_rate": 3.4214704069283064e-08,
+      "loss": 0.1006,
+      "step": 7456
+    },
+    {
+      "epoch": 3.9057097957045572,
+      "loss_partial": 0.051513671875,
+      "step": 7456
+    },
+    {
+      "epoch": 3.9057097957045572,
+      "loss_partial": 0.057373046875,
+      "step": 7456
+    },
+    {
+      "epoch": 3.9062336301728653,
+      "grad_norm": 0.6983863711357117,
+      "learning_rate": 3.3837898184052585e-08,
+      "loss": 0.0969,
+      "step": 7457
+    },
+    {
+      "epoch": 3.9062336301728653,
+      "loss_partial": 0.046630859375,
+      "step": 7457
+    },
+    {
+      "epoch": 3.9062336301728653,
+      "loss_partial": 0.051025390625,
+      "step": 7457
+    },
+    {
+      "epoch": 3.9067574646411733,
+      "grad_norm": 0.7741601467132568,
+      "learning_rate": 3.346317516507269e-08,
+      "loss": 0.0974,
+      "step": 7458
+    },
+    {
+      "epoch": 3.9067574646411733,
+      "loss_partial": 0.048095703125,
+      "step": 7458
+    },
+    {
+      "epoch": 3.9067574646411733,
+      "loss_partial": 0.041748046875,
+      "step": 7458
+    },
+    {
+      "epoch": 3.9072812991094814,
+      "grad_norm": 0.7657503485679626,
+      "learning_rate": 3.30905350906574e-08,
+      "loss": 0.0991,
+      "step": 7459
+    },
+    {
+      "epoch": 3.9072812991094814,
+      "loss_partial": 0.055419921875,
+      "step": 7459
+    },
+    {
+      "epoch": 3.9072812991094814,
+      "loss_partial": 0.050537109375,
+      "step": 7459
+    },
+    {
+      "epoch": 3.9078051335777895,
+      "grad_norm": 0.815895140171051,
+      "learning_rate": 3.271997803868665e-08,
+      "loss": 0.1015,
+      "step": 7460
+    },
+    {
+      "epoch": 3.9078051335777895,
+      "loss_partial": 0.04833984375,
+      "step": 7460
+    },
+    {
+      "epoch": 3.9078051335777895,
+      "loss_partial": 0.04541015625,
+      "step": 7460
+    },
+    {
+      "epoch": 3.9083289680460975,
+      "grad_norm": 0.8854469656944275,
+      "learning_rate": 3.2351504086605147e-08,
+      "loss": 0.0984,
+      "step": 7461
+    },
+    {
+      "epoch": 3.9083289680460975,
+      "loss_partial": 0.056396484375,
+      "step": 7461
+    },
+    {
+      "epoch": 3.9083289680460975,
+      "loss_partial": 0.04443359375,
+      "step": 7461
+    },
+    {
+      "epoch": 3.9088528025144056,
+      "grad_norm": 0.814610481262207,
+      "learning_rate": 3.198511331142018e-08,
+      "loss": 0.1012,
+      "step": 7462
+    },
+    {
+      "epoch": 3.9088528025144056,
+      "loss_partial": 0.04443359375,
+      "step": 7462
+    },
+    {
+      "epoch": 3.9088528025144056,
+      "loss_partial": 0.04443359375,
+      "step": 7462
+    },
+    {
+      "epoch": 3.9093766369827136,
+      "grad_norm": 0.8384259939193726,
+      "learning_rate": 3.1620805789707163e-08,
+      "loss": 0.0965,
+      "step": 7463
+    },
+    {
+      "epoch": 3.9093766369827136,
+      "loss_partial": 0.04931640625,
+      "step": 7463
+    },
+    {
+      "epoch": 3.9093766369827136,
+      "loss_partial": 0.05078125,
+      "step": 7463
+    },
+    {
+      "epoch": 3.9099004714510217,
+      "grad_norm": 0.9181491136550903,
+      "learning_rate": 3.125858159760298e-08,
+      "loss": 0.0991,
+      "step": 7464
+    },
+    {
+      "epoch": 3.9099004714510217,
+      "loss_partial": 0.05517578125,
+      "step": 7464
+    },
+    {
+      "epoch": 3.9099004714510217,
+      "loss_partial": 0.0546875,
+      "step": 7464
+    },
+    {
+      "epoch": 3.9104243059193298,
+      "grad_norm": 0.8382053971290588,
+      "learning_rate": 3.0898440810810395e-08,
+      "loss": 0.1,
+      "step": 7465
+    },
+    {
+      "epoch": 3.9104243059193298,
+      "loss_partial": 0.0537109375,
+      "step": 7465
+    },
+    {
+      "epoch": 3.9104243059193298,
+      "loss_partial": 0.055419921875,
+      "step": 7465
+    },
+    {
+      "epoch": 3.9109481403876374,
+      "grad_norm": 0.7944844365119934,
+      "learning_rate": 3.054038350459587e-08,
+      "loss": 0.1001,
+      "step": 7466
+    },
+    {
+      "epoch": 3.9109481403876374,
+      "loss_partial": 0.0498046875,
+      "step": 7466
+    },
+    {
+      "epoch": 3.9109481403876374,
+      "loss_partial": 0.046142578125,
+      "step": 7466
+    },
+    {
+      "epoch": 3.9114719748559454,
+      "grad_norm": 0.716349184513092,
+      "learning_rate": 3.0184409753792886e-08,
+      "loss": 0.1008,
+      "step": 7467
+    },
+    {
+      "epoch": 3.9114719748559454,
+      "loss_partial": 0.048583984375,
+      "step": 7467
+    },
+    {
+      "epoch": 3.9114719748559454,
+      "loss_partial": 0.044189453125,
+      "step": 7467
+    },
+    {
+      "epoch": 3.9119958093242535,
+      "grad_norm": 0.8530163168907166,
+      "learning_rate": 2.983051963279637e-08,
+      "loss": 0.1005,
+      "step": 7468
+    },
+    {
+      "epoch": 3.9119958093242535,
+      "loss_partial": 0.053466796875,
+      "step": 7468
+    },
+    {
+      "epoch": 3.9119958093242535,
+      "loss_partial": 0.04833984375,
+      "step": 7468
+    },
+    {
+      "epoch": 3.9125196437925616,
+      "grad_norm": 0.7377022504806519,
+      "learning_rate": 2.9478713215568277e-08,
+      "loss": 0.0986,
+      "step": 7469
+    },
+    {
+      "epoch": 3.9125196437925616,
+      "loss_partial": 0.044677734375,
+      "step": 7469
+    },
+    {
+      "epoch": 3.9125196437925616,
+      "loss_partial": 0.046875,
+      "step": 7469
+    },
+    {
+      "epoch": 3.9130434782608696,
+      "grad_norm": 0.7750663161277771,
+      "learning_rate": 2.912899057563201e-08,
+      "loss": 0.0998,
+      "step": 7470
+    },
+    {
+      "epoch": 3.9130434782608696,
+      "loss_partial": 0.051025390625,
+      "step": 7470
+    },
+    {
+      "epoch": 3.9130434782608696,
+      "loss_partial": 0.048583984375,
+      "step": 7470
+    },
+    {
+      "epoch": 3.9135673127291777,
+      "grad_norm": 0.7765111923217773,
+      "learning_rate": 2.8781351786079104e-08,
+      "loss": 0.1016,
+      "step": 7471
+    },
+    {
+      "epoch": 3.9135673127291777,
+      "loss_partial": 0.040283203125,
+      "step": 7471
+    },
+    {
+      "epoch": 3.9135673127291777,
+      "loss_partial": 0.0458984375,
+      "step": 7471
+    },
+    {
+      "epoch": 3.9140911471974857,
+      "grad_norm": 0.7307941317558289,
+      "learning_rate": 2.8435796919564774e-08,
+      "loss": 0.1019,
+      "step": 7472
+    },
+    {
+      "epoch": 3.9140911471974857,
+      "loss_partial": 0.044921875,
+      "step": 7472
+    },
+    {
+      "epoch": 3.9140911471974857,
+      "loss_partial": 0.044921875,
+      "step": 7472
+    },
+    {
+      "epoch": 3.9146149816657934,
+      "grad_norm": 0.8438915014266968,
+      "learning_rate": 2.8092326048305696e-08,
+      "loss": 0.1021,
+      "step": 7473
+    },
+    {
+      "epoch": 3.9146149816657934,
+      "loss_partial": 0.049072265625,
+      "step": 7473
+    },
+    {
+      "epoch": 3.9146149816657934,
+      "loss_partial": 0.04248046875,
+      "step": 7473
+    },
+    {
+      "epoch": 3.9151388161341014,
+      "grad_norm": 0.8223075270652771,
+      "learning_rate": 2.7750939244085562e-08,
+      "loss": 0.1025,
+      "step": 7474
+    },
+    {
+      "epoch": 3.9151388161341014,
+      "loss_partial": 0.04443359375,
+      "step": 7474
+    },
+    {
+      "epoch": 3.9151388161341014,
+      "loss_partial": 0.05322265625,
+      "step": 7474
+    },
+    {
+      "epoch": 3.9156626506024095,
+      "grad_norm": 0.8423779010772705,
+      "learning_rate": 2.7411636578252853e-08,
+      "loss": 0.1,
+      "step": 7475
+    },
+    {
+      "epoch": 3.9156626506024095,
+      "loss_partial": 0.050048828125,
+      "step": 7475
+    },
+    {
+      "epoch": 3.9156626506024095,
+      "loss_partial": 0.054931640625,
+      "step": 7475
+    },
+    {
+      "epoch": 3.9161864850707175,
+      "grad_norm": 0.8773342967033386,
+      "learning_rate": 2.7074418121719735e-08,
+      "loss": 0.0984,
+      "step": 7476
+    },
+    {
+      "epoch": 3.9161864850707175,
+      "loss_partial": 0.043212890625,
+      "step": 7476
+    },
+    {
+      "epoch": 3.9161864850707175,
+      "loss_partial": 0.050537109375,
+      "step": 7476
+    },
+    {
+      "epoch": 3.9167103195390256,
+      "grad_norm": 0.8518958687782288,
+      "learning_rate": 2.673928394496206e-08,
+      "loss": 0.1018,
+      "step": 7477
+    },
+    {
+      "epoch": 3.9167103195390256,
+      "loss_partial": 0.04736328125,
+      "step": 7477
+    },
+    {
+      "epoch": 3.9167103195390256,
+      "loss_partial": 0.057373046875,
+      "step": 7477
+    },
+    {
+      "epoch": 3.9172341540073337,
+      "grad_norm": 1.2604581117630005,
+      "learning_rate": 2.6406234118021567e-08,
+      "loss": 0.0959,
+      "step": 7478
+    },
+    {
+      "epoch": 3.9172341540073337,
+      "loss_partial": 0.052978515625,
+      "step": 7478
+    },
+    {
+      "epoch": 3.9172341540073337,
+      "loss_partial": 0.05517578125,
+      "step": 7478
+    },
+    {
+      "epoch": 3.9177579884756417,
+      "grad_norm": 0.608333170413971,
+      "learning_rate": 2.60752687105037e-08,
+      "loss": 0.0992,
+      "step": 7479
+    },
+    {
+      "epoch": 3.9177579884756417,
+      "loss_partial": 0.049072265625,
+      "step": 7479
+    },
+    {
+      "epoch": 3.9177579884756417,
+      "loss_partial": 0.043212890625,
+      "step": 7479
+    },
+    {
+      "epoch": 3.9182818229439498,
+      "grad_norm": 0.8695808053016663,
+      "learning_rate": 2.5746387791577567e-08,
+      "loss": 0.1006,
+      "step": 7480
+    },
+    {
+      "epoch": 3.9182818229439498,
+      "loss_partial": 0.057373046875,
+      "step": 7480
+    },
+    {
+      "epoch": 3.9182818229439498,
+      "loss_partial": 0.0576171875,
+      "step": 7480
+    },
+    {
+      "epoch": 3.918805657412258,
+      "grad_norm": 0.6728467345237732,
+      "learning_rate": 2.5419591429977076e-08,
+      "loss": 0.1041,
+      "step": 7481
+    },
+    {
+      "epoch": 3.918805657412258,
+      "loss_partial": 0.0576171875,
+      "step": 7481
+    },
+    {
+      "epoch": 3.918805657412258,
+      "loss_partial": 0.057373046875,
+      "step": 7481
+    },
+    {
+      "epoch": 3.919329491880566,
+      "grad_norm": 0.9517044425010681,
+      "learning_rate": 2.509487969400093e-08,
+      "loss": 0.1023,
+      "step": 7482
+    },
+    {
+      "epoch": 3.919329491880566,
+      "loss_partial": 0.046630859375,
+      "step": 7482
+    },
+    {
+      "epoch": 3.919329491880566,
+      "loss_partial": 0.048095703125,
+      "step": 7482
+    },
+    {
+      "epoch": 3.919853326348874,
+      "grad_norm": 0.8169007897377014,
+      "learning_rate": 2.477225265151262e-08,
+      "loss": 0.0981,
+      "step": 7483
+    },
+    {
+      "epoch": 3.919853326348874,
+      "loss_partial": 0.048828125,
+      "step": 7483
+    },
+    {
+      "epoch": 3.919853326348874,
+      "loss_partial": 0.043212890625,
+      "step": 7483
+    },
+    {
+      "epoch": 3.920377160817182,
+      "grad_norm": 0.7503048181533813,
+      "learning_rate": 2.4451710369938208e-08,
+      "loss": 0.0973,
+      "step": 7484
+    },
+    {
+      "epoch": 3.920377160817182,
+      "loss_partial": 0.044677734375,
+      "step": 7484
+    },
+    {
+      "epoch": 3.920377160817182,
+      "loss_partial": 0.046875,
+      "step": 7484
+    },
+    {
+      "epoch": 3.9209009952854896,
+      "grad_norm": 0.8505650758743286,
+      "learning_rate": 2.4133252916268556e-08,
+      "loss": 0.0981,
+      "step": 7485
+    },
+    {
+      "epoch": 3.9209009952854896,
+      "loss_partial": 0.04541015625,
+      "step": 7485
+    },
+    {
+      "epoch": 3.9209009952854896,
+      "loss_partial": 0.055419921875,
+      "step": 7485
+    },
+    {
+      "epoch": 3.9214248297537977,
+      "grad_norm": 0.7495592832565308,
+      "learning_rate": 2.3816880357061533e-08,
+      "loss": 0.0976,
+      "step": 7486
+    },
+    {
+      "epoch": 3.9214248297537977,
+      "loss_partial": 0.047119140625,
+      "step": 7486
+    },
+    {
+      "epoch": 3.9214248297537977,
+      "loss_partial": 0.0419921875,
+      "step": 7486
+    },
+    {
+      "epoch": 3.9219486642221058,
+      "grad_norm": 0.7338898181915283,
+      "learning_rate": 2.350259275843536e-08,
+      "loss": 0.1024,
+      "step": 7487
+    },
+    {
+      "epoch": 3.9219486642221058,
+      "loss_partial": 0.04736328125,
+      "step": 7487
+    },
+    {
+      "epoch": 3.9219486642221058,
+      "loss_partial": 0.06298828125,
+      "step": 7487
+    },
+    {
+      "epoch": 3.922472498690414,
+      "grad_norm": 0.8027750849723816,
+      "learning_rate": 2.3190390186074163e-08,
+      "loss": 0.1029,
+      "step": 7488
+    },
+    {
+      "epoch": 3.922472498690414,
+      "loss_partial": 0.053466796875,
+      "step": 7488
+    },
+    {
+      "epoch": 3.922472498690414,
+      "loss_partial": 0.049560546875,
+      "step": 7488
+    },
+    {
+      "epoch": 3.922996333158722,
+      "grad_norm": 0.660892903804779,
+      "learning_rate": 2.2880272705225746e-08,
+      "loss": 0.1014,
+      "step": 7489
+    },
+    {
+      "epoch": 3.922996333158722,
+      "loss_partial": 0.047607421875,
+      "step": 7489
+    },
+    {
+      "epoch": 3.922996333158722,
+      "loss_partial": 0.05078125,
+      "step": 7489
+    },
+    {
+      "epoch": 3.92352016762703,
+      "grad_norm": 0.8565353155136108,
+      "learning_rate": 2.2572240380704936e-08,
+      "loss": 0.0983,
+      "step": 7490
+    },
+    {
+      "epoch": 3.92352016762703,
+      "loss_partial": 0.045654296875,
+      "step": 7490
+    },
+    {
+      "epoch": 3.92352016762703,
+      "loss_partial": 0.051513671875,
+      "step": 7490
+    },
+    {
+      "epoch": 3.924044002095338,
+      "grad_norm": 0.6864820718765259,
+      "learning_rate": 2.2266293276886898e-08,
+      "loss": 0.1015,
+      "step": 7491
+    },
+    {
+      "epoch": 3.924044002095338,
+      "loss_partial": 0.048095703125,
+      "step": 7491
+    },
+    {
+      "epoch": 3.924044002095338,
+      "loss_partial": 0.04248046875,
+      "step": 7491
+    },
+    {
+      "epoch": 3.924567836563646,
+      "grad_norm": 0.7240925431251526,
+      "learning_rate": 2.1962431457713818e-08,
+      "loss": 0.1008,
+      "step": 7492
+    },
+    {
+      "epoch": 3.924567836563646,
+      "loss_partial": 0.06005859375,
+      "step": 7492
+    },
+    {
+      "epoch": 3.924567836563646,
+      "loss_partial": 0.045654296875,
+      "step": 7492
+    },
+    {
+      "epoch": 3.9250916710319537,
+      "grad_norm": 0.7718127965927124,
+      "learning_rate": 2.1660654986689344e-08,
+      "loss": 0.0993,
+      "step": 7493
+    },
+    {
+      "epoch": 3.9250916710319537,
+      "loss_partial": 0.052490234375,
+      "step": 7493
+    },
+    {
+      "epoch": 3.9250916710319537,
+      "loss_partial": 0.043701171875,
+      "step": 7493
+    },
+    {
+      "epoch": 3.9256155055002617,
+      "grad_norm": 0.8829641342163086,
+      "learning_rate": 2.1360963926884136e-08,
+      "loss": 0.0977,
+      "step": 7494
+    },
+    {
+      "epoch": 3.9256155055002617,
+      "loss_partial": 0.043701171875,
+      "step": 7494
+    },
+    {
+      "epoch": 3.9256155055002617,
+      "loss_partial": 0.0439453125,
+      "step": 7494
+    },
+    {
+      "epoch": 3.92613933996857,
+      "grad_norm": 0.7499809861183167,
+      "learning_rate": 2.106335834093143e-08,
+      "loss": 0.0999,
+      "step": 7495
+    },
+    {
+      "epoch": 3.92613933996857,
+      "loss_partial": 0.0458984375,
+      "step": 7495
+    },
+    {
+      "epoch": 3.92613933996857,
+      "loss_partial": 0.046142578125,
+      "step": 7495
+    },
+    {
+      "epoch": 3.926663174436878,
+      "grad_norm": 0.8745001554489136,
+      "learning_rate": 2.076783829102924e-08,
+      "loss": 0.0968,
+      "step": 7496
+    },
+    {
+      "epoch": 3.926663174436878,
+      "loss_partial": 0.040771484375,
+      "step": 7496
+    },
+    {
+      "epoch": 3.926663174436878,
+      "loss_partial": 0.053466796875,
+      "step": 7496
+    },
+    {
+      "epoch": 3.927187008905186,
+      "grad_norm": 0.5545993447303772,
+      "learning_rate": 2.0474403838939284e-08,
+      "loss": 0.1021,
+      "step": 7497
+    },
+    {
+      "epoch": 3.927187008905186,
+      "loss_partial": 0.055908203125,
+      "step": 7497
+    },
+    {
+      "epoch": 3.927187008905186,
+      "loss_partial": 0.06005859375,
+      "step": 7497
+    },
+    {
+      "epoch": 3.927710843373494,
+      "grad_norm": 0.9434733986854553,
+      "learning_rate": 2.0183055045986944e-08,
+      "loss": 0.0992,
+      "step": 7498
+    },
+    {
+      "epoch": 3.927710843373494,
+      "loss_partial": 0.04248046875,
+      "step": 7498
+    },
+    {
+      "epoch": 3.927710843373494,
+      "loss_partial": 0.041748046875,
+      "step": 7498
+    },
+    {
+      "epoch": 3.928234677841802,
+      "grad_norm": 0.8782018423080444,
+      "learning_rate": 1.9893791973062403e-08,
+      "loss": 0.0956,
+      "step": 7499
+    },
+    {
+      "epoch": 3.928234677841802,
+      "loss_partial": 0.052734375,
+      "step": 7499
+    },
+    {
+      "epoch": 3.928234677841802,
+      "loss_partial": 0.04833984375,
+      "step": 7499
+    },
+    {
+      "epoch": 3.92875851231011,
+      "grad_norm": 0.8260599970817566,
+      "learning_rate": 1.960661468062064e-08,
+      "loss": 0.1013,
+      "step": 7500
+    },
+    {
+      "epoch": 3.92875851231011,
+      "loss_partial": 0.046142578125,
+      "step": 7500
+    },
+    {
+      "epoch": 3.92875851231011,
+      "loss_partial": 0.0537109375,
+      "step": 7500
+    },
+    {
+      "epoch": 3.929282346778418,
+      "grad_norm": 0.774155855178833,
+      "learning_rate": 1.9321523228679195e-08,
+      "loss": 0.098,
+      "step": 7501
+    },
+    {
+      "epoch": 3.929282346778418,
+      "loss_partial": 0.041259765625,
+      "step": 7501
+    },
+    {
+      "epoch": 3.929282346778418,
+      "loss_partial": 0.04931640625,
+      "step": 7501
+    },
+    {
+      "epoch": 3.929806181246726,
+      "grad_norm": 0.8384050130844116,
+      "learning_rate": 1.9038517676821522e-08,
+      "loss": 0.0995,
+      "step": 7502
+    },
+    {
+      "epoch": 3.929806181246726,
+      "loss_partial": 0.050048828125,
+      "step": 7502
+    },
+    {
+      "epoch": 3.929806181246726,
+      "loss_partial": 0.057373046875,
+      "step": 7502
+    },
+    {
+      "epoch": 3.9303300157150343,
+      "grad_norm": 0.6486390829086304,
+      "learning_rate": 1.875759808419253e-08,
+      "loss": 0.0978,
+      "step": 7503
+    },
+    {
+      "epoch": 3.9303300157150343,
+      "loss_partial": 0.0419921875,
+      "step": 7503
+    },
+    {
+      "epoch": 3.9303300157150343,
+      "loss_partial": 0.058349609375,
+      "step": 7503
+    },
+    {
+      "epoch": 3.9308538501833423,
+      "grad_norm": 0.8752291798591614,
+      "learning_rate": 1.8478764509503034e-08,
+      "loss": 0.0999,
+      "step": 7504
+    },
+    {
+      "epoch": 3.9308538501833423,
+      "loss_partial": 0.0693359375,
+      "step": 7504
+    },
+    {
+      "epoch": 3.9308538501833423,
+      "loss_partial": 0.05517578125,
+      "step": 7504
+    },
+    {
+      "epoch": 3.93137768465165,
+      "grad_norm": 0.756162166595459,
+      "learning_rate": 1.820201701102864e-08,
+      "loss": 0.1022,
+      "step": 7505
+    },
+    {
+      "epoch": 3.93137768465165,
+      "loss_partial": 0.056884765625,
+      "step": 7505
+    },
+    {
+      "epoch": 3.93137768465165,
+      "loss_partial": 0.04833984375,
+      "step": 7505
+    },
+    {
+      "epoch": 3.931901519119958,
+      "grad_norm": 0.6006596684455872,
+      "learning_rate": 1.7927355646607525e-08,
+      "loss": 0.1027,
+      "step": 7506
+    },
+    {
+      "epoch": 3.931901519119958,
+      "loss_partial": 0.0419921875,
+      "step": 7506
+    },
+    {
+      "epoch": 3.931901519119958,
+      "loss_partial": 0.056396484375,
+      "step": 7506
+    },
+    {
+      "epoch": 3.932425353588266,
+      "grad_norm": 0.795560896396637,
+      "learning_rate": 1.7654780473640443e-08,
+      "loss": 0.1015,
+      "step": 7507
+    },
+    {
+      "epoch": 3.932425353588266,
+      "loss_partial": 0.047607421875,
+      "step": 7507
+    },
+    {
+      "epoch": 3.932425353588266,
+      "loss_partial": 0.046142578125,
+      "step": 7507
+    },
+    {
+      "epoch": 3.932949188056574,
+      "grad_norm": 0.7227761745452881,
+      "learning_rate": 1.7384291549096267e-08,
+      "loss": 0.101,
+      "step": 7508
+    },
+    {
+      "epoch": 3.932949188056574,
+      "loss_partial": 0.04150390625,
+      "step": 7508
+    },
+    {
+      "epoch": 3.932949188056574,
+      "loss_partial": 0.048828125,
+      "step": 7508
+    },
+    {
+      "epoch": 3.933473022524882,
+      "grad_norm": 0.6727795004844666,
+      "learning_rate": 1.711588892950422e-08,
+      "loss": 0.0988,
+      "step": 7509
+    },
+    {
+      "epoch": 3.933473022524882,
+      "loss_partial": 0.054931640625,
+      "step": 7509
+    },
+    {
+      "epoch": 3.933473022524882,
+      "loss_partial": 0.051513671875,
+      "step": 7509
+    },
+    {
+      "epoch": 3.9339968569931902,
+      "grad_norm": 0.8389805555343628,
+      "learning_rate": 1.6849572670959435e-08,
+      "loss": 0.1016,
+      "step": 7510
+    },
+    {
+      "epoch": 3.9339968569931902,
+      "loss_partial": 0.06103515625,
+      "step": 7510
+    },
+    {
+      "epoch": 3.9339968569931902,
+      "loss_partial": 0.044921875,
+      "step": 7510
+    },
+    {
+      "epoch": 3.9345206914614983,
+      "grad_norm": 0.7419790625572205,
+      "learning_rate": 1.658534282911961e-08,
+      "loss": 0.0999,
+      "step": 7511
+    },
+    {
+      "epoch": 3.9345206914614983,
+      "loss_partial": 0.052734375,
+      "step": 7511
+    },
+    {
+      "epoch": 3.9345206914614983,
+      "loss_partial": 0.054443359375,
+      "step": 7511
+    },
+    {
+      "epoch": 3.935044525929806,
+      "grad_norm": 0.6833216547966003,
+      "learning_rate": 1.6323199459208348e-08,
+      "loss": 0.0953,
+      "step": 7512
+    },
+    {
+      "epoch": 3.935044525929806,
+      "loss_partial": 0.0498046875,
+      "step": 7512
+    },
+    {
+      "epoch": 3.935044525929806,
+      "loss_partial": 0.038330078125,
+      "step": 7512
+    },
+    {
+      "epoch": 3.935568360398114,
+      "grad_norm": 0.7752898931503296,
+      "learning_rate": 1.606314261600961e-08,
+      "loss": 0.0967,
+      "step": 7513
+    },
+    {
+      "epoch": 3.935568360398114,
+      "loss_partial": 0.051513671875,
+      "step": 7513
+    },
+    {
+      "epoch": 3.935568360398114,
+      "loss_partial": 0.044921875,
+      "step": 7513
+    },
+    {
+      "epoch": 3.936092194866422,
+      "grad_norm": 0.7641606330871582,
+      "learning_rate": 1.5805172353876575e-08,
+      "loss": 0.0982,
+      "step": 7514
+    },
+    {
+      "epoch": 3.936092194866422,
+      "loss_partial": 0.052001953125,
+      "step": 7514
+    },
+    {
+      "epoch": 3.936092194866422,
+      "loss_partial": 0.043701171875,
+      "step": 7514
+    },
+    {
+      "epoch": 3.93661602933473,
+      "grad_norm": 0.9675564169883728,
+      "learning_rate": 1.554928872672168e-08,
+      "loss": 0.0967,
+      "step": 7515
+    },
+    {
+      "epoch": 3.93661602933473,
+      "loss_partial": 0.04541015625,
+      "step": 7515
+    },
+    {
+      "epoch": 3.93661602933473,
+      "loss_partial": 0.060791015625,
+      "step": 7515
+    },
+    {
+      "epoch": 3.937139863803038,
+      "grad_norm": 0.8284558653831482,
+      "learning_rate": 1.5295491788023252e-08,
+      "loss": 0.1014,
+      "step": 7516
+    },
+    {
+      "epoch": 3.937139863803038,
+      "loss_partial": 0.0419921875,
+      "step": 7516
+    },
+    {
+      "epoch": 3.937139863803038,
+      "loss_partial": 0.0517578125,
+      "step": 7516
+    },
+    {
+      "epoch": 3.937663698271346,
+      "grad_norm": 0.8937148451805115,
+      "learning_rate": 1.5043781590823313e-08,
+      "loss": 0.101,
+      "step": 7517
+    },
+    {
+      "epoch": 3.937663698271346,
+      "loss_partial": 0.05712890625,
+      "step": 7517
+    },
+    {
+      "epoch": 3.937663698271346,
+      "loss_partial": 0.046875,
+      "step": 7517
+    },
+    {
+      "epoch": 3.9381875327396543,
+      "grad_norm": 0.795762836933136,
+      "learning_rate": 1.4794158187728669e-08,
+      "loss": 0.0996,
+      "step": 7518
+    },
+    {
+      "epoch": 3.9381875327396543,
+      "loss_partial": 0.050048828125,
+      "step": 7518
+    },
+    {
+      "epoch": 3.9381875327396543,
+      "loss_partial": 0.05126953125,
+      "step": 7518
+    },
+    {
+      "epoch": 3.9387113672079623,
+      "grad_norm": 0.7504622936248779,
+      "learning_rate": 1.4546621630907587e-08,
+      "loss": 0.1,
+      "step": 7519
+    },
+    {
+      "epoch": 3.9387113672079623,
+      "loss_partial": 0.0439453125,
+      "step": 7519
+    },
+    {
+      "epoch": 3.9387113672079623,
+      "loss_partial": 0.047607421875,
+      "step": 7519
+    },
+    {
+      "epoch": 3.9392352016762704,
+      "grad_norm": 0.9499995112419128,
+      "learning_rate": 1.4301171972095352e-08,
+      "loss": 0.1025,
+      "step": 7520
+    },
+    {
+      "epoch": 3.9392352016762704,
+      "loss_partial": 0.057373046875,
+      "step": 7520
+    },
+    {
+      "epoch": 3.9392352016762704,
+      "loss_partial": 0.04248046875,
+      "step": 7520
+    },
+    {
+      "epoch": 3.9397590361445785,
+      "grad_norm": 0.8368024826049805,
+      "learning_rate": 1.4057809262587597e-08,
+      "loss": 0.096,
+      "step": 7521
+    },
+    {
+      "epoch": 3.9397590361445785,
+      "loss_partial": 0.05224609375,
+      "step": 7521
+    },
+    {
+      "epoch": 3.9397590361445785,
+      "loss_partial": 0.0458984375,
+      "step": 7521
+    },
+    {
+      "epoch": 3.9402828706128865,
+      "grad_norm": 0.672549843788147,
+      "learning_rate": 1.3816533553246968e-08,
+      "loss": 0.0999,
+      "step": 7522
+    },
+    {
+      "epoch": 3.9402828706128865,
+      "loss_partial": 0.045654296875,
+      "step": 7522
+    },
+    {
+      "epoch": 3.9402828706128865,
+      "loss_partial": 0.039794921875,
+      "step": 7522
+    },
+    {
+      "epoch": 3.9408067050811946,
+      "grad_norm": 0.9134852886199951,
+      "learning_rate": 1.357734489449869e-08,
+      "loss": 0.0992,
+      "step": 7523
+    },
+    {
+      "epoch": 3.9408067050811946,
+      "loss_partial": 0.046142578125,
+      "step": 7523
+    },
+    {
+      "epoch": 3.9408067050811946,
+      "loss_partial": 0.048583984375,
+      "step": 7523
+    },
+    {
+      "epoch": 3.9413305395495026,
+      "grad_norm": 0.8000752329826355,
+      "learning_rate": 1.3340243336329439e-08,
+      "loss": 0.0949,
+      "step": 7524
+    },
+    {
+      "epoch": 3.9413305395495026,
+      "loss_partial": 0.044189453125,
+      "step": 7524
+    },
+    {
+      "epoch": 3.9413305395495026,
+      "loss_partial": 0.04150390625,
+      "step": 7524
+    },
+    {
+      "epoch": 3.9418543740178102,
+      "grad_norm": 1.1022899150848389,
+      "learning_rate": 1.3105228928295133e-08,
+      "loss": 0.0974,
+      "step": 7525
+    },
+    {
+      "epoch": 3.9418543740178102,
+      "loss_partial": 0.056396484375,
+      "step": 7525
+    },
+    {
+      "epoch": 3.9418543740178102,
+      "loss_partial": 0.049072265625,
+      "step": 7525
+    },
+    {
+      "epoch": 3.9423782084861183,
+      "grad_norm": 1.1346567869186401,
+      "learning_rate": 1.2872301719509817e-08,
+      "loss": 0.1019,
+      "step": 7526
+    },
+    {
+      "epoch": 3.9423782084861183,
+      "loss_partial": 0.053466796875,
+      "step": 7526
+    },
+    {
+      "epoch": 3.9423782084861183,
+      "loss_partial": 0.051513671875,
+      "step": 7526
+    },
+    {
+      "epoch": 3.9429020429544264,
+      "grad_norm": 0.8515775799751282,
+      "learning_rate": 1.264146175865566e-08,
+      "loss": 0.0981,
+      "step": 7527
+    },
+    {
+      "epoch": 3.9429020429544264,
+      "loss_partial": 0.044677734375,
+      "step": 7527
+    },
+    {
+      "epoch": 3.9429020429544264,
+      "loss_partial": 0.048095703125,
+      "step": 7527
+    },
+    {
+      "epoch": 3.9434258774227344,
+      "grad_norm": 0.9797539114952087,
+      "learning_rate": 1.2412709093976294e-08,
+      "loss": 0.0998,
+      "step": 7528
+    },
+    {
+      "epoch": 3.9434258774227344,
+      "loss_partial": 0.046875,
+      "step": 7528
+    },
+    {
+      "epoch": 3.9434258774227344,
+      "loss_partial": 0.04833984375,
+      "step": 7528
+    },
+    {
+      "epoch": 3.9439497118910425,
+      "grad_norm": 0.7499280571937561,
+      "learning_rate": 1.2186043773279032e-08,
+      "loss": 0.1003,
+      "step": 7529
+    },
+    {
+      "epoch": 3.9439497118910425,
+      "loss_partial": 0.04931640625,
+      "step": 7529
+    },
+    {
+      "epoch": 3.9439497118910425,
+      "loss_partial": 0.056396484375,
+      "step": 7529
+    },
+    {
+      "epoch": 3.9444735463593505,
+      "grad_norm": 0.7897759675979614,
+      "learning_rate": 1.1961465843934873e-08,
+      "loss": 0.1003,
+      "step": 7530
+    },
+    {
+      "epoch": 3.9444735463593505,
+      "loss_partial": 0.0595703125,
+      "step": 7530
+    },
+    {
+      "epoch": 3.9444735463593505,
+      "loss_partial": 0.04541015625,
+      "step": 7530
+    },
+    {
+      "epoch": 3.9449973808276586,
+      "grad_norm": 0.7705312967300415,
+      "learning_rate": 1.1738975352881821e-08,
+      "loss": 0.1,
+      "step": 7531
+    },
+    {
+      "epoch": 3.9449973808276586,
+      "loss_partial": 0.043212890625,
+      "step": 7531
+    },
+    {
+      "epoch": 3.9449973808276586,
+      "loss_partial": 0.048095703125,
+      "step": 7531
+    },
+    {
+      "epoch": 3.945521215295966,
+      "grad_norm": 1.1104475259780884,
+      "learning_rate": 1.151857234661602e-08,
+      "loss": 0.1022,
+      "step": 7532
+    },
+    {
+      "epoch": 3.945521215295966,
+      "loss_partial": 0.0380859375,
+      "step": 7532
+    },
+    {
+      "epoch": 3.945521215295966,
+      "loss_partial": 0.054443359375,
+      "step": 7532
+    },
+    {
+      "epoch": 3.9460450497642743,
+      "grad_norm": 0.8418075442314148,
+      "learning_rate": 1.1300256871202841e-08,
+      "loss": 0.0986,
+      "step": 7533
+    },
+    {
+      "epoch": 3.9460450497642743,
+      "loss_partial": 0.0537109375,
+      "step": 7533
+    },
+    {
+      "epoch": 3.9460450497642743,
+      "loss_partial": 0.0595703125,
+      "step": 7533
+    },
+    {
+      "epoch": 3.9465688842325823,
+      "grad_norm": 0.6862663626670837,
+      "learning_rate": 1.1084028972266903e-08,
+      "loss": 0.0977,
+      "step": 7534
+    },
+    {
+      "epoch": 3.9465688842325823,
+      "loss_partial": 0.04541015625,
+      "step": 7534
+    },
+    {
+      "epoch": 3.9465688842325823,
+      "loss_partial": 0.06103515625,
+      "step": 7534
+    },
+    {
+      "epoch": 3.9470927187008904,
+      "grad_norm": 0.7843126654624939,
+      "learning_rate": 1.086988869499983e-08,
+      "loss": 0.0998,
+      "step": 7535
+    },
+    {
+      "epoch": 3.9470927187008904,
+      "loss_partial": 0.0439453125,
+      "step": 7535
+    },
+    {
+      "epoch": 3.9470927187008904,
+      "loss_partial": 0.052978515625,
+      "step": 7535
+    },
+    {
+      "epoch": 3.9476165531691985,
+      "grad_norm": 0.7287627458572388,
+      "learning_rate": 1.0657836084154716e-08,
+      "loss": 0.0994,
+      "step": 7536
+    },
+    {
+      "epoch": 3.9476165531691985,
+      "loss_partial": 0.05224609375,
+      "step": 7536
+    },
+    {
+      "epoch": 3.9476165531691985,
+      "loss_partial": 0.055419921875,
+      "step": 7536
+    },
+    {
+      "epoch": 3.9481403876375065,
+      "grad_norm": 0.7571084499359131,
+      "learning_rate": 1.044787118404944e-08,
+      "loss": 0.1008,
+      "step": 7537
+    },
+    {
+      "epoch": 3.9481403876375065,
+      "loss_partial": 0.046875,
+      "step": 7537
+    },
+    {
+      "epoch": 3.9481403876375065,
+      "loss_partial": 0.057861328125,
+      "step": 7537
+    },
+    {
+      "epoch": 3.9486642221058146,
+      "grad_norm": 0.7516546845436096,
+      "learning_rate": 1.023999403856557e-08,
+      "loss": 0.1022,
+      "step": 7538
+    },
+    {
+      "epoch": 3.9486642221058146,
+      "loss_partial": 0.046875,
+      "step": 7538
+    },
+    {
+      "epoch": 3.9486642221058146,
+      "loss_partial": 0.04638671875,
+      "step": 7538
+    },
+    {
+      "epoch": 3.9491880565741226,
+      "grad_norm": 0.7086932063102722,
+      "learning_rate": 1.0034204691148353e-08,
+      "loss": 0.099,
+      "step": 7539
+    },
+    {
+      "epoch": 3.9491880565741226,
+      "loss_partial": 0.05517578125,
+      "step": 7539
+    },
+    {
+      "epoch": 3.9491880565741226,
+      "loss_partial": 0.04638671875,
+      "step": 7539
+    },
+    {
+      "epoch": 3.9497118910424307,
+      "grad_norm": 0.808996319770813,
+      "learning_rate": 9.830503184806716e-09,
+      "loss": 0.1013,
+      "step": 7540
+    },
+    {
+      "epoch": 3.9497118910424307,
+      "loss_partial": 0.0498046875,
+      "step": 7540
+    },
+    {
+      "epoch": 3.9497118910424307,
+      "loss_partial": 0.0537109375,
+      "step": 7540
+    },
+    {
+      "epoch": 3.9502357255107388,
+      "grad_norm": 0.6219347715377808,
+      "learning_rate": 9.628889562112164e-09,
+      "loss": 0.1006,
+      "step": 7541
+    },
+    {
+      "epoch": 3.9502357255107388,
+      "loss_partial": 0.045654296875,
+      "step": 7541
+    },
+    {
+      "epoch": 3.9502357255107388,
+      "loss_partial": 0.04931640625,
+      "step": 7541
+    },
+    {
+      "epoch": 3.950759559979047,
+      "grad_norm": 0.7188477516174316,
+      "learning_rate": 9.429363865200991e-09,
+      "loss": 0.098,
+      "step": 7542
+    },
+    {
+      "epoch": 3.950759559979047,
+      "loss_partial": 0.052978515625,
+      "step": 7542
+    },
+    {
+      "epoch": 3.950759559979047,
+      "loss_partial": 0.052001953125,
+      "step": 7542
+    },
+    {
+      "epoch": 3.951283394447355,
+      "grad_norm": 0.671321451663971,
+      "learning_rate": 9.231926135773173e-09,
+      "loss": 0.0974,
+      "step": 7543
+    },
+    {
+      "epoch": 3.951283394447355,
+      "loss_partial": 0.05908203125,
+      "step": 7543
+    },
+    {
+      "epoch": 3.951283394447355,
+      "loss_partial": 0.0537109375,
+      "step": 7543
+    },
+    {
+      "epoch": 3.9518072289156625,
+      "grad_norm": 0.7600880861282349,
+      "learning_rate": 9.036576415091258e-09,
+      "loss": 0.1015,
+      "step": 7544
+    },
+    {
+      "epoch": 3.9518072289156625,
+      "loss_partial": 0.05517578125,
+      "step": 7544
+    },
+    {
+      "epoch": 3.9518072289156625,
+      "loss_partial": 0.040771484375,
+      "step": 7544
+    },
+    {
+      "epoch": 3.9523310633839706,
+      "grad_norm": 0.7488523125648499,
+      "learning_rate": 8.843314743983699e-09,
+      "loss": 0.0958,
+      "step": 7545
+    },
+    {
+      "epoch": 3.9523310633839706,
+      "loss_partial": 0.052734375,
+      "step": 7545
+    },
+    {
+      "epoch": 3.9523310633839706,
+      "loss_partial": 0.044677734375,
+      "step": 7545
+    },
+    {
+      "epoch": 3.9528548978522786,
+      "grad_norm": 0.9170558452606201,
+      "learning_rate": 8.6521411628393e-09,
+      "loss": 0.0956,
+      "step": 7546
+    },
+    {
+      "epoch": 3.9528548978522786,
+      "loss_partial": 0.044921875,
+      "step": 7546
+    },
+    {
+      "epoch": 3.9528548978522786,
+      "loss_partial": 0.05908203125,
+      "step": 7546
+    },
+    {
+      "epoch": 3.9533787323205867,
+      "grad_norm": 0.8463433384895325,
+      "learning_rate": 8.463055711611656e-09,
+      "loss": 0.1017,
+      "step": 7547
+    },
+    {
+      "epoch": 3.9533787323205867,
+      "loss_partial": 0.06396484375,
+      "step": 7547
+    },
+    {
+      "epoch": 3.9533787323205867,
+      "loss_partial": 0.053955078125,
+      "step": 7547
+    },
+    {
+      "epoch": 3.9539025667888947,
+      "grad_norm": 0.7828716039657593,
+      "learning_rate": 8.276058429821376e-09,
+      "loss": 0.1015,
+      "step": 7548
+    },
+    {
+      "epoch": 3.9539025667888947,
+      "loss_partial": 0.04931640625,
+      "step": 7548
+    },
+    {
+      "epoch": 3.9539025667888947,
+      "loss_partial": 0.046875,
+      "step": 7548
+    },
+    {
+      "epoch": 3.954426401257203,
+      "grad_norm": 0.9353117346763611,
+      "learning_rate": 8.0911493565472e-09,
+      "loss": 0.0989,
+      "step": 7549
+    },
+    {
+      "epoch": 3.954426401257203,
+      "loss_partial": 0.05615234375,
+      "step": 7549
+    },
+    {
+      "epoch": 3.954426401257203,
+      "loss_partial": 0.044677734375,
+      "step": 7549
+    },
+    {
+      "epoch": 3.954950235725511,
+      "grad_norm": 0.7148250937461853,
+      "learning_rate": 7.908328530433774e-09,
+      "loss": 0.1017,
+      "step": 7550
+    },
+    {
+      "epoch": 3.954950235725511,
+      "loss_partial": 0.045166015625,
+      "step": 7550
+    },
+    {
+      "epoch": 3.954950235725511,
+      "loss_partial": 0.050537109375,
+      "step": 7550
+    },
+    {
+      "epoch": 3.955474070193819,
+      "grad_norm": 0.8874078392982483,
+      "learning_rate": 7.727595989691639e-09,
+      "loss": 0.1017,
+      "step": 7551
+    },
+    {
+      "epoch": 3.955474070193819,
+      "loss_partial": 0.048828125,
+      "step": 7551
+    },
+    {
+      "epoch": 3.955474070193819,
+      "loss_partial": 0.06201171875,
+      "step": 7551
+    },
+    {
+      "epoch": 3.9559979046621265,
+      "grad_norm": 0.9588555097579956,
+      "learning_rate": 7.548951772091695e-09,
+      "loss": 0.0989,
+      "step": 7552
+    },
+    {
+      "epoch": 3.9559979046621265,
+      "loss_partial": 0.0654296875,
+      "step": 7552
+    },
+    {
+      "epoch": 3.9559979046621265,
+      "loss_partial": 0.052734375,
+      "step": 7552
+    },
+    {
+      "epoch": 3.9565217391304346,
+      "grad_norm": 0.9444665908813477,
+      "learning_rate": 7.372395914968522e-09,
+      "loss": 0.1038,
+      "step": 7553
+    },
+    {
+      "epoch": 3.9565217391304346,
+      "loss_partial": 0.04150390625,
+      "step": 7553
+    },
+    {
+      "epoch": 3.9565217391304346,
+      "loss_partial": 0.05126953125,
+      "step": 7553
+    },
+    {
+      "epoch": 3.9570455735987426,
+      "grad_norm": 0.8401587009429932,
+      "learning_rate": 7.197928455222602e-09,
+      "loss": 0.1004,
+      "step": 7554
+    },
+    {
+      "epoch": 3.9570455735987426,
+      "loss_partial": 0.0458984375,
+      "step": 7554
+    },
+    {
+      "epoch": 3.9570455735987426,
+      "loss_partial": 0.0556640625,
+      "step": 7554
+    },
+    {
+      "epoch": 3.9575694080670507,
+      "grad_norm": 0.9958390593528748,
+      "learning_rate": 7.025549429315881e-09,
+      "loss": 0.0995,
+      "step": 7555
+    },
+    {
+      "epoch": 3.9575694080670507,
+      "loss_partial": 0.0517578125,
+      "step": 7555
+    },
+    {
+      "epoch": 3.9575694080670507,
+      "loss_partial": 0.053955078125,
+      "step": 7555
+    },
+    {
+      "epoch": 3.9580932425353588,
+      "grad_norm": 0.8315263390541077,
+      "learning_rate": 6.8552588732739845e-09,
+      "loss": 0.0967,
+      "step": 7556
+    },
+    {
+      "epoch": 3.9580932425353588,
+      "loss_partial": 0.042724609375,
+      "step": 7556
+    },
+    {
+      "epoch": 3.9580932425353588,
+      "loss_partial": 0.050537109375,
+      "step": 7556
+    },
+    {
+      "epoch": 3.958617077003667,
+      "grad_norm": 0.8001790046691895,
+      "learning_rate": 6.687056822688442e-09,
+      "loss": 0.1009,
+      "step": 7557
+    },
+    {
+      "epoch": 3.958617077003667,
+      "loss_partial": 0.0498046875,
+      "step": 7557
+    },
+    {
+      "epoch": 3.958617077003667,
+      "loss_partial": 0.043701171875,
+      "step": 7557
+    },
+    {
+      "epoch": 3.959140911471975,
+      "grad_norm": 0.7900808453559875,
+      "learning_rate": 6.5209433127100266e-09,
+      "loss": 0.1001,
+      "step": 7558
+    },
+    {
+      "epoch": 3.959140911471975,
+      "loss_partial": 0.0654296875,
+      "step": 7558
+    },
+    {
+      "epoch": 3.959140911471975,
+      "loss_partial": 0.043701171875,
+      "step": 7558
+    },
+    {
+      "epoch": 3.959664745940283,
+      "grad_norm": 0.6815615296363831,
+      "learning_rate": 6.3569183780576304e-09,
+      "loss": 0.0992,
+      "step": 7559
+    },
+    {
+      "epoch": 3.959664745940283,
+      "loss_partial": 0.041748046875,
+      "step": 7559
+    },
+    {
+      "epoch": 3.959664745940283,
+      "loss_partial": 0.0517578125,
+      "step": 7559
+    },
+    {
+      "epoch": 3.960188580408591,
+      "grad_norm": 0.8155821561813354,
+      "learning_rate": 6.194982053010501e-09,
+      "loss": 0.0989,
+      "step": 7560
+    },
+    {
+      "epoch": 3.960188580408591,
+      "loss_partial": 0.0478515625,
+      "step": 7560
+    },
+    {
+      "epoch": 3.960188580408591,
+      "loss_partial": 0.03955078125,
+      "step": 7560
+    },
+    {
+      "epoch": 3.960712414876899,
+      "grad_norm": 0.6966478228569031,
+      "learning_rate": 6.035134371411567e-09,
+      "loss": 0.1028,
+      "step": 7561
+    },
+    {
+      "epoch": 3.960712414876899,
+      "loss_partial": 0.04296875,
+      "step": 7561
+    },
+    {
+      "epoch": 3.960712414876899,
+      "loss_partial": 0.048095703125,
+      "step": 7561
+    },
+    {
+      "epoch": 3.961236249345207,
+      "grad_norm": 0.6976132988929749,
+      "learning_rate": 5.877375366668547e-09,
+      "loss": 0.0989,
+      "step": 7562
+    },
+    {
+      "epoch": 3.961236249345207,
+      "loss_partial": 0.047119140625,
+      "step": 7562
+    },
+    {
+      "epoch": 3.961236249345207,
+      "loss_partial": 0.06103515625,
+      "step": 7562
+    },
+    {
+      "epoch": 3.961760083813515,
+      "grad_norm": 0.6995316743850708,
+      "learning_rate": 5.721705071752848e-09,
+      "loss": 0.1033,
+      "step": 7563
+    },
+    {
+      "epoch": 3.961760083813515,
+      "loss_partial": 0.042236328125,
+      "step": 7563
+    },
+    {
+      "epoch": 3.961760083813515,
+      "loss_partial": 0.05029296875,
+      "step": 7563
+    },
+    {
+      "epoch": 3.962283918281823,
+      "grad_norm": 0.7094444036483765,
+      "learning_rate": 5.568123519198443e-09,
+      "loss": 0.0978,
+      "step": 7564
+    },
+    {
+      "epoch": 3.962283918281823,
+      "loss_partial": 0.051513671875,
+      "step": 7564
+    },
+    {
+      "epoch": 3.962283918281823,
+      "loss_partial": 0.04736328125,
+      "step": 7564
+    },
+    {
+      "epoch": 3.962807752750131,
+      "grad_norm": 0.8601605296134949,
+      "learning_rate": 5.41663074110188e-09,
+      "loss": 0.0997,
+      "step": 7565
+    },
+    {
+      "epoch": 3.962807752750131,
+      "loss_partial": 0.0625,
+      "step": 7565
+    },
+    {
+      "epoch": 3.962807752750131,
+      "loss_partial": 0.044921875,
+      "step": 7565
+    },
+    {
+      "epoch": 3.963331587218439,
+      "grad_norm": 1.1084810495376587,
+      "learning_rate": 5.2672267691245006e-09,
+      "loss": 0.1049,
+      "step": 7566
+    },
+    {
+      "epoch": 3.963331587218439,
+      "loss_partial": 0.0537109375,
+      "step": 7566
+    },
+    {
+      "epoch": 3.963331587218439,
+      "loss_partial": 0.0517578125,
+      "step": 7566
+    },
+    {
+      "epoch": 3.963855421686747,
+      "grad_norm": 0.8608137965202332,
+      "learning_rate": 5.119911634492436e-09,
+      "loss": 0.1,
+      "step": 7567
+    },
+    {
+      "epoch": 3.963855421686747,
+      "loss_partial": 0.049560546875,
+      "step": 7567
+    },
+    {
+      "epoch": 3.963855421686747,
+      "loss_partial": 0.058837890625,
+      "step": 7567
+    },
+    {
+      "epoch": 3.964379256155055,
+      "grad_norm": 0.7746254205703735,
+      "learning_rate": 4.974685367991061e-09,
+      "loss": 0.1021,
+      "step": 7568
+    },
+    {
+      "epoch": 3.964379256155055,
+      "loss_partial": 0.045166015625,
+      "step": 7568
+    },
+    {
+      "epoch": 3.964379256155055,
+      "loss_partial": 0.047119140625,
+      "step": 7568
+    },
+    {
+      "epoch": 3.964903090623363,
+      "grad_norm": 0.8044475317001343,
+      "learning_rate": 4.831547999973873e-09,
+      "loss": 0.0983,
+      "step": 7569
+    },
+    {
+      "epoch": 3.964903090623363,
+      "loss_partial": 0.05029296875,
+      "step": 7569
+    },
+    {
+      "epoch": 3.964903090623363,
+      "loss_partial": 0.048095703125,
+      "step": 7569
+    },
+    {
+      "epoch": 3.965426925091671,
+      "grad_norm": 0.7759712338447571,
+      "learning_rate": 4.690499560354722e-09,
+      "loss": 0.0958,
+      "step": 7570
+    },
+    {
+      "epoch": 3.965426925091671,
+      "loss_partial": 0.052734375,
+      "step": 7570
+    },
+    {
+      "epoch": 3.965426925091671,
+      "loss_partial": 0.05029296875,
+      "step": 7570
+    },
+    {
+      "epoch": 3.9659507595599788,
+      "grad_norm": 0.8472874760627747,
+      "learning_rate": 4.551540078611138e-09,
+      "loss": 0.0961,
+      "step": 7571
+    },
+    {
+      "epoch": 3.9659507595599788,
+      "loss_partial": 0.0419921875,
+      "step": 7571
+    },
+    {
+      "epoch": 3.9659507595599788,
+      "loss_partial": 0.051513671875,
+      "step": 7571
+    },
+    {
+      "epoch": 3.966474594028287,
+      "grad_norm": 0.898427426815033,
+      "learning_rate": 4.414669583786557e-09,
+      "loss": 0.0973,
+      "step": 7572
+    },
+    {
+      "epoch": 3.966474594028287,
+      "loss_partial": 0.04541015625,
+      "step": 7572
+    },
+    {
+      "epoch": 3.966474594028287,
+      "loss_partial": 0.06005859375,
+      "step": 7572
+    },
+    {
+      "epoch": 3.966998428496595,
+      "grad_norm": 0.7093219757080078,
+      "learning_rate": 4.279888104485874e-09,
+      "loss": 0.0984,
+      "step": 7573
+    },
+    {
+      "epoch": 3.966998428496595,
+      "loss_partial": 0.04443359375,
+      "step": 7573
+    },
+    {
+      "epoch": 3.966998428496595,
+      "loss_partial": 0.05517578125,
+      "step": 7573
+    },
+    {
+      "epoch": 3.967522262964903,
+      "grad_norm": 0.9659728407859802,
+      "learning_rate": 4.147195668875448e-09,
+      "loss": 0.0966,
+      "step": 7574
+    },
+    {
+      "epoch": 3.967522262964903,
+      "loss_partial": 0.052734375,
+      "step": 7574
+    },
+    {
+      "epoch": 3.967522262964903,
+      "loss_partial": 0.046142578125,
+      "step": 7574
+    },
+    {
+      "epoch": 3.968046097433211,
+      "grad_norm": 0.7202832102775574,
+      "learning_rate": 4.01659230468976e-09,
+      "loss": 0.098,
+      "step": 7575
+    },
+    {
+      "epoch": 3.968046097433211,
+      "loss_partial": 0.04638671875,
+      "step": 7575
+    },
+    {
+      "epoch": 3.968046097433211,
+      "loss_partial": 0.0517578125,
+      "step": 7575
+    },
+    {
+      "epoch": 3.968569931901519,
+      "grad_norm": 0.8546369671821594,
+      "learning_rate": 3.888078039222531e-09,
+      "loss": 0.1011,
+      "step": 7576
+    },
+    {
+      "epoch": 3.968569931901519,
+      "loss_partial": 0.068359375,
+      "step": 7576
+    },
+    {
+      "epoch": 3.968569931901519,
+      "loss_partial": 0.0556640625,
+      "step": 7576
+    },
+    {
+      "epoch": 3.969093766369827,
+      "grad_norm": 1.0174510478973389,
+      "learning_rate": 3.76165289933228e-09,
+      "loss": 0.1039,
+      "step": 7577
+    },
+    {
+      "epoch": 3.969093766369827,
+      "loss_partial": 0.04541015625,
+      "step": 7577
+    },
+    {
+      "epoch": 3.969093766369827,
+      "loss_partial": 0.05126953125,
+      "step": 7577
+    },
+    {
+      "epoch": 3.969617600838135,
+      "grad_norm": 1.020121455192566,
+      "learning_rate": 3.6373169114423123e-09,
+      "loss": 0.0977,
+      "step": 7578
+    },
+    {
+      "epoch": 3.969617600838135,
+      "loss_partial": 0.049560546875,
+      "step": 7578
+    },
+    {
+      "epoch": 3.969617600838135,
+      "loss_partial": 0.055908203125,
+      "step": 7578
+    },
+    {
+      "epoch": 3.9701414353064433,
+      "grad_norm": 0.7715235352516174,
+      "learning_rate": 3.515070101537399e-09,
+      "loss": 0.0992,
+      "step": 7579
+    },
+    {
+      "epoch": 3.9701414353064433,
+      "loss_partial": 0.049560546875,
+      "step": 7579
+    },
+    {
+      "epoch": 3.9701414353064433,
+      "loss_partial": 0.053466796875,
+      "step": 7579
+    },
+    {
+      "epoch": 3.9706652697747513,
+      "grad_norm": 0.8511924743652344,
+      "learning_rate": 3.394912495167102e-09,
+      "loss": 0.0976,
+      "step": 7580
+    },
+    {
+      "epoch": 3.9706652697747513,
+      "loss_partial": 0.05126953125,
+      "step": 7580
+    },
+    {
+      "epoch": 3.9706652697747513,
+      "loss_partial": 0.052978515625,
+      "step": 7580
+    },
+    {
+      "epoch": 3.9711891042430594,
+      "grad_norm": 0.8237831592559814,
+      "learning_rate": 3.276844117442446e-09,
+      "loss": 0.0986,
+      "step": 7581
+    },
+    {
+      "epoch": 3.9711891042430594,
+      "loss_partial": 0.045166015625,
+      "step": 7581
+    },
+    {
+      "epoch": 3.9711891042430594,
+      "loss_partial": 0.044677734375,
+      "step": 7581
+    },
+    {
+      "epoch": 3.9717129387113674,
+      "grad_norm": 0.8737230896949768,
+      "learning_rate": 3.1608649930392476e-09,
+      "loss": 0.0972,
+      "step": 7582
+    },
+    {
+      "epoch": 3.9717129387113674,
+      "loss_partial": 0.0439453125,
+      "step": 7582
+    },
+    {
+      "epoch": 3.9717129387113674,
+      "loss_partial": 0.036865234375,
+      "step": 7582
+    },
+    {
+      "epoch": 3.9722367731796755,
+      "grad_norm": 0.808079719543457,
+      "learning_rate": 3.0469751461970064e-09,
+      "loss": 0.0972,
+      "step": 7583
+    },
+    {
+      "epoch": 3.9722367731796755,
+      "loss_partial": 0.0439453125,
+      "step": 7583
+    },
+    {
+      "epoch": 3.9722367731796755,
+      "loss_partial": 0.038330078125,
+      "step": 7583
+    },
+    {
+      "epoch": 3.972760607647983,
+      "grad_norm": 0.8623876571655273,
+      "learning_rate": 2.935174600717794e-09,
+      "loss": 0.097,
+      "step": 7584
+    },
+    {
+      "epoch": 3.972760607647983,
+      "loss_partial": 0.053466796875,
+      "step": 7584
+    },
+    {
+      "epoch": 3.972760607647983,
+      "loss_partial": 0.05615234375,
+      "step": 7584
+    },
+    {
+      "epoch": 3.973284442116291,
+      "grad_norm": 0.9334010481834412,
+      "learning_rate": 2.8254633799662536e-09,
+      "loss": 0.0995,
+      "step": 7585
+    },
+    {
+      "epoch": 3.973284442116291,
+      "loss_partial": 0.0478515625,
+      "step": 7585
+    },
+    {
+      "epoch": 3.973284442116291,
+      "loss_partial": 0.04541015625,
+      "step": 7585
+    },
+    {
+      "epoch": 3.9738082765845992,
+      "grad_norm": 0.791857898235321,
+      "learning_rate": 2.7178415068729315e-09,
+      "loss": 0.0977,
+      "step": 7586
+    },
+    {
+      "epoch": 3.9738082765845992,
+      "loss_partial": 0.04931640625,
+      "step": 7586
+    },
+    {
+      "epoch": 3.9738082765845992,
+      "loss_partial": 0.05126953125,
+      "step": 7586
+    },
+    {
+      "epoch": 3.9743321110529073,
+      "grad_norm": 0.7613568902015686,
+      "learning_rate": 2.612309003928726e-09,
+      "loss": 0.0994,
+      "step": 7587
+    },
+    {
+      "epoch": 3.9743321110529073,
+      "loss_partial": 0.057373046875,
+      "step": 7587
+    },
+    {
+      "epoch": 3.9743321110529073,
+      "loss_partial": 0.049072265625,
+      "step": 7587
+    },
+    {
+      "epoch": 3.9748559455212154,
+      "grad_norm": 0.8664255738258362,
+      "learning_rate": 2.5088658931904376e-09,
+      "loss": 0.0975,
+      "step": 7588
+    },
+    {
+      "epoch": 3.9748559455212154,
+      "loss_partial": 0.057373046875,
+      "step": 7588
+    },
+    {
+      "epoch": 3.9748559455212154,
+      "loss_partial": 0.047119140625,
+      "step": 7588
+    },
+    {
+      "epoch": 3.9753797799895234,
+      "grad_norm": 1.008812665939331,
+      "learning_rate": 2.4075121962763294e-09,
+      "loss": 0.1028,
+      "step": 7589
+    },
+    {
+      "epoch": 3.9753797799895234,
+      "loss_partial": 0.056396484375,
+      "step": 7589
+    },
+    {
+      "epoch": 3.9753797799895234,
+      "loss_partial": 0.049560546875,
+      "step": 7589
+    },
+    {
+      "epoch": 3.9759036144578315,
+      "grad_norm": 0.810235857963562,
+      "learning_rate": 2.3082479343694565e-09,
+      "loss": 0.0995,
+      "step": 7590
+    },
+    {
+      "epoch": 3.9759036144578315,
+      "loss_partial": 0.052490234375,
+      "step": 7590
+    },
+    {
+      "epoch": 3.9759036144578315,
+      "loss_partial": 0.0546875,
+      "step": 7590
+    },
+    {
+      "epoch": 3.976427448926139,
+      "grad_norm": 0.7794727087020874,
+      "learning_rate": 2.2110731282143362e-09,
+      "loss": 0.1015,
+      "step": 7591
+    },
+    {
+      "epoch": 3.976427448926139,
+      "loss_partial": 0.057373046875,
+      "step": 7591
+    },
+    {
+      "epoch": 3.976427448926139,
+      "loss_partial": 0.05224609375,
+      "step": 7591
+    },
+    {
+      "epoch": 3.976951283394447,
+      "grad_norm": 0.7231674194335938,
+      "learning_rate": 2.1159877981213884e-09,
+      "loss": 0.0978,
+      "step": 7592
+    },
+    {
+      "epoch": 3.976951283394447,
+      "loss_partial": 0.051025390625,
+      "step": 7592
+    },
+    {
+      "epoch": 3.976951283394447,
+      "loss_partial": 0.054931640625,
+      "step": 7592
+    },
+    {
+      "epoch": 3.977475117862755,
+      "grad_norm": 0.7244316935539246,
+      "learning_rate": 2.022991963961385e-09,
+      "loss": 0.0969,
+      "step": 7593
+    },
+    {
+      "epoch": 3.977475117862755,
+      "loss_partial": 0.045166015625,
+      "step": 7593
+    },
+    {
+      "epoch": 3.977475117862755,
+      "loss_partial": 0.05615234375,
+      "step": 7593
+    },
+    {
+      "epoch": 3.9779989523310633,
+      "grad_norm": 0.8165163993835449,
+      "learning_rate": 1.93208564516989e-09,
+      "loss": 0.097,
+      "step": 7594
+    },
+    {
+      "epoch": 3.9779989523310633,
+      "loss_partial": 0.052490234375,
+      "step": 7594
+    },
+    {
+      "epoch": 3.9779989523310633,
+      "loss_partial": 0.051025390625,
+      "step": 7594
+    },
+    {
+      "epoch": 3.9785227867993713,
+      "grad_norm": 0.8599740862846375,
+      "learning_rate": 1.8432688607472604e-09,
+      "loss": 0.0973,
+      "step": 7595
+    },
+    {
+      "epoch": 3.9785227867993713,
+      "loss_partial": 0.05126953125,
+      "step": 7595
+    },
+    {
+      "epoch": 3.9785227867993713,
+      "loss_partial": 0.04248046875,
+      "step": 7595
+    },
+    {
+      "epoch": 3.9790466212676794,
+      "grad_norm": 0.754027247428894,
+      "learning_rate": 1.7565416292542047e-09,
+      "loss": 0.0988,
+      "step": 7596
+    },
+    {
+      "epoch": 3.9790466212676794,
+      "loss_partial": 0.052001953125,
+      "step": 7596
+    },
+    {
+      "epoch": 3.9790466212676794,
+      "loss_partial": 0.05322265625,
+      "step": 7596
+    },
+    {
+      "epoch": 3.9795704557359874,
+      "grad_norm": 0.7069061398506165,
+      "learning_rate": 1.6719039688162242e-09,
+      "loss": 0.0965,
+      "step": 7597
+    },
+    {
+      "epoch": 3.9795704557359874,
+      "loss_partial": 0.0576171875,
+      "step": 7597
+    },
+    {
+      "epoch": 3.9795704557359874,
+      "loss_partial": 0.037109375,
+      "step": 7597
+    },
+    {
+      "epoch": 3.9800942902042955,
+      "grad_norm": 1.1084941625595093,
+      "learning_rate": 1.589355897122502e-09,
+      "loss": 0.1009,
+      "step": 7598
+    },
+    {
+      "epoch": 3.9800942902042955,
+      "loss_partial": 0.05029296875,
+      "step": 7598
+    },
+    {
+      "epoch": 3.9800942902042955,
+      "loss_partial": 0.04296875,
+      "step": 7598
+    },
+    {
+      "epoch": 3.9806181246726036,
+      "grad_norm": 0.9505265355110168,
+      "learning_rate": 1.5088974314259042e-09,
+      "loss": 0.0988,
+      "step": 7599
+    },
+    {
+      "epoch": 3.9806181246726036,
+      "loss_partial": 0.048095703125,
+      "step": 7599
+    },
+    {
+      "epoch": 3.9806181246726036,
+      "loss_partial": 0.042724609375,
+      "step": 7599
+    },
+    {
+      "epoch": 3.9811419591409116,
+      "grad_norm": 0.9960018396377563,
+      "learning_rate": 1.4305285885407582e-09,
+      "loss": 0.1007,
+      "step": 7600
+    },
+    {
+      "epoch": 3.9811419591409116,
+      "loss_partial": 0.04541015625,
+      "step": 7600
+    },
+    {
+      "epoch": 3.9811419591409116,
+      "loss_partial": 0.043212890625,
+      "step": 7600
+    },
+    {
+      "epoch": 3.9816657936092197,
+      "grad_norm": 0.8858935832977295,
+      "learning_rate": 1.3542493848450745e-09,
+      "loss": 0.096,
+      "step": 7601
+    },
+    {
+      "epoch": 3.9816657936092197,
+      "loss_partial": 0.046142578125,
+      "step": 7601
+    },
+    {
+      "epoch": 3.9816657936092197,
+      "loss_partial": 0.0478515625,
+      "step": 7601
+    },
+    {
+      "epoch": 3.9821896280775277,
+      "grad_norm": 0.9269049763679504,
+      "learning_rate": 1.2800598362827654e-09,
+      "loss": 0.0992,
+      "step": 7602
+    },
+    {
+      "epoch": 3.9821896280775277,
+      "loss_partial": 0.049560546875,
+      "step": 7602
+    },
+    {
+      "epoch": 3.9821896280775277,
+      "loss_partial": 0.046142578125,
+      "step": 7602
+    },
+    {
+      "epoch": 3.9827134625458354,
+      "grad_norm": 0.9292490482330322,
+      "learning_rate": 1.2079599583569857e-09,
+      "loss": 0.1014,
+      "step": 7603
+    },
+    {
+      "epoch": 3.9827134625458354,
+      "loss_partial": 0.0498046875,
+      "step": 7603
+    },
+    {
+      "epoch": 3.9827134625458354,
+      "loss_partial": 0.054443359375,
+      "step": 7603
+    },
+    {
+      "epoch": 3.9832372970141434,
+      "grad_norm": 0.9025018811225891,
+      "learning_rate": 1.1379497661367921e-09,
+      "loss": 0.1012,
+      "step": 7604
+    },
+    {
+      "epoch": 3.9832372970141434,
+      "loss_partial": 0.058837890625,
+      "step": 7604
+    },
+    {
+      "epoch": 3.9832372970141434,
+      "loss_partial": 0.04443359375,
+      "step": 7604
+    },
+    {
+      "epoch": 3.9837611314824515,
+      "grad_norm": 0.7914414405822754,
+      "learning_rate": 1.0700292742538143e-09,
+      "loss": 0.0981,
+      "step": 7605
+    },
+    {
+      "epoch": 3.9837611314824515,
+      "loss_partial": 0.0537109375,
+      "step": 7605
+    },
+    {
+      "epoch": 3.9837611314824515,
+      "loss_partial": 0.047607421875,
+      "step": 7605
+    },
+    {
+      "epoch": 3.9842849659507595,
+      "grad_norm": 0.9393589496612549,
+      "learning_rate": 1.0041984969033636e-09,
+      "loss": 0.1009,
+      "step": 7606
+    },
+    {
+      "epoch": 3.9842849659507595,
+      "loss_partial": 0.05078125,
+      "step": 7606
+    },
+    {
+      "epoch": 3.9842849659507595,
+      "loss_partial": 0.06591796875,
+      "step": 7606
+    },
+    {
+      "epoch": 3.9848088004190676,
+      "grad_norm": 1.1176114082336426,
+      "learning_rate": 9.404574478433237e-10,
+      "loss": 0.1,
+      "step": 7607
+    },
+    {
+      "epoch": 3.9848088004190676,
+      "loss_partial": 0.05224609375,
+      "step": 7607
+    },
+    {
+      "epoch": 3.9848088004190676,
+      "loss_partial": 0.044677734375,
+      "step": 7607
+    },
+    {
+      "epoch": 3.9853326348873757,
+      "grad_norm": 0.9115298390388489,
+      "learning_rate": 8.78806140396371e-10,
+      "loss": 0.098,
+      "step": 7608
+    },
+    {
+      "epoch": 3.9853326348873757,
+      "loss_partial": 0.041748046875,
+      "step": 7608
+    },
+    {
+      "epoch": 3.9853326348873757,
+      "loss_partial": 0.052001953125,
+      "step": 7608
+    },
+    {
+      "epoch": 3.9858564693556837,
+      "grad_norm": 0.7862516641616821,
+      "learning_rate": 8.192445874455334e-10,
+      "loss": 0.0994,
+      "step": 7609
+    },
+    {
+      "epoch": 3.9858564693556837,
+      "loss_partial": 0.05126953125,
+      "step": 7609
+    },
+    {
+      "epoch": 3.9858564693556837,
+      "loss_partial": 0.045166015625,
+      "step": 7609
+    },
+    {
+      "epoch": 3.9863803038239918,
+      "grad_norm": 0.8927168250083923,
+      "learning_rate": 7.617728014397419e-10,
+      "loss": 0.1012,
+      "step": 7610
+    },
+    {
+      "epoch": 3.9863803038239918,
+      "loss_partial": 0.053955078125,
+      "step": 7610
+    },
+    {
+      "epoch": 3.9863803038239918,
+      "loss_partial": 0.04638671875,
+      "step": 7610
+    },
+    {
+      "epoch": 3.9869041382922994,
+      "grad_norm": 0.7834617495536804,
+      "learning_rate": 7.063907943893888e-10,
+      "loss": 0.1023,
+      "step": 7611
+    },
+    {
+      "epoch": 3.9869041382922994,
+      "loss_partial": 0.05224609375,
+      "step": 7611
+    },
+    {
+      "epoch": 3.9869041382922994,
+      "loss_partial": 0.060302734375,
+      "step": 7611
+    },
+    {
+      "epoch": 3.9874279727606075,
+      "grad_norm": 0.9548975229263306,
+      "learning_rate": 6.530985778696596e-10,
+      "loss": 0.1024,
+      "step": 7612
+    },
+    {
+      "epoch": 3.9874279727606075,
+      "loss_partial": 0.060791015625,
+      "step": 7612
+    },
+    {
+      "epoch": 3.9874279727606075,
+      "loss_partial": 0.04541015625,
+      "step": 7612
+    },
+    {
+      "epoch": 3.9879518072289155,
+      "grad_norm": 0.7578810453414917,
+      "learning_rate": 6.018961630183118e-10,
+      "loss": 0.099,
+      "step": 7613
+    },
+    {
+      "epoch": 3.9879518072289155,
+      "loss_partial": 0.060546875,
+      "step": 7613
+    },
+    {
+      "epoch": 3.9879518072289155,
+      "loss_partial": 0.050048828125,
+      "step": 7613
+    },
+    {
+      "epoch": 3.9884756416972236,
+      "grad_norm": 0.9137651920318604,
+      "learning_rate": 5.527835605367849e-10,
+      "loss": 0.1034,
+      "step": 7614
+    },
+    {
+      "epoch": 3.9884756416972236,
+      "loss_partial": 0.047119140625,
+      "step": 7614
+    },
+    {
+      "epoch": 3.9884756416972236,
+      "loss_partial": 0.044189453125,
+      "step": 7614
+    },
+    {
+      "epoch": 3.9889994761655316,
+      "grad_norm": 0.848575234413147,
+      "learning_rate": 5.057607806890907e-10,
+      "loss": 0.0999,
+      "step": 7615
+    },
+    {
+      "epoch": 3.9889994761655316,
+      "loss_partial": 0.044921875,
+      "step": 7615
+    },
+    {
+      "epoch": 3.9889994761655316,
+      "loss_partial": 0.046875,
+      "step": 7615
+    },
+    {
+      "epoch": 3.9895233106338397,
+      "grad_norm": 0.811093807220459,
+      "learning_rate": 4.608278333018135e-10,
+      "loss": 0.1033,
+      "step": 7616
+    },
+    {
+      "epoch": 3.9895233106338397,
+      "loss_partial": 0.051025390625,
+      "step": 7616
+    },
+    {
+      "epoch": 3.9895233106338397,
+      "loss_partial": 0.04248046875,
+      "step": 7616
+    },
+    {
+      "epoch": 3.9900471451021478,
+      "grad_norm": 0.8746946454048157,
+      "learning_rate": 4.179847277652194e-10,
+      "loss": 0.0984,
+      "step": 7617
+    },
+    {
+      "epoch": 3.9900471451021478,
+      "loss_partial": 0.058349609375,
+      "step": 7617
+    },
+    {
+      "epoch": 3.9900471451021478,
+      "loss_partial": 0.0517578125,
+      "step": 7617
+    },
+    {
+      "epoch": 3.990570979570456,
+      "grad_norm": 0.9759007096290588,
+      "learning_rate": 3.772314730354776e-10,
+      "loss": 0.1001,
+      "step": 7618
+    },
+    {
+      "epoch": 3.990570979570456,
+      "loss_partial": 0.044677734375,
+      "step": 7618
+    },
+    {
+      "epoch": 3.990570979570456,
+      "loss_partial": 0.0390625,
+      "step": 7618
+    },
+    {
+      "epoch": 3.991094814038764,
+      "grad_norm": 0.8763273358345032,
+      "learning_rate": 3.385680776279987e-10,
+      "loss": 0.0993,
+      "step": 7619
+    },
+    {
+      "epoch": 3.991094814038764,
+      "loss_partial": 0.057861328125,
+      "step": 7619
+    },
+    {
+      "epoch": 3.991094814038764,
+      "loss_partial": 0.049072265625,
+      "step": 7619
+    },
+    {
+      "epoch": 3.991618648507072,
+      "grad_norm": 1.105895757675171,
+      "learning_rate": 3.0199454962409616e-10,
+      "loss": 0.0989,
+      "step": 7620
+    },
+    {
+      "epoch": 3.991618648507072,
+      "loss_partial": 0.0576171875,
+      "step": 7620
+    },
+    {
+      "epoch": 3.991618648507072,
+      "loss_partial": 0.051025390625,
+      "step": 7620
+    },
+    {
+      "epoch": 3.99214248297538,
+      "grad_norm": 1.0545778274536133,
+      "learning_rate": 2.675108966676554e-10,
+      "loss": 0.1005,
+      "step": 7621
+    },
+    {
+      "epoch": 3.99214248297538,
+      "loss_partial": 0.055908203125,
+      "step": 7621
+    },
+    {
+      "epoch": 3.99214248297538,
+      "loss_partial": 0.046630859375,
+      "step": 7621
+    },
+    {
+      "epoch": 3.992666317443688,
+      "grad_norm": 0.7945617437362671,
+      "learning_rate": 2.351171259640239e-10,
+      "loss": 0.1029,
+      "step": 7622
+    },
+    {
+      "epoch": 3.992666317443688,
+      "loss_partial": 0.051513671875,
+      "step": 7622
+    },
+    {
+      "epoch": 3.992666317443688,
+      "loss_partial": 0.046630859375,
+      "step": 7622
+    },
+    {
+      "epoch": 3.9931901519119957,
+      "grad_norm": 0.7807433009147644,
+      "learning_rate": 2.0481324428445193e-10,
+      "loss": 0.0978,
+      "step": 7623
+    },
+    {
+      "epoch": 3.9931901519119957,
+      "loss_partial": 0.05322265625,
+      "step": 7623
+    },
+    {
+      "epoch": 3.9931901519119957,
+      "loss_partial": 0.052001953125,
+      "step": 7623
+    },
+    {
+      "epoch": 3.9937139863803037,
+      "grad_norm": 0.7863500714302063,
+      "learning_rate": 1.765992579627618e-10,
+      "loss": 0.1032,
+      "step": 7624
+    },
+    {
+      "epoch": 3.9937139863803037,
+      "loss_partial": 0.038818359375,
+      "step": 7624
+    },
+    {
+      "epoch": 3.9937139863803037,
+      "loss_partial": 0.0556640625,
+      "step": 7624
+    },
+    {
+      "epoch": 3.994237820848612,
+      "grad_norm": 0.7371606826782227,
+      "learning_rate": 1.5047517289312752e-10,
+      "loss": 0.1012,
+      "step": 7625
+    },
+    {
+      "epoch": 3.994237820848612,
+      "loss_partial": 0.04541015625,
+      "step": 7625
+    },
+    {
+      "epoch": 3.994237820848612,
+      "loss_partial": 0.04931640625,
+      "step": 7625
+    },
+    {
+      "epoch": 3.99476165531692,
+      "grad_norm": 0.8089141845703125,
+      "learning_rate": 1.2644099453895663e-10,
+      "loss": 0.1013,
+      "step": 7626
+    },
+    {
+      "epoch": 3.99476165531692,
+      "loss_partial": 0.03564453125,
+      "step": 7626
+    },
+    {
+      "epoch": 3.99476165531692,
+      "loss_partial": 0.046630859375,
+      "step": 7626
+    },
+    {
+      "epoch": 3.995285489785228,
+      "grad_norm": 0.6819777488708496,
+      "learning_rate": 1.0449672791956744e-10,
+      "loss": 0.1003,
+      "step": 7627
+    },
+    {
+      "epoch": 3.995285489785228,
+      "loss_partial": 0.060546875,
+      "step": 7627
+    },
+    {
+      "epoch": 3.995285489785228,
+      "loss_partial": 0.04833984375,
+      "step": 7627
+    },
+    {
+      "epoch": 3.995809324253536,
+      "grad_norm": 0.8706820011138916,
+      "learning_rate": 8.464237762351169e-11,
+      "loss": 0.0957,
+      "step": 7628
+    },
+    {
+      "epoch": 3.995809324253536,
+      "loss_partial": 0.053955078125,
+      "step": 7628
+    },
+    {
+      "epoch": 3.995809324253536,
+      "loss_partial": 0.052734375,
+      "step": 7628
+    },
+    {
+      "epoch": 3.996333158721844,
+      "grad_norm": 0.5770823955535889,
+      "learning_rate": 6.687794779969281e-11,
+      "loss": 0.1014,
+      "step": 7629
+    },
+    {
+      "epoch": 3.996333158721844,
+      "loss_partial": 0.04150390625,
+      "step": 7629
+    },
+    {
+      "epoch": 3.996333158721844,
+      "loss_partial": 0.054931640625,
+      "step": 7629
+    },
+    {
+      "epoch": 3.9968569931901516,
+      "grad_norm": 0.7526847720146179,
+      "learning_rate": 5.1203442160696615e-11,
+      "loss": 0.0964,
+      "step": 7630
+    },
+    {
+      "epoch": 3.9968569931901516,
+      "loss_partial": 0.057373046875,
+      "step": 7630
+    },
+    {
+      "epoch": 3.9968569931901516,
+      "loss_partial": 0.048583984375,
+      "step": 7630
+    },
+    {
+      "epoch": 3.9973808276584597,
+      "grad_norm": 0.8004710078239441,
+      "learning_rate": 3.761886398279124e-11,
+      "loss": 0.099,
+      "step": 7631
+    },
+    {
+      "epoch": 3.9973808276584597,
+      "loss_partial": 0.057373046875,
+      "step": 7631
+    },
+    {
+      "epoch": 3.9973808276584597,
+      "loss_partial": 0.05224609375,
+      "step": 7631
+    },
+    {
+      "epoch": 3.9979046621267678,
+      "grad_norm": 0.9281145334243774,
+      "learning_rate": 2.6124216103706745e-11,
+      "loss": 0.1006,
+      "step": 7632
+    },
+    {
+      "epoch": 3.9979046621267678,
+      "loss_partial": 0.050537109375,
+      "step": 7632
+    },
+    {
+      "epoch": 3.9979046621267678,
+      "loss_partial": 0.046875,
+      "step": 7632
+    },
+    {
+      "epoch": 3.998428496595076,
+      "grad_norm": 0.8053582310676575,
+      "learning_rate": 1.671950092707597e-11,
+      "loss": 0.0994,
+      "step": 7633
+    },
+    {
+      "epoch": 3.998428496595076,
+      "loss_partial": 0.051025390625,
+      "step": 7633
+    },
+    {
+      "epoch": 3.998428496595076,
+      "loss_partial": 0.0498046875,
+      "step": 7633
+    },
+    {
+      "epoch": 3.998952331063384,
+      "grad_norm": 0.7907680869102478,
+      "learning_rate": 9.404720417993673e-12,
+      "loss": 0.1022,
+      "step": 7634
+    },
+    {
+      "epoch": 3.998952331063384,
+      "loss_partial": 0.041748046875,
+      "step": 7634
+    },
+    {
+      "epoch": 3.998952331063384,
+      "loss_partial": 0.05126953125,
+      "step": 7634
+    },
+    {
+      "epoch": 3.999476165531692,
+      "grad_norm": 0.6177612543106079,
+      "learning_rate": 4.179876105236957e-12,
+      "loss": 0.101,
+      "step": 7635
+    },
+    {
+      "epoch": 3.999476165531692,
+      "loss_partial": 0.045166015625,
+      "step": 7635
+    },
+    {
+      "epoch": 3.999476165531692,
+      "loss_partial": 0.048095703125,
+      "step": 7635
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.7417303323745728,
+      "learning_rate": 1.044969081265279e-12,
+      "loss": 0.0977,
+      "step": 7636
+    },
+    {
+      "epoch": 4.0,
+      "step": 7636,
+      "total_flos": 4.198012256466806e+20,
+      "train_loss": 0.10746196828755239,
+      "train_runtime": 24592.7538,
+      "train_samples_per_second": 158.965,
+      "train_steps_per_second": 0.31
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 7636,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.198012256466806e+20,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}