diff --git "a/final/trainer_state.json" "b/final/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/final/trainer_state.json"
@@ -0,0 +1,4185 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.393593739637907,
+  "eval_steps": 500,
+  "global_step": 5935,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0006631739505272233,
+      "grad_norm": 1.1596795320510864,
+      "learning_rate": 2.910112359550562e-06,
+      "loss": 0.9517,
+      "step": 10
+    },
+    {
+      "epoch": 0.0013263479010544466,
+      "grad_norm": 1.0129684209823608,
+      "learning_rate": 3.921348314606742e-06,
+      "loss": 0.9185,
+      "step": 20
+    },
+    {
+      "epoch": 0.00198952185158167,
+      "grad_norm": 1.0162608623504639,
+      "learning_rate": 4.932584269662922e-06,
+      "loss": 0.8679,
+      "step": 30
+    },
+    {
+      "epoch": 0.0026526958021088932,
+      "grad_norm": 0.8911998271942139,
+      "learning_rate": 5.943820224719102e-06,
+      "loss": 0.8628,
+      "step": 40
+    },
+    {
+      "epoch": 0.0033158697526361166,
+      "grad_norm": 0.8088529706001282,
+      "learning_rate": 6.955056179775282e-06,
+      "loss": 0.8746,
+      "step": 50
+    },
+    {
+      "epoch": 0.00397904370316334,
+      "grad_norm": 0.8642328977584839,
+      "learning_rate": 7.966292134831462e-06,
+      "loss": 0.8327,
+      "step": 60
+    },
+    {
+      "epoch": 0.004642217653690563,
+      "grad_norm": 0.8503588438034058,
+      "learning_rate": 8.977528089887641e-06,
+      "loss": 0.8412,
+      "step": 70
+    },
+    {
+      "epoch": 0.0053053916042177865,
+      "grad_norm": 0.8862006664276123,
+      "learning_rate": 9.988764044943822e-06,
+      "loss": 0.8231,
+      "step": 80
+    },
+    {
+      "epoch": 0.00596856555474501,
+      "grad_norm": 0.9262124300003052,
+      "learning_rate": 1.1000000000000001e-05,
+      "loss": 0.8087,
+      "step": 90
+    },
+    {
+      "epoch": 0.006631739505272233,
+      "grad_norm": 0.9693242907524109,
+      "learning_rate": 1.2011235955056182e-05,
+      "loss": 0.8197,
+      "step": 100
+    },
+    {
+      "epoch": 0.007294913455799456,
+      "grad_norm": 0.8964329361915588,
+      "learning_rate": 1.3022471910112362e-05,
+      "loss": 0.7927,
+      "step": 110
+    },
+    {
+      "epoch": 0.00795808740632668,
+      "grad_norm": 0.8408939838409424,
+      "learning_rate": 1.403370786516854e-05,
+      "loss": 0.8229,
+      "step": 120
+    },
+    {
+      "epoch": 0.008621261356853902,
+      "grad_norm": 0.8564504981040955,
+      "learning_rate": 1.504494382022472e-05,
+      "loss": 0.7813,
+      "step": 130
+    },
+    {
+      "epoch": 0.009284435307381126,
+      "grad_norm": 0.9394861459732056,
+      "learning_rate": 1.60561797752809e-05,
+      "loss": 0.8178,
+      "step": 140
+    },
+    {
+      "epoch": 0.009947609257908349,
+      "grad_norm": 0.8412289023399353,
+      "learning_rate": 1.706741573033708e-05,
+      "loss": 0.7863,
+      "step": 150
+    },
+    {
+      "epoch": 0.010610783208435573,
+      "grad_norm": 0.8816506266593933,
+      "learning_rate": 1.8078651685393256e-05,
+      "loss": 0.7883,
+      "step": 160
+    },
+    {
+      "epoch": 0.011273957158962795,
+      "grad_norm": 0.8023780584335327,
+      "learning_rate": 1.908988764044944e-05,
+      "loss": 0.7992,
+      "step": 170
+    },
+    {
+      "epoch": 0.01193713110949002,
+      "grad_norm": 0.9203169941902161,
+      "learning_rate": 2e-05,
+      "loss": 0.8106,
+      "step": 180
+    },
+    {
+      "epoch": 0.012600305060017242,
+      "grad_norm": 0.8466951251029968,
+      "learning_rate": 2e-05,
+      "loss": 0.7944,
+      "step": 190
+    },
+    {
+      "epoch": 0.013263479010544466,
+      "grad_norm": 0.807468593120575,
+      "learning_rate": 2e-05,
+      "loss": 0.7964,
+      "step": 200
+    },
+    {
+      "epoch": 0.013926652961071689,
+      "grad_norm": 0.74900221824646,
+      "learning_rate": 2e-05,
+      "loss": 0.7963,
+      "step": 210
+    },
+    {
+      "epoch": 0.014589826911598913,
+      "grad_norm": 0.8038553595542908,
+      "learning_rate": 2e-05,
+      "loss": 0.7799,
+      "step": 220
+    },
+    {
+      "epoch": 0.015253000862126135,
+      "grad_norm": 0.6701945662498474,
+      "learning_rate": 2e-05,
+      "loss": 0.8199,
+      "step": 230
+    },
+    {
+      "epoch": 0.01591617481265336,
+      "grad_norm": 0.649372398853302,
+      "learning_rate": 2e-05,
+      "loss": 0.7925,
+      "step": 240
+    },
+    {
+      "epoch": 0.016579348763180582,
+      "grad_norm": 0.6375704407691956,
+      "learning_rate": 2e-05,
+      "loss": 0.8243,
+      "step": 250
+    },
+    {
+      "epoch": 0.017242522713707804,
+      "grad_norm": 0.660631537437439,
+      "learning_rate": 2e-05,
+      "loss": 0.7956,
+      "step": 260
+    },
+    {
+      "epoch": 0.01790569666423503,
+      "grad_norm": 0.620918869972229,
+      "learning_rate": 2e-05,
+      "loss": 0.7858,
+      "step": 270
+    },
+    {
+      "epoch": 0.018568870614762253,
+      "grad_norm": 0.5827407240867615,
+      "learning_rate": 2e-05,
+      "loss": 0.7998,
+      "step": 280
+    },
+    {
+      "epoch": 0.019232044565289475,
+      "grad_norm": 0.578879177570343,
+      "learning_rate": 2e-05,
+      "loss": 0.7597,
+      "step": 290
+    },
+    {
+      "epoch": 0.019895218515816698,
+      "grad_norm": 0.5654085874557495,
+      "learning_rate": 2e-05,
+      "loss": 0.7533,
+      "step": 300
+    },
+    {
+      "epoch": 0.020558392466343924,
+      "grad_norm": 0.5633082985877991,
+      "learning_rate": 2e-05,
+      "loss": 0.7592,
+      "step": 310
+    },
+    {
+      "epoch": 0.021221566416871146,
+      "grad_norm": 0.6228300333023071,
+      "learning_rate": 2e-05,
+      "loss": 0.7857,
+      "step": 320
+    },
+    {
+      "epoch": 0.02188474036739837,
+      "grad_norm": 0.5642741918563843,
+      "learning_rate": 2e-05,
+      "loss": 0.7652,
+      "step": 330
+    },
+    {
+      "epoch": 0.02254791431792559,
+      "grad_norm": 0.5845516920089722,
+      "learning_rate": 2e-05,
+      "loss": 0.7792,
+      "step": 340
+    },
+    {
+      "epoch": 0.023211088268452817,
+      "grad_norm": 0.5533984899520874,
+      "learning_rate": 2e-05,
+      "loss": 0.7662,
+      "step": 350
+    },
+    {
+      "epoch": 0.02387426221898004,
+      "grad_norm": 1.1515164375305176,
+      "learning_rate": 2e-05,
+      "loss": 0.7853,
+      "step": 360
+    },
+    {
+      "epoch": 0.02453743616950726,
+      "grad_norm": 0.5596652030944824,
+      "learning_rate": 2e-05,
+      "loss": 0.7939,
+      "step": 370
+    },
+    {
+      "epoch": 0.025200610120034484,
+      "grad_norm": 0.5617090463638306,
+      "learning_rate": 2e-05,
+      "loss": 0.7896,
+      "step": 380
+    },
+    {
+      "epoch": 0.02586378407056171,
+      "grad_norm": 0.80217444896698,
+      "learning_rate": 2e-05,
+      "loss": 0.7616,
+      "step": 390
+    },
+    {
+      "epoch": 0.026526958021088932,
+      "grad_norm": 0.8099656701087952,
+      "learning_rate": 2e-05,
+      "loss": 0.7716,
+      "step": 400
+    },
+    {
+      "epoch": 0.027190131971616155,
+      "grad_norm": 0.5415481925010681,
+      "learning_rate": 2e-05,
+      "loss": 0.7764,
+      "step": 410
+    },
+    {
+      "epoch": 0.027853305922143377,
+      "grad_norm": 0.7534944415092468,
+      "learning_rate": 2e-05,
+      "loss": 0.7879,
+      "step": 420
+    },
+    {
+      "epoch": 0.0285164798726706,
+      "grad_norm": 0.602469801902771,
+      "learning_rate": 2e-05,
+      "loss": 0.771,
+      "step": 430
+    },
+    {
+      "epoch": 0.029179653823197826,
+      "grad_norm": 0.5386860370635986,
+      "learning_rate": 2e-05,
+      "loss": 0.791,
+      "step": 440
+    },
+    {
+      "epoch": 0.029842827773725048,
+      "grad_norm": 0.5772424936294556,
+      "learning_rate": 2e-05,
+      "loss": 0.7806,
+      "step": 450
+    },
+    {
+      "epoch": 0.03050600172425227,
+      "grad_norm": 0.5686757564544678,
+      "learning_rate": 2e-05,
+      "loss": 0.7729,
+      "step": 460
+    },
+    {
+      "epoch": 0.031169175674779493,
+      "grad_norm": 0.5157404541969299,
+      "learning_rate": 2e-05,
+      "loss": 0.7387,
+      "step": 470
+    },
+    {
+      "epoch": 0.03183234962530672,
+      "grad_norm": 0.5490089654922485,
+      "learning_rate": 2e-05,
+      "loss": 0.7653,
+      "step": 480
+    },
+    {
+      "epoch": 0.03249552357583394,
+      "grad_norm": 0.5582148432731628,
+      "learning_rate": 2e-05,
+      "loss": 0.7827,
+      "step": 490
+    },
+    {
+      "epoch": 0.033158697526361164,
+      "grad_norm": 0.5726275444030762,
+      "learning_rate": 2e-05,
+      "loss": 0.7942,
+      "step": 500
+    },
+    {
+      "epoch": 0.033821871476888386,
+      "grad_norm": 1.6204034090042114,
+      "learning_rate": 2e-05,
+      "loss": 0.7893,
+      "step": 510
+    },
+    {
+      "epoch": 0.03448504542741561,
+      "grad_norm": 0.5208715796470642,
+      "learning_rate": 2e-05,
+      "loss": 0.7714,
+      "step": 520
+    },
+    {
+      "epoch": 0.03514821937794283,
+      "grad_norm": 0.5310713052749634,
+      "learning_rate": 2e-05,
+      "loss": 0.7802,
+      "step": 530
+    },
+    {
+      "epoch": 0.03581139332847006,
+      "grad_norm": 0.5323279500007629,
+      "learning_rate": 2e-05,
+      "loss": 0.7548,
+      "step": 540
+    },
+    {
+      "epoch": 0.03647456727899728,
+      "grad_norm": 0.48549482226371765,
+      "learning_rate": 2e-05,
+      "loss": 0.7649,
+      "step": 550
+    },
+    {
+      "epoch": 0.037137741229524505,
+      "grad_norm": 0.5288492441177368,
+      "learning_rate": 2e-05,
+      "loss": 0.7777,
+      "step": 560
+    },
+    {
+      "epoch": 0.03780091518005173,
+      "grad_norm": 0.4877089560031891,
+      "learning_rate": 2e-05,
+      "loss": 0.7802,
+      "step": 570
+    },
+    {
+      "epoch": 0.03846408913057895,
+      "grad_norm": 0.5317091345787048,
+      "learning_rate": 2e-05,
+      "loss": 0.7833,
+      "step": 580
+    },
+    {
+      "epoch": 0.03912726308110617,
+      "grad_norm": 0.5533668994903564,
+      "learning_rate": 2e-05,
+      "loss": 0.7702,
+      "step": 590
+    },
+    {
+      "epoch": 0.039790437031633395,
+      "grad_norm": 0.5711390972137451,
+      "learning_rate": 2e-05,
+      "loss": 0.7801,
+      "step": 600
+    },
+    {
+      "epoch": 0.04045361098216062,
+      "grad_norm": 0.5276714563369751,
+      "learning_rate": 2e-05,
+      "loss": 0.7846,
+      "step": 610
+    },
+    {
+      "epoch": 0.04111678493268785,
+      "grad_norm": 0.5505608320236206,
+      "learning_rate": 2e-05,
+      "loss": 0.7575,
+      "step": 620
+    },
+    {
+      "epoch": 0.04177995888321507,
+      "grad_norm": 0.49005311727523804,
+      "learning_rate": 2e-05,
+      "loss": 0.7698,
+      "step": 630
+    },
+    {
+      "epoch": 0.04244313283374229,
+      "grad_norm": 0.4874991178512573,
+      "learning_rate": 2e-05,
+      "loss": 0.7716,
+      "step": 640
+    },
+    {
+      "epoch": 0.043106306784269514,
+      "grad_norm": 0.5269560217857361,
+      "learning_rate": 2e-05,
+      "loss": 0.7607,
+      "step": 650
+    },
+    {
+      "epoch": 0.04376948073479674,
+      "grad_norm": 0.50971519947052,
+      "learning_rate": 2e-05,
+      "loss": 0.7473,
+      "step": 660
+    },
+    {
+      "epoch": 0.04443265468532396,
+      "grad_norm": 0.6813302636146545,
+      "learning_rate": 2e-05,
+      "loss": 0.7437,
+      "step": 670
+    },
+    {
+      "epoch": 0.04509582863585118,
+      "grad_norm": 0.4699794352054596,
+      "learning_rate": 2e-05,
+      "loss": 0.7794,
+      "step": 680
+    },
+    {
+      "epoch": 0.045759002586378404,
+      "grad_norm": 0.4931847155094147,
+      "learning_rate": 2e-05,
+      "loss": 0.7788,
+      "step": 690
+    },
+    {
+      "epoch": 0.046422176536905634,
+      "grad_norm": 0.527800977230072,
+      "learning_rate": 2e-05,
+      "loss": 0.788,
+      "step": 700
+    },
+    {
+      "epoch": 0.047085350487432856,
+      "grad_norm": 0.48107194900512695,
+      "learning_rate": 2e-05,
+      "loss": 0.763,
+      "step": 710
+    },
+    {
+      "epoch": 0.04774852443796008,
+      "grad_norm": 0.48204314708709717,
+      "learning_rate": 2e-05,
+      "loss": 0.7595,
+      "step": 720
+    },
+    {
+      "epoch": 0.0484116983884873,
+      "grad_norm": 0.48960623145103455,
+      "learning_rate": 2e-05,
+      "loss": 0.7647,
+      "step": 730
+    },
+    {
+      "epoch": 0.04907487233901452,
+      "grad_norm": 0.4727330505847931,
+      "learning_rate": 2e-05,
+      "loss": 0.7618,
+      "step": 740
+    },
+    {
+      "epoch": 0.049738046289541746,
+      "grad_norm": 0.48611021041870117,
+      "learning_rate": 2e-05,
+      "loss": 0.7773,
+      "step": 750
+    },
+    {
+      "epoch": 0.05040122024006897,
+      "grad_norm": 0.481209397315979,
+      "learning_rate": 2e-05,
+      "loss": 0.7909,
+      "step": 760
+    },
+    {
+      "epoch": 0.05106439419059619,
+      "grad_norm": 0.5047401785850525,
+      "learning_rate": 2e-05,
+      "loss": 0.7613,
+      "step": 770
+    },
+    {
+      "epoch": 0.05172756814112342,
+      "grad_norm": 0.5307384729385376,
+      "learning_rate": 2e-05,
+      "loss": 0.7672,
+      "step": 780
+    },
+    {
+      "epoch": 0.05239074209165064,
+      "grad_norm": 0.5360110402107239,
+      "learning_rate": 2e-05,
+      "loss": 0.7441,
+      "step": 790
+    },
+    {
+      "epoch": 0.053053916042177865,
+      "grad_norm": 0.4864613711833954,
+      "learning_rate": 2e-05,
+      "loss": 0.7748,
+      "step": 800
+    },
+    {
+      "epoch": 0.05371708999270509,
+      "grad_norm": 0.4954187572002411,
+      "learning_rate": 2e-05,
+      "loss": 0.7528,
+      "step": 810
+    },
+    {
+      "epoch": 0.05438026394323231,
+      "grad_norm": 0.504557192325592,
+      "learning_rate": 2e-05,
+      "loss": 0.7494,
+      "step": 820
+    },
+    {
+      "epoch": 0.05504343789375953,
+      "grad_norm": 0.48328864574432373,
+      "learning_rate": 2e-05,
+      "loss": 0.7764,
+      "step": 830
+    },
+    {
+      "epoch": 0.055706611844286755,
+      "grad_norm": 0.47445255517959595,
+      "learning_rate": 2e-05,
+      "loss": 0.754,
+      "step": 840
+    },
+    {
+      "epoch": 0.05636978579481398,
+      "grad_norm": 0.5029422044754028,
+      "learning_rate": 2e-05,
+      "loss": 0.77,
+      "step": 850
+    },
+    {
+      "epoch": 0.0570329597453412,
+      "grad_norm": 0.5073195695877075,
+      "learning_rate": 2e-05,
+      "loss": 0.7597,
+      "step": 860
+    },
+    {
+      "epoch": 0.05769613369586843,
+      "grad_norm": 0.4880322813987732,
+      "learning_rate": 2e-05,
+      "loss": 0.7595,
+      "step": 870
+    },
+    {
+      "epoch": 0.05835930764639565,
+      "grad_norm": 0.513581395149231,
+      "learning_rate": 2e-05,
+      "loss": 0.7818,
+      "step": 880
+    },
+    {
+      "epoch": 0.059022481596922874,
+      "grad_norm": 0.4951395094394684,
+      "learning_rate": 2e-05,
+      "loss": 0.7881,
+      "step": 890
+    },
+    {
+      "epoch": 0.059685655547450096,
+      "grad_norm": 0.45365604758262634,
+      "learning_rate": 2e-05,
+      "loss": 0.7577,
+      "step": 900
+    },
+    {
+      "epoch": 0.06034882949797732,
+      "grad_norm": 0.4641544818878174,
+      "learning_rate": 2e-05,
+      "loss": 0.7685,
+      "step": 910
+    },
+    {
+      "epoch": 0.06101200344850454,
+      "grad_norm": 0.48189839720726013,
+      "learning_rate": 2e-05,
+      "loss": 0.7767,
+      "step": 920
+    },
+    {
+      "epoch": 0.061675177399031764,
+      "grad_norm": 0.4569750428199768,
+      "learning_rate": 2e-05,
+      "loss": 0.7469,
+      "step": 930
+    },
+    {
+      "epoch": 0.062338351349558986,
+      "grad_norm": 0.4633541703224182,
+      "learning_rate": 2e-05,
+      "loss": 0.76,
+      "step": 940
+    },
+    {
+      "epoch": 0.06300152530008621,
+      "grad_norm": 0.4753730595111847,
+      "learning_rate": 2e-05,
+      "loss": 0.7528,
+      "step": 950
+    },
+    {
+      "epoch": 0.06366469925061344,
+      "grad_norm": 0.4677918553352356,
+      "learning_rate": 2e-05,
+      "loss": 0.7569,
+      "step": 960
+    },
+    {
+      "epoch": 0.06432787320114065,
+      "grad_norm": 0.4754522442817688,
+      "learning_rate": 2e-05,
+      "loss": 0.7659,
+      "step": 970
+    },
+    {
+      "epoch": 0.06499104715166788,
+      "grad_norm": 0.5752569437026978,
+      "learning_rate": 2e-05,
+      "loss": 0.7599,
+      "step": 980
+    },
+    {
+      "epoch": 0.06565422110219511,
+      "grad_norm": 0.5086958408355713,
+      "learning_rate": 2e-05,
+      "loss": 0.7631,
+      "step": 990
+    },
+    {
+      "epoch": 0.06631739505272233,
+      "grad_norm": 0.5891866087913513,
+      "learning_rate": 2e-05,
+      "loss": 0.7597,
+      "step": 1000
+    },
+    {
+      "epoch": 0.06698056900324956,
+      "grad_norm": 0.46219590306282043,
+      "learning_rate": 2e-05,
+      "loss": 0.7579,
+      "step": 1010
+    },
+    {
+      "epoch": 0.06764374295377677,
+      "grad_norm": 0.4759126305580139,
+      "learning_rate": 2e-05,
+      "loss": 0.7554,
+      "step": 1020
+    },
+    {
+      "epoch": 0.068306916904304,
+      "grad_norm": 0.5318647027015686,
+      "learning_rate": 2e-05,
+      "loss": 0.7481,
+      "step": 1030
+    },
+    {
+      "epoch": 0.06897009085483122,
+      "grad_norm": 0.4814603924751282,
+      "learning_rate": 2e-05,
+      "loss": 0.7729,
+      "step": 1040
+    },
+    {
+      "epoch": 0.06963326480535845,
+      "grad_norm": 0.5173375010490417,
+      "learning_rate": 2e-05,
+      "loss": 0.7589,
+      "step": 1050
+    },
+    {
+      "epoch": 0.07029643875588566,
+      "grad_norm": 0.5993748903274536,
+      "learning_rate": 2e-05,
+      "loss": 0.7736,
+      "step": 1060
+    },
+    {
+      "epoch": 0.07095961270641289,
+      "grad_norm": 0.5081777572631836,
+      "learning_rate": 2e-05,
+      "loss": 0.7522,
+      "step": 1070
+    },
+    {
+      "epoch": 0.07162278665694012,
+      "grad_norm": 0.45993325114250183,
+      "learning_rate": 2e-05,
+      "loss": 0.7801,
+      "step": 1080
+    },
+    {
+      "epoch": 0.07228596060746734,
+      "grad_norm": 0.4642050564289093,
+      "learning_rate": 2e-05,
+      "loss": 0.7594,
+      "step": 1090
+    },
+    {
+      "epoch": 0.07294913455799457,
+      "grad_norm": 0.45918765664100647,
+      "learning_rate": 2e-05,
+      "loss": 0.7292,
+      "step": 1100
+    },
+    {
+      "epoch": 0.07361230850852178,
+      "grad_norm": 0.4907342493534088,
+      "learning_rate": 2e-05,
+      "loss": 0.768,
+      "step": 1110
+    },
+    {
+      "epoch": 0.07427548245904901,
+      "grad_norm": 0.46477046608924866,
+      "learning_rate": 2e-05,
+      "loss": 0.7546,
+      "step": 1120
+    },
+    {
+      "epoch": 0.07493865640957623,
+      "grad_norm": 0.5048196911811829,
+      "learning_rate": 2e-05,
+      "loss": 0.7762,
+      "step": 1130
+    },
+    {
+      "epoch": 0.07560183036010346,
+      "grad_norm": 0.48439109325408936,
+      "learning_rate": 2e-05,
+      "loss": 0.7385,
+      "step": 1140
+    },
+    {
+      "epoch": 0.07626500431063069,
+      "grad_norm": 0.47210803627967834,
+      "learning_rate": 2e-05,
+      "loss": 0.7694,
+      "step": 1150
+    },
+    {
+      "epoch": 0.0769281782611579,
+      "grad_norm": 0.4778144359588623,
+      "learning_rate": 2e-05,
+      "loss": 0.7382,
+      "step": 1160
+    },
+    {
+      "epoch": 0.07759135221168513,
+      "grad_norm": 0.5094350576400757,
+      "learning_rate": 2e-05,
+      "loss": 0.7488,
+      "step": 1170
+    },
+    {
+      "epoch": 0.07825452616221235,
+      "grad_norm": 0.5327107310295105,
+      "learning_rate": 2e-05,
+      "loss": 0.7457,
+      "step": 1180
+    },
+    {
+      "epoch": 0.07891770011273957,
+      "grad_norm": 0.4854157567024231,
+      "learning_rate": 2e-05,
+      "loss": 0.7531,
+      "step": 1190
+    },
+    {
+      "epoch": 0.07958087406326679,
+      "grad_norm": 0.8898469805717468,
+      "learning_rate": 2e-05,
+      "loss": 0.7552,
+      "step": 1200
+    },
+    {
+      "epoch": 0.08024404801379402,
+      "grad_norm": 0.4582747519016266,
+      "learning_rate": 2e-05,
+      "loss": 0.7346,
+      "step": 1210
+    },
+    {
+      "epoch": 0.08090722196432124,
+      "grad_norm": 0.4968053102493286,
+      "learning_rate": 2e-05,
+      "loss": 0.7464,
+      "step": 1220
+    },
+    {
+      "epoch": 0.08157039591484846,
+      "grad_norm": 0.5363363027572632,
+      "learning_rate": 2e-05,
+      "loss": 0.7607,
+      "step": 1230
+    },
+    {
+      "epoch": 0.0822335698653757,
+      "grad_norm": 0.5088996887207031,
+      "learning_rate": 2e-05,
+      "loss": 0.7482,
+      "step": 1240
+    },
+    {
+      "epoch": 0.08289674381590291,
+      "grad_norm": 0.4793884754180908,
+      "learning_rate": 2e-05,
+      "loss": 0.7469,
+      "step": 1250
+    },
+    {
+      "epoch": 0.08355991776643014,
+      "grad_norm": 1.361615538597107,
+      "learning_rate": 2e-05,
+      "loss": 0.7451,
+      "step": 1260
+    },
+    {
+      "epoch": 0.08422309171695735,
+      "grad_norm": 0.47477322816848755,
+      "learning_rate": 2e-05,
+      "loss": 0.7631,
+      "step": 1270
+    },
+    {
+      "epoch": 0.08488626566748458,
+      "grad_norm": 0.49219903349876404,
+      "learning_rate": 2e-05,
+      "loss": 0.7709,
+      "step": 1280
+    },
+    {
+      "epoch": 0.0855494396180118,
+      "grad_norm": 0.46074602007865906,
+      "learning_rate": 2e-05,
+      "loss": 0.75,
+      "step": 1290
+    },
+    {
+      "epoch": 0.08621261356853903,
+      "grad_norm": 0.4657946228981018,
+      "learning_rate": 2e-05,
+      "loss": 0.7494,
+      "step": 1300
+    },
+    {
+      "epoch": 0.08687578751906624,
+      "grad_norm": 1.8625658750534058,
+      "learning_rate": 2e-05,
+      "loss": 0.7213,
+      "step": 1310
+    },
+    {
+      "epoch": 0.08753896146959347,
+      "grad_norm": 0.4719648063182831,
+      "learning_rate": 2e-05,
+      "loss": 0.7651,
+      "step": 1320
+    },
+    {
+      "epoch": 0.0882021354201207,
+      "grad_norm": 0.45125117897987366,
+      "learning_rate": 2e-05,
+      "loss": 0.7575,
+      "step": 1330
+    },
+    {
+      "epoch": 0.08886530937064792,
+      "grad_norm": 0.4656333029270172,
+      "learning_rate": 2e-05,
+      "loss": 0.7473,
+      "step": 1340
+    },
+    {
+      "epoch": 0.08952848332117515,
+      "grad_norm": 0.4801299273967743,
+      "learning_rate": 2e-05,
+      "loss": 0.7496,
+      "step": 1350
+    },
+    {
+      "epoch": 0.09019165727170236,
+      "grad_norm": 0.4366644620895386,
+      "learning_rate": 2e-05,
+      "loss": 0.7361,
+      "step": 1360
+    },
+    {
+      "epoch": 0.09085483122222959,
+      "grad_norm": 0.43298009037971497,
+      "learning_rate": 2e-05,
+      "loss": 0.7358,
+      "step": 1370
+    },
+    {
+      "epoch": 0.09151800517275681,
+      "grad_norm": 0.45950010418891907,
+      "learning_rate": 2e-05,
+      "loss": 0.7431,
+      "step": 1380
+    },
+    {
+      "epoch": 0.09218117912328404,
+      "grad_norm": 0.4419451057910919,
+      "learning_rate": 2e-05,
+      "loss": 0.7559,
+      "step": 1390
+    },
+    {
+      "epoch": 0.09284435307381127,
+      "grad_norm": 0.5011473894119263,
+      "learning_rate": 2e-05,
+      "loss": 0.747,
+      "step": 1400
+    },
+    {
+      "epoch": 0.09350752702433848,
+      "grad_norm": 0.4722769260406494,
+      "learning_rate": 2e-05,
+      "loss": 0.7598,
+      "step": 1410
+    },
+    {
+      "epoch": 0.09417070097486571,
+      "grad_norm": 0.4498542547225952,
+      "learning_rate": 2e-05,
+      "loss": 0.7429,
+      "step": 1420
+    },
+    {
+      "epoch": 0.09483387492539293,
+      "grad_norm": 0.464199960231781,
+      "learning_rate": 2e-05,
+      "loss": 0.7684,
+      "step": 1430
+    },
+    {
+      "epoch": 0.09549704887592016,
+      "grad_norm": 0.5018302202224731,
+      "learning_rate": 2e-05,
+      "loss": 0.7698,
+      "step": 1440
+    },
+    {
+      "epoch": 0.09616022282644737,
+      "grad_norm": 0.4524178206920624,
+      "learning_rate": 2e-05,
+      "loss": 0.7644,
+      "step": 1450
+    },
+    {
+      "epoch": 0.0968233967769746,
+      "grad_norm": 0.46702370047569275,
+      "learning_rate": 2e-05,
+      "loss": 0.7526,
+      "step": 1460
+    },
+    {
+      "epoch": 0.09748657072750182,
+      "grad_norm": 0.4664076864719391,
+      "learning_rate": 2e-05,
+      "loss": 0.7638,
+      "step": 1470
+    },
+    {
+      "epoch": 0.09814974467802905,
+      "grad_norm": 0.48490601778030396,
+      "learning_rate": 2e-05,
+      "loss": 0.7535,
+      "step": 1480
+    },
+    {
+      "epoch": 0.09881291862855628,
+      "grad_norm": 0.5127236247062683,
+      "learning_rate": 2e-05,
+      "loss": 0.7365,
+      "step": 1490
+    },
+    {
+      "epoch": 0.09947609257908349,
+      "grad_norm": 0.46068698167800903,
+      "learning_rate": 2e-05,
+      "loss": 0.735,
+      "step": 1500
+    },
+    {
+      "epoch": 0.10013926652961072,
+      "grad_norm": 0.44757401943206787,
+      "learning_rate": 2e-05,
+      "loss": 0.7587,
+      "step": 1510
+    },
+    {
+      "epoch": 0.10080244048013794,
+      "grad_norm": 0.475946843624115,
+      "learning_rate": 2e-05,
+      "loss": 0.7474,
+      "step": 1520
+    },
+    {
+      "epoch": 0.10146561443066517,
+      "grad_norm": 0.4839072823524475,
+      "learning_rate": 2e-05,
+      "loss": 0.75,
+      "step": 1530
+    },
+    {
+      "epoch": 0.10212878838119238,
+      "grad_norm": 0.477365106344223,
+      "learning_rate": 2e-05,
+      "loss": 0.7487,
+      "step": 1540
+    },
+    {
+      "epoch": 0.10279196233171961,
+      "grad_norm": 0.48756539821624756,
+      "learning_rate": 2e-05,
+      "loss": 0.7446,
+      "step": 1550
+    },
+    {
+      "epoch": 0.10345513628224684,
+      "grad_norm": 0.46172550320625305,
+      "learning_rate": 2e-05,
+      "loss": 0.7336,
+      "step": 1560
+    },
+    {
+      "epoch": 0.10411831023277406,
+      "grad_norm": 0.4753364026546478,
+      "learning_rate": 2e-05,
+      "loss": 0.7267,
+      "step": 1570
+    },
+    {
+      "epoch": 0.10478148418330128,
+      "grad_norm": 0.4389295279979706,
+      "learning_rate": 2e-05,
+      "loss": 0.7494,
+      "step": 1580
+    },
+    {
+      "epoch": 0.1054446581338285,
+      "grad_norm": 0.44949832558631897,
+      "learning_rate": 2e-05,
+      "loss": 0.7334,
+      "step": 1590
+    },
+    {
+      "epoch": 0.10610783208435573,
+      "grad_norm": 0.46615809202194214,
+      "learning_rate": 2e-05,
+      "loss": 0.7669,
+      "step": 1600
+    },
+    {
+      "epoch": 0.10677100603488295,
+      "grad_norm": 0.4528898000717163,
+      "learning_rate": 2e-05,
+      "loss": 0.7503,
+      "step": 1610
+    },
+    {
+      "epoch": 0.10743417998541017,
+      "grad_norm": 0.571483314037323,
+      "learning_rate": 2e-05,
+      "loss": 0.7383,
+      "step": 1620
+    },
+    {
+      "epoch": 0.10809735393593739,
+      "grad_norm": 0.4646868109703064,
+      "learning_rate": 2e-05,
+      "loss": 0.7512,
+      "step": 1630
+    },
+    {
+      "epoch": 0.10876052788646462,
+      "grad_norm": 0.4529063105583191,
+      "learning_rate": 2e-05,
+      "loss": 0.7255,
+      "step": 1640
+    },
+    {
+      "epoch": 0.10942370183699185,
+      "grad_norm": 0.5024081468582153,
+      "learning_rate": 2e-05,
+      "loss": 0.739,
+      "step": 1650
+    },
+    {
+      "epoch": 0.11008687578751906,
+      "grad_norm": 0.4648872911930084,
+      "learning_rate": 2e-05,
+      "loss": 0.746,
+      "step": 1660
+    },
+    {
+      "epoch": 0.1107500497380463,
+      "grad_norm": 0.488041490316391,
+      "learning_rate": 2e-05,
+      "loss": 0.738,
+      "step": 1670
+    },
+    {
+      "epoch": 0.11141322368857351,
+      "grad_norm": 0.4653259515762329,
+      "learning_rate": 2e-05,
+      "loss": 0.7256,
+      "step": 1680
+    },
+    {
+      "epoch": 0.11207639763910074,
+      "grad_norm": 0.48732203245162964,
+      "learning_rate": 2e-05,
+      "loss": 0.7429,
+      "step": 1690
+    },
+    {
+      "epoch": 0.11273957158962795,
+      "grad_norm": 0.572994589805603,
+      "learning_rate": 2e-05,
+      "loss": 0.7573,
+      "step": 1700
+    },
+    {
+      "epoch": 0.11340274554015518,
+      "grad_norm": 0.4531424045562744,
+      "learning_rate": 2e-05,
+      "loss": 0.7422,
+      "step": 1710
+    },
+    {
+      "epoch": 0.1140659194906824,
+      "grad_norm": 0.512623131275177,
+      "learning_rate": 2e-05,
+      "loss": 0.7292,
+      "step": 1720
+    },
+    {
+      "epoch": 0.11472909344120963,
+      "grad_norm": 0.4822714328765869,
+      "learning_rate": 2e-05,
+      "loss": 0.7415,
+      "step": 1730
+    },
+    {
+      "epoch": 0.11539226739173686,
+      "grad_norm": 0.4871460497379303,
+      "learning_rate": 2e-05,
+      "loss": 0.7417,
+      "step": 1740
+    },
+    {
+      "epoch": 0.11605544134226407,
+      "grad_norm": 1.012222409248352,
+      "learning_rate": 2e-05,
+      "loss": 0.736,
+      "step": 1750
+    },
+    {
+      "epoch": 0.1167186152927913,
+      "grad_norm": 0.530349612236023,
+      "learning_rate": 2e-05,
+      "loss": 0.728,
+      "step": 1760
+    },
+    {
+      "epoch": 0.11738178924331852,
+      "grad_norm": 0.46132221817970276,
+      "learning_rate": 2e-05,
+      "loss": 0.7216,
+      "step": 1770
+    },
+    {
+      "epoch": 0.11804496319384575,
+      "grad_norm": 0.4651985168457031,
+      "learning_rate": 2e-05,
+      "loss": 0.7562,
+      "step": 1780
+    },
+    {
+      "epoch": 0.11870813714437296,
+      "grad_norm": 0.44336026906967163,
+      "learning_rate": 2e-05,
+      "loss": 0.7441,
+      "step": 1790
+    },
+    {
+      "epoch": 0.11937131109490019,
+      "grad_norm": 0.4634976089000702,
+      "learning_rate": 2e-05,
+      "loss": 0.7746,
+      "step": 1800
+    },
+    {
+      "epoch": 0.12003448504542742,
+      "grad_norm": 0.6048588752746582,
+      "learning_rate": 2e-05,
+      "loss": 0.7485,
+      "step": 1810
+    },
+    {
+      "epoch": 0.12069765899595464,
+      "grad_norm": 0.47721830010414124,
+      "learning_rate": 2e-05,
+      "loss": 0.7484,
+      "step": 1820
+    },
+    {
+      "epoch": 0.12136083294648187,
+      "grad_norm": 0.464031457901001,
+      "learning_rate": 2e-05,
+      "loss": 0.7176,
+      "step": 1830
+    },
+    {
+      "epoch": 0.12202400689700908,
+      "grad_norm": 0.4706140458583832,
+      "learning_rate": 2e-05,
+      "loss": 0.7097,
+      "step": 1840
+    },
+    {
+      "epoch": 0.12268718084753631,
+      "grad_norm": 0.4383341670036316,
+      "learning_rate": 2e-05,
+      "loss": 0.7444,
+      "step": 1850
+    },
+    {
+      "epoch": 0.12335035479806353,
+      "grad_norm": 0.4670451581478119,
+      "learning_rate": 2e-05,
+      "loss": 0.7596,
+      "step": 1860
+    },
+    {
+      "epoch": 0.12401352874859076,
+      "grad_norm": 0.4673280417919159,
+      "learning_rate": 2e-05,
+      "loss": 0.7292,
+      "step": 1870
+    },
+    {
+      "epoch": 0.12467670269911797,
+      "grad_norm": 0.49208033084869385,
+      "learning_rate": 2e-05,
+      "loss": 0.7655,
+      "step": 1880
+    },
+    {
+      "epoch": 0.12533987664964522,
+      "grad_norm": 0.44742557406425476,
+      "learning_rate": 2e-05,
+      "loss": 0.7431,
+      "step": 1890
+    },
+    {
+      "epoch": 0.12600305060017242,
+      "grad_norm": 0.47594326734542847,
+      "learning_rate": 2e-05,
+      "loss": 0.7524,
+      "step": 1900
+    },
+    {
+      "epoch": 0.12666622455069965,
+      "grad_norm": 0.49186161160469055,
+      "learning_rate": 2e-05,
+      "loss": 0.7371,
+      "step": 1910
+    },
+    {
+      "epoch": 0.12732939850122688,
+      "grad_norm": 0.455226331949234,
+      "learning_rate": 2e-05,
+      "loss": 0.7535,
+      "step": 1920
+    },
+    {
+      "epoch": 0.1279925724517541,
+      "grad_norm": 0.49632561206817627,
+      "learning_rate": 2e-05,
+      "loss": 0.7362,
+      "step": 1930
+    },
+    {
+      "epoch": 0.1286557464022813,
+      "grad_norm": 0.4459867477416992,
+      "learning_rate": 2e-05,
+      "loss": 0.7312,
+      "step": 1940
+    },
+    {
+      "epoch": 0.12931892035280854,
+      "grad_norm": 0.48936307430267334,
+      "learning_rate": 2e-05,
+      "loss": 0.7296,
+      "step": 1950
+    },
+    {
+      "epoch": 0.12998209430333577,
+      "grad_norm": 0.46001771092414856,
+      "learning_rate": 2e-05,
+      "loss": 0.7452,
+      "step": 1960
+    },
+    {
+      "epoch": 0.130645268253863,
+      "grad_norm": 0.4681333601474762,
+      "learning_rate": 2e-05,
+      "loss": 0.73,
+      "step": 1970
+    },
+    {
+      "epoch": 0.13130844220439022,
+      "grad_norm": 0.476224422454834,
+      "learning_rate": 2e-05,
+      "loss": 0.7481,
+      "step": 1980
+    },
+    {
+      "epoch": 0.13197161615491743,
+      "grad_norm": 0.45250940322875977,
+      "learning_rate": 2e-05,
+      "loss": 0.7424,
+      "step": 1990
+    },
+    {
+      "epoch": 0.13263479010544466,
+      "grad_norm": 0.49830418825149536,
+      "learning_rate": 2e-05,
+      "loss": 0.7553,
+      "step": 2000
+    },
+    {
+      "epoch": 0.13329796405597188,
+      "grad_norm": 0.4795004725456238,
+      "learning_rate": 2e-05,
+      "loss": 0.7209,
+      "step": 2010
+    },
+    {
+      "epoch": 0.13396113800649911,
+      "grad_norm": 0.805539608001709,
+      "learning_rate": 2e-05,
+      "loss": 0.7467,
+      "step": 2020
+    },
+    {
+      "epoch": 0.13462431195702632,
+      "grad_norm": 0.6310796737670898,
+      "learning_rate": 2e-05,
+      "loss": 0.7566,
+      "step": 2030
+    },
+    {
+      "epoch": 0.13528748590755355,
+      "grad_norm": 0.43787965178489685,
+      "learning_rate": 2e-05,
+      "loss": 0.7198,
+      "step": 2040
+    },
+    {
+      "epoch": 0.13595065985808077,
+      "grad_norm": 0.6050550937652588,
+      "learning_rate": 2e-05,
+      "loss": 0.7439,
+      "step": 2050
+    },
+    {
+      "epoch": 0.136613833808608,
+      "grad_norm": 0.4821194112300873,
+      "learning_rate": 2e-05,
+      "loss": 0.7681,
+      "step": 2060
+    },
+    {
+      "epoch": 0.13727700775913523,
+      "grad_norm": 0.46896836161613464,
+      "learning_rate": 2e-05,
+      "loss": 0.7288,
+      "step": 2070
+    },
+    {
+      "epoch": 0.13794018170966243,
+      "grad_norm": 0.4994990825653076,
+      "learning_rate": 2e-05,
+      "loss": 0.7446,
+      "step": 2080
+    },
+    {
+      "epoch": 0.13860335566018966,
+      "grad_norm": 0.4725533127784729,
+      "learning_rate": 2e-05,
+      "loss": 0.7301,
+      "step": 2090
+    },
+    {
+      "epoch": 0.1392665296107169,
+      "grad_norm": 0.4225281774997711,
+      "learning_rate": 2e-05,
+      "loss": 0.7342,
+      "step": 2100
+    },
+    {
+      "epoch": 0.13992970356124412,
+      "grad_norm": 0.4997630715370178,
+      "learning_rate": 2e-05,
+      "loss": 0.7455,
+      "step": 2110
+    },
+    {
+      "epoch": 0.14059287751177132,
+      "grad_norm": 0.4552975594997406,
+      "learning_rate": 2e-05,
+      "loss": 0.75,
+      "step": 2120
+    },
+    {
+      "epoch": 0.14125605146229855,
+      "grad_norm": 0.4498264491558075,
+      "learning_rate": 2e-05,
+      "loss": 0.7462,
+      "step": 2130
+    },
+    {
+      "epoch": 0.14191922541282578,
+      "grad_norm": 0.49117112159729004,
+      "learning_rate": 2e-05,
+      "loss": 0.7541,
+      "step": 2140
+    },
+    {
+      "epoch": 0.142582399363353,
+      "grad_norm": 0.49303096532821655,
+      "learning_rate": 2e-05,
+      "loss": 0.7301,
+      "step": 2150
+    },
+    {
+      "epoch": 0.14324557331388024,
+      "grad_norm": 0.45166993141174316,
+      "learning_rate": 2e-05,
+      "loss": 0.7597,
+      "step": 2160
+    },
+    {
+      "epoch": 0.14390874726440744,
+      "grad_norm": 0.468921035528183,
+      "learning_rate": 2e-05,
+      "loss": 0.7409,
+      "step": 2170
+    },
+    {
+      "epoch": 0.14457192121493467,
+      "grad_norm": 0.46468544006347656,
+      "learning_rate": 2e-05,
+      "loss": 0.731,
+      "step": 2180
+    },
+    {
+      "epoch": 0.1452350951654619,
+      "grad_norm": 0.4980156719684601,
+      "learning_rate": 2e-05,
+      "loss": 0.7343,
+      "step": 2190
+    },
+    {
+      "epoch": 0.14589826911598913,
+      "grad_norm": 0.4513563811779022,
+      "learning_rate": 2e-05,
+      "loss": 0.7248,
+      "step": 2200
+    },
+    {
+      "epoch": 0.14656144306651633,
+      "grad_norm": 0.43859392404556274,
+      "learning_rate": 2e-05,
+      "loss": 0.7293,
+      "step": 2210
+    },
+    {
+      "epoch": 0.14722461701704356,
+      "grad_norm": 0.4643859565258026,
+      "learning_rate": 2e-05,
+      "loss": 0.7475,
+      "step": 2220
+    },
+    {
+      "epoch": 0.1478877909675708,
+      "grad_norm": 0.5044799447059631,
+      "learning_rate": 2e-05,
+      "loss": 0.7288,
+      "step": 2230
+    },
+    {
+      "epoch": 0.14855096491809802,
+      "grad_norm": 0.5265306234359741,
+      "learning_rate": 2e-05,
+      "loss": 0.718,
+      "step": 2240
+    },
+    {
+      "epoch": 0.14921413886862525,
+      "grad_norm": 0.4421214461326599,
+      "learning_rate": 2e-05,
+      "loss": 0.7334,
+      "step": 2250
+    },
+    {
+      "epoch": 0.14987731281915245,
+      "grad_norm": 0.46880438923835754,
+      "learning_rate": 2e-05,
+      "loss": 0.7198,
+      "step": 2260
+    },
+    {
+      "epoch": 0.15054048676967968,
+      "grad_norm": 0.5459731817245483,
+      "learning_rate": 2e-05,
+      "loss": 0.7662,
+      "step": 2270
+    },
+    {
+      "epoch": 0.1512036607202069,
+      "grad_norm": 0.47586730122566223,
+      "learning_rate": 2e-05,
+      "loss": 0.7404,
+      "step": 2280
+    },
+    {
+      "epoch": 0.15186683467073414,
+      "grad_norm": 0.4461500346660614,
+      "learning_rate": 2e-05,
+      "loss": 0.7448,
+      "step": 2290
+    },
+    {
+      "epoch": 0.15253000862126137,
+      "grad_norm": 0.4607698917388916,
+      "learning_rate": 2e-05,
+      "loss": 0.7544,
+      "step": 2300
+    },
+    {
+      "epoch": 0.15319318257178857,
+      "grad_norm": 0.44479936361312866,
+      "learning_rate": 2e-05,
+      "loss": 0.7432,
+      "step": 2310
+    },
+    {
+      "epoch": 0.1538563565223158,
+      "grad_norm": 0.4355045557022095,
+      "learning_rate": 2e-05,
+      "loss": 0.7222,
+      "step": 2320
+    },
+    {
+      "epoch": 0.15451953047284303,
+      "grad_norm": 0.43298035860061646,
+      "learning_rate": 2e-05,
+      "loss": 0.7492,
+      "step": 2330
+    },
+    {
+      "epoch": 0.15518270442337026,
+      "grad_norm": 0.42725154757499695,
+      "learning_rate": 2e-05,
+      "loss": 0.7576,
+      "step": 2340
+    },
+    {
+      "epoch": 0.15584587837389746,
+      "grad_norm": 0.4170139729976654,
+      "learning_rate": 2e-05,
+      "loss": 0.7321,
+      "step": 2350
+    },
+    {
+      "epoch": 0.1565090523244247,
+      "grad_norm": 0.48317450284957886,
+      "learning_rate": 2e-05,
+      "loss": 0.7337,
+      "step": 2360
+    },
+    {
+      "epoch": 0.15717222627495192,
+      "grad_norm": 0.47873780131340027,
+      "learning_rate": 2e-05,
+      "loss": 0.7572,
+      "step": 2370
+    },
+    {
+      "epoch": 0.15783540022547915,
+      "grad_norm": 0.47494009137153625,
+      "learning_rate": 2e-05,
+      "loss": 0.7197,
+      "step": 2380
+    },
+    {
+      "epoch": 0.15849857417600638,
+      "grad_norm": 0.45792391896247864,
+      "learning_rate": 2e-05,
+      "loss": 0.7682,
+      "step": 2390
+    },
+    {
+      "epoch": 0.15916174812653358,
+      "grad_norm": 0.4486759603023529,
+      "learning_rate": 2e-05,
+      "loss": 0.7336,
+      "step": 2400
+    },
+    {
+      "epoch": 0.1598249220770608,
+      "grad_norm": 0.47603151202201843,
+      "learning_rate": 2e-05,
+      "loss": 0.737,
+      "step": 2410
+    },
+    {
+      "epoch": 0.16048809602758804,
+      "grad_norm": 0.45742881298065186,
+      "learning_rate": 2e-05,
+      "loss": 0.7217,
+      "step": 2420
+    },
+    {
+      "epoch": 0.16115126997811527,
+      "grad_norm": 0.4568350911140442,
+      "learning_rate": 2e-05,
+      "loss": 0.7698,
+      "step": 2430
+    },
+    {
+      "epoch": 0.16181444392864247,
+      "grad_norm": 0.4493866562843323,
+      "learning_rate": 2e-05,
+      "loss": 0.7422,
+      "step": 2440
+    },
+    {
+      "epoch": 0.1624776178791697,
+      "grad_norm": 0.6067711114883423,
+      "learning_rate": 2e-05,
+      "loss": 0.7471,
+      "step": 2450
+    },
+    {
+      "epoch": 0.16314079182969693,
+      "grad_norm": 0.4765845239162445,
+      "learning_rate": 2e-05,
+      "loss": 0.7413,
+      "step": 2460
+    },
+    {
+      "epoch": 0.16380396578022416,
+      "grad_norm": 1.5535005331039429,
+      "learning_rate": 2e-05,
+      "loss": 0.7536,
+      "step": 2470
+    },
+    {
+      "epoch": 0.1644671397307514,
+      "grad_norm": 0.47922787070274353,
+      "learning_rate": 2e-05,
+      "loss": 0.7366,
+      "step": 2480
+    },
+    {
+      "epoch": 0.1651303136812786,
+      "grad_norm": 0.4495774805545807,
+      "learning_rate": 2e-05,
+      "loss": 0.7182,
+      "step": 2490
+    },
+    {
+      "epoch": 0.16579348763180582,
+      "grad_norm": 0.556162416934967,
+      "learning_rate": 2e-05,
+      "loss": 0.7367,
+      "step": 2500
+    },
+    {
+      "epoch": 0.16645666158233305,
+      "grad_norm": 0.4780755043029785,
+      "learning_rate": 2e-05,
+      "loss": 0.7219,
+      "step": 2510
+    },
+    {
+      "epoch": 0.16711983553286028,
+      "grad_norm": 0.4964953660964966,
+      "learning_rate": 2e-05,
+      "loss": 0.7197,
+      "step": 2520
+    },
+    {
+      "epoch": 0.16778300948338748,
+      "grad_norm": 0.44503602385520935,
+      "learning_rate": 2e-05,
+      "loss": 0.7241,
+      "step": 2530
+    },
+    {
+      "epoch": 0.1684461834339147,
+      "grad_norm": 0.4624878168106079,
+      "learning_rate": 2e-05,
+      "loss": 0.7477,
+      "step": 2540
+    },
+    {
+      "epoch": 0.16910935738444194,
+      "grad_norm": 0.49463391304016113,
+      "learning_rate": 2e-05,
+      "loss": 0.732,
+      "step": 2550
+    },
+    {
+      "epoch": 0.16977253133496917,
+      "grad_norm": 0.4355657696723938,
+      "learning_rate": 2e-05,
+      "loss": 0.7229,
+      "step": 2560
+    },
+    {
+      "epoch": 0.1704357052854964,
+      "grad_norm": 0.43896934390068054,
+      "learning_rate": 2e-05,
+      "loss": 0.7187,
+      "step": 2570
+    },
+    {
+      "epoch": 0.1710988792360236,
+      "grad_norm": 0.4886817932128906,
+      "learning_rate": 2e-05,
+      "loss": 0.7052,
+      "step": 2580
+    },
+    {
+      "epoch": 0.17176205318655083,
+      "grad_norm": 0.41980621218681335,
+      "learning_rate": 2e-05,
+      "loss": 0.7442,
+      "step": 2590
+    },
+    {
+      "epoch": 0.17242522713707806,
+      "grad_norm": 0.4909554719924927,
+      "learning_rate": 2e-05,
+      "loss": 0.7359,
+      "step": 2600
+    },
+    {
+      "epoch": 0.1730884010876053,
+      "grad_norm": 0.4882756173610687,
+      "learning_rate": 2e-05,
+      "loss": 0.722,
+      "step": 2610
+    },
+    {
+      "epoch": 0.1737515750381325,
+      "grad_norm": 0.6686330437660217,
+      "learning_rate": 2e-05,
+      "loss": 0.7238,
+      "step": 2620
+    },
+    {
+      "epoch": 0.17441474898865972,
+      "grad_norm": 0.5391948819160461,
+      "learning_rate": 2e-05,
+      "loss": 0.7068,
+      "step": 2630
+    },
+    {
+      "epoch": 0.17507792293918695,
+      "grad_norm": 0.48737525939941406,
+      "learning_rate": 2e-05,
+      "loss": 0.73,
+      "step": 2640
+    },
+    {
+      "epoch": 0.17574109688971418,
+      "grad_norm": 0.45221763849258423,
+      "learning_rate": 2e-05,
+      "loss": 0.7374,
+      "step": 2650
+    },
+    {
+      "epoch": 0.1764042708402414,
+      "grad_norm": 0.5481429696083069,
+      "learning_rate": 2e-05,
+      "loss": 0.749,
+      "step": 2660
+    },
+    {
+      "epoch": 0.1770674447907686,
+      "grad_norm": 0.4444698691368103,
+      "learning_rate": 2e-05,
+      "loss": 0.722,
+      "step": 2670
+    },
+    {
+      "epoch": 0.17773061874129584,
+      "grad_norm": 0.4412497580051422,
+      "learning_rate": 2e-05,
+      "loss": 0.7407,
+      "step": 2680
+    },
+    {
+      "epoch": 0.17839379269182307,
+      "grad_norm": 0.47715237736701965,
+      "learning_rate": 2e-05,
+      "loss": 0.7104,
+      "step": 2690
+    },
+    {
+      "epoch": 0.1790569666423503,
+      "grad_norm": 0.45212236046791077,
+      "learning_rate": 2e-05,
+      "loss": 0.7091,
+      "step": 2700
+    },
+    {
+      "epoch": 0.17972014059287753,
+      "grad_norm": 0.4850757420063019,
+      "learning_rate": 2e-05,
+      "loss": 0.7348,
+      "step": 2710
+    },
+    {
+      "epoch": 0.18038331454340473,
+      "grad_norm": 0.4519940912723541,
+      "learning_rate": 2e-05,
+      "loss": 0.7019,
+      "step": 2720
+    },
+    {
+      "epoch": 0.18104648849393196,
+      "grad_norm": 0.46579867601394653,
+      "learning_rate": 2e-05,
+      "loss": 0.7382,
+      "step": 2730
+    },
+    {
+      "epoch": 0.18170966244445919,
+      "grad_norm": 0.46576786041259766,
+      "learning_rate": 2e-05,
+      "loss": 0.7478,
+      "step": 2740
+    },
+    {
+      "epoch": 0.18237283639498642,
+      "grad_norm": 0.47330838441848755,
+      "learning_rate": 2e-05,
+      "loss": 0.7189,
+      "step": 2750
+    },
+    {
+      "epoch": 0.18303601034551362,
+      "grad_norm": 0.4114646315574646,
+      "learning_rate": 2e-05,
+      "loss": 0.7129,
+      "step": 2760
+    },
+    {
+      "epoch": 0.18369918429604085,
+      "grad_norm": 0.43425285816192627,
+      "learning_rate": 2e-05,
+      "loss": 0.7216,
+      "step": 2770
+    },
+    {
+      "epoch": 0.18436235824656808,
+      "grad_norm": 0.43742886185646057,
+      "learning_rate": 2e-05,
+      "loss": 0.7447,
+      "step": 2780
+    },
+    {
+      "epoch": 0.1850255321970953,
+      "grad_norm": 0.4592493176460266,
+      "learning_rate": 2e-05,
+      "loss": 0.7266,
+      "step": 2790
+    },
+    {
+      "epoch": 0.18568870614762253,
+      "grad_norm": 0.4362889528274536,
+      "learning_rate": 2e-05,
+      "loss": 0.7243,
+      "step": 2800
+    },
+    {
+      "epoch": 0.18635188009814974,
+      "grad_norm": 0.4235585629940033,
+      "learning_rate": 2e-05,
+      "loss": 0.7247,
+      "step": 2810
+    },
+    {
+      "epoch": 0.18701505404867697,
+      "grad_norm": 0.41314443945884705,
+      "learning_rate": 2e-05,
+      "loss": 0.7169,
+      "step": 2820
+    },
+    {
+      "epoch": 0.1876782279992042,
+      "grad_norm": 0.43876609206199646,
+      "learning_rate": 2e-05,
+      "loss": 0.7344,
+      "step": 2830
+    },
+    {
+      "epoch": 0.18834140194973142,
+      "grad_norm": 0.4480982720851898,
+      "learning_rate": 2e-05,
+      "loss": 0.7357,
+      "step": 2840
+    },
+    {
+      "epoch": 0.18900457590025863,
+      "grad_norm": 0.5882231593132019,
+      "learning_rate": 2e-05,
+      "loss": 0.7164,
+      "step": 2850
+    },
+    {
+      "epoch": 0.18966774985078586,
+      "grad_norm": 0.4759451746940613,
+      "learning_rate": 2e-05,
+      "loss": 0.7087,
+      "step": 2860
+    },
+    {
+      "epoch": 0.19033092380131308,
+      "grad_norm": 0.5012674331665039,
+      "learning_rate": 2e-05,
+      "loss": 0.723,
+      "step": 2870
+    },
+    {
+      "epoch": 0.1909940977518403,
+      "grad_norm": 0.4789334833621979,
+      "learning_rate": 2e-05,
+      "loss": 0.7324,
+      "step": 2880
+    },
+    {
+      "epoch": 0.19165727170236754,
+      "grad_norm": 0.45085087418556213,
+      "learning_rate": 2e-05,
+      "loss": 0.7121,
+      "step": 2890
+    },
+    {
+      "epoch": 0.19232044565289474,
+      "grad_norm": 0.5018136501312256,
+      "learning_rate": 2e-05,
+      "loss": 0.7415,
+      "step": 2900
+    },
+    {
+      "epoch": 0.19298361960342197,
+      "grad_norm": 0.4381272494792938,
+      "learning_rate": 2e-05,
+      "loss": 0.7324,
+      "step": 2910
+    },
+    {
+      "epoch": 0.1936467935539492,
+      "grad_norm": 0.49085962772369385,
+      "learning_rate": 2e-05,
+      "loss": 0.728,
+      "step": 2920
+    },
+    {
+      "epoch": 0.19430996750447643,
+      "grad_norm": 0.46983602643013,
+      "learning_rate": 2e-05,
+      "loss": 0.7255,
+      "step": 2930
+    },
+    {
+      "epoch": 0.19497314145500363,
+      "grad_norm": 0.4413234293460846,
+      "learning_rate": 2e-05,
+      "loss": 0.7295,
+      "step": 2940
+    },
+    {
+      "epoch": 0.19563631540553086,
+      "grad_norm": 0.4224623441696167,
+      "learning_rate": 2e-05,
+      "loss": 0.7185,
+      "step": 2950
+    },
+    {
+      "epoch": 0.1962994893560581,
+      "grad_norm": 0.44557204842567444,
+      "learning_rate": 2e-05,
+      "loss": 0.7498,
+      "step": 2960
+    },
+    {
+      "epoch": 0.19696266330658532,
+      "grad_norm": 0.4442812502384186,
+      "learning_rate": 2e-05,
+      "loss": 0.7385,
+      "step": 2970
+    },
+    {
+      "epoch": 0.19762583725711255,
+      "grad_norm": 0.45735692977905273,
+      "learning_rate": 2e-05,
+      "loss": 0.7358,
+      "step": 2980
+    },
+    {
+      "epoch": 0.19828901120763975,
+      "grad_norm": 0.4599263072013855,
+      "learning_rate": 2e-05,
+      "loss": 0.7343,
+      "step": 2990
+    },
+    {
+      "epoch": 0.19895218515816698,
+      "grad_norm": 0.4946780502796173,
+      "learning_rate": 2e-05,
+      "loss": 0.7208,
+      "step": 3000
+    },
+    {
+      "epoch": 0.1996153591086942,
+      "grad_norm": 0.44876787066459656,
+      "learning_rate": 2e-05,
+      "loss": 0.7484,
+      "step": 3010
+    },
+    {
+      "epoch": 0.20027853305922144,
+      "grad_norm": 0.4538794755935669,
+      "learning_rate": 2e-05,
+      "loss": 0.7071,
+      "step": 3020
+    },
+    {
+      "epoch": 0.20094170700974864,
+      "grad_norm": 0.4586864709854126,
+      "learning_rate": 2e-05,
+      "loss": 0.7235,
+      "step": 3030
+    },
+    {
+      "epoch": 0.20160488096027587,
+      "grad_norm": 1.1874250173568726,
+      "learning_rate": 2e-05,
+      "loss": 0.7386,
+      "step": 3040
+    },
+    {
+      "epoch": 0.2022680549108031,
+      "grad_norm": 0.43232521414756775,
+      "learning_rate": 2e-05,
+      "loss": 0.7197,
+      "step": 3050
+    },
+    {
+      "epoch": 0.20293122886133033,
+      "grad_norm": 0.4386948347091675,
+      "learning_rate": 2e-05,
+      "loss": 0.7427,
+      "step": 3060
+    },
+    {
+      "epoch": 0.20359440281185756,
+      "grad_norm": 0.4672735333442688,
+      "learning_rate": 2e-05,
+      "loss": 0.7359,
+      "step": 3070
+    },
+    {
+      "epoch": 0.20425757676238476,
+      "grad_norm": 0.4516846835613251,
+      "learning_rate": 2e-05,
+      "loss": 0.7044,
+      "step": 3080
+    },
+    {
+      "epoch": 0.204920750712912,
+      "grad_norm": 0.4240630567073822,
+      "learning_rate": 2e-05,
+      "loss": 0.7136,
+      "step": 3090
+    },
+    {
+      "epoch": 0.20558392466343922,
+      "grad_norm": 0.4282251298427582,
+      "learning_rate": 2e-05,
+      "loss": 0.7328,
+      "step": 3100
+    },
+    {
+      "epoch": 0.20624709861396645,
+      "grad_norm": 0.40108299255371094,
+      "learning_rate": 2e-05,
+      "loss": 0.7519,
+      "step": 3110
+    },
+    {
+      "epoch": 0.20691027256449368,
+      "grad_norm": 0.4190857708454132,
+      "learning_rate": 2e-05,
+      "loss": 0.7169,
+      "step": 3120
+    },
+    {
+      "epoch": 0.20757344651502088,
+      "grad_norm": 0.44728732109069824,
+      "learning_rate": 2e-05,
+      "loss": 0.7496,
+      "step": 3130
+    },
+    {
+      "epoch": 0.2082366204655481,
+      "grad_norm": 0.41856780648231506,
+      "learning_rate": 2e-05,
+      "loss": 0.7089,
+      "step": 3140
+    },
+    {
+      "epoch": 0.20889979441607534,
+      "grad_norm": 0.49630051851272583,
+      "learning_rate": 2e-05,
+      "loss": 0.7568,
+      "step": 3150
+    },
+    {
+      "epoch": 0.20956296836660257,
+      "grad_norm": 0.44060710072517395,
+      "learning_rate": 2e-05,
+      "loss": 0.6966,
+      "step": 3160
+    },
+    {
+      "epoch": 0.21022614231712977,
+      "grad_norm": 0.45406848192214966,
+      "learning_rate": 2e-05,
+      "loss": 0.7144,
+      "step": 3170
+    },
+    {
+      "epoch": 0.210889316267657,
+      "grad_norm": 0.42265212535858154,
+      "learning_rate": 2e-05,
+      "loss": 0.7275,
+      "step": 3180
+    },
+    {
+      "epoch": 0.21155249021818423,
+      "grad_norm": 0.44354790449142456,
+      "learning_rate": 2e-05,
+      "loss": 0.7276,
+      "step": 3190
+    },
+    {
+      "epoch": 0.21221566416871146,
+      "grad_norm": 0.44552451372146606,
+      "learning_rate": 2e-05,
+      "loss": 0.7052,
+      "step": 3200
+    },
+    {
+      "epoch": 0.2128788381192387,
+      "grad_norm": 0.4217117428779602,
+      "learning_rate": 2e-05,
+      "loss": 0.7117,
+      "step": 3210
+    },
+    {
+      "epoch": 0.2135420120697659,
+      "grad_norm": 0.42038828134536743,
+      "learning_rate": 2e-05,
+      "loss": 0.7047,
+      "step": 3220
+    },
+    {
+      "epoch": 0.21420518602029312,
+      "grad_norm": 0.42352569103240967,
+      "learning_rate": 2e-05,
+      "loss": 0.7033,
+      "step": 3230
+    },
+    {
+      "epoch": 0.21486835997082035,
+      "grad_norm": 0.483313649892807,
+      "learning_rate": 2e-05,
+      "loss": 0.7326,
+      "step": 3240
+    },
+    {
+      "epoch": 0.21553153392134758,
+      "grad_norm": 0.45873650908470154,
+      "learning_rate": 2e-05,
+      "loss": 0.7421,
+      "step": 3250
+    },
+    {
+      "epoch": 0.21619470787187478,
+      "grad_norm": 0.45367300510406494,
+      "learning_rate": 2e-05,
+      "loss": 0.7123,
+      "step": 3260
+    },
+    {
+      "epoch": 0.216857881822402,
+      "grad_norm": 0.4540068805217743,
+      "learning_rate": 2e-05,
+      "loss": 0.7194,
+      "step": 3270
+    },
+    {
+      "epoch": 0.21752105577292924,
+      "grad_norm": 0.4581505358219147,
+      "learning_rate": 2e-05,
+      "loss": 0.7064,
+      "step": 3280
+    },
+    {
+      "epoch": 0.21818422972345647,
+      "grad_norm": 0.4296276569366455,
+      "learning_rate": 2e-05,
+      "loss": 0.7394,
+      "step": 3290
+    },
+    {
+      "epoch": 0.2188474036739837,
+      "grad_norm": 0.4123869240283966,
+      "learning_rate": 2e-05,
+      "loss": 0.736,
+      "step": 3300
+    },
+    {
+      "epoch": 0.2195105776245109,
+      "grad_norm": 0.44229796528816223,
+      "learning_rate": 2e-05,
+      "loss": 0.7202,
+      "step": 3310
+    },
+    {
+      "epoch": 0.22017375157503813,
+      "grad_norm": 0.4266810715198517,
+      "learning_rate": 2e-05,
+      "loss": 0.7189,
+      "step": 3320
+    },
+    {
+      "epoch": 0.22083692552556536,
+      "grad_norm": 0.4784625470638275,
+      "learning_rate": 2e-05,
+      "loss": 0.726,
+      "step": 3330
+    },
+    {
+      "epoch": 0.2215000994760926,
+      "grad_norm": 0.422902911901474,
+      "learning_rate": 2e-05,
+      "loss": 0.7193,
+      "step": 3340
+    },
+    {
+      "epoch": 0.2221632734266198,
+      "grad_norm": 0.4550988972187042,
+      "learning_rate": 2e-05,
+      "loss": 0.7174,
+      "step": 3350
+    },
+    {
+      "epoch": 0.22282644737714702,
+      "grad_norm": 0.44610080122947693,
+      "learning_rate": 2e-05,
+      "loss": 0.7273,
+      "step": 3360
+    },
+    {
+      "epoch": 0.22348962132767425,
+      "grad_norm": 0.43774133920669556,
+      "learning_rate": 2e-05,
+      "loss": 0.7213,
+      "step": 3370
+    },
+    {
+      "epoch": 0.22415279527820148,
+      "grad_norm": 0.4690759778022766,
+      "learning_rate": 2e-05,
+      "loss": 0.7181,
+      "step": 3380
+    },
+    {
+      "epoch": 0.2248159692287287,
+      "grad_norm": 0.42695504426956177,
+      "learning_rate": 2e-05,
+      "loss": 0.7526,
+      "step": 3390
+    },
+    {
+      "epoch": 0.2254791431792559,
+      "grad_norm": 0.4128272831439972,
+      "learning_rate": 2e-05,
+      "loss": 0.7179,
+      "step": 3400
+    },
+    {
+      "epoch": 0.22614231712978314,
+      "grad_norm": 0.4729095697402954,
+      "learning_rate": 2e-05,
+      "loss": 0.7162,
+      "step": 3410
+    },
+    {
+      "epoch": 0.22680549108031037,
+      "grad_norm": 0.5577989220619202,
+      "learning_rate": 2e-05,
+      "loss": 0.7287,
+      "step": 3420
+    },
+    {
+      "epoch": 0.2274686650308376,
+      "grad_norm": 0.445027232170105,
+      "learning_rate": 2e-05,
+      "loss": 0.7147,
+      "step": 3430
+    },
+    {
+      "epoch": 0.2281318389813648,
+      "grad_norm": 0.42570748925209045,
+      "learning_rate": 2e-05,
+      "loss": 0.7271,
+      "step": 3440
+    },
+    {
+      "epoch": 0.22879501293189203,
+      "grad_norm": 0.46589401364326477,
+      "learning_rate": 2e-05,
+      "loss": 0.6928,
+      "step": 3450
+    },
+    {
+      "epoch": 0.22945818688241926,
+      "grad_norm": 0.4657121002674103,
+      "learning_rate": 2e-05,
+      "loss": 0.7428,
+      "step": 3460
+    },
+    {
+      "epoch": 0.2301213608329465,
+      "grad_norm": 0.4366071820259094,
+      "learning_rate": 2e-05,
+      "loss": 0.7061,
+      "step": 3470
+    },
+    {
+      "epoch": 0.23078453478347372,
+      "grad_norm": 0.4379141628742218,
+      "learning_rate": 2e-05,
+      "loss": 0.7312,
+      "step": 3480
+    },
+    {
+      "epoch": 0.23144770873400092,
+      "grad_norm": 0.4345671832561493,
+      "learning_rate": 2e-05,
+      "loss": 0.7337,
+      "step": 3490
+    },
+    {
+      "epoch": 0.23211088268452815,
+      "grad_norm": 0.4744890630245209,
+      "learning_rate": 2e-05,
+      "loss": 0.6989,
+      "step": 3500
+    },
+    {
+      "epoch": 0.23277405663505538,
+      "grad_norm": 0.4590144455432892,
+      "learning_rate": 2e-05,
+      "loss": 0.7419,
+      "step": 3510
+    },
+    {
+      "epoch": 0.2334372305855826,
+      "grad_norm": 0.43060076236724854,
+      "learning_rate": 2e-05,
+      "loss": 0.75,
+      "step": 3520
+    },
+    {
+      "epoch": 0.23410040453610984,
+      "grad_norm": 0.443794846534729,
+      "learning_rate": 2e-05,
+      "loss": 0.7265,
+      "step": 3530
+    },
+    {
+      "epoch": 0.23476357848663704,
+      "grad_norm": 0.41293781995773315,
+      "learning_rate": 2e-05,
+      "loss": 0.6978,
+      "step": 3540
+    },
+    {
+      "epoch": 0.23542675243716427,
+      "grad_norm": 0.4318286180496216,
+      "learning_rate": 2e-05,
+      "loss": 0.7027,
+      "step": 3550
+    },
+    {
+      "epoch": 0.2360899263876915,
+      "grad_norm": 0.44508758187294006,
+      "learning_rate": 2e-05,
+      "loss": 0.7116,
+      "step": 3560
+    },
+    {
+      "epoch": 0.23675310033821872,
+      "grad_norm": 0.44815289974212646,
+      "learning_rate": 2e-05,
+      "loss": 0.7196,
+      "step": 3570
+    },
+    {
+      "epoch": 0.23741627428874593,
+      "grad_norm": 0.44028452038764954,
+      "learning_rate": 2e-05,
+      "loss": 0.7062,
+      "step": 3580
+    },
+    {
+      "epoch": 0.23807944823927316,
+      "grad_norm": 0.4268244802951813,
+      "learning_rate": 2e-05,
+      "loss": 0.7231,
+      "step": 3590
+    },
+    {
+      "epoch": 0.23874262218980039,
+      "grad_norm": 0.43976491689682007,
+      "learning_rate": 2e-05,
+      "loss": 0.7083,
+      "step": 3600
+    },
+    {
+      "epoch": 0.23940579614032761,
+      "grad_norm": 0.45280003547668457,
+      "learning_rate": 2e-05,
+      "loss": 0.7081,
+      "step": 3610
+    },
+    {
+      "epoch": 0.24006897009085484,
+      "grad_norm": 0.5611624717712402,
+      "learning_rate": 2e-05,
+      "loss": 0.7373,
+      "step": 3620
+    },
+    {
+      "epoch": 0.24073214404138205,
+      "grad_norm": 0.45173415541648865,
+      "learning_rate": 2e-05,
+      "loss": 0.7446,
+      "step": 3630
+    },
+    {
+      "epoch": 0.24139531799190928,
+      "grad_norm": 0.4839193820953369,
+      "learning_rate": 2e-05,
+      "loss": 0.7108,
+      "step": 3640
+    },
+    {
+      "epoch": 0.2420584919424365,
+      "grad_norm": 0.48864683508872986,
+      "learning_rate": 2e-05,
+      "loss": 0.7502,
+      "step": 3650
+    },
+    {
+      "epoch": 0.24272166589296373,
+      "grad_norm": 6.257961273193359,
+      "learning_rate": 2e-05,
+      "loss": 0.7874,
+      "step": 3660
+    },
+    {
+      "epoch": 0.24338483984349094,
+      "grad_norm": 0.42597684264183044,
+      "learning_rate": 2e-05,
+      "loss": 0.7081,
+      "step": 3670
+    },
+    {
+      "epoch": 0.24404801379401816,
+      "grad_norm": 0.4785451889038086,
+      "learning_rate": 2e-05,
+      "loss": 0.7285,
+      "step": 3680
+    },
+    {
+      "epoch": 0.2447111877445454,
+      "grad_norm": 0.4379217028617859,
+      "learning_rate": 2e-05,
+      "loss": 0.7114,
+      "step": 3690
+    },
+    {
+      "epoch": 0.24537436169507262,
+      "grad_norm": 0.47318464517593384,
+      "learning_rate": 2e-05,
+      "loss": 0.7031,
+      "step": 3700
+    },
+    {
+      "epoch": 0.24603753564559985,
+      "grad_norm": 0.43900227546691895,
+      "learning_rate": 2e-05,
+      "loss": 0.7165,
+      "step": 3710
+    },
+    {
+      "epoch": 0.24670070959612705,
+      "grad_norm": 0.435876727104187,
+      "learning_rate": 2e-05,
+      "loss": 0.7227,
+      "step": 3720
+    },
+    {
+      "epoch": 0.24736388354665428,
+      "grad_norm": 0.43884292244911194,
+      "learning_rate": 2e-05,
+      "loss": 0.7259,
+      "step": 3730
+    },
+    {
+      "epoch": 0.2480270574971815,
+      "grad_norm": 0.44448748230934143,
+      "learning_rate": 2e-05,
+      "loss": 0.7317,
+      "step": 3740
+    },
+    {
+      "epoch": 0.24869023144770874,
+      "grad_norm": 0.5201537609100342,
+      "learning_rate": 2e-05,
+      "loss": 0.7032,
+      "step": 3750
+    },
+    {
+      "epoch": 0.24935340539823594,
+      "grad_norm": 0.40599125623703003,
+      "learning_rate": 2e-05,
+      "loss": 0.7267,
+      "step": 3760
+    },
+    {
+      "epoch": 0.2500165793487632,
+      "grad_norm": 0.4755355715751648,
+      "learning_rate": 2e-05,
+      "loss": 0.7407,
+      "step": 3770
+    },
+    {
+      "epoch": 0.25067975329929043,
+      "grad_norm": 0.44619736075401306,
+      "learning_rate": 2e-05,
+      "loss": 0.7305,
+      "step": 3780
+    },
+    {
+      "epoch": 0.2513429272498176,
+      "grad_norm": 0.43437689542770386,
+      "learning_rate": 2e-05,
+      "loss": 0.7043,
+      "step": 3790
+    },
+    {
+      "epoch": 0.25200610120034483,
+      "grad_norm": 0.4214416742324829,
+      "learning_rate": 2e-05,
+      "loss": 0.738,
+      "step": 3800
+    },
+    {
+      "epoch": 0.25266927515087206,
+      "grad_norm": 0.4558330178260803,
+      "learning_rate": 2e-05,
+      "loss": 0.6979,
+      "step": 3810
+    },
+    {
+      "epoch": 0.2533324491013993,
+      "grad_norm": 0.43614012002944946,
+      "learning_rate": 2e-05,
+      "loss": 0.7248,
+      "step": 3820
+    },
+    {
+      "epoch": 0.2539956230519265,
+      "grad_norm": 0.4879645109176636,
+      "learning_rate": 2e-05,
+      "loss": 0.7049,
+      "step": 3830
+    },
+    {
+      "epoch": 0.25465879700245375,
+      "grad_norm": 0.40824270248413086,
+      "learning_rate": 2e-05,
+      "loss": 0.7119,
+      "step": 3840
+    },
+    {
+      "epoch": 0.255321970952981,
+      "grad_norm": 0.44525161385536194,
+      "learning_rate": 2e-05,
+      "loss": 0.7238,
+      "step": 3850
+    },
+    {
+      "epoch": 0.2559851449035082,
+      "grad_norm": 0.4895830452442169,
+      "learning_rate": 2e-05,
+      "loss": 0.7061,
+      "step": 3860
+    },
+    {
+      "epoch": 0.25664831885403544,
+      "grad_norm": 0.4618409276008606,
+      "learning_rate": 2e-05,
+      "loss": 0.7099,
+      "step": 3870
+    },
+    {
+      "epoch": 0.2573114928045626,
+      "grad_norm": 0.40779945254325867,
+      "learning_rate": 2e-05,
+      "loss": 0.7273,
+      "step": 3880
+    },
+    {
+      "epoch": 0.25797466675508984,
+      "grad_norm": 0.44511210918426514,
+      "learning_rate": 2e-05,
+      "loss": 0.702,
+      "step": 3890
+    },
+    {
+      "epoch": 0.2586378407056171,
+      "grad_norm": 0.45794036984443665,
+      "learning_rate": 2e-05,
+      "loss": 0.7281,
+      "step": 3900
+    },
+    {
+      "epoch": 0.2593010146561443,
+      "grad_norm": 0.4685095548629761,
+      "learning_rate": 2e-05,
+      "loss": 0.7446,
+      "step": 3910
+    },
+    {
+      "epoch": 0.25996418860667153,
+      "grad_norm": 0.4546467959880829,
+      "learning_rate": 2e-05,
+      "loss": 0.728,
+      "step": 3920
+    },
+    {
+      "epoch": 0.26062736255719876,
+      "grad_norm": 0.43525663018226624,
+      "learning_rate": 2e-05,
+      "loss": 0.7127,
+      "step": 3930
+    },
+    {
+      "epoch": 0.261290536507726,
+      "grad_norm": 0.49840015172958374,
+      "learning_rate": 2e-05,
+      "loss": 0.7344,
+      "step": 3940
+    },
+    {
+      "epoch": 0.2619537104582532,
+      "grad_norm": 0.44585609436035156,
+      "learning_rate": 2e-05,
+      "loss": 0.7385,
+      "step": 3950
+    },
+    {
+      "epoch": 0.26261688440878045,
+      "grad_norm": 0.44912222027778625,
+      "learning_rate": 2e-05,
+      "loss": 0.7332,
+      "step": 3960
+    },
+    {
+      "epoch": 0.2632800583593076,
+      "grad_norm": 0.4206899106502533,
+      "learning_rate": 2e-05,
+      "loss": 0.7133,
+      "step": 3970
+    },
+    {
+      "epoch": 0.26394323230983485,
+      "grad_norm": 0.4299122393131256,
+      "learning_rate": 2e-05,
+      "loss": 0.7172,
+      "step": 3980
+    },
+    {
+      "epoch": 0.2646064062603621,
+      "grad_norm": 0.41612157225608826,
+      "learning_rate": 2e-05,
+      "loss": 0.7169,
+      "step": 3990
+    },
+    {
+      "epoch": 0.2652695802108893,
+      "grad_norm": 0.4067608416080475,
+      "learning_rate": 2e-05,
+      "loss": 0.7051,
+      "step": 4000
+    },
+    {
+      "epoch": 0.26593275416141654,
+      "grad_norm": 0.46370911598205566,
+      "learning_rate": 2e-05,
+      "loss": 0.7152,
+      "step": 4010
+    },
+    {
+      "epoch": 0.26659592811194377,
+      "grad_norm": 0.45039990544319153,
+      "learning_rate": 2e-05,
+      "loss": 0.7003,
+      "step": 4020
+    },
+    {
+      "epoch": 0.267259102062471,
+      "grad_norm": 0.4659903347492218,
+      "learning_rate": 2e-05,
+      "loss": 0.7071,
+      "step": 4030
+    },
+    {
+      "epoch": 0.26792227601299823,
+      "grad_norm": 0.45298582315444946,
+      "learning_rate": 2e-05,
+      "loss": 0.7462,
+      "step": 4040
+    },
+    {
+      "epoch": 0.26858544996352546,
+      "grad_norm": 0.4633753299713135,
+      "learning_rate": 2e-05,
+      "loss": 0.7148,
+      "step": 4050
+    },
+    {
+      "epoch": 0.26924862391405263,
+      "grad_norm": 0.4284294545650482,
+      "learning_rate": 2e-05,
+      "loss": 0.7063,
+      "step": 4060
+    },
+    {
+      "epoch": 0.26991179786457986,
+      "grad_norm": 0.440674751996994,
+      "learning_rate": 2e-05,
+      "loss": 0.7159,
+      "step": 4070
+    },
+    {
+      "epoch": 0.2705749718151071,
+      "grad_norm": 0.45898178219795227,
+      "learning_rate": 2e-05,
+      "loss": 0.7172,
+      "step": 4080
+    },
+    {
+      "epoch": 0.2712381457656343,
+      "grad_norm": 0.42899852991104126,
+      "learning_rate": 2e-05,
+      "loss": 0.7107,
+      "step": 4090
+    },
+    {
+      "epoch": 0.27190131971616155,
+      "grad_norm": 0.4513186812400818,
+      "learning_rate": 2e-05,
+      "loss": 0.7014,
+      "step": 4100
+    },
+    {
+      "epoch": 0.2725644936666888,
+      "grad_norm": 0.4538447856903076,
+      "learning_rate": 2e-05,
+      "loss": 0.7412,
+      "step": 4110
+    },
+    {
+      "epoch": 0.273227667617216,
+      "grad_norm": 0.40217098593711853,
+      "learning_rate": 2e-05,
+      "loss": 0.7188,
+      "step": 4120
+    },
+    {
+      "epoch": 0.27389084156774324,
+      "grad_norm": 0.4167708456516266,
+      "learning_rate": 2e-05,
+      "loss": 0.7247,
+      "step": 4130
+    },
+    {
+      "epoch": 0.27455401551827047,
+      "grad_norm": 0.40658578276634216,
+      "learning_rate": 2e-05,
+      "loss": 0.7214,
+      "step": 4140
+    },
+    {
+      "epoch": 0.27521718946879764,
+      "grad_norm": 0.43742382526397705,
+      "learning_rate": 2e-05,
+      "loss": 0.733,
+      "step": 4150
+    },
+    {
+      "epoch": 0.27588036341932487,
+      "grad_norm": 0.41422826051712036,
+      "learning_rate": 2e-05,
+      "loss": 0.715,
+      "step": 4160
+    },
+    {
+      "epoch": 0.2765435373698521,
+      "grad_norm": 0.41707736253738403,
+      "learning_rate": 2e-05,
+      "loss": 0.7095,
+      "step": 4170
+    },
+    {
+      "epoch": 0.27720671132037933,
+      "grad_norm": 0.40865060687065125,
+      "learning_rate": 2e-05,
+      "loss": 0.7164,
+      "step": 4180
+    },
+    {
+      "epoch": 0.27786988527090656,
+      "grad_norm": 0.4409600496292114,
+      "learning_rate": 2e-05,
+      "loss": 0.7287,
+      "step": 4190
+    },
+    {
+      "epoch": 0.2785330592214338,
+      "grad_norm": 0.47225630283355713,
+      "learning_rate": 2e-05,
+      "loss": 0.7136,
+      "step": 4200
+    },
+    {
+      "epoch": 0.279196233171961,
+      "grad_norm": 0.4420481324195862,
+      "learning_rate": 2e-05,
+      "loss": 0.7319,
+      "step": 4210
+    },
+    {
+      "epoch": 0.27985940712248825,
+      "grad_norm": 0.422795832157135,
+      "learning_rate": 2e-05,
+      "loss": 0.7214,
+      "step": 4220
+    },
+    {
+      "epoch": 0.2805225810730155,
+      "grad_norm": 0.4085385501384735,
+      "learning_rate": 2e-05,
+      "loss": 0.7193,
+      "step": 4230
+    },
+    {
+      "epoch": 0.28118575502354265,
+      "grad_norm": 0.44029298424720764,
+      "learning_rate": 2e-05,
+      "loss": 0.7279,
+      "step": 4240
+    },
+    {
+      "epoch": 0.2818489289740699,
+      "grad_norm": 0.4251239597797394,
+      "learning_rate": 2e-05,
+      "loss": 0.744,
+      "step": 4250
+    },
+    {
+      "epoch": 0.2825121029245971,
+      "grad_norm": 0.4440610706806183,
+      "learning_rate": 2e-05,
+      "loss": 0.7224,
+      "step": 4260
+    },
+    {
+      "epoch": 0.28317527687512434,
+      "grad_norm": 0.6336266994476318,
+      "learning_rate": 2e-05,
+      "loss": 0.7077,
+      "step": 4270
+    },
+    {
+      "epoch": 0.28383845082565157,
+      "grad_norm": 0.40553924441337585,
+      "learning_rate": 2e-05,
+      "loss": 0.7134,
+      "step": 4280
+    },
+    {
+      "epoch": 0.2845016247761788,
+      "grad_norm": 0.4212879538536072,
+      "learning_rate": 2e-05,
+      "loss": 0.7126,
+      "step": 4290
+    },
+    {
+      "epoch": 0.285164798726706,
+      "grad_norm": 0.4236951470375061,
+      "learning_rate": 2e-05,
+      "loss": 0.7055,
+      "step": 4300
+    },
+    {
+      "epoch": 0.28582797267723326,
+      "grad_norm": 0.4220391809940338,
+      "learning_rate": 2e-05,
+      "loss": 0.7137,
+      "step": 4310
+    },
+    {
+      "epoch": 0.2864911466277605,
+      "grad_norm": 0.4318537712097168,
+      "learning_rate": 2e-05,
+      "loss": 0.727,
+      "step": 4320
+    },
+    {
+      "epoch": 0.28715432057828766,
+      "grad_norm": 0.5137214064598083,
+      "learning_rate": 2e-05,
+      "loss": 0.7071,
+      "step": 4330
+    },
+    {
+      "epoch": 0.2878174945288149,
+      "grad_norm": 0.41365182399749756,
+      "learning_rate": 2e-05,
+      "loss": 0.7075,
+      "step": 4340
+    },
+    {
+      "epoch": 0.2884806684793421,
+      "grad_norm": 0.4234631359577179,
+      "learning_rate": 2e-05,
+      "loss": 0.7134,
+      "step": 4350
+    },
+    {
+      "epoch": 0.28914384242986935,
+      "grad_norm": 0.44713494181632996,
+      "learning_rate": 2e-05,
+      "loss": 0.7075,
+      "step": 4360
+    },
+    {
+      "epoch": 0.2898070163803966,
+      "grad_norm": 0.454023152589798,
+      "learning_rate": 2e-05,
+      "loss": 0.6971,
+      "step": 4370
+    },
+    {
+      "epoch": 0.2904701903309238,
+      "grad_norm": 0.5051842927932739,
+      "learning_rate": 2e-05,
+      "loss": 0.7326,
+      "step": 4380
+    },
+    {
+      "epoch": 0.29113336428145103,
+      "grad_norm": 0.5243337750434875,
+      "learning_rate": 2e-05,
+      "loss": 0.7078,
+      "step": 4390
+    },
+    {
+      "epoch": 0.29179653823197826,
+      "grad_norm": 0.42526450753211975,
+      "learning_rate": 2e-05,
+      "loss": 0.7035,
+      "step": 4400
+    },
+    {
+      "epoch": 0.2924597121825055,
+      "grad_norm": 0.4312632381916046,
+      "learning_rate": 2e-05,
+      "loss": 0.7195,
+      "step": 4410
+    },
+    {
+      "epoch": 0.29312288613303267,
+      "grad_norm": 0.4477943181991577,
+      "learning_rate": 2e-05,
+      "loss": 0.7168,
+      "step": 4420
+    },
+    {
+      "epoch": 0.2937860600835599,
+      "grad_norm": 0.4300823211669922,
+      "learning_rate": 2e-05,
+      "loss": 0.7174,
+      "step": 4430
+    },
+    {
+      "epoch": 0.2944492340340871,
+      "grad_norm": 0.4278593361377716,
+      "learning_rate": 2e-05,
+      "loss": 0.7012,
+      "step": 4440
+    },
+    {
+      "epoch": 0.29511240798461436,
+      "grad_norm": 0.473036527633667,
+      "learning_rate": 2e-05,
+      "loss": 0.7072,
+      "step": 4450
+    },
+    {
+      "epoch": 0.2957755819351416,
+      "grad_norm": 0.5350799560546875,
+      "learning_rate": 2e-05,
+      "loss": 0.724,
+      "step": 4460
+    },
+    {
+      "epoch": 0.2964387558856688,
+      "grad_norm": 0.40777650475502014,
+      "learning_rate": 2e-05,
+      "loss": 0.6892,
+      "step": 4470
+    },
+    {
+      "epoch": 0.29710192983619604,
+      "grad_norm": 0.4430738091468811,
+      "learning_rate": 2e-05,
+      "loss": 0.7112,
+      "step": 4480
+    },
+    {
+      "epoch": 0.2977651037867233,
+      "grad_norm": 0.5116422772407532,
+      "learning_rate": 2e-05,
+      "loss": 0.7102,
+      "step": 4490
+    },
+    {
+      "epoch": 0.2984282777372505,
+      "grad_norm": 0.4413306713104248,
+      "learning_rate": 2e-05,
+      "loss": 0.703,
+      "step": 4500
+    },
+    {
+      "epoch": 0.29909145168777773,
+      "grad_norm": 0.44744178652763367,
+      "learning_rate": 2e-05,
+      "loss": 0.7224,
+      "step": 4510
+    },
+    {
+      "epoch": 0.2997546256383049,
+      "grad_norm": 0.4297047257423401,
+      "learning_rate": 2e-05,
+      "loss": 0.7257,
+      "step": 4520
+    },
+    {
+      "epoch": 0.30041779958883214,
+      "grad_norm": 0.4384410083293915,
+      "learning_rate": 2e-05,
+      "loss": 0.6988,
+      "step": 4530
+    },
+    {
+      "epoch": 0.30108097353935936,
+      "grad_norm": 0.4327126145362854,
+      "learning_rate": 2e-05,
+      "loss": 0.7354,
+      "step": 4540
+    },
+    {
+      "epoch": 0.3017441474898866,
+      "grad_norm": 0.4421708285808563,
+      "learning_rate": 2e-05,
+      "loss": 0.7248,
+      "step": 4550
+    },
+    {
+      "epoch": 0.3024073214404138,
+      "grad_norm": 0.4230192303657532,
+      "learning_rate": 2e-05,
+      "loss": 0.7068,
+      "step": 4560
+    },
+    {
+      "epoch": 0.30307049539094105,
+      "grad_norm": 0.4164934456348419,
+      "learning_rate": 2e-05,
+      "loss": 0.7064,
+      "step": 4570
+    },
+    {
+      "epoch": 0.3037336693414683,
+      "grad_norm": 0.42152565717697144,
+      "learning_rate": 2e-05,
+      "loss": 0.7301,
+      "step": 4580
+    },
+    {
+      "epoch": 0.3043968432919955,
+      "grad_norm": 0.4739789068698883,
+      "learning_rate": 2e-05,
+      "loss": 0.6957,
+      "step": 4590
+    },
+    {
+      "epoch": 0.30506001724252274,
+      "grad_norm": 0.42717763781547546,
+      "learning_rate": 2e-05,
+      "loss": 0.7124,
+      "step": 4600
+    },
+    {
+      "epoch": 0.3057231911930499,
+      "grad_norm": 0.5130007266998291,
+      "learning_rate": 2e-05,
+      "loss": 0.7385,
+      "step": 4610
+    },
+    {
+      "epoch": 0.30638636514357714,
+      "grad_norm": 0.4164990186691284,
+      "learning_rate": 2e-05,
+      "loss": 0.7122,
+      "step": 4620
+    },
+    {
+      "epoch": 0.3070495390941044,
+      "grad_norm": 0.4283609986305237,
+      "learning_rate": 2e-05,
+      "loss": 0.6891,
+      "step": 4630
+    },
+    {
+      "epoch": 0.3077127130446316,
+      "grad_norm": 0.41688019037246704,
+      "learning_rate": 2e-05,
+      "loss": 0.7143,
+      "step": 4640
+    },
+    {
+      "epoch": 0.30837588699515883,
+      "grad_norm": 0.4387054443359375,
+      "learning_rate": 2e-05,
+      "loss": 0.7008,
+      "step": 4650
+    },
+    {
+      "epoch": 0.30903906094568606,
+      "grad_norm": 0.4529563784599304,
+      "learning_rate": 2e-05,
+      "loss": 0.6975,
+      "step": 4660
+    },
+    {
+      "epoch": 0.3097022348962133,
+      "grad_norm": 0.45100653171539307,
+      "learning_rate": 2e-05,
+      "loss": 0.7173,
+      "step": 4670
+    },
+    {
+      "epoch": 0.3103654088467405,
+      "grad_norm": 0.43708083033561707,
+      "learning_rate": 2e-05,
+      "loss": 0.7306,
+      "step": 4680
+    },
+    {
+      "epoch": 0.31102858279726775,
+      "grad_norm": 0.44238895177841187,
+      "learning_rate": 2e-05,
+      "loss": 0.7065,
+      "step": 4690
+    },
+    {
+      "epoch": 0.3116917567477949,
+      "grad_norm": 0.41368401050567627,
+      "learning_rate": 2e-05,
+      "loss": 0.6941,
+      "step": 4700
+    },
+    {
+      "epoch": 0.31235493069832215,
+      "grad_norm": 0.4319513738155365,
+      "learning_rate": 2e-05,
+      "loss": 0.6883,
+      "step": 4710
+    },
+    {
+      "epoch": 0.3130181046488494,
+      "grad_norm": 0.44077882170677185,
+      "learning_rate": 2e-05,
+      "loss": 0.7037,
+      "step": 4720
+    },
+    {
+      "epoch": 0.3136812785993766,
+      "grad_norm": 0.4406428337097168,
+      "learning_rate": 2e-05,
+      "loss": 0.7042,
+      "step": 4730
+    },
+    {
+      "epoch": 0.31434445254990384,
+      "grad_norm": 0.4604378044605255,
+      "learning_rate": 2e-05,
+      "loss": 0.6969,
+      "step": 4740
+    },
+    {
+      "epoch": 0.31500762650043107,
+      "grad_norm": 0.4481877088546753,
+      "learning_rate": 2e-05,
+      "loss": 0.6937,
+      "step": 4750
+    },
+    {
+      "epoch": 0.3156708004509583,
+      "grad_norm": 0.4246469736099243,
+      "learning_rate": 2e-05,
+      "loss": 0.7249,
+      "step": 4760
+    },
+    {
+      "epoch": 0.31633397440148553,
+      "grad_norm": 0.399565190076828,
+      "learning_rate": 2e-05,
+      "loss": 0.7022,
+      "step": 4770
+    },
+    {
+      "epoch": 0.31699714835201276,
+      "grad_norm": 0.4086315929889679,
+      "learning_rate": 2e-05,
+      "loss": 0.7241,
+      "step": 4780
+    },
+    {
+      "epoch": 0.31766032230253993,
+      "grad_norm": 0.4032701253890991,
+      "learning_rate": 2e-05,
+      "loss": 0.7258,
+      "step": 4790
+    },
+    {
+      "epoch": 0.31832349625306716,
+      "grad_norm": 0.39910489320755005,
+      "learning_rate": 2e-05,
+      "loss": 0.7256,
+      "step": 4800
+    },
+    {
+      "epoch": 0.3189866702035944,
+      "grad_norm": 0.4245343506336212,
+      "learning_rate": 2e-05,
+      "loss": 0.7212,
+      "step": 4810
+    },
+    {
+      "epoch": 0.3196498441541216,
+      "grad_norm": 0.42083439230918884,
+      "learning_rate": 2e-05,
+      "loss": 0.725,
+      "step": 4820
+    },
+    {
+      "epoch": 0.32031301810464885,
+      "grad_norm": 0.8856384754180908,
+      "learning_rate": 2e-05,
+      "loss": 0.7128,
+      "step": 4830
+    },
+    {
+      "epoch": 0.3209761920551761,
+      "grad_norm": 0.42265793681144714,
+      "learning_rate": 2e-05,
+      "loss": 0.7149,
+      "step": 4840
+    },
+    {
+      "epoch": 0.3216393660057033,
+      "grad_norm": 0.418274849653244,
+      "learning_rate": 2e-05,
+      "loss": 0.7008,
+      "step": 4850
+    },
+    {
+      "epoch": 0.32230253995623054,
+      "grad_norm": 0.44719722867012024,
+      "learning_rate": 2e-05,
+      "loss": 0.7169,
+      "step": 4860
+    },
+    {
+      "epoch": 0.32296571390675777,
+      "grad_norm": 0.4416983127593994,
+      "learning_rate": 2e-05,
+      "loss": 0.693,
+      "step": 4870
+    },
+    {
+      "epoch": 0.32362888785728494,
+      "grad_norm": 0.42274999618530273,
+      "learning_rate": 2e-05,
+      "loss": 0.7383,
+      "step": 4880
+    },
+    {
+      "epoch": 0.32429206180781217,
+      "grad_norm": 0.49183934926986694,
+      "learning_rate": 2e-05,
+      "loss": 0.721,
+      "step": 4890
+    },
+    {
+      "epoch": 0.3249552357583394,
+      "grad_norm": 0.4221515953540802,
+      "learning_rate": 2e-05,
+      "loss": 0.7014,
+      "step": 4900
+    },
+    {
+      "epoch": 0.32561840970886663,
+      "grad_norm": 0.4345867335796356,
+      "learning_rate": 2e-05,
+      "loss": 0.6958,
+      "step": 4910
+    },
+    {
+      "epoch": 0.32628158365939386,
+      "grad_norm": 0.4104120135307312,
+      "learning_rate": 2e-05,
+      "loss": 0.707,
+      "step": 4920
+    },
+    {
+      "epoch": 0.3269447576099211,
+      "grad_norm": 0.480258584022522,
+      "learning_rate": 2e-05,
+      "loss": 0.6942,
+      "step": 4930
+    },
+    {
+      "epoch": 0.3276079315604483,
+      "grad_norm": 0.5035107135772705,
+      "learning_rate": 2e-05,
+      "loss": 0.7108,
+      "step": 4940
+    },
+    {
+      "epoch": 0.32827110551097555,
+      "grad_norm": 0.4675476551055908,
+      "learning_rate": 2e-05,
+      "loss": 0.7051,
+      "step": 4950
+    },
+    {
+      "epoch": 0.3289342794615028,
+      "grad_norm": 0.4497123956680298,
+      "learning_rate": 2e-05,
+      "loss": 0.7192,
+      "step": 4960
+    },
+    {
+      "epoch": 0.32959745341202995,
+      "grad_norm": 0.4247696101665497,
+      "learning_rate": 2e-05,
+      "loss": 0.718,
+      "step": 4970
+    },
+    {
+      "epoch": 0.3302606273625572,
+      "grad_norm": 0.4365120530128479,
+      "learning_rate": 2e-05,
+      "loss": 0.7034,
+      "step": 4980
+    },
+    {
+      "epoch": 0.3309238013130844,
+      "grad_norm": 0.4657589793205261,
+      "learning_rate": 2e-05,
+      "loss": 0.7221,
+      "step": 4990
+    },
+    {
+      "epoch": 0.33158697526361164,
+      "grad_norm": 0.4138410985469818,
+      "learning_rate": 2e-05,
+      "loss": 0.6798,
+      "step": 5000
+    },
+    {
+      "epoch": 0.33225014921413887,
+      "grad_norm": 0.4804857671260834,
+      "learning_rate": 2e-05,
+      "loss": 0.7127,
+      "step": 5010
+    },
+    {
+      "epoch": 0.3329133231646661,
+      "grad_norm": 0.43275633454322815,
+      "learning_rate": 2e-05,
+      "loss": 0.7141,
+      "step": 5020
+    },
+    {
+      "epoch": 0.3335764971151933,
+      "grad_norm": 0.4018505811691284,
+      "learning_rate": 2e-05,
+      "loss": 0.7067,
+      "step": 5030
+    },
+    {
+      "epoch": 0.33423967106572056,
+      "grad_norm": 0.4772813320159912,
+      "learning_rate": 2e-05,
+      "loss": 0.6925,
+      "step": 5040
+    },
+    {
+      "epoch": 0.3349028450162478,
+      "grad_norm": 0.4661034345626831,
+      "learning_rate": 2e-05,
+      "loss": 0.693,
+      "step": 5050
+    },
+    {
+      "epoch": 0.33556601896677496,
+      "grad_norm": 0.3940356969833374,
+      "learning_rate": 2e-05,
+      "loss": 0.7052,
+      "step": 5060
+    },
+    {
+      "epoch": 0.3362291929173022,
+      "grad_norm": 0.40755006670951843,
+      "learning_rate": 2e-05,
+      "loss": 0.6905,
+      "step": 5070
+    },
+    {
+      "epoch": 0.3368923668678294,
+      "grad_norm": 0.4546791911125183,
+      "learning_rate": 2e-05,
+      "loss": 0.7117,
+      "step": 5080
+    },
+    {
+      "epoch": 0.33755554081835665,
+      "grad_norm": 0.4277106821537018,
+      "learning_rate": 2e-05,
+      "loss": 0.7144,
+      "step": 5090
+    },
+    {
+      "epoch": 0.3382187147688839,
+      "grad_norm": 0.4051600992679596,
+      "learning_rate": 2e-05,
+      "loss": 0.6937,
+      "step": 5100
+    },
+    {
+      "epoch": 0.3388818887194111,
+      "grad_norm": 0.43261224031448364,
+      "learning_rate": 2e-05,
+      "loss": 0.6999,
+      "step": 5110
+    },
+    {
+      "epoch": 0.33954506266993834,
+      "grad_norm": 0.42372557520866394,
+      "learning_rate": 2e-05,
+      "loss": 0.7105,
+      "step": 5120
+    },
+    {
+      "epoch": 0.34020823662046557,
+      "grad_norm": 0.49420851469039917,
+      "learning_rate": 2e-05,
+      "loss": 0.7107,
+      "step": 5130
+    },
+    {
+      "epoch": 0.3408714105709928,
+      "grad_norm": 0.4319324493408203,
+      "learning_rate": 2e-05,
+      "loss": 0.6903,
+      "step": 5140
+    },
+    {
+      "epoch": 0.34153458452151997,
+      "grad_norm": 0.4231835901737213,
+      "learning_rate": 2e-05,
+      "loss": 0.7101,
+      "step": 5150
+    },
+    {
+      "epoch": 0.3421977584720472,
+      "grad_norm": 0.41818317770957947,
+      "learning_rate": 2e-05,
+      "loss": 0.719,
+      "step": 5160
+    },
+    {
+      "epoch": 0.3428609324225744,
+      "grad_norm": 0.4805700480937958,
+      "learning_rate": 2e-05,
+      "loss": 0.7006,
+      "step": 5170
+    },
+    {
+      "epoch": 0.34352410637310166,
+      "grad_norm": 0.4527558982372284,
+      "learning_rate": 2e-05,
+      "loss": 0.7188,
+      "step": 5180
+    },
+    {
+      "epoch": 0.3441872803236289,
+      "grad_norm": 0.5771711468696594,
+      "learning_rate": 2e-05,
+      "loss": 0.7117,
+      "step": 5190
+    },
+    {
+      "epoch": 0.3448504542741561,
+      "grad_norm": 0.43666353821754456,
+      "learning_rate": 2e-05,
+      "loss": 0.7206,
+      "step": 5200
+    },
+    {
+      "epoch": 0.34551362822468334,
+      "grad_norm": 0.42155739665031433,
+      "learning_rate": 2e-05,
+      "loss": 0.71,
+      "step": 5210
+    },
+    {
+      "epoch": 0.3461768021752106,
+      "grad_norm": 0.40986326336860657,
+      "learning_rate": 2e-05,
+      "loss": 0.6981,
+      "step": 5220
+    },
+    {
+      "epoch": 0.3468399761257378,
+      "grad_norm": 0.4199829399585724,
+      "learning_rate": 2e-05,
+      "loss": 0.7053,
+      "step": 5230
+    },
+    {
+      "epoch": 0.347503150076265,
+      "grad_norm": 0.41298285126686096,
+      "learning_rate": 2e-05,
+      "loss": 0.6849,
+      "step": 5240
+    },
+    {
+      "epoch": 0.3481663240267922,
+      "grad_norm": 0.42672210931777954,
+      "learning_rate": 2e-05,
+      "loss": 0.691,
+      "step": 5250
+    },
+    {
+      "epoch": 0.34882949797731944,
+      "grad_norm": 0.43835774064064026,
+      "learning_rate": 2e-05,
+      "loss": 0.6876,
+      "step": 5260
+    },
+    {
+      "epoch": 0.34949267192784667,
+      "grad_norm": 0.40993231534957886,
+      "learning_rate": 2e-05,
+      "loss": 0.7297,
+      "step": 5270
+    },
+    {
+      "epoch": 0.3501558458783739,
+      "grad_norm": 0.45235201716423035,
+      "learning_rate": 2e-05,
+      "loss": 0.699,
+      "step": 5280
+    },
+    {
+      "epoch": 0.3508190198289011,
+      "grad_norm": 0.4141416549682617,
+      "learning_rate": 2e-05,
+      "loss": 0.7048,
+      "step": 5290
+    },
+    {
+      "epoch": 0.35148219377942835,
+      "grad_norm": 0.4252654016017914,
+      "learning_rate": 2e-05,
+      "loss": 0.7101,
+      "step": 5300
+    },
+    {
+      "epoch": 0.3521453677299556,
+      "grad_norm": 0.43203818798065186,
+      "learning_rate": 2e-05,
+      "loss": 0.6998,
+      "step": 5310
+    },
+    {
+      "epoch": 0.3528085416804828,
+      "grad_norm": 0.40562939643859863,
+      "learning_rate": 2e-05,
+      "loss": 0.6865,
+      "step": 5320
+    },
+    {
+      "epoch": 0.35347171563101004,
+      "grad_norm": 0.4484749734401703,
+      "learning_rate": 2e-05,
+      "loss": 0.7036,
+      "step": 5330
+    },
+    {
+      "epoch": 0.3541348895815372,
+      "grad_norm": 0.5492318868637085,
+      "learning_rate": 2e-05,
+      "loss": 0.7002,
+      "step": 5340
+    },
+    {
+      "epoch": 0.35479806353206444,
+      "grad_norm": 0.5349329113960266,
+      "learning_rate": 1.999194520388485e-05,
+      "loss": 0.6855,
+      "step": 5350
+    },
+    {
+      "epoch": 0.3554612374825917,
+      "grad_norm": 0.470889151096344,
+      "learning_rate": 1.9959247303157763e-05,
+      "loss": 0.7176,
+      "step": 5360
+    },
+    {
+      "epoch": 0.3561244114331189,
+      "grad_norm": 0.41299134492874146,
+      "learning_rate": 1.990149424459817e-05,
+      "loss": 0.7035,
+      "step": 5370
+    },
+    {
+      "epoch": 0.35678758538364613,
+      "grad_norm": 0.450900137424469,
+      "learning_rate": 1.9818847538648688e-05,
+      "loss": 0.693,
+      "step": 5380
+    },
+    {
+      "epoch": 0.35745075933417336,
+      "grad_norm": 0.42808613181114197,
+      "learning_rate": 1.971153831256721e-05,
+      "loss": 0.6777,
+      "step": 5390
+    },
+    {
+      "epoch": 0.3581139332847006,
+      "grad_norm": 0.43397775292396545,
+      "learning_rate": 1.9579866664063435e-05,
+      "loss": 0.7049,
+      "step": 5400
+    },
+    {
+      "epoch": 0.3587771072352278,
+      "grad_norm": 0.4414255619049072,
+      "learning_rate": 1.9424200822054782e-05,
+      "loss": 0.7204,
+      "step": 5410
+    },
+    {
+      "epoch": 0.35944028118575505,
+      "grad_norm": 0.4406629204750061,
+      "learning_rate": 1.9244976116888626e-05,
+      "loss": 0.7162,
+      "step": 5420
+    },
+    {
+      "epoch": 0.3601034551362822,
+      "grad_norm": 0.43205147981643677,
+      "learning_rate": 1.904269376291071e-05,
+      "loss": 0.7101,
+      "step": 5430
+    },
+    {
+      "epoch": 0.36076662908680945,
+      "grad_norm": 0.4646110236644745,
+      "learning_rate": 1.88179194567844e-05,
+      "loss": 0.7074,
+      "step": 5440
+    },
+    {
+      "epoch": 0.3614298030373367,
+      "grad_norm": 0.4257560968399048,
+      "learning_rate": 1.8571281795480632e-05,
+      "loss": 0.7213,
+      "step": 5450
+    },
+    {
+      "epoch": 0.3620929769878639,
+      "grad_norm": 0.4763351082801819,
+      "learning_rate": 1.8303470518362808e-05,
+      "loss": 0.6993,
+      "step": 5460
+    },
+    {
+      "epoch": 0.36275615093839114,
+      "grad_norm": 0.4287746846675873,
+      "learning_rate": 1.801523457828271e-05,
+      "loss": 0.6837,
+      "step": 5470
+    },
+    {
+      "epoch": 0.36341932488891837,
+      "grad_norm": 0.41130346059799194,
+      "learning_rate": 1.7707380047081796e-05,
+      "loss": 0.6783,
+      "step": 5480
+    },
+    {
+      "epoch": 0.3640824988394456,
+      "grad_norm": 0.42945313453674316,
+      "learning_rate": 1.7380767861355252e-05,
+      "loss": 0.6811,
+      "step": 5490
+    },
+    {
+      "epoch": 0.36474567278997283,
+      "grad_norm": 0.4491105377674103,
+      "learning_rate": 1.703631141478299e-05,
+      "loss": 0.718,
+      "step": 5500
+    },
+    {
+      "epoch": 0.36540884674050006,
+      "grad_norm": 0.44828975200653076,
+      "learning_rate": 1.6674974003760705e-05,
+      "loss": 0.6942,
+      "step": 5510
+    },
+    {
+      "epoch": 0.36607202069102723,
+      "grad_norm": 0.4078657925128937,
+      "learning_rate": 1.6297766133474556e-05,
+      "loss": 0.7016,
+      "step": 5520
+    },
+    {
+      "epoch": 0.36673519464155446,
+      "grad_norm": 0.479805588722229,
+      "learning_rate": 1.5905742691953186e-05,
+      "loss": 0.6724,
+      "step": 5530
+    },
+    {
+      "epoch": 0.3673983685920817,
+      "grad_norm": 0.4268275797367096,
+      "learning_rate": 1.55e-05,
+      "loss": 0.7033,
+      "step": 5540
+    },
+    {
+      "epoch": 0.3680615425426089,
+      "grad_norm": 0.4907868504524231,
+      "learning_rate": 1.5081672745255862e-05,
+      "loss": 0.7065,
+      "step": 5550
+    },
+    {
+      "epoch": 0.36872471649313615,
+      "grad_norm": 0.4120388329029083,
+      "learning_rate": 1.4651930808966206e-05,
+      "loss": 0.672,
+      "step": 5560
+    },
+    {
+      "epoch": 0.3693878904436634,
+      "grad_norm": 0.4049204885959625,
+      "learning_rate": 1.4211975994326848e-05,
+      "loss": 0.681,
+      "step": 5570
+    },
+    {
+      "epoch": 0.3700510643941906,
+      "grad_norm": 0.39874541759490967,
+      "learning_rate": 1.376303866555779e-05,
+      "loss": 0.7047,
+      "step": 5580
+    },
+    {
+      "epoch": 0.37071423834471784,
+      "grad_norm": 0.44057223200798035,
+      "learning_rate": 1.3306374307104164e-05,
+      "loss": 0.6986,
+      "step": 5590
+    },
+    {
+      "epoch": 0.37137741229524507,
+      "grad_norm": 0.46499621868133545,
+      "learning_rate": 1.2843260012586718e-05,
+      "loss": 0.7048,
+      "step": 5600
+    },
+    {
+      "epoch": 0.37204058624577224,
+      "grad_norm": 0.39177006483078003,
+      "learning_rate": 1.2374990913320726e-05,
+      "loss": 0.705,
+      "step": 5610
+    },
+    {
+      "epoch": 0.37270376019629947,
+      "grad_norm": 0.4431237578392029,
+      "learning_rate": 1.1902876556391247e-05,
+      "loss": 0.6792,
+      "step": 5620
+    },
+    {
+      "epoch": 0.3733669341468267,
+      "grad_norm": 0.47281622886657715,
+      "learning_rate": 1.1428237242413685e-05,
+      "loss": 0.6847,
+      "step": 5630
+    },
+    {
+      "epoch": 0.37403010809735393,
+      "grad_norm": 0.43715614080429077,
+      "learning_rate": 1.095240033322132e-05,
+      "loss": 0.6952,
+      "step": 5640
+    },
+    {
+      "epoch": 0.37469328204788116,
+      "grad_norm": 0.6536780595779419,
+      "learning_rate": 1.047669653980572e-05,
+      "loss": 0.6789,
+      "step": 5650
+    },
+    {
+      "epoch": 0.3753564559984084,
+      "grad_norm": 0.42568010091781616,
+      "learning_rate": 1.0002456200890902e-05,
+      "loss": 0.6734,
+      "step": 5660
+    },
+    {
+      "epoch": 0.3760196299489356,
+      "grad_norm": 0.4551185071468353,
+      "learning_rate": 9.531005562548668e-06,
+      "loss": 0.691,
+      "step": 5670
+    },
+    {
+      "epoch": 0.37668280389946285,
+      "grad_norm": 0.4053123891353607,
+      "learning_rate": 9.06366306925925e-06,
+      "loss": 0.6905,
+      "step": 5680
+    },
+    {
+      "epoch": 0.3773459778499901,
+      "grad_norm": 0.44013091921806335,
+      "learning_rate": 8.601735676789686e-06,
+      "loss": 0.7033,
+      "step": 5690
+    },
+    {
+      "epoch": 0.37800915180051725,
+      "grad_norm": 0.4361964762210846,
+      "learning_rate": 8.14651519720118e-06,
+      "loss": 0.7004,
+      "step": 5700
+    },
+    {
+      "epoch": 0.3786723257510445,
+      "grad_norm": 0.4564845561981201,
+      "learning_rate": 7.699274686206866e-06,
+      "loss": 0.6906,
+      "step": 5710
+    },
+    {
+      "epoch": 0.3793354997015717,
+      "grad_norm": 0.4334389865398407,
+      "learning_rate": 7.261264882983024e-06,
+      "loss": 0.6925,
+      "step": 5720
+    },
+    {
+      "epoch": 0.37999867365209894,
+      "grad_norm": 0.596119225025177,
+      "learning_rate": 6.8337107123900386e-06,
+      "loss": 0.7364,
+      "step": 5730
+    },
+    {
+      "epoch": 0.38066184760262617,
+      "grad_norm": 0.3939472734928131,
+      "learning_rate": 6.417807859384881e-06,
+      "loss": 0.6926,
+      "step": 5740
+    },
+    {
+      "epoch": 0.3813250215531534,
+      "grad_norm": 0.4060012698173523,
+      "learning_rate": 6.014719425205009e-06,
+      "loss": 0.6877,
+      "step": 5750
+    },
+    {
+      "epoch": 0.3819881955036806,
+      "grad_norm": 0.40652188658714294,
+      "learning_rate": 5.6255726746749275e-06,
+      "loss": 0.6857,
+      "step": 5760
+    },
+    {
+      "epoch": 0.38265136945420786,
+      "grad_norm": 0.40667375922203064,
+      "learning_rate": 5.251455883731728e-06,
+      "loss": 0.6827,
+      "step": 5770
+    },
+    {
+      "epoch": 0.3833145434047351,
+      "grad_norm": 0.4259006083011627,
+      "learning_rate": 4.893415295985812e-06,
+      "loss": 0.6984,
+      "step": 5780
+    },
+    {
+      "epoch": 0.38397771735526226,
+      "grad_norm": 0.4109838306903839,
+      "learning_rate": 4.552452196827885e-06,
+      "loss": 0.6843,
+      "step": 5790
+    },
+    {
+      "epoch": 0.3846408913057895,
+      "grad_norm": 0.4440414607524872,
+      "learning_rate": 4.229520113264785e-06,
+      "loss": 0.6688,
+      "step": 5800
+    },
+    {
+      "epoch": 0.3853040652563167,
+      "grad_norm": 0.43259701132774353,
+      "learning_rate": 3.925522147314915e-06,
+      "loss": 0.6997,
+      "step": 5810
+    },
+    {
+      "epoch": 0.38596723920684395,
+      "grad_norm": 0.4748006761074066,
+      "learning_rate": 3.641308450420675e-06,
+      "loss": 0.683,
+      "step": 5820
+    },
+    {
+      "epoch": 0.3866304131573712,
+      "grad_norm": 0.4007572531700134,
+      "learning_rate": 3.37767384594087e-06,
+      "loss": 0.7087,
+      "step": 5830
+    },
+    {
+      "epoch": 0.3872935871078984,
+      "grad_norm": 0.39761948585510254,
+      "learning_rate": 3.1353556063719357e-06,
+      "loss": 0.6765,
+      "step": 5840
+    },
+    {
+      "epoch": 0.38795676105842564,
+      "grad_norm": 0.4246253967285156,
+      "learning_rate": 2.9150313915141474e-06,
+      "loss": 0.6728,
+      "step": 5850
+    },
+    {
+      "epoch": 0.38861993500895287,
+      "grad_norm": 0.45218634605407715,
+      "learning_rate": 2.717317353348891e-06,
+      "loss": 0.7055,
+      "step": 5860
+    },
+    {
+      "epoch": 0.3892831089594801,
+      "grad_norm": 0.4528957009315491,
+      "learning_rate": 2.5427664129268253e-06,
+      "loss": 0.6893,
+      "step": 5870
+    },
+    {
+      "epoch": 0.38994628291000727,
+      "grad_norm": 0.4231392741203308,
+      "learning_rate": 2.391866714085717e-06,
+      "loss": 0.6885,
+      "step": 5880
+    },
+    {
+      "epoch": 0.3906094568605345,
+      "grad_norm": 0.5117971301078796,
+      "learning_rate": 2.2650402583222753e-06,
+      "loss": 0.6926,
+      "step": 5890
+    },
+    {
+      "epoch": 0.39127263081106173,
+      "grad_norm": 0.4415339529514313,
+      "learning_rate": 2.1626417246356398e-06,
+      "loss": 0.6778,
+      "step": 5900
+    },
+    {
+      "epoch": 0.39193580476158896,
+      "grad_norm": 0.4275270700454712,
+      "learning_rate": 2.084957477642894e-06,
+      "loss": 0.7156,
+      "step": 5910
+    },
+    {
+      "epoch": 0.3925989787121162,
+      "grad_norm": 0.4181906282901764,
+      "learning_rate": 2.0322047667405147e-06,
+      "loss": 0.6977,
+      "step": 5920
+    },
+    {
+      "epoch": 0.3932621526626434,
+      "grad_norm": 0.43344050645828247,
+      "learning_rate": 2.0045311185513342e-06,
+      "loss": 0.6995,
+      "step": 5930
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 5935,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 2035,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.6494000762504348e+19,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}