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": 0.27790784791159395,
+  "eval_steps": 500,
+  "global_step": 5935,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0004682524817381532,
+      "grad_norm": 0.7815176248550415,
+      "learning_rate": 2.910112359550562e-06,
+      "loss": 0.8415,
+      "step": 10
+    },
+    {
+      "epoch": 0.0009365049634763064,
+      "grad_norm": 0.8108879327774048,
+      "learning_rate": 3.921348314606742e-06,
+      "loss": 0.8485,
+      "step": 20
+    },
+    {
+      "epoch": 0.0014047574452144597,
+      "grad_norm": 0.8000619411468506,
+      "learning_rate": 4.932584269662922e-06,
+      "loss": 0.83,
+      "step": 30
+    },
+    {
+      "epoch": 0.0018730099269526129,
+      "grad_norm": 0.7511084675788879,
+      "learning_rate": 5.943820224719102e-06,
+      "loss": 0.8517,
+      "step": 40
+    },
+    {
+      "epoch": 0.002341262408690766,
+      "grad_norm": 0.7430337071418762,
+      "learning_rate": 6.955056179775282e-06,
+      "loss": 0.847,
+      "step": 50
+    },
+    {
+      "epoch": 0.0028095148904289193,
+      "grad_norm": 0.8726410269737244,
+      "learning_rate": 7.966292134831462e-06,
+      "loss": 0.8172,
+      "step": 60
+    },
+    {
+      "epoch": 0.0032777673721670723,
+      "grad_norm": 0.8567508459091187,
+      "learning_rate": 8.977528089887641e-06,
+      "loss": 0.8377,
+      "step": 70
+    },
+    {
+      "epoch": 0.0037460198539052258,
+      "grad_norm": 0.7639268040657043,
+      "learning_rate": 9.988764044943822e-06,
+      "loss": 0.8184,
+      "step": 80
+    },
+    {
+      "epoch": 0.004214272335643379,
+      "grad_norm": 0.9645025134086609,
+      "learning_rate": 1.1000000000000001e-05,
+      "loss": 0.8178,
+      "step": 90
+    },
+    {
+      "epoch": 0.004682524817381532,
+      "grad_norm": 0.801426887512207,
+      "learning_rate": 1.2011235955056182e-05,
+      "loss": 0.8351,
+      "step": 100
+    },
+    {
+      "epoch": 0.005150777299119685,
+      "grad_norm": 0.8018251061439514,
+      "learning_rate": 1.3022471910112362e-05,
+      "loss": 0.8493,
+      "step": 110
+    },
+    {
+      "epoch": 0.005619029780857839,
+      "grad_norm": 1.6218419075012207,
+      "learning_rate": 1.403370786516854e-05,
+      "loss": 0.8367,
+      "step": 120
+    },
+    {
+      "epoch": 0.006087282262595992,
+      "grad_norm": 0.738893449306488,
+      "learning_rate": 1.504494382022472e-05,
+      "loss": 0.8622,
+      "step": 130
+    },
+    {
+      "epoch": 0.006555534744334145,
+      "grad_norm": 0.9413527846336365,
+      "learning_rate": 1.60561797752809e-05,
+      "loss": 0.8024,
+      "step": 140
+    },
+    {
+      "epoch": 0.007023787226072298,
+      "grad_norm": 0.8190211653709412,
+      "learning_rate": 1.706741573033708e-05,
+      "loss": 0.8216,
+      "step": 150
+    },
+    {
+      "epoch": 0.0074920397078104516,
+      "grad_norm": 0.6878272294998169,
+      "learning_rate": 1.8078651685393256e-05,
+      "loss": 0.8609,
+      "step": 160
+    },
+    {
+      "epoch": 0.007960292189548605,
+      "grad_norm": 0.7031309008598328,
+      "learning_rate": 1.908988764044944e-05,
+      "loss": 0.8954,
+      "step": 170
+    },
+    {
+      "epoch": 0.008428544671286758,
+      "grad_norm": 0.7149655818939209,
+      "learning_rate": 2e-05,
+      "loss": 0.8378,
+      "step": 180
+    },
+    {
+      "epoch": 0.008896797153024912,
+      "grad_norm": 0.6629250645637512,
+      "learning_rate": 2e-05,
+      "loss": 0.8002,
+      "step": 190
+    },
+    {
+      "epoch": 0.009365049634763064,
+      "grad_norm": 0.6915798783302307,
+      "learning_rate": 2e-05,
+      "loss": 0.8358,
+      "step": 200
+    },
+    {
+      "epoch": 0.009833302116501217,
+      "grad_norm": 0.6538642644882202,
+      "learning_rate": 2e-05,
+      "loss": 0.8535,
+      "step": 210
+    },
+    {
+      "epoch": 0.01030155459823937,
+      "grad_norm": 0.584514856338501,
+      "learning_rate": 2e-05,
+      "loss": 0.8362,
+      "step": 220
+    },
+    {
+      "epoch": 0.010769807079977524,
+      "grad_norm": 1.3603894710540771,
+      "learning_rate": 2e-05,
+      "loss": 0.8086,
+      "step": 230
+    },
+    {
+      "epoch": 0.011238059561715677,
+      "grad_norm": 0.5758199691772461,
+      "learning_rate": 2e-05,
+      "loss": 0.8336,
+      "step": 240
+    },
+    {
+      "epoch": 0.01170631204345383,
+      "grad_norm": 0.6881021857261658,
+      "learning_rate": 2e-05,
+      "loss": 0.8444,
+      "step": 250
+    },
+    {
+      "epoch": 0.012174564525191984,
+      "grad_norm": 1.9581451416015625,
+      "learning_rate": 2e-05,
+      "loss": 0.8433,
+      "step": 260
+    },
+    {
+      "epoch": 0.012642817006930136,
+      "grad_norm": 0.5425940752029419,
+      "learning_rate": 2e-05,
+      "loss": 0.8213,
+      "step": 270
+    },
+    {
+      "epoch": 0.01311106948866829,
+      "grad_norm": 0.5425044298171997,
+      "learning_rate": 2e-05,
+      "loss": 0.8179,
+      "step": 280
+    },
+    {
+      "epoch": 0.013579321970406443,
+      "grad_norm": 0.657163143157959,
+      "learning_rate": 2e-05,
+      "loss": 0.8649,
+      "step": 290
+    },
+    {
+      "epoch": 0.014047574452144596,
+      "grad_norm": 0.569369375705719,
+      "learning_rate": 2e-05,
+      "loss": 0.8438,
+      "step": 300
+    },
+    {
+      "epoch": 0.01451582693388275,
+      "grad_norm": 0.554995059967041,
+      "learning_rate": 2e-05,
+      "loss": 0.8085,
+      "step": 310
+    },
+    {
+      "epoch": 0.014984079415620903,
+      "grad_norm": 0.5484607219696045,
+      "learning_rate": 2e-05,
+      "loss": 0.8089,
+      "step": 320
+    },
+    {
+      "epoch": 0.015452331897359057,
+      "grad_norm": 0.5753210186958313,
+      "learning_rate": 2e-05,
+      "loss": 0.8142,
+      "step": 330
+    },
+    {
+      "epoch": 0.01592058437909721,
+      "grad_norm": 0.7434160113334656,
+      "learning_rate": 2e-05,
+      "loss": 0.8188,
+      "step": 340
+    },
+    {
+      "epoch": 0.01638883686083536,
+      "grad_norm": 0.574074923992157,
+      "learning_rate": 2e-05,
+      "loss": 0.8563,
+      "step": 350
+    },
+    {
+      "epoch": 0.016857089342573517,
+      "grad_norm": 0.5934783816337585,
+      "learning_rate": 2e-05,
+      "loss": 0.839,
+      "step": 360
+    },
+    {
+      "epoch": 0.01732534182431167,
+      "grad_norm": 2.1430094242095947,
+      "learning_rate": 2e-05,
+      "loss": 0.8244,
+      "step": 370
+    },
+    {
+      "epoch": 0.017793594306049824,
+      "grad_norm": 0.5127322673797607,
+      "learning_rate": 2e-05,
+      "loss": 0.7979,
+      "step": 380
+    },
+    {
+      "epoch": 0.018261846787787975,
+      "grad_norm": 0.5187241435050964,
+      "learning_rate": 2e-05,
+      "loss": 0.8122,
+      "step": 390
+    },
+    {
+      "epoch": 0.018730099269526127,
+      "grad_norm": 0.5075030326843262,
+      "learning_rate": 2e-05,
+      "loss": 0.8645,
+      "step": 400
+    },
+    {
+      "epoch": 0.019198351751264282,
+      "grad_norm": 0.4950161576271057,
+      "learning_rate": 2e-05,
+      "loss": 0.8418,
+      "step": 410
+    },
+    {
+      "epoch": 0.019666604233002434,
+      "grad_norm": 0.5396143794059753,
+      "learning_rate": 2e-05,
+      "loss": 0.8152,
+      "step": 420
+    },
+    {
+      "epoch": 0.02013485671474059,
+      "grad_norm": 0.5515366196632385,
+      "learning_rate": 2e-05,
+      "loss": 0.7911,
+      "step": 430
+    },
+    {
+      "epoch": 0.02060310919647874,
+      "grad_norm": 0.5184299945831299,
+      "learning_rate": 2e-05,
+      "loss": 0.8403,
+      "step": 440
+    },
+    {
+      "epoch": 0.021071361678216896,
+      "grad_norm": 0.4794081449508667,
+      "learning_rate": 2e-05,
+      "loss": 0.8233,
+      "step": 450
+    },
+    {
+      "epoch": 0.021539614159955048,
+      "grad_norm": 0.7160170674324036,
+      "learning_rate": 2e-05,
+      "loss": 0.8302,
+      "step": 460
+    },
+    {
+      "epoch": 0.0220078666416932,
+      "grad_norm": 0.6666805148124695,
+      "learning_rate": 2e-05,
+      "loss": 0.8202,
+      "step": 470
+    },
+    {
+      "epoch": 0.022476119123431355,
+      "grad_norm": 0.5356866717338562,
+      "learning_rate": 2e-05,
+      "loss": 0.8418,
+      "step": 480
+    },
+    {
+      "epoch": 0.022944371605169506,
+      "grad_norm": 0.4900036156177521,
+      "learning_rate": 2e-05,
+      "loss": 0.829,
+      "step": 490
+    },
+    {
+      "epoch": 0.02341262408690766,
+      "grad_norm": 0.47087252140045166,
+      "learning_rate": 2e-05,
+      "loss": 0.8418,
+      "step": 500
+    },
+    {
+      "epoch": 0.023880876568645813,
+      "grad_norm": 0.5223408341407776,
+      "learning_rate": 2e-05,
+      "loss": 0.7874,
+      "step": 510
+    },
+    {
+      "epoch": 0.02434912905038397,
+      "grad_norm": 0.5723909735679626,
+      "learning_rate": 2e-05,
+      "loss": 0.8552,
+      "step": 520
+    },
+    {
+      "epoch": 0.02481738153212212,
+      "grad_norm": 0.4538605213165283,
+      "learning_rate": 2e-05,
+      "loss": 0.7783,
+      "step": 530
+    },
+    {
+      "epoch": 0.025285634013860272,
+      "grad_norm": 0.4878973960876465,
+      "learning_rate": 2e-05,
+      "loss": 0.8085,
+      "step": 540
+    },
+    {
+      "epoch": 0.025753886495598427,
+      "grad_norm": 0.5102939605712891,
+      "learning_rate": 2e-05,
+      "loss": 0.799,
+      "step": 550
+    },
+    {
+      "epoch": 0.02622213897733658,
+      "grad_norm": 0.5069881081581116,
+      "learning_rate": 2e-05,
+      "loss": 0.8201,
+      "step": 560
+    },
+    {
+      "epoch": 0.026690391459074734,
+      "grad_norm": 0.45733290910720825,
+      "learning_rate": 2e-05,
+      "loss": 0.844,
+      "step": 570
+    },
+    {
+      "epoch": 0.027158643940812886,
+      "grad_norm": 0.4658640921115875,
+      "learning_rate": 2e-05,
+      "loss": 0.8291,
+      "step": 580
+    },
+    {
+      "epoch": 0.02762689642255104,
+      "grad_norm": 0.48104459047317505,
+      "learning_rate": 2e-05,
+      "loss": 0.8432,
+      "step": 590
+    },
+    {
+      "epoch": 0.028095148904289192,
+      "grad_norm": 0.46297943592071533,
+      "learning_rate": 2e-05,
+      "loss": 0.8192,
+      "step": 600
+    },
+    {
+      "epoch": 0.028563401386027348,
+      "grad_norm": 0.450825959444046,
+      "learning_rate": 2e-05,
+      "loss": 0.7987,
+      "step": 610
+    },
+    {
+      "epoch": 0.0290316538677655,
+      "grad_norm": 0.4797734022140503,
+      "learning_rate": 2e-05,
+      "loss": 0.8325,
+      "step": 620
+    },
+    {
+      "epoch": 0.02949990634950365,
+      "grad_norm": 0.481893390417099,
+      "learning_rate": 2e-05,
+      "loss": 0.7837,
+      "step": 630
+    },
+    {
+      "epoch": 0.029968158831241806,
+      "grad_norm": 0.47477421164512634,
+      "learning_rate": 2e-05,
+      "loss": 0.8278,
+      "step": 640
+    },
+    {
+      "epoch": 0.030436411312979958,
+      "grad_norm": 0.5964585542678833,
+      "learning_rate": 2e-05,
+      "loss": 0.7969,
+      "step": 650
+    },
+    {
+      "epoch": 0.030904663794718113,
+      "grad_norm": 0.46593981981277466,
+      "learning_rate": 2e-05,
+      "loss": 0.8307,
+      "step": 660
+    },
+    {
+      "epoch": 0.03137291627645627,
+      "grad_norm": 0.48121365904808044,
+      "learning_rate": 2e-05,
+      "loss": 0.8174,
+      "step": 670
+    },
+    {
+      "epoch": 0.03184116875819442,
+      "grad_norm": 0.5038956999778748,
+      "learning_rate": 2e-05,
+      "loss": 0.8339,
+      "step": 680
+    },
+    {
+      "epoch": 0.03230942123993257,
+      "grad_norm": 0.5365706086158752,
+      "learning_rate": 2e-05,
+      "loss": 0.8291,
+      "step": 690
+    },
+    {
+      "epoch": 0.03277767372167072,
+      "grad_norm": 0.5319100618362427,
+      "learning_rate": 2e-05,
+      "loss": 0.7953,
+      "step": 700
+    },
+    {
+      "epoch": 0.033245926203408875,
+      "grad_norm": 0.4542364776134491,
+      "learning_rate": 2e-05,
+      "loss": 0.8263,
+      "step": 710
+    },
+    {
+      "epoch": 0.033714178685147034,
+      "grad_norm": 0.47826236486434937,
+      "learning_rate": 2e-05,
+      "loss": 0.8321,
+      "step": 720
+    },
+    {
+      "epoch": 0.034182431166885185,
+      "grad_norm": 0.44108861684799194,
+      "learning_rate": 2e-05,
+      "loss": 0.7989,
+      "step": 730
+    },
+    {
+      "epoch": 0.03465068364862334,
+      "grad_norm": 0.4653085172176361,
+      "learning_rate": 2e-05,
+      "loss": 0.7994,
+      "step": 740
+    },
+    {
+      "epoch": 0.03511893613036149,
+      "grad_norm": 0.5099039077758789,
+      "learning_rate": 2e-05,
+      "loss": 0.7591,
+      "step": 750
+    },
+    {
+      "epoch": 0.03558718861209965,
+      "grad_norm": 0.48215553164482117,
+      "learning_rate": 2e-05,
+      "loss": 0.7921,
+      "step": 760
+    },
+    {
+      "epoch": 0.0360554410938378,
+      "grad_norm": 0.46062806248664856,
+      "learning_rate": 2e-05,
+      "loss": 0.8244,
+      "step": 770
+    },
+    {
+      "epoch": 0.03652369357557595,
+      "grad_norm": 0.5108964443206787,
+      "learning_rate": 2e-05,
+      "loss": 0.8667,
+      "step": 780
+    },
+    {
+      "epoch": 0.0369919460573141,
+      "grad_norm": 0.47737154364585876,
+      "learning_rate": 2e-05,
+      "loss": 0.8246,
+      "step": 790
+    },
+    {
+      "epoch": 0.037460198539052254,
+      "grad_norm": 0.5114936828613281,
+      "learning_rate": 2e-05,
+      "loss": 0.8188,
+      "step": 800
+    },
+    {
+      "epoch": 0.03792845102079041,
+      "grad_norm": 0.47046583890914917,
+      "learning_rate": 2e-05,
+      "loss": 0.8393,
+      "step": 810
+    },
+    {
+      "epoch": 0.038396703502528565,
+      "grad_norm": 0.4314744472503662,
+      "learning_rate": 2e-05,
+      "loss": 0.7953,
+      "step": 820
+    },
+    {
+      "epoch": 0.038864955984266716,
+      "grad_norm": 0.4930090308189392,
+      "learning_rate": 2e-05,
+      "loss": 0.8048,
+      "step": 830
+    },
+    {
+      "epoch": 0.03933320846600487,
+      "grad_norm": 0.4331071078777313,
+      "learning_rate": 2e-05,
+      "loss": 0.8389,
+      "step": 840
+    },
+    {
+      "epoch": 0.03980146094774302,
+      "grad_norm": 0.4713480472564697,
+      "learning_rate": 2e-05,
+      "loss": 0.8165,
+      "step": 850
+    },
+    {
+      "epoch": 0.04026971342948118,
+      "grad_norm": 0.4679141938686371,
+      "learning_rate": 2e-05,
+      "loss": 0.8198,
+      "step": 860
+    },
+    {
+      "epoch": 0.04073796591121933,
+      "grad_norm": 0.7808457612991333,
+      "learning_rate": 2e-05,
+      "loss": 0.8528,
+      "step": 870
+    },
+    {
+      "epoch": 0.04120621839295748,
+      "grad_norm": 0.44788986444473267,
+      "learning_rate": 2e-05,
+      "loss": 0.8225,
+      "step": 880
+    },
+    {
+      "epoch": 0.041674470874695634,
+      "grad_norm": 0.480165034532547,
+      "learning_rate": 2e-05,
+      "loss": 0.8185,
+      "step": 890
+    },
+    {
+      "epoch": 0.04214272335643379,
+      "grad_norm": 0.46314555406570435,
+      "learning_rate": 2e-05,
+      "loss": 0.8116,
+      "step": 900
+    },
+    {
+      "epoch": 0.042610975838171944,
+      "grad_norm": 0.6263248920440674,
+      "learning_rate": 2e-05,
+      "loss": 0.8418,
+      "step": 910
+    },
+    {
+      "epoch": 0.043079228319910096,
+      "grad_norm": 0.4619986414909363,
+      "learning_rate": 2e-05,
+      "loss": 0.8378,
+      "step": 920
+    },
+    {
+      "epoch": 0.04354748080164825,
+      "grad_norm": 0.4389607906341553,
+      "learning_rate": 2e-05,
+      "loss": 0.8153,
+      "step": 930
+    },
+    {
+      "epoch": 0.0440157332833864,
+      "grad_norm": 0.5322599411010742,
+      "learning_rate": 2e-05,
+      "loss": 0.8049,
+      "step": 940
+    },
+    {
+      "epoch": 0.04448398576512456,
+      "grad_norm": 0.4362427890300751,
+      "learning_rate": 2e-05,
+      "loss": 0.7895,
+      "step": 950
+    },
+    {
+      "epoch": 0.04495223824686271,
+      "grad_norm": 0.560736358165741,
+      "learning_rate": 2e-05,
+      "loss": 0.7834,
+      "step": 960
+    },
+    {
+      "epoch": 0.04542049072860086,
+      "grad_norm": 0.49655383825302124,
+      "learning_rate": 2e-05,
+      "loss": 0.8419,
+      "step": 970
+    },
+    {
+      "epoch": 0.04588874321033901,
+      "grad_norm": 0.447719931602478,
+      "learning_rate": 2e-05,
+      "loss": 0.8268,
+      "step": 980
+    },
+    {
+      "epoch": 0.04635699569207717,
+      "grad_norm": 0.472380131483078,
+      "learning_rate": 2e-05,
+      "loss": 0.797,
+      "step": 990
+    },
+    {
+      "epoch": 0.04682524817381532,
+      "grad_norm": 0.4404251277446747,
+      "learning_rate": 2e-05,
+      "loss": 0.8216,
+      "step": 1000
+    },
+    {
+      "epoch": 0.047293500655553475,
+      "grad_norm": 0.8469703793525696,
+      "learning_rate": 2e-05,
+      "loss": 0.8061,
+      "step": 1010
+    },
+    {
+      "epoch": 0.047761753137291627,
+      "grad_norm": 0.49690261483192444,
+      "learning_rate": 2e-05,
+      "loss": 0.8739,
+      "step": 1020
+    },
+    {
+      "epoch": 0.04823000561902978,
+      "grad_norm": 0.4546132981777191,
+      "learning_rate": 2e-05,
+      "loss": 0.8611,
+      "step": 1030
+    },
+    {
+      "epoch": 0.04869825810076794,
+      "grad_norm": 0.9852249026298523,
+      "learning_rate": 2e-05,
+      "loss": 0.8151,
+      "step": 1040
+    },
+    {
+      "epoch": 0.04916651058250609,
+      "grad_norm": 0.4317361116409302,
+      "learning_rate": 2e-05,
+      "loss": 0.8253,
+      "step": 1050
+    },
+    {
+      "epoch": 0.04963476306424424,
+      "grad_norm": 0.48968589305877686,
+      "learning_rate": 2e-05,
+      "loss": 0.8241,
+      "step": 1060
+    },
+    {
+      "epoch": 0.05010301554598239,
+      "grad_norm": 0.5702976584434509,
+      "learning_rate": 2e-05,
+      "loss": 0.788,
+      "step": 1070
+    },
+    {
+      "epoch": 0.050571268027720544,
+      "grad_norm": 0.4381064474582672,
+      "learning_rate": 2e-05,
+      "loss": 0.7815,
+      "step": 1080
+    },
+    {
+      "epoch": 0.0510395205094587,
+      "grad_norm": 0.48637041449546814,
+      "learning_rate": 2e-05,
+      "loss": 0.8424,
+      "step": 1090
+    },
+    {
+      "epoch": 0.051507772991196854,
+      "grad_norm": 0.4259801506996155,
+      "learning_rate": 2e-05,
+      "loss": 0.817,
+      "step": 1100
+    },
+    {
+      "epoch": 0.051976025472935006,
+      "grad_norm": 0.4507712423801422,
+      "learning_rate": 2e-05,
+      "loss": 0.856,
+      "step": 1110
+    },
+    {
+      "epoch": 0.05244427795467316,
+      "grad_norm": 0.4860295057296753,
+      "learning_rate": 2e-05,
+      "loss": 0.8139,
+      "step": 1120
+    },
+    {
+      "epoch": 0.052912530436411316,
+      "grad_norm": 0.5244073271751404,
+      "learning_rate": 2e-05,
+      "loss": 0.8165,
+      "step": 1130
+    },
+    {
+      "epoch": 0.05338078291814947,
+      "grad_norm": 0.4211806654930115,
+      "learning_rate": 2e-05,
+      "loss": 0.7845,
+      "step": 1140
+    },
+    {
+      "epoch": 0.05384903539988762,
+      "grad_norm": 0.4540943205356598,
+      "learning_rate": 2e-05,
+      "loss": 0.7812,
+      "step": 1150
+    },
+    {
+      "epoch": 0.05431728788162577,
+      "grad_norm": 0.4845108985900879,
+      "learning_rate": 2e-05,
+      "loss": 0.8138,
+      "step": 1160
+    },
+    {
+      "epoch": 0.05478554036336392,
+      "grad_norm": 0.46812689304351807,
+      "learning_rate": 2e-05,
+      "loss": 0.8065,
+      "step": 1170
+    },
+    {
+      "epoch": 0.05525379284510208,
+      "grad_norm": 0.42959365248680115,
+      "learning_rate": 2e-05,
+      "loss": 0.8089,
+      "step": 1180
+    },
+    {
+      "epoch": 0.05572204532684023,
+      "grad_norm": 0.7998138070106506,
+      "learning_rate": 2e-05,
+      "loss": 0.8137,
+      "step": 1190
+    },
+    {
+      "epoch": 0.056190297808578385,
+      "grad_norm": 0.46024176478385925,
+      "learning_rate": 2e-05,
+      "loss": 0.8341,
+      "step": 1200
+    },
+    {
+      "epoch": 0.05665855029031654,
+      "grad_norm": 0.4920409023761749,
+      "learning_rate": 2e-05,
+      "loss": 0.8328,
+      "step": 1210
+    },
+    {
+      "epoch": 0.057126802772054695,
+      "grad_norm": 0.44345328211784363,
+      "learning_rate": 2e-05,
+      "loss": 0.8242,
+      "step": 1220
+    },
+    {
+      "epoch": 0.05759505525379285,
+      "grad_norm": 0.4638558030128479,
+      "learning_rate": 2e-05,
+      "loss": 0.8111,
+      "step": 1230
+    },
+    {
+      "epoch": 0.058063307735531,
+      "grad_norm": 0.6508961319923401,
+      "learning_rate": 2e-05,
+      "loss": 0.809,
+      "step": 1240
+    },
+    {
+      "epoch": 0.05853156021726915,
+      "grad_norm": 0.4465144872665405,
+      "learning_rate": 2e-05,
+      "loss": 0.8195,
+      "step": 1250
+    },
+    {
+      "epoch": 0.0589998126990073,
+      "grad_norm": 0.4405592978000641,
+      "learning_rate": 2e-05,
+      "loss": 0.8352,
+      "step": 1260
+    },
+    {
+      "epoch": 0.05946806518074546,
+      "grad_norm": 0.43743807077407837,
+      "learning_rate": 2e-05,
+      "loss": 0.8149,
+      "step": 1270
+    },
+    {
+      "epoch": 0.05993631766248361,
+      "grad_norm": 0.5173885226249695,
+      "learning_rate": 2e-05,
+      "loss": 0.8421,
+      "step": 1280
+    },
+    {
+      "epoch": 0.060404570144221764,
+      "grad_norm": 0.41101977229118347,
+      "learning_rate": 2e-05,
+      "loss": 0.8493,
+      "step": 1290
+    },
+    {
+      "epoch": 0.060872822625959916,
+      "grad_norm": 0.41480135917663574,
+      "learning_rate": 2e-05,
+      "loss": 0.7971,
+      "step": 1300
+    },
+    {
+      "epoch": 0.06134107510769807,
+      "grad_norm": 0.5274170637130737,
+      "learning_rate": 2e-05,
+      "loss": 0.8106,
+      "step": 1310
+    },
+    {
+      "epoch": 0.061809327589436226,
+      "grad_norm": 0.4326211214065552,
+      "learning_rate": 2e-05,
+      "loss": 0.7927,
+      "step": 1320
+    },
+    {
+      "epoch": 0.06227758007117438,
+      "grad_norm": 0.4117155373096466,
+      "learning_rate": 2e-05,
+      "loss": 0.7877,
+      "step": 1330
+    },
+    {
+      "epoch": 0.06274583255291254,
+      "grad_norm": 0.4453534483909607,
+      "learning_rate": 2e-05,
+      "loss": 0.8417,
+      "step": 1340
+    },
+    {
+      "epoch": 0.06321408503465069,
+      "grad_norm": 0.6158928871154785,
+      "learning_rate": 2e-05,
+      "loss": 0.7797,
+      "step": 1350
+    },
+    {
+      "epoch": 0.06368233751638884,
+      "grad_norm": 0.43932247161865234,
+      "learning_rate": 2e-05,
+      "loss": 0.8042,
+      "step": 1360
+    },
+    {
+      "epoch": 0.06415058999812699,
+      "grad_norm": 0.5323647856712341,
+      "learning_rate": 2e-05,
+      "loss": 0.798,
+      "step": 1370
+    },
+    {
+      "epoch": 0.06461884247986514,
+      "grad_norm": 0.4197414517402649,
+      "learning_rate": 2e-05,
+      "loss": 0.8223,
+      "step": 1380
+    },
+    {
+      "epoch": 0.0650870949616033,
+      "grad_norm": 0.48507630825042725,
+      "learning_rate": 2e-05,
+      "loss": 0.8105,
+      "step": 1390
+    },
+    {
+      "epoch": 0.06555534744334145,
+      "grad_norm": 0.47112739086151123,
+      "learning_rate": 2e-05,
+      "loss": 0.8249,
+      "step": 1400
+    },
+    {
+      "epoch": 0.0660235999250796,
+      "grad_norm": 0.48371198773384094,
+      "learning_rate": 2e-05,
+      "loss": 0.8105,
+      "step": 1410
+    },
+    {
+      "epoch": 0.06649185240681775,
+      "grad_norm": 0.43702569603919983,
+      "learning_rate": 2e-05,
+      "loss": 0.7906,
+      "step": 1420
+    },
+    {
+      "epoch": 0.06696010488855592,
+      "grad_norm": 0.44430628418922424,
+      "learning_rate": 2e-05,
+      "loss": 0.8313,
+      "step": 1430
+    },
+    {
+      "epoch": 0.06742835737029407,
+      "grad_norm": 0.4579251706600189,
+      "learning_rate": 2e-05,
+      "loss": 0.8415,
+      "step": 1440
+    },
+    {
+      "epoch": 0.06789660985203222,
+      "grad_norm": 0.48323583602905273,
+      "learning_rate": 2e-05,
+      "loss": 0.8341,
+      "step": 1450
+    },
+    {
+      "epoch": 0.06836486233377037,
+      "grad_norm": 0.427022248506546,
+      "learning_rate": 2e-05,
+      "loss": 0.8007,
+      "step": 1460
+    },
+    {
+      "epoch": 0.06883311481550852,
+      "grad_norm": 0.4505695104598999,
+      "learning_rate": 2e-05,
+      "loss": 0.8095,
+      "step": 1470
+    },
+    {
+      "epoch": 0.06930136729724667,
+      "grad_norm": 0.45807015895843506,
+      "learning_rate": 2e-05,
+      "loss": 0.7785,
+      "step": 1480
+    },
+    {
+      "epoch": 0.06976961977898483,
+      "grad_norm": 0.5102551579475403,
+      "learning_rate": 2e-05,
+      "loss": 0.806,
+      "step": 1490
+    },
+    {
+      "epoch": 0.07023787226072298,
+      "grad_norm": 0.42989158630371094,
+      "learning_rate": 2e-05,
+      "loss": 0.8295,
+      "step": 1500
+    },
+    {
+      "epoch": 0.07070612474246113,
+      "grad_norm": 0.5128605961799622,
+      "learning_rate": 2e-05,
+      "loss": 0.8218,
+      "step": 1510
+    },
+    {
+      "epoch": 0.0711743772241993,
+      "grad_norm": 0.4889835715293884,
+      "learning_rate": 2e-05,
+      "loss": 0.7938,
+      "step": 1520
+    },
+    {
+      "epoch": 0.07164262970593745,
+      "grad_norm": 0.4175378680229187,
+      "learning_rate": 2e-05,
+      "loss": 0.8033,
+      "step": 1530
+    },
+    {
+      "epoch": 0.0721108821876756,
+      "grad_norm": 0.4378540813922882,
+      "learning_rate": 2e-05,
+      "loss": 0.8042,
+      "step": 1540
+    },
+    {
+      "epoch": 0.07257913466941375,
+      "grad_norm": 0.4375297427177429,
+      "learning_rate": 2e-05,
+      "loss": 0.8218,
+      "step": 1550
+    },
+    {
+      "epoch": 0.0730473871511519,
+      "grad_norm": 0.49845996499061584,
+      "learning_rate": 2e-05,
+      "loss": 0.8044,
+      "step": 1560
+    },
+    {
+      "epoch": 0.07351563963289005,
+      "grad_norm": 0.46285390853881836,
+      "learning_rate": 2e-05,
+      "loss": 0.7989,
+      "step": 1570
+    },
+    {
+      "epoch": 0.0739838921146282,
+      "grad_norm": 0.46297210454940796,
+      "learning_rate": 2e-05,
+      "loss": 0.8065,
+      "step": 1580
+    },
+    {
+      "epoch": 0.07445214459636636,
+      "grad_norm": 0.5150440335273743,
+      "learning_rate": 2e-05,
+      "loss": 0.8284,
+      "step": 1590
+    },
+    {
+      "epoch": 0.07492039707810451,
+      "grad_norm": 0.47179800271987915,
+      "learning_rate": 2e-05,
+      "loss": 0.8059,
+      "step": 1600
+    },
+    {
+      "epoch": 0.07538864955984266,
+      "grad_norm": 0.4652591347694397,
+      "learning_rate": 2e-05,
+      "loss": 0.8229,
+      "step": 1610
+    },
+    {
+      "epoch": 0.07585690204158083,
+      "grad_norm": 0.4540858864784241,
+      "learning_rate": 2e-05,
+      "loss": 0.7697,
+      "step": 1620
+    },
+    {
+      "epoch": 0.07632515452331898,
+      "grad_norm": 0.4078024923801422,
+      "learning_rate": 2e-05,
+      "loss": 0.8106,
+      "step": 1630
+    },
+    {
+      "epoch": 0.07679340700505713,
+      "grad_norm": 0.44727492332458496,
+      "learning_rate": 2e-05,
+      "loss": 0.8216,
+      "step": 1640
+    },
+    {
+      "epoch": 0.07726165948679528,
+      "grad_norm": 0.46308594942092896,
+      "learning_rate": 2e-05,
+      "loss": 0.8184,
+      "step": 1650
+    },
+    {
+      "epoch": 0.07772991196853343,
+      "grad_norm": 0.4770665764808655,
+      "learning_rate": 2e-05,
+      "loss": 0.7977,
+      "step": 1660
+    },
+    {
+      "epoch": 0.07819816445027158,
+      "grad_norm": 0.4776230752468109,
+      "learning_rate": 2e-05,
+      "loss": 0.7921,
+      "step": 1670
+    },
+    {
+      "epoch": 0.07866641693200974,
+      "grad_norm": 0.5174942016601562,
+      "learning_rate": 2e-05,
+      "loss": 0.8066,
+      "step": 1680
+    },
+    {
+      "epoch": 0.07913466941374789,
+      "grad_norm": 0.43850815296173096,
+      "learning_rate": 2e-05,
+      "loss": 0.8068,
+      "step": 1690
+    },
+    {
+      "epoch": 0.07960292189548604,
+      "grad_norm": 1.0964583158493042,
+      "learning_rate": 2e-05,
+      "loss": 0.7769,
+      "step": 1700
+    },
+    {
+      "epoch": 0.0800711743772242,
+      "grad_norm": 0.4482307434082031,
+      "learning_rate": 2e-05,
+      "loss": 0.7822,
+      "step": 1710
+    },
+    {
+      "epoch": 0.08053942685896236,
+      "grad_norm": 0.5412369966506958,
+      "learning_rate": 2e-05,
+      "loss": 0.7915,
+      "step": 1720
+    },
+    {
+      "epoch": 0.08100767934070051,
+      "grad_norm": 0.4515257179737091,
+      "learning_rate": 2e-05,
+      "loss": 0.8215,
+      "step": 1730
+    },
+    {
+      "epoch": 0.08147593182243866,
+      "grad_norm": 0.4404001533985138,
+      "learning_rate": 2e-05,
+      "loss": 0.8162,
+      "step": 1740
+    },
+    {
+      "epoch": 0.08194418430417681,
+      "grad_norm": 0.4400689899921417,
+      "learning_rate": 2e-05,
+      "loss": 0.7876,
+      "step": 1750
+    },
+    {
+      "epoch": 0.08241243678591496,
+      "grad_norm": 0.4302707314491272,
+      "learning_rate": 2e-05,
+      "loss": 0.8352,
+      "step": 1760
+    },
+    {
+      "epoch": 0.08288068926765312,
+      "grad_norm": 0.4659354090690613,
+      "learning_rate": 2e-05,
+      "loss": 0.7758,
+      "step": 1770
+    },
+    {
+      "epoch": 0.08334894174939127,
+      "grad_norm": 0.4827459752559662,
+      "learning_rate": 2e-05,
+      "loss": 0.7892,
+      "step": 1780
+    },
+    {
+      "epoch": 0.08381719423112942,
+      "grad_norm": 1.0690308809280396,
+      "learning_rate": 2e-05,
+      "loss": 0.8228,
+      "step": 1790
+    },
+    {
+      "epoch": 0.08428544671286758,
+      "grad_norm": 0.4485591650009155,
+      "learning_rate": 2e-05,
+      "loss": 0.7583,
+      "step": 1800
+    },
+    {
+      "epoch": 0.08475369919460574,
+      "grad_norm": 0.44925209879875183,
+      "learning_rate": 2e-05,
+      "loss": 0.7587,
+      "step": 1810
+    },
+    {
+      "epoch": 0.08522195167634389,
+      "grad_norm": 0.4562433063983917,
+      "learning_rate": 2e-05,
+      "loss": 0.7677,
+      "step": 1820
+    },
+    {
+      "epoch": 0.08569020415808204,
+      "grad_norm": 0.45940324664115906,
+      "learning_rate": 2e-05,
+      "loss": 0.7985,
+      "step": 1830
+    },
+    {
+      "epoch": 0.08615845663982019,
+      "grad_norm": 0.4248304069042206,
+      "learning_rate": 2e-05,
+      "loss": 0.7445,
+      "step": 1840
+    },
+    {
+      "epoch": 0.08662670912155834,
+      "grad_norm": 0.48499906063079834,
+      "learning_rate": 2e-05,
+      "loss": 0.8037,
+      "step": 1850
+    },
+    {
+      "epoch": 0.0870949616032965,
+      "grad_norm": 0.43459802865982056,
+      "learning_rate": 2e-05,
+      "loss": 0.7893,
+      "step": 1860
+    },
+    {
+      "epoch": 0.08756321408503465,
+      "grad_norm": 0.521557629108429,
+      "learning_rate": 2e-05,
+      "loss": 0.79,
+      "step": 1870
+    },
+    {
+      "epoch": 0.0880314665667728,
+      "grad_norm": 0.424723744392395,
+      "learning_rate": 2e-05,
+      "loss": 0.7598,
+      "step": 1880
+    },
+    {
+      "epoch": 0.08849971904851096,
+      "grad_norm": 0.4430510401725769,
+      "learning_rate": 2e-05,
+      "loss": 0.8156,
+      "step": 1890
+    },
+    {
+      "epoch": 0.08896797153024912,
+      "grad_norm": 0.43703287839889526,
+      "learning_rate": 2e-05,
+      "loss": 0.8031,
+      "step": 1900
+    },
+    {
+      "epoch": 0.08943622401198727,
+      "grad_norm": 0.46375030279159546,
+      "learning_rate": 2e-05,
+      "loss": 0.7903,
+      "step": 1910
+    },
+    {
+      "epoch": 0.08990447649372542,
+      "grad_norm": 0.46522295475006104,
+      "learning_rate": 2e-05,
+      "loss": 0.7418,
+      "step": 1920
+    },
+    {
+      "epoch": 0.09037272897546357,
+      "grad_norm": 0.40295276045799255,
+      "learning_rate": 2e-05,
+      "loss": 0.781,
+      "step": 1930
+    },
+    {
+      "epoch": 0.09084098145720172,
+      "grad_norm": 0.42140400409698486,
+      "learning_rate": 2e-05,
+      "loss": 0.7899,
+      "step": 1940
+    },
+    {
+      "epoch": 0.09130923393893987,
+      "grad_norm": 0.6773248314857483,
+      "learning_rate": 2e-05,
+      "loss": 0.8514,
+      "step": 1950
+    },
+    {
+      "epoch": 0.09177748642067803,
+      "grad_norm": 0.46820682287216187,
+      "learning_rate": 2e-05,
+      "loss": 0.8016,
+      "step": 1960
+    },
+    {
+      "epoch": 0.09224573890241618,
+      "grad_norm": 0.4235709011554718,
+      "learning_rate": 2e-05,
+      "loss": 0.7933,
+      "step": 1970
+    },
+    {
+      "epoch": 0.09271399138415434,
+      "grad_norm": 0.43302154541015625,
+      "learning_rate": 2e-05,
+      "loss": 0.7481,
+      "step": 1980
+    },
+    {
+      "epoch": 0.0931822438658925,
+      "grad_norm": 0.4129074215888977,
+      "learning_rate": 2e-05,
+      "loss": 0.7644,
+      "step": 1990
+    },
+    {
+      "epoch": 0.09365049634763065,
+      "grad_norm": 0.47755202651023865,
+      "learning_rate": 2e-05,
+      "loss": 0.7979,
+      "step": 2000
+    },
+    {
+      "epoch": 0.0941187488293688,
+      "grad_norm": 0.4275096356868744,
+      "learning_rate": 2e-05,
+      "loss": 0.7869,
+      "step": 2010
+    },
+    {
+      "epoch": 0.09458700131110695,
+      "grad_norm": 0.41398802399635315,
+      "learning_rate": 2e-05,
+      "loss": 0.7833,
+      "step": 2020
+    },
+    {
+      "epoch": 0.0950552537928451,
+      "grad_norm": 0.4200058877468109,
+      "learning_rate": 2e-05,
+      "loss": 0.791,
+      "step": 2030
+    },
+    {
+      "epoch": 0.09552350627458325,
+      "grad_norm": 0.4077129364013672,
+      "learning_rate": 2e-05,
+      "loss": 0.7645,
+      "step": 2040
+    },
+    {
+      "epoch": 0.0959917587563214,
+      "grad_norm": 0.416229248046875,
+      "learning_rate": 2e-05,
+      "loss": 0.7874,
+      "step": 2050
+    },
+    {
+      "epoch": 0.09646001123805956,
+      "grad_norm": 0.4201897382736206,
+      "learning_rate": 2e-05,
+      "loss": 0.7936,
+      "step": 2060
+    },
+    {
+      "epoch": 0.09692826371979771,
+      "grad_norm": 0.44387736916542053,
+      "learning_rate": 2e-05,
+      "loss": 0.7972,
+      "step": 2070
+    },
+    {
+      "epoch": 0.09739651620153587,
+      "grad_norm": 0.4564218521118164,
+      "learning_rate": 2e-05,
+      "loss": 0.8083,
+      "step": 2080
+    },
+    {
+      "epoch": 0.09786476868327403,
+      "grad_norm": 0.4306756854057312,
+      "learning_rate": 2e-05,
+      "loss": 0.7695,
+      "step": 2090
+    },
+    {
+      "epoch": 0.09833302116501218,
+      "grad_norm": 0.47157105803489685,
+      "learning_rate": 2e-05,
+      "loss": 0.7916,
+      "step": 2100
+    },
+    {
+      "epoch": 0.09880127364675033,
+      "grad_norm": 0.5025231838226318,
+      "learning_rate": 2e-05,
+      "loss": 0.7783,
+      "step": 2110
+    },
+    {
+      "epoch": 0.09926952612848848,
+      "grad_norm": 0.49327343702316284,
+      "learning_rate": 2e-05,
+      "loss": 0.7915,
+      "step": 2120
+    },
+    {
+      "epoch": 0.09973777861022663,
+      "grad_norm": 0.4390990734100342,
+      "learning_rate": 2e-05,
+      "loss": 0.7971,
+      "step": 2130
+    },
+    {
+      "epoch": 0.10020603109196478,
+      "grad_norm": 0.4255847930908203,
+      "learning_rate": 2e-05,
+      "loss": 0.7539,
+      "step": 2140
+    },
+    {
+      "epoch": 0.10067428357370294,
+      "grad_norm": 0.4064248502254486,
+      "learning_rate": 2e-05,
+      "loss": 0.8041,
+      "step": 2150
+    },
+    {
+      "epoch": 0.10114253605544109,
+      "grad_norm": 0.4446522891521454,
+      "learning_rate": 2e-05,
+      "loss": 0.7974,
+      "step": 2160
+    },
+    {
+      "epoch": 0.10161078853717925,
+      "grad_norm": 0.4395385980606079,
+      "learning_rate": 2e-05,
+      "loss": 0.7771,
+      "step": 2170
+    },
+    {
+      "epoch": 0.1020790410189174,
+      "grad_norm": 0.4272850751876831,
+      "learning_rate": 2e-05,
+      "loss": 0.8522,
+      "step": 2180
+    },
+    {
+      "epoch": 0.10254729350065556,
+      "grad_norm": 0.3931410014629364,
+      "learning_rate": 2e-05,
+      "loss": 0.8056,
+      "step": 2190
+    },
+    {
+      "epoch": 0.10301554598239371,
+      "grad_norm": 0.46096065640449524,
+      "learning_rate": 2e-05,
+      "loss": 0.7735,
+      "step": 2200
+    },
+    {
+      "epoch": 0.10348379846413186,
+      "grad_norm": 0.458321750164032,
+      "learning_rate": 2e-05,
+      "loss": 0.8037,
+      "step": 2210
+    },
+    {
+      "epoch": 0.10395205094587001,
+      "grad_norm": 0.4780912697315216,
+      "learning_rate": 2e-05,
+      "loss": 0.8117,
+      "step": 2220
+    },
+    {
+      "epoch": 0.10442030342760816,
+      "grad_norm": 0.4787966012954712,
+      "learning_rate": 2e-05,
+      "loss": 0.7722,
+      "step": 2230
+    },
+    {
+      "epoch": 0.10488855590934631,
+      "grad_norm": 0.44335323572158813,
+      "learning_rate": 2e-05,
+      "loss": 0.8006,
+      "step": 2240
+    },
+    {
+      "epoch": 0.10535680839108447,
+      "grad_norm": 0.46708425879478455,
+      "learning_rate": 2e-05,
+      "loss": 0.8177,
+      "step": 2250
+    },
+    {
+      "epoch": 0.10582506087282263,
+      "grad_norm": 0.41492828726768494,
+      "learning_rate": 2e-05,
+      "loss": 0.8078,
+      "step": 2260
+    },
+    {
+      "epoch": 0.10629331335456078,
+      "grad_norm": 0.4242907762527466,
+      "learning_rate": 2e-05,
+      "loss": 0.8162,
+      "step": 2270
+    },
+    {
+      "epoch": 0.10676156583629894,
+      "grad_norm": 0.43097883462905884,
+      "learning_rate": 2e-05,
+      "loss": 0.7563,
+      "step": 2280
+    },
+    {
+      "epoch": 0.10722981831803709,
+      "grad_norm": 0.3991326689720154,
+      "learning_rate": 2e-05,
+      "loss": 0.7908,
+      "step": 2290
+    },
+    {
+      "epoch": 0.10769807079977524,
+      "grad_norm": 0.5473389625549316,
+      "learning_rate": 2e-05,
+      "loss": 0.7945,
+      "step": 2300
+    },
+    {
+      "epoch": 0.10816632328151339,
+      "grad_norm": 0.46721047163009644,
+      "learning_rate": 2e-05,
+      "loss": 0.798,
+      "step": 2310
+    },
+    {
+      "epoch": 0.10863457576325154,
+      "grad_norm": 0.4458387792110443,
+      "learning_rate": 2e-05,
+      "loss": 0.8095,
+      "step": 2320
+    },
+    {
+      "epoch": 0.1091028282449897,
+      "grad_norm": 0.5851238369941711,
+      "learning_rate": 2e-05,
+      "loss": 0.79,
+      "step": 2330
+    },
+    {
+      "epoch": 0.10957108072672785,
+      "grad_norm": 0.46909311413764954,
+      "learning_rate": 2e-05,
+      "loss": 0.7908,
+      "step": 2340
+    },
+    {
+      "epoch": 0.11003933320846601,
+      "grad_norm": 0.5781607031822205,
+      "learning_rate": 2e-05,
+      "loss": 0.8256,
+      "step": 2350
+    },
+    {
+      "epoch": 0.11050758569020416,
+      "grad_norm": 0.4696139693260193,
+      "learning_rate": 2e-05,
+      "loss": 0.8044,
+      "step": 2360
+    },
+    {
+      "epoch": 0.11097583817194231,
+      "grad_norm": 0.4228568971157074,
+      "learning_rate": 2e-05,
+      "loss": 0.7684,
+      "step": 2370
+    },
+    {
+      "epoch": 0.11144409065368047,
+      "grad_norm": 0.4293256402015686,
+      "learning_rate": 2e-05,
+      "loss": 0.8011,
+      "step": 2380
+    },
+    {
+      "epoch": 0.11191234313541862,
+      "grad_norm": 0.4462338984012604,
+      "learning_rate": 2e-05,
+      "loss": 0.7655,
+      "step": 2390
+    },
+    {
+      "epoch": 0.11238059561715677,
+      "grad_norm": 0.47912371158599854,
+      "learning_rate": 2e-05,
+      "loss": 0.7984,
+      "step": 2400
+    },
+    {
+      "epoch": 0.11284884809889492,
+      "grad_norm": 0.9592866897583008,
+      "learning_rate": 2e-05,
+      "loss": 0.7766,
+      "step": 2410
+    },
+    {
+      "epoch": 0.11331710058063307,
+      "grad_norm": 0.39810967445373535,
+      "learning_rate": 2e-05,
+      "loss": 0.7604,
+      "step": 2420
+    },
+    {
+      "epoch": 0.11378535306237123,
+      "grad_norm": 0.45152178406715393,
+      "learning_rate": 2e-05,
+      "loss": 0.79,
+      "step": 2430
+    },
+    {
+      "epoch": 0.11425360554410939,
+      "grad_norm": 0.44523119926452637,
+      "learning_rate": 2e-05,
+      "loss": 0.8067,
+      "step": 2440
+    },
+    {
+      "epoch": 0.11472185802584754,
+      "grad_norm": 0.47627031803131104,
+      "learning_rate": 2e-05,
+      "loss": 0.7797,
+      "step": 2450
+    },
+    {
+      "epoch": 0.1151901105075857,
+      "grad_norm": 0.45713964104652405,
+      "learning_rate": 2e-05,
+      "loss": 0.7613,
+      "step": 2460
+    },
+    {
+      "epoch": 0.11565836298932385,
+      "grad_norm": 0.4603721499443054,
+      "learning_rate": 2e-05,
+      "loss": 0.7985,
+      "step": 2470
+    },
+    {
+      "epoch": 0.116126615471062,
+      "grad_norm": 0.5172587633132935,
+      "learning_rate": 2e-05,
+      "loss": 0.7631,
+      "step": 2480
+    },
+    {
+      "epoch": 0.11659486795280015,
+      "grad_norm": 0.4585573971271515,
+      "learning_rate": 2e-05,
+      "loss": 0.7675,
+      "step": 2490
+    },
+    {
+      "epoch": 0.1170631204345383,
+      "grad_norm": 0.4653816223144531,
+      "learning_rate": 2e-05,
+      "loss": 0.8136,
+      "step": 2500
+    },
+    {
+      "epoch": 0.11753137291627645,
+      "grad_norm": 0.45800113677978516,
+      "learning_rate": 2e-05,
+      "loss": 0.7873,
+      "step": 2510
+    },
+    {
+      "epoch": 0.1179996253980146,
+      "grad_norm": 1.0582246780395508,
+      "learning_rate": 2e-05,
+      "loss": 0.7743,
+      "step": 2520
+    },
+    {
+      "epoch": 0.11846787787975276,
+      "grad_norm": 0.4848767817020416,
+      "learning_rate": 2e-05,
+      "loss": 0.7817,
+      "step": 2530
+    },
+    {
+      "epoch": 0.11893613036149092,
+      "grad_norm": 0.45420584082603455,
+      "learning_rate": 2e-05,
+      "loss": 0.7787,
+      "step": 2540
+    },
+    {
+      "epoch": 0.11940438284322907,
+      "grad_norm": 0.446957528591156,
+      "learning_rate": 2e-05,
+      "loss": 0.7691,
+      "step": 2550
+    },
+    {
+      "epoch": 0.11987263532496722,
+      "grad_norm": 0.4325239956378937,
+      "learning_rate": 2e-05,
+      "loss": 0.7725,
+      "step": 2560
+    },
+    {
+      "epoch": 0.12034088780670538,
+      "grad_norm": 0.3984229266643524,
+      "learning_rate": 2e-05,
+      "loss": 0.7517,
+      "step": 2570
+    },
+    {
+      "epoch": 0.12080914028844353,
+      "grad_norm": 0.5047958493232727,
+      "learning_rate": 2e-05,
+      "loss": 0.7832,
+      "step": 2580
+    },
+    {
+      "epoch": 0.12127739277018168,
+      "grad_norm": 0.4237649142742157,
+      "learning_rate": 2e-05,
+      "loss": 0.7554,
+      "step": 2590
+    },
+    {
+      "epoch": 0.12174564525191983,
+      "grad_norm": 0.5109700560569763,
+      "learning_rate": 2e-05,
+      "loss": 0.791,
+      "step": 2600
+    },
+    {
+      "epoch": 0.12221389773365798,
+      "grad_norm": 0.47834986448287964,
+      "learning_rate": 2e-05,
+      "loss": 0.7781,
+      "step": 2610
+    },
+    {
+      "epoch": 0.12268215021539614,
+      "grad_norm": 0.43558385968208313,
+      "learning_rate": 2e-05,
+      "loss": 0.7887,
+      "step": 2620
+    },
+    {
+      "epoch": 0.1231504026971343,
+      "grad_norm": 0.46023640036582947,
+      "learning_rate": 2e-05,
+      "loss": 0.8184,
+      "step": 2630
+    },
+    {
+      "epoch": 0.12361865517887245,
+      "grad_norm": 0.43065863847732544,
+      "learning_rate": 2e-05,
+      "loss": 0.8087,
+      "step": 2640
+    },
+    {
+      "epoch": 0.1240869076606106,
+      "grad_norm": 0.46921634674072266,
+      "learning_rate": 2e-05,
+      "loss": 0.8023,
+      "step": 2650
+    },
+    {
+      "epoch": 0.12455516014234876,
+      "grad_norm": 0.45447105169296265,
+      "learning_rate": 2e-05,
+      "loss": 0.7865,
+      "step": 2660
+    },
+    {
+      "epoch": 0.12502341262408692,
+      "grad_norm": 0.5031870007514954,
+      "learning_rate": 2e-05,
+      "loss": 0.7982,
+      "step": 2670
+    },
+    {
+      "epoch": 0.12549166510582507,
+      "grad_norm": 0.4168241024017334,
+      "learning_rate": 2e-05,
+      "loss": 0.7755,
+      "step": 2680
+    },
+    {
+      "epoch": 0.12595991758756322,
+      "grad_norm": 0.489377498626709,
+      "learning_rate": 2e-05,
+      "loss": 0.797,
+      "step": 2690
+    },
+    {
+      "epoch": 0.12642817006930138,
+      "grad_norm": 0.42968428134918213,
+      "learning_rate": 2e-05,
+      "loss": 0.7752,
+      "step": 2700
+    },
+    {
+      "epoch": 0.12689642255103953,
+      "grad_norm": 0.45984914898872375,
+      "learning_rate": 2e-05,
+      "loss": 0.7716,
+      "step": 2710
+    },
+    {
+      "epoch": 0.12736467503277768,
+      "grad_norm": 0.47864603996276855,
+      "learning_rate": 2e-05,
+      "loss": 0.7735,
+      "step": 2720
+    },
+    {
+      "epoch": 0.12783292751451583,
+      "grad_norm": 0.4695641100406647,
+      "learning_rate": 2e-05,
+      "loss": 0.7924,
+      "step": 2730
+    },
+    {
+      "epoch": 0.12830117999625398,
+      "grad_norm": 0.42772170901298523,
+      "learning_rate": 2e-05,
+      "loss": 0.7798,
+      "step": 2740
+    },
+    {
+      "epoch": 0.12876943247799214,
+      "grad_norm": 0.526038408279419,
+      "learning_rate": 2e-05,
+      "loss": 0.8099,
+      "step": 2750
+    },
+    {
+      "epoch": 0.1292376849597303,
+      "grad_norm": 0.48115813732147217,
+      "learning_rate": 2e-05,
+      "loss": 0.7725,
+      "step": 2760
+    },
+    {
+      "epoch": 0.12970593744146844,
+      "grad_norm": 0.8196518421173096,
+      "learning_rate": 2e-05,
+      "loss": 0.7476,
+      "step": 2770
+    },
+    {
+      "epoch": 0.1301741899232066,
+      "grad_norm": 0.4384530186653137,
+      "learning_rate": 2e-05,
+      "loss": 0.7528,
+      "step": 2780
+    },
+    {
+      "epoch": 0.13064244240494474,
+      "grad_norm": 0.48463067412376404,
+      "learning_rate": 2e-05,
+      "loss": 0.7801,
+      "step": 2790
+    },
+    {
+      "epoch": 0.1311106948866829,
+      "grad_norm": 0.431721955537796,
+      "learning_rate": 2e-05,
+      "loss": 0.7843,
+      "step": 2800
+    },
+    {
+      "epoch": 0.13157894736842105,
+      "grad_norm": 0.4804224371910095,
+      "learning_rate": 2e-05,
+      "loss": 0.8155,
+      "step": 2810
+    },
+    {
+      "epoch": 0.1320471998501592,
+      "grad_norm": 0.4451720118522644,
+      "learning_rate": 2e-05,
+      "loss": 0.7953,
+      "step": 2820
+    },
+    {
+      "epoch": 0.13251545233189735,
+      "grad_norm": 0.5056591629981995,
+      "learning_rate": 2e-05,
+      "loss": 0.7696,
+      "step": 2830
+    },
+    {
+      "epoch": 0.1329837048136355,
+      "grad_norm": 0.40849485993385315,
+      "learning_rate": 2e-05,
+      "loss": 0.7782,
+      "step": 2840
+    },
+    {
+      "epoch": 0.13345195729537365,
+      "grad_norm": 0.4380452036857605,
+      "learning_rate": 2e-05,
+      "loss": 0.7444,
+      "step": 2850
+    },
+    {
+      "epoch": 0.13392020977711183,
+      "grad_norm": 0.4316031336784363,
+      "learning_rate": 2e-05,
+      "loss": 0.7728,
+      "step": 2860
+    },
+    {
+      "epoch": 0.13438846225884998,
+      "grad_norm": 0.46132344007492065,
+      "learning_rate": 2e-05,
+      "loss": 0.7649,
+      "step": 2870
+    },
+    {
+      "epoch": 0.13485671474058814,
+      "grad_norm": 0.4467066526412964,
+      "learning_rate": 2e-05,
+      "loss": 0.7802,
+      "step": 2880
+    },
+    {
+      "epoch": 0.1353249672223263,
+      "grad_norm": 0.4437260031700134,
+      "learning_rate": 2e-05,
+      "loss": 0.7728,
+      "step": 2890
+    },
+    {
+      "epoch": 0.13579321970406444,
+      "grad_norm": 0.41009271144866943,
+      "learning_rate": 2e-05,
+      "loss": 0.785,
+      "step": 2900
+    },
+    {
+      "epoch": 0.1362614721858026,
+      "grad_norm": 0.4567055106163025,
+      "learning_rate": 2e-05,
+      "loss": 0.7977,
+      "step": 2910
+    },
+    {
+      "epoch": 0.13672972466754074,
+      "grad_norm": 0.4779447019100189,
+      "learning_rate": 2e-05,
+      "loss": 0.786,
+      "step": 2920
+    },
+    {
+      "epoch": 0.1371979771492789,
+      "grad_norm": 0.4813325107097626,
+      "learning_rate": 2e-05,
+      "loss": 0.7602,
+      "step": 2930
+    },
+    {
+      "epoch": 0.13766622963101705,
+      "grad_norm": 0.46978309750556946,
+      "learning_rate": 2e-05,
+      "loss": 0.8055,
+      "step": 2940
+    },
+    {
+      "epoch": 0.1381344821127552,
+      "grad_norm": 0.436978816986084,
+      "learning_rate": 2e-05,
+      "loss": 0.7712,
+      "step": 2950
+    },
+    {
+      "epoch": 0.13860273459449335,
+      "grad_norm": 0.4569421410560608,
+      "learning_rate": 2e-05,
+      "loss": 0.7704,
+      "step": 2960
+    },
+    {
+      "epoch": 0.1390709870762315,
+      "grad_norm": 0.4956076443195343,
+      "learning_rate": 2e-05,
+      "loss": 0.7934,
+      "step": 2970
+    },
+    {
+      "epoch": 0.13953923955796965,
+      "grad_norm": 0.5110732913017273,
+      "learning_rate": 2e-05,
+      "loss": 0.7768,
+      "step": 2980
+    },
+    {
+      "epoch": 0.1400074920397078,
+      "grad_norm": 0.4133819043636322,
+      "learning_rate": 2e-05,
+      "loss": 0.7859,
+      "step": 2990
+    },
+    {
+      "epoch": 0.14047574452144596,
+      "grad_norm": 0.5259489417076111,
+      "learning_rate": 2e-05,
+      "loss": 0.7861,
+      "step": 3000
+    },
+    {
+      "epoch": 0.1409439970031841,
+      "grad_norm": 0.4838019013404846,
+      "learning_rate": 2e-05,
+      "loss": 0.7816,
+      "step": 3010
+    },
+    {
+      "epoch": 0.14141224948492226,
+      "grad_norm": 0.43243408203125,
+      "learning_rate": 2e-05,
+      "loss": 0.7951,
+      "step": 3020
+    },
+    {
+      "epoch": 0.1418805019666604,
+      "grad_norm": 0.4447523355484009,
+      "learning_rate": 2e-05,
+      "loss": 0.7745,
+      "step": 3030
+    },
+    {
+      "epoch": 0.1423487544483986,
+      "grad_norm": 0.426066130399704,
+      "learning_rate": 2e-05,
+      "loss": 0.7306,
+      "step": 3040
+    },
+    {
+      "epoch": 0.14281700693013674,
+      "grad_norm": 0.4635065495967865,
+      "learning_rate": 2e-05,
+      "loss": 0.7769,
+      "step": 3050
+    },
+    {
+      "epoch": 0.1432852594118749,
+      "grad_norm": 0.4404882788658142,
+      "learning_rate": 2e-05,
+      "loss": 0.7852,
+      "step": 3060
+    },
+    {
+      "epoch": 0.14375351189361305,
+      "grad_norm": 0.4820006489753723,
+      "learning_rate": 2e-05,
+      "loss": 0.7805,
+      "step": 3070
+    },
+    {
+      "epoch": 0.1442217643753512,
+      "grad_norm": 1.9450064897537231,
+      "learning_rate": 2e-05,
+      "loss": 0.7459,
+      "step": 3080
+    },
+    {
+      "epoch": 0.14469001685708935,
+      "grad_norm": 0.4669210612773895,
+      "learning_rate": 2e-05,
+      "loss": 0.7594,
+      "step": 3090
+    },
+    {
+      "epoch": 0.1451582693388275,
+      "grad_norm": 0.4511769711971283,
+      "learning_rate": 2e-05,
+      "loss": 0.7574,
+      "step": 3100
+    },
+    {
+      "epoch": 0.14562652182056565,
+      "grad_norm": 0.4666540026664734,
+      "learning_rate": 2e-05,
+      "loss": 0.7624,
+      "step": 3110
+    },
+    {
+      "epoch": 0.1460947743023038,
+      "grad_norm": 0.4351053535938263,
+      "learning_rate": 2e-05,
+      "loss": 0.7786,
+      "step": 3120
+    },
+    {
+      "epoch": 0.14656302678404196,
+      "grad_norm": 0.48357275128364563,
+      "learning_rate": 2e-05,
+      "loss": 0.7756,
+      "step": 3130
+    },
+    {
+      "epoch": 0.1470312792657801,
+      "grad_norm": 0.4528772830963135,
+      "learning_rate": 2e-05,
+      "loss": 0.7812,
+      "step": 3140
+    },
+    {
+      "epoch": 0.14749953174751826,
+      "grad_norm": 0.43779534101486206,
+      "learning_rate": 2e-05,
+      "loss": 0.7698,
+      "step": 3150
+    },
+    {
+      "epoch": 0.1479677842292564,
+      "grad_norm": 0.4632822275161743,
+      "learning_rate": 2e-05,
+      "loss": 0.7414,
+      "step": 3160
+    },
+    {
+      "epoch": 0.14843603671099456,
+      "grad_norm": 0.4203107953071594,
+      "learning_rate": 2e-05,
+      "loss": 0.786,
+      "step": 3170
+    },
+    {
+      "epoch": 0.14890428919273271,
+      "grad_norm": 0.4951493740081787,
+      "learning_rate": 2e-05,
+      "loss": 0.7653,
+      "step": 3180
+    },
+    {
+      "epoch": 0.14937254167447087,
+      "grad_norm": 0.5392319560050964,
+      "learning_rate": 2e-05,
+      "loss": 0.7691,
+      "step": 3190
+    },
+    {
+      "epoch": 0.14984079415620902,
+      "grad_norm": 0.4392766058444977,
+      "learning_rate": 2e-05,
+      "loss": 0.7664,
+      "step": 3200
+    },
+    {
+      "epoch": 0.15030904663794717,
+      "grad_norm": 0.5064263939857483,
+      "learning_rate": 2e-05,
+      "loss": 0.7751,
+      "step": 3210
+    },
+    {
+      "epoch": 0.15077729911968532,
+      "grad_norm": 0.4974293112754822,
+      "learning_rate": 2e-05,
+      "loss": 0.7846,
+      "step": 3220
+    },
+    {
+      "epoch": 0.1512455516014235,
+      "grad_norm": 0.5040820240974426,
+      "learning_rate": 2e-05,
+      "loss": 0.7406,
+      "step": 3230
+    },
+    {
+      "epoch": 0.15171380408316165,
+      "grad_norm": 0.4319583475589752,
+      "learning_rate": 2e-05,
+      "loss": 0.7487,
+      "step": 3240
+    },
+    {
+      "epoch": 0.1521820565648998,
+      "grad_norm": 0.46588271856307983,
+      "learning_rate": 2e-05,
+      "loss": 0.773,
+      "step": 3250
+    },
+    {
+      "epoch": 0.15265030904663796,
+      "grad_norm": 0.41532087326049805,
+      "learning_rate": 2e-05,
+      "loss": 0.7661,
+      "step": 3260
+    },
+    {
+      "epoch": 0.1531185615283761,
+      "grad_norm": 0.5403257608413696,
+      "learning_rate": 2e-05,
+      "loss": 0.8033,
+      "step": 3270
+    },
+    {
+      "epoch": 0.15358681401011426,
+      "grad_norm": 0.4624592363834381,
+      "learning_rate": 2e-05,
+      "loss": 0.8036,
+      "step": 3280
+    },
+    {
+      "epoch": 0.1540550664918524,
+      "grad_norm": 0.4184800088405609,
+      "learning_rate": 2e-05,
+      "loss": 0.8141,
+      "step": 3290
+    },
+    {
+      "epoch": 0.15452331897359056,
+      "grad_norm": 0.41161057353019714,
+      "learning_rate": 2e-05,
+      "loss": 0.7622,
+      "step": 3300
+    },
+    {
+      "epoch": 0.1549915714553287,
+      "grad_norm": 0.41965046525001526,
+      "learning_rate": 2e-05,
+      "loss": 0.743,
+      "step": 3310
+    },
+    {
+      "epoch": 0.15545982393706687,
+      "grad_norm": 0.547100841999054,
+      "learning_rate": 2e-05,
+      "loss": 0.739,
+      "step": 3320
+    },
+    {
+      "epoch": 0.15592807641880502,
+      "grad_norm": 0.4471840262413025,
+      "learning_rate": 2e-05,
+      "loss": 0.7939,
+      "step": 3330
+    },
+    {
+      "epoch": 0.15639632890054317,
+      "grad_norm": 0.40980273485183716,
+      "learning_rate": 2e-05,
+      "loss": 0.7598,
+      "step": 3340
+    },
+    {
+      "epoch": 0.15686458138228132,
+      "grad_norm": 0.5068540573120117,
+      "learning_rate": 2e-05,
+      "loss": 0.7419,
+      "step": 3350
+    },
+    {
+      "epoch": 0.15733283386401947,
+      "grad_norm": 0.4310580790042877,
+      "learning_rate": 2e-05,
+      "loss": 0.7807,
+      "step": 3360
+    },
+    {
+      "epoch": 0.15780108634575762,
+      "grad_norm": 0.44191500544548035,
+      "learning_rate": 2e-05,
+      "loss": 0.7844,
+      "step": 3370
+    },
+    {
+      "epoch": 0.15826933882749578,
+      "grad_norm": 0.492794930934906,
+      "learning_rate": 2e-05,
+      "loss": 0.7829,
+      "step": 3380
+    },
+    {
+      "epoch": 0.15873759130923393,
+      "grad_norm": 0.46001243591308594,
+      "learning_rate": 2e-05,
+      "loss": 0.7616,
+      "step": 3390
+    },
+    {
+      "epoch": 0.15920584379097208,
+      "grad_norm": 0.452966570854187,
+      "learning_rate": 2e-05,
+      "loss": 0.8051,
+      "step": 3400
+    },
+    {
+      "epoch": 0.15967409627271026,
+      "grad_norm": 0.3855198621749878,
+      "learning_rate": 2e-05,
+      "loss": 0.7608,
+      "step": 3410
+    },
+    {
+      "epoch": 0.1601423487544484,
+      "grad_norm": 0.4697616398334503,
+      "learning_rate": 2e-05,
+      "loss": 0.7637,
+      "step": 3420
+    },
+    {
+      "epoch": 0.16061060123618656,
+      "grad_norm": 0.47525501251220703,
+      "learning_rate": 2e-05,
+      "loss": 0.7569,
+      "step": 3430
+    },
+    {
+      "epoch": 0.1610788537179247,
+      "grad_norm": 0.49193453788757324,
+      "learning_rate": 2e-05,
+      "loss": 0.7792,
+      "step": 3440
+    },
+    {
+      "epoch": 0.16154710619966287,
+      "grad_norm": 0.5018397569656372,
+      "learning_rate": 2e-05,
+      "loss": 0.8023,
+      "step": 3450
+    },
+    {
+      "epoch": 0.16201535868140102,
+      "grad_norm": 0.4693327248096466,
+      "learning_rate": 2e-05,
+      "loss": 0.7602,
+      "step": 3460
+    },
+    {
+      "epoch": 0.16248361116313917,
+      "grad_norm": 0.4147368371486664,
+      "learning_rate": 2e-05,
+      "loss": 0.7711,
+      "step": 3470
+    },
+    {
+      "epoch": 0.16295186364487732,
+      "grad_norm": 0.48835280537605286,
+      "learning_rate": 2e-05,
+      "loss": 0.8112,
+      "step": 3480
+    },
+    {
+      "epoch": 0.16342011612661547,
+      "grad_norm": 0.4873306155204773,
+      "learning_rate": 2e-05,
+      "loss": 0.8052,
+      "step": 3490
+    },
+    {
+      "epoch": 0.16388836860835362,
+      "grad_norm": 0.5845130681991577,
+      "learning_rate": 2e-05,
+      "loss": 0.7896,
+      "step": 3500
+    },
+    {
+      "epoch": 0.16435662109009178,
+      "grad_norm": 0.5305046439170837,
+      "learning_rate": 2e-05,
+      "loss": 0.7635,
+      "step": 3510
+    },
+    {
+      "epoch": 0.16482487357182993,
+      "grad_norm": 0.4971988797187805,
+      "learning_rate": 2e-05,
+      "loss": 0.7448,
+      "step": 3520
+    },
+    {
+      "epoch": 0.16529312605356808,
+      "grad_norm": 0.4730629026889801,
+      "learning_rate": 2e-05,
+      "loss": 0.7166,
+      "step": 3530
+    },
+    {
+      "epoch": 0.16576137853530623,
+      "grad_norm": 0.4115239381790161,
+      "learning_rate": 2e-05,
+      "loss": 0.7846,
+      "step": 3540
+    },
+    {
+      "epoch": 0.16622963101704438,
+      "grad_norm": 0.4584404528141022,
+      "learning_rate": 2e-05,
+      "loss": 0.744,
+      "step": 3550
+    },
+    {
+      "epoch": 0.16669788349878253,
+      "grad_norm": 0.4739014208316803,
+      "learning_rate": 2e-05,
+      "loss": 0.8232,
+      "step": 3560
+    },
+    {
+      "epoch": 0.16716613598052069,
+      "grad_norm": 0.44108620285987854,
+      "learning_rate": 2e-05,
+      "loss": 0.7595,
+      "step": 3570
+    },
+    {
+      "epoch": 0.16763438846225884,
+      "grad_norm": 0.4143202006816864,
+      "learning_rate": 2e-05,
+      "loss": 0.7571,
+      "step": 3580
+    },
+    {
+      "epoch": 0.168102640943997,
+      "grad_norm": 0.5289865136146545,
+      "learning_rate": 2e-05,
+      "loss": 0.7487,
+      "step": 3590
+    },
+    {
+      "epoch": 0.16857089342573517,
+      "grad_norm": 0.453819215297699,
+      "learning_rate": 2e-05,
+      "loss": 0.7516,
+      "step": 3600
+    },
+    {
+      "epoch": 0.16903914590747332,
+      "grad_norm": 0.4348923861980438,
+      "learning_rate": 2e-05,
+      "loss": 0.7767,
+      "step": 3610
+    },
+    {
+      "epoch": 0.16950739838921147,
+      "grad_norm": 0.4643234610557556,
+      "learning_rate": 2e-05,
+      "loss": 0.7945,
+      "step": 3620
+    },
+    {
+      "epoch": 0.16997565087094962,
+      "grad_norm": 0.47802820801734924,
+      "learning_rate": 2e-05,
+      "loss": 0.7327,
+      "step": 3630
+    },
+    {
+      "epoch": 0.17044390335268778,
+      "grad_norm": 0.4367299973964691,
+      "learning_rate": 2e-05,
+      "loss": 0.7923,
+      "step": 3640
+    },
+    {
+      "epoch": 0.17091215583442593,
+      "grad_norm": 0.4553154408931732,
+      "learning_rate": 2e-05,
+      "loss": 0.7655,
+      "step": 3650
+    },
+    {
+      "epoch": 0.17138040831616408,
+      "grad_norm": 0.4764273464679718,
+      "learning_rate": 2e-05,
+      "loss": 0.7554,
+      "step": 3660
+    },
+    {
+      "epoch": 0.17184866079790223,
+      "grad_norm": 0.44732391834259033,
+      "learning_rate": 2e-05,
+      "loss": 0.7692,
+      "step": 3670
+    },
+    {
+      "epoch": 0.17231691327964038,
+      "grad_norm": 0.5650924444198608,
+      "learning_rate": 2e-05,
+      "loss": 0.7731,
+      "step": 3680
+    },
+    {
+      "epoch": 0.17278516576137853,
+      "grad_norm": 0.6082085371017456,
+      "learning_rate": 2e-05,
+      "loss": 0.7769,
+      "step": 3690
+    },
+    {
+      "epoch": 0.17325341824311669,
+      "grad_norm": 0.5217570066452026,
+      "learning_rate": 2e-05,
+      "loss": 0.7568,
+      "step": 3700
+    },
+    {
+      "epoch": 0.17372167072485484,
+      "grad_norm": 0.46141839027404785,
+      "learning_rate": 2e-05,
+      "loss": 0.7528,
+      "step": 3710
+    },
+    {
+      "epoch": 0.174189923206593,
+      "grad_norm": 0.48429611325263977,
+      "learning_rate": 2e-05,
+      "loss": 0.797,
+      "step": 3720
+    },
+    {
+      "epoch": 0.17465817568833114,
+      "grad_norm": 0.5406231880187988,
+      "learning_rate": 2e-05,
+      "loss": 0.7687,
+      "step": 3730
+    },
+    {
+      "epoch": 0.1751264281700693,
+      "grad_norm": 0.42686790227890015,
+      "learning_rate": 2e-05,
+      "loss": 0.7706,
+      "step": 3740
+    },
+    {
+      "epoch": 0.17559468065180744,
+      "grad_norm": 0.44343850016593933,
+      "learning_rate": 2e-05,
+      "loss": 0.7863,
+      "step": 3750
+    },
+    {
+      "epoch": 0.1760629331335456,
+      "grad_norm": 0.4633735716342926,
+      "learning_rate": 2e-05,
+      "loss": 0.7569,
+      "step": 3760
+    },
+    {
+      "epoch": 0.17653118561528375,
+      "grad_norm": 0.4800039529800415,
+      "learning_rate": 2e-05,
+      "loss": 0.7502,
+      "step": 3770
+    },
+    {
+      "epoch": 0.17699943809702193,
+      "grad_norm": 0.4586217701435089,
+      "learning_rate": 2e-05,
+      "loss": 0.8227,
+      "step": 3780
+    },
+    {
+      "epoch": 0.17746769057876008,
+      "grad_norm": 0.4920169413089752,
+      "learning_rate": 2e-05,
+      "loss": 0.7901,
+      "step": 3790
+    },
+    {
+      "epoch": 0.17793594306049823,
+      "grad_norm": 0.47960785031318665,
+      "learning_rate": 2e-05,
+      "loss": 0.7633,
+      "step": 3800
+    },
+    {
+      "epoch": 0.17840419554223638,
+      "grad_norm": 0.4717206656932831,
+      "learning_rate": 2e-05,
+      "loss": 0.819,
+      "step": 3810
+    },
+    {
+      "epoch": 0.17887244802397453,
+      "grad_norm": 0.511841893196106,
+      "learning_rate": 2e-05,
+      "loss": 0.7732,
+      "step": 3820
+    },
+    {
+      "epoch": 0.17934070050571269,
+      "grad_norm": 0.47291046380996704,
+      "learning_rate": 2e-05,
+      "loss": 0.7682,
+      "step": 3830
+    },
+    {
+      "epoch": 0.17980895298745084,
+      "grad_norm": 0.40757039189338684,
+      "learning_rate": 2e-05,
+      "loss": 0.7546,
+      "step": 3840
+    },
+    {
+      "epoch": 0.180277205469189,
+      "grad_norm": 0.5135296583175659,
+      "learning_rate": 2e-05,
+      "loss": 0.7602,
+      "step": 3850
+    },
+    {
+      "epoch": 0.18074545795092714,
+      "grad_norm": 0.4503514766693115,
+      "learning_rate": 2e-05,
+      "loss": 0.738,
+      "step": 3860
+    },
+    {
+      "epoch": 0.1812137104326653,
+      "grad_norm": 0.49815234541893005,
+      "learning_rate": 2e-05,
+      "loss": 0.7765,
+      "step": 3870
+    },
+    {
+      "epoch": 0.18168196291440344,
+      "grad_norm": 0.4232896864414215,
+      "learning_rate": 2e-05,
+      "loss": 0.7836,
+      "step": 3880
+    },
+    {
+      "epoch": 0.1821502153961416,
+      "grad_norm": 0.4334942102432251,
+      "learning_rate": 2e-05,
+      "loss": 0.7253,
+      "step": 3890
+    },
+    {
+      "epoch": 0.18261846787787975,
+      "grad_norm": 0.43816936016082764,
+      "learning_rate": 2e-05,
+      "loss": 0.7814,
+      "step": 3900
+    },
+    {
+      "epoch": 0.1830867203596179,
+      "grad_norm": 0.4376625418663025,
+      "learning_rate": 2e-05,
+      "loss": 0.7547,
+      "step": 3910
+    },
+    {
+      "epoch": 0.18355497284135605,
+      "grad_norm": 0.4537307620048523,
+      "learning_rate": 2e-05,
+      "loss": 0.7475,
+      "step": 3920
+    },
+    {
+      "epoch": 0.1840232253230942,
+      "grad_norm": 0.4813627302646637,
+      "learning_rate": 2e-05,
+      "loss": 0.7681,
+      "step": 3930
+    },
+    {
+      "epoch": 0.18449147780483235,
+      "grad_norm": 0.5152831077575684,
+      "learning_rate": 2e-05,
+      "loss": 0.7151,
+      "step": 3940
+    },
+    {
+      "epoch": 0.1849597302865705,
+      "grad_norm": 0.4522545337677002,
+      "learning_rate": 2e-05,
+      "loss": 0.7779,
+      "step": 3950
+    },
+    {
+      "epoch": 0.18542798276830869,
+      "grad_norm": 0.6297553777694702,
+      "learning_rate": 2e-05,
+      "loss": 0.7902,
+      "step": 3960
+    },
+    {
+      "epoch": 0.18589623525004684,
+      "grad_norm": 0.4071296751499176,
+      "learning_rate": 2e-05,
+      "loss": 0.7563,
+      "step": 3970
+    },
+    {
+      "epoch": 0.186364487731785,
+      "grad_norm": 0.4333728849887848,
+      "learning_rate": 2e-05,
+      "loss": 0.7854,
+      "step": 3980
+    },
+    {
+      "epoch": 0.18683274021352314,
+      "grad_norm": 0.44967952370643616,
+      "learning_rate": 2e-05,
+      "loss": 0.7522,
+      "step": 3990
+    },
+    {
+      "epoch": 0.1873009926952613,
+      "grad_norm": 1.3324648141860962,
+      "learning_rate": 2e-05,
+      "loss": 0.7475,
+      "step": 4000
+    },
+    {
+      "epoch": 0.18776924517699944,
+      "grad_norm": 0.42225560545921326,
+      "learning_rate": 2e-05,
+      "loss": 0.7515,
+      "step": 4010
+    },
+    {
+      "epoch": 0.1882374976587376,
+      "grad_norm": 0.4295327365398407,
+      "learning_rate": 2e-05,
+      "loss": 0.7989,
+      "step": 4020
+    },
+    {
+      "epoch": 0.18870575014047575,
+      "grad_norm": 0.4431644380092621,
+      "learning_rate": 2e-05,
+      "loss": 0.7446,
+      "step": 4030
+    },
+    {
+      "epoch": 0.1891740026222139,
+      "grad_norm": 0.4686296880245209,
+      "learning_rate": 2e-05,
+      "loss": 0.7488,
+      "step": 4040
+    },
+    {
+      "epoch": 0.18964225510395205,
+      "grad_norm": 0.41146647930145264,
+      "learning_rate": 2e-05,
+      "loss": 0.7627,
+      "step": 4050
+    },
+    {
+      "epoch": 0.1901105075856902,
+      "grad_norm": 0.5990108847618103,
+      "learning_rate": 2e-05,
+      "loss": 0.7834,
+      "step": 4060
+    },
+    {
+      "epoch": 0.19057876006742835,
+      "grad_norm": 0.4420095682144165,
+      "learning_rate": 2e-05,
+      "loss": 0.7197,
+      "step": 4070
+    },
+    {
+      "epoch": 0.1910470125491665,
+      "grad_norm": 0.46126800775527954,
+      "learning_rate": 2e-05,
+      "loss": 0.7522,
+      "step": 4080
+    },
+    {
+      "epoch": 0.19151526503090466,
+      "grad_norm": 0.4785154163837433,
+      "learning_rate": 2e-05,
+      "loss": 0.7555,
+      "step": 4090
+    },
+    {
+      "epoch": 0.1919835175126428,
+      "grad_norm": 0.5068645477294922,
+      "learning_rate": 2e-05,
+      "loss": 0.7717,
+      "step": 4100
+    },
+    {
+      "epoch": 0.19245176999438096,
+      "grad_norm": 0.5095046758651733,
+      "learning_rate": 2e-05,
+      "loss": 0.7928,
+      "step": 4110
+    },
+    {
+      "epoch": 0.1929200224761191,
+      "grad_norm": 0.4467774033546448,
+      "learning_rate": 2e-05,
+      "loss": 0.7303,
+      "step": 4120
+    },
+    {
+      "epoch": 0.19338827495785726,
+      "grad_norm": 0.4577523469924927,
+      "learning_rate": 2e-05,
+      "loss": 0.733,
+      "step": 4130
+    },
+    {
+      "epoch": 0.19385652743959542,
+      "grad_norm": 0.4515482783317566,
+      "learning_rate": 2e-05,
+      "loss": 0.7537,
+      "step": 4140
+    },
+    {
+      "epoch": 0.1943247799213336,
+      "grad_norm": 0.5130146741867065,
+      "learning_rate": 2e-05,
+      "loss": 0.761,
+      "step": 4150
+    },
+    {
+      "epoch": 0.19479303240307175,
+      "grad_norm": 0.47396641969680786,
+      "learning_rate": 2e-05,
+      "loss": 0.7459,
+      "step": 4160
+    },
+    {
+      "epoch": 0.1952612848848099,
+      "grad_norm": 0.4405386447906494,
+      "learning_rate": 2e-05,
+      "loss": 0.7798,
+      "step": 4170
+    },
+    {
+      "epoch": 0.19572953736654805,
+      "grad_norm": 0.4735771715641022,
+      "learning_rate": 2e-05,
+      "loss": 0.7695,
+      "step": 4180
+    },
+    {
+      "epoch": 0.1961977898482862,
+      "grad_norm": 0.4145168364048004,
+      "learning_rate": 2e-05,
+      "loss": 0.8006,
+      "step": 4190
+    },
+    {
+      "epoch": 0.19666604233002435,
+      "grad_norm": 0.5201475024223328,
+      "learning_rate": 2e-05,
+      "loss": 0.7508,
+      "step": 4200
+    },
+    {
+      "epoch": 0.1971342948117625,
+      "grad_norm": 0.44904908537864685,
+      "learning_rate": 2e-05,
+      "loss": 0.7819,
+      "step": 4210
+    },
+    {
+      "epoch": 0.19760254729350066,
+      "grad_norm": 0.4453062415122986,
+      "learning_rate": 2e-05,
+      "loss": 0.7273,
+      "step": 4220
+    },
+    {
+      "epoch": 0.1980707997752388,
+      "grad_norm": 0.4952180087566376,
+      "learning_rate": 2e-05,
+      "loss": 0.7513,
+      "step": 4230
+    },
+    {
+      "epoch": 0.19853905225697696,
+      "grad_norm": 0.6102429032325745,
+      "learning_rate": 2e-05,
+      "loss": 0.7505,
+      "step": 4240
+    },
+    {
+      "epoch": 0.1990073047387151,
+      "grad_norm": 0.4875331223011017,
+      "learning_rate": 2e-05,
+      "loss": 0.7632,
+      "step": 4250
+    },
+    {
+      "epoch": 0.19947555722045326,
+      "grad_norm": 0.5022391676902771,
+      "learning_rate": 2e-05,
+      "loss": 0.7639,
+      "step": 4260
+    },
+    {
+      "epoch": 0.19994380970219142,
+      "grad_norm": 0.44564130902290344,
+      "learning_rate": 2e-05,
+      "loss": 0.7377,
+      "step": 4270
+    },
+    {
+      "epoch": 0.20041206218392957,
+      "grad_norm": 0.509864091873169,
+      "learning_rate": 2e-05,
+      "loss": 0.7791,
+      "step": 4280
+    },
+    {
+      "epoch": 0.20088031466566772,
+      "grad_norm": 0.45319506525993347,
+      "learning_rate": 2e-05,
+      "loss": 0.7693,
+      "step": 4290
+    },
+    {
+      "epoch": 0.20134856714740587,
+      "grad_norm": 0.4179646968841553,
+      "learning_rate": 2e-05,
+      "loss": 0.7724,
+      "step": 4300
+    },
+    {
+      "epoch": 0.20181681962914402,
+      "grad_norm": 0.44443565607070923,
+      "learning_rate": 2e-05,
+      "loss": 0.7748,
+      "step": 4310
+    },
+    {
+      "epoch": 0.20228507211088217,
+      "grad_norm": 0.5083085894584656,
+      "learning_rate": 2e-05,
+      "loss": 0.74,
+      "step": 4320
+    },
+    {
+      "epoch": 0.20275332459262035,
+      "grad_norm": 0.43336138129234314,
+      "learning_rate": 2e-05,
+      "loss": 0.7701,
+      "step": 4330
+    },
+    {
+      "epoch": 0.2032215770743585,
+      "grad_norm": 0.48234793543815613,
+      "learning_rate": 2e-05,
+      "loss": 0.7723,
+      "step": 4340
+    },
+    {
+      "epoch": 0.20368982955609666,
+      "grad_norm": 0.6275402307510376,
+      "learning_rate": 2e-05,
+      "loss": 0.7683,
+      "step": 4350
+    },
+    {
+      "epoch": 0.2041580820378348,
+      "grad_norm": 0.4710438847541809,
+      "learning_rate": 2e-05,
+      "loss": 0.7461,
+      "step": 4360
+    },
+    {
+      "epoch": 0.20462633451957296,
+      "grad_norm": 0.4537172317504883,
+      "learning_rate": 2e-05,
+      "loss": 0.734,
+      "step": 4370
+    },
+    {
+      "epoch": 0.2050945870013111,
+      "grad_norm": 0.4454044699668884,
+      "learning_rate": 2e-05,
+      "loss": 0.7699,
+      "step": 4380
+    },
+    {
+      "epoch": 0.20556283948304926,
+      "grad_norm": 0.455642968416214,
+      "learning_rate": 2e-05,
+      "loss": 0.7548,
+      "step": 4390
+    },
+    {
+      "epoch": 0.20603109196478742,
+      "grad_norm": 0.45310547947883606,
+      "learning_rate": 2e-05,
+      "loss": 0.7754,
+      "step": 4400
+    },
+    {
+      "epoch": 0.20649934444652557,
+      "grad_norm": 0.4452892541885376,
+      "learning_rate": 2e-05,
+      "loss": 0.7774,
+      "step": 4410
+    },
+    {
+      "epoch": 0.20696759692826372,
+      "grad_norm": 0.5954875946044922,
+      "learning_rate": 2e-05,
+      "loss": 0.7175,
+      "step": 4420
+    },
+    {
+      "epoch": 0.20743584941000187,
+      "grad_norm": 0.43730244040489197,
+      "learning_rate": 2e-05,
+      "loss": 0.7432,
+      "step": 4430
+    },
+    {
+      "epoch": 0.20790410189174002,
+      "grad_norm": 0.4403478801250458,
+      "learning_rate": 2e-05,
+      "loss": 0.7893,
+      "step": 4440
+    },
+    {
+      "epoch": 0.20837235437347817,
+      "grad_norm": 0.47121667861938477,
+      "learning_rate": 2e-05,
+      "loss": 0.7813,
+      "step": 4450
+    },
+    {
+      "epoch": 0.20884060685521633,
+      "grad_norm": 0.442446768283844,
+      "learning_rate": 2e-05,
+      "loss": 0.8086,
+      "step": 4460
+    },
+    {
+      "epoch": 0.20930885933695448,
+      "grad_norm": 0.4204050898551941,
+      "learning_rate": 2e-05,
+      "loss": 0.7554,
+      "step": 4470
+    },
+    {
+      "epoch": 0.20977711181869263,
+      "grad_norm": 0.47490179538726807,
+      "learning_rate": 2e-05,
+      "loss": 0.7572,
+      "step": 4480
+    },
+    {
+      "epoch": 0.21024536430043078,
+      "grad_norm": 0.433151513338089,
+      "learning_rate": 2e-05,
+      "loss": 0.7459,
+      "step": 4490
+    },
+    {
+      "epoch": 0.21071361678216893,
+      "grad_norm": 0.47067561745643616,
+      "learning_rate": 2e-05,
+      "loss": 0.7485,
+      "step": 4500
+    },
+    {
+      "epoch": 0.21118186926390708,
+      "grad_norm": 0.6863535046577454,
+      "learning_rate": 2e-05,
+      "loss": 0.7935,
+      "step": 4510
+    },
+    {
+      "epoch": 0.21165012174564526,
+      "grad_norm": 0.531719446182251,
+      "learning_rate": 2e-05,
+      "loss": 0.7552,
+      "step": 4520
+    },
+    {
+      "epoch": 0.21211837422738342,
+      "grad_norm": 0.4008723795413971,
+      "learning_rate": 2e-05,
+      "loss": 0.7481,
+      "step": 4530
+    },
+    {
+      "epoch": 0.21258662670912157,
+      "grad_norm": 0.48098883032798767,
+      "learning_rate": 2e-05,
+      "loss": 0.7483,
+      "step": 4540
+    },
+    {
+      "epoch": 0.21305487919085972,
+      "grad_norm": 0.4323178231716156,
+      "learning_rate": 2e-05,
+      "loss": 0.7779,
+      "step": 4550
+    },
+    {
+      "epoch": 0.21352313167259787,
+      "grad_norm": 0.487679123878479,
+      "learning_rate": 2e-05,
+      "loss": 0.7696,
+      "step": 4560
+    },
+    {
+      "epoch": 0.21399138415433602,
+      "grad_norm": 0.705156147480011,
+      "learning_rate": 2e-05,
+      "loss": 0.7455,
+      "step": 4570
+    },
+    {
+      "epoch": 0.21445963663607417,
+      "grad_norm": 0.4571705162525177,
+      "learning_rate": 2e-05,
+      "loss": 0.732,
+      "step": 4580
+    },
+    {
+      "epoch": 0.21492788911781233,
+      "grad_norm": 0.5622167587280273,
+      "learning_rate": 2e-05,
+      "loss": 0.7546,
+      "step": 4590
+    },
+    {
+      "epoch": 0.21539614159955048,
+      "grad_norm": 0.42841002345085144,
+      "learning_rate": 2e-05,
+      "loss": 0.7311,
+      "step": 4600
+    },
+    {
+      "epoch": 0.21586439408128863,
+      "grad_norm": 0.4392237365245819,
+      "learning_rate": 2e-05,
+      "loss": 0.7591,
+      "step": 4610
+    },
+    {
+      "epoch": 0.21633264656302678,
+      "grad_norm": 0.4754595458507538,
+      "learning_rate": 2e-05,
+      "loss": 0.7412,
+      "step": 4620
+    },
+    {
+      "epoch": 0.21680089904476493,
+      "grad_norm": 0.4496627748012543,
+      "learning_rate": 2e-05,
+      "loss": 0.7846,
+      "step": 4630
+    },
+    {
+      "epoch": 0.21726915152650308,
+      "grad_norm": 0.4356529414653778,
+      "learning_rate": 2e-05,
+      "loss": 0.7621,
+      "step": 4640
+    },
+    {
+      "epoch": 0.21773740400824124,
+      "grad_norm": 0.4718836545944214,
+      "learning_rate": 2e-05,
+      "loss": 0.7044,
+      "step": 4650
+    },
+    {
+      "epoch": 0.2182056564899794,
+      "grad_norm": 0.4545374810695648,
+      "learning_rate": 2e-05,
+      "loss": 0.7429,
+      "step": 4660
+    },
+    {
+      "epoch": 0.21867390897171754,
+      "grad_norm": 0.4936463534832001,
+      "learning_rate": 2e-05,
+      "loss": 0.7759,
+      "step": 4670
+    },
+    {
+      "epoch": 0.2191421614534557,
+      "grad_norm": 0.4476785659790039,
+      "learning_rate": 2e-05,
+      "loss": 0.7955,
+      "step": 4680
+    },
+    {
+      "epoch": 0.21961041393519384,
+      "grad_norm": 0.4549963176250458,
+      "learning_rate": 2e-05,
+      "loss": 0.7898,
+      "step": 4690
+    },
+    {
+      "epoch": 0.22007866641693202,
+      "grad_norm": 0.5105022192001343,
+      "learning_rate": 2e-05,
+      "loss": 0.7665,
+      "step": 4700
+    },
+    {
+      "epoch": 0.22054691889867017,
+      "grad_norm": 0.4230107367038727,
+      "learning_rate": 2e-05,
+      "loss": 0.7159,
+      "step": 4710
+    },
+    {
+      "epoch": 0.22101517138040833,
+      "grad_norm": 0.46920329332351685,
+      "learning_rate": 2e-05,
+      "loss": 0.7735,
+      "step": 4720
+    },
+    {
+      "epoch": 0.22148342386214648,
+      "grad_norm": 0.5480507612228394,
+      "learning_rate": 2e-05,
+      "loss": 0.7712,
+      "step": 4730
+    },
+    {
+      "epoch": 0.22195167634388463,
+      "grad_norm": 0.44134387373924255,
+      "learning_rate": 2e-05,
+      "loss": 0.7407,
+      "step": 4740
+    },
+    {
+      "epoch": 0.22241992882562278,
+      "grad_norm": 0.4618501365184784,
+      "learning_rate": 2e-05,
+      "loss": 0.7359,
+      "step": 4750
+    },
+    {
+      "epoch": 0.22288818130736093,
+      "grad_norm": 0.46679604053497314,
+      "learning_rate": 2e-05,
+      "loss": 0.7411,
+      "step": 4760
+    },
+    {
+      "epoch": 0.22335643378909908,
+      "grad_norm": 0.61300128698349,
+      "learning_rate": 2e-05,
+      "loss": 0.7259,
+      "step": 4770
+    },
+    {
+      "epoch": 0.22382468627083724,
+      "grad_norm": 0.5050122141838074,
+      "learning_rate": 2e-05,
+      "loss": 0.7551,
+      "step": 4780
+    },
+    {
+      "epoch": 0.2242929387525754,
+      "grad_norm": 0.44730883836746216,
+      "learning_rate": 2e-05,
+      "loss": 0.7244,
+      "step": 4790
+    },
+    {
+      "epoch": 0.22476119123431354,
+      "grad_norm": 0.45580169558525085,
+      "learning_rate": 2e-05,
+      "loss": 0.7631,
+      "step": 4800
+    },
+    {
+      "epoch": 0.2252294437160517,
+      "grad_norm": 0.4521130621433258,
+      "learning_rate": 2e-05,
+      "loss": 0.7858,
+      "step": 4810
+    },
+    {
+      "epoch": 0.22569769619778984,
+      "grad_norm": 0.5105381608009338,
+      "learning_rate": 2e-05,
+      "loss": 0.7518,
+      "step": 4820
+    },
+    {
+      "epoch": 0.226165948679528,
+      "grad_norm": 0.4405578374862671,
+      "learning_rate": 2e-05,
+      "loss": 0.7585,
+      "step": 4830
+    },
+    {
+      "epoch": 0.22663420116126615,
+      "grad_norm": 0.44144561886787415,
+      "learning_rate": 2e-05,
+      "loss": 0.7283,
+      "step": 4840
+    },
+    {
+      "epoch": 0.2271024536430043,
+      "grad_norm": 0.4783679246902466,
+      "learning_rate": 2e-05,
+      "loss": 0.7595,
+      "step": 4850
+    },
+    {
+      "epoch": 0.22757070612474245,
+      "grad_norm": 0.4455217123031616,
+      "learning_rate": 2e-05,
+      "loss": 0.7416,
+      "step": 4860
+    },
+    {
+      "epoch": 0.2280389586064806,
+      "grad_norm": 0.4736783504486084,
+      "learning_rate": 2e-05,
+      "loss": 0.7295,
+      "step": 4870
+    },
+    {
+      "epoch": 0.22850721108821878,
+      "grad_norm": 0.4496026635169983,
+      "learning_rate": 2e-05,
+      "loss": 0.7768,
+      "step": 4880
+    },
+    {
+      "epoch": 0.22897546356995693,
+      "grad_norm": 0.4254869520664215,
+      "learning_rate": 2e-05,
+      "loss": 0.7569,
+      "step": 4890
+    },
+    {
+      "epoch": 0.22944371605169508,
+      "grad_norm": 0.5083935260772705,
+      "learning_rate": 2e-05,
+      "loss": 0.7119,
+      "step": 4900
+    },
+    {
+      "epoch": 0.22991196853343324,
+      "grad_norm": 0.5440414547920227,
+      "learning_rate": 2e-05,
+      "loss": 0.7408,
+      "step": 4910
+    },
+    {
+      "epoch": 0.2303802210151714,
+      "grad_norm": 0.4203755259513855,
+      "learning_rate": 2e-05,
+      "loss": 0.7569,
+      "step": 4920
+    },
+    {
+      "epoch": 0.23084847349690954,
+      "grad_norm": 0.43250465393066406,
+      "learning_rate": 2e-05,
+      "loss": 0.7208,
+      "step": 4930
+    },
+    {
+      "epoch": 0.2313167259786477,
+      "grad_norm": 0.40719085931777954,
+      "learning_rate": 2e-05,
+      "loss": 0.753,
+      "step": 4940
+    },
+    {
+      "epoch": 0.23178497846038584,
+      "grad_norm": 0.5324130058288574,
+      "learning_rate": 2e-05,
+      "loss": 0.7013,
+      "step": 4950
+    },
+    {
+      "epoch": 0.232253230942124,
+      "grad_norm": 0.4467906057834625,
+      "learning_rate": 2e-05,
+      "loss": 0.7145,
+      "step": 4960
+    },
+    {
+      "epoch": 0.23272148342386215,
+      "grad_norm": 0.41979438066482544,
+      "learning_rate": 2e-05,
+      "loss": 0.7139,
+      "step": 4970
+    },
+    {
+      "epoch": 0.2331897359056003,
+      "grad_norm": 0.4772839844226837,
+      "learning_rate": 2e-05,
+      "loss": 0.767,
+      "step": 4980
+    },
+    {
+      "epoch": 0.23365798838733845,
+      "grad_norm": 0.5461247563362122,
+      "learning_rate": 2e-05,
+      "loss": 0.7394,
+      "step": 4990
+    },
+    {
+      "epoch": 0.2341262408690766,
+      "grad_norm": 0.47626641392707825,
+      "learning_rate": 2e-05,
+      "loss": 0.7319,
+      "step": 5000
+    },
+    {
+      "epoch": 0.23459449335081475,
+      "grad_norm": 0.53751540184021,
+      "learning_rate": 2e-05,
+      "loss": 0.6977,
+      "step": 5010
+    },
+    {
+      "epoch": 0.2350627458325529,
+      "grad_norm": 0.448814332485199,
+      "learning_rate": 2e-05,
+      "loss": 0.767,
+      "step": 5020
+    },
+    {
+      "epoch": 0.23553099831429106,
+      "grad_norm": 0.5457051992416382,
+      "learning_rate": 2e-05,
+      "loss": 0.7231,
+      "step": 5030
+    },
+    {
+      "epoch": 0.2359992507960292,
+      "grad_norm": 0.5122005343437195,
+      "learning_rate": 2e-05,
+      "loss": 0.7569,
+      "step": 5040
+    },
+    {
+      "epoch": 0.23646750327776736,
+      "grad_norm": 0.47288933396339417,
+      "learning_rate": 2e-05,
+      "loss": 0.7804,
+      "step": 5050
+    },
+    {
+      "epoch": 0.2369357557595055,
+      "grad_norm": 0.5332443118095398,
+      "learning_rate": 2e-05,
+      "loss": 0.7425,
+      "step": 5060
+    },
+    {
+      "epoch": 0.2374040082412437,
+      "grad_norm": 0.47076839208602905,
+      "learning_rate": 2e-05,
+      "loss": 0.7612,
+      "step": 5070
+    },
+    {
+      "epoch": 0.23787226072298184,
+      "grad_norm": 0.4321301579475403,
+      "learning_rate": 2e-05,
+      "loss": 0.7663,
+      "step": 5080
+    },
+    {
+      "epoch": 0.23834051320472,
+      "grad_norm": 0.5092326998710632,
+      "learning_rate": 2e-05,
+      "loss": 0.733,
+      "step": 5090
+    },
+    {
+      "epoch": 0.23880876568645815,
+      "grad_norm": 0.45297569036483765,
+      "learning_rate": 2e-05,
+      "loss": 0.7389,
+      "step": 5100
+    },
+    {
+      "epoch": 0.2392770181681963,
+      "grad_norm": 0.44341638684272766,
+      "learning_rate": 2e-05,
+      "loss": 0.7514,
+      "step": 5110
+    },
+    {
+      "epoch": 0.23974527064993445,
+      "grad_norm": 0.4600866436958313,
+      "learning_rate": 2e-05,
+      "loss": 0.7186,
+      "step": 5120
+    },
+    {
+      "epoch": 0.2402135231316726,
+      "grad_norm": 0.6210048794746399,
+      "learning_rate": 2e-05,
+      "loss": 0.7353,
+      "step": 5130
+    },
+    {
+      "epoch": 0.24068177561341075,
+      "grad_norm": 0.4804288148880005,
+      "learning_rate": 2e-05,
+      "loss": 0.7576,
+      "step": 5140
+    },
+    {
+      "epoch": 0.2411500280951489,
+      "grad_norm": 0.5150524973869324,
+      "learning_rate": 2e-05,
+      "loss": 0.7246,
+      "step": 5150
+    },
+    {
+      "epoch": 0.24161828057688706,
+      "grad_norm": 0.4765327572822571,
+      "learning_rate": 2e-05,
+      "loss": 0.7414,
+      "step": 5160
+    },
+    {
+      "epoch": 0.2420865330586252,
+      "grad_norm": 0.4479826092720032,
+      "learning_rate": 2e-05,
+      "loss": 0.7266,
+      "step": 5170
+    },
+    {
+      "epoch": 0.24255478554036336,
+      "grad_norm": 0.43373504281044006,
+      "learning_rate": 2e-05,
+      "loss": 0.7458,
+      "step": 5180
+    },
+    {
+      "epoch": 0.2430230380221015,
+      "grad_norm": 0.46008190512657166,
+      "learning_rate": 2e-05,
+      "loss": 0.7516,
+      "step": 5190
+    },
+    {
+      "epoch": 0.24349129050383966,
+      "grad_norm": 0.4433428943157196,
+      "learning_rate": 2e-05,
+      "loss": 0.7338,
+      "step": 5200
+    },
+    {
+      "epoch": 0.24395954298557782,
+      "grad_norm": 0.4364524185657501,
+      "learning_rate": 2e-05,
+      "loss": 0.7555,
+      "step": 5210
+    },
+    {
+      "epoch": 0.24442779546731597,
+      "grad_norm": 0.5086431503295898,
+      "learning_rate": 2e-05,
+      "loss": 0.7453,
+      "step": 5220
+    },
+    {
+      "epoch": 0.24489604794905412,
+      "grad_norm": 0.5044958591461182,
+      "learning_rate": 2e-05,
+      "loss": 0.7489,
+      "step": 5230
+    },
+    {
+      "epoch": 0.24536430043079227,
+      "grad_norm": 0.4961252212524414,
+      "learning_rate": 2e-05,
+      "loss": 0.7634,
+      "step": 5240
+    },
+    {
+      "epoch": 0.24583255291253045,
+      "grad_norm": 0.42669373750686646,
+      "learning_rate": 2e-05,
+      "loss": 0.7296,
+      "step": 5250
+    },
+    {
+      "epoch": 0.2463008053942686,
+      "grad_norm": 0.4691210985183716,
+      "learning_rate": 2e-05,
+      "loss": 0.7411,
+      "step": 5260
+    },
+    {
+      "epoch": 0.24676905787600675,
+      "grad_norm": 0.6449697017669678,
+      "learning_rate": 2e-05,
+      "loss": 0.7307,
+      "step": 5270
+    },
+    {
+      "epoch": 0.2472373103577449,
+      "grad_norm": 0.49102461338043213,
+      "learning_rate": 2e-05,
+      "loss": 0.7721,
+      "step": 5280
+    },
+    {
+      "epoch": 0.24770556283948306,
+      "grad_norm": 0.41137269139289856,
+      "learning_rate": 2e-05,
+      "loss": 0.761,
+      "step": 5290
+    },
+    {
+      "epoch": 0.2481738153212212,
+      "grad_norm": 0.4368797242641449,
+      "learning_rate": 2e-05,
+      "loss": 0.7406,
+      "step": 5300
+    },
+    {
+      "epoch": 0.24864206780295936,
+      "grad_norm": 0.5086705088615417,
+      "learning_rate": 2e-05,
+      "loss": 0.7588,
+      "step": 5310
+    },
+    {
+      "epoch": 0.2491103202846975,
+      "grad_norm": 0.5293259620666504,
+      "learning_rate": 2e-05,
+      "loss": 0.7583,
+      "step": 5320
+    },
+    {
+      "epoch": 0.24957857276643566,
+      "grad_norm": 0.43683844804763794,
+      "learning_rate": 2e-05,
+      "loss": 0.7231,
+      "step": 5330
+    },
+    {
+      "epoch": 0.25004682524817384,
+      "grad_norm": 0.45331719517707825,
+      "learning_rate": 2e-05,
+      "loss": 0.6957,
+      "step": 5340
+    },
+    {
+      "epoch": 0.250515077729912,
+      "grad_norm": 0.465107798576355,
+      "learning_rate": 1.999194520388485e-05,
+      "loss": 0.7405,
+      "step": 5350
+    },
+    {
+      "epoch": 0.25098333021165015,
+      "grad_norm": 0.43206900358200073,
+      "learning_rate": 1.9959247303157763e-05,
+      "loss": 0.7123,
+      "step": 5360
+    },
+    {
+      "epoch": 0.2514515826933883,
+      "grad_norm": 0.43132132291793823,
+      "learning_rate": 1.990149424459817e-05,
+      "loss": 0.7311,
+      "step": 5370
+    },
+    {
+      "epoch": 0.25191983517512645,
+      "grad_norm": 0.48592984676361084,
+      "learning_rate": 1.9818847538648688e-05,
+      "loss": 0.714,
+      "step": 5380
+    },
+    {
+      "epoch": 0.2523880876568646,
+      "grad_norm": 0.6471542716026306,
+      "learning_rate": 1.971153831256721e-05,
+      "loss": 0.7647,
+      "step": 5390
+    },
+    {
+      "epoch": 0.25285634013860275,
+      "grad_norm": 0.48665595054626465,
+      "learning_rate": 1.9579866664063435e-05,
+      "loss": 0.7294,
+      "step": 5400
+    },
+    {
+      "epoch": 0.2533245926203409,
+      "grad_norm": 0.4819086492061615,
+      "learning_rate": 1.9424200822054782e-05,
+      "loss": 0.7224,
+      "step": 5410
+    },
+    {
+      "epoch": 0.25379284510207906,
+      "grad_norm": 0.5288002490997314,
+      "learning_rate": 1.9244976116888626e-05,
+      "loss": 0.7389,
+      "step": 5420
+    },
+    {
+      "epoch": 0.2542610975838172,
+      "grad_norm": 0.5115023255348206,
+      "learning_rate": 1.904269376291071e-05,
+      "loss": 0.7439,
+      "step": 5430
+    },
+    {
+      "epoch": 0.25472935006555536,
+      "grad_norm": 0.4631025195121765,
+      "learning_rate": 1.88179194567844e-05,
+      "loss": 0.7255,
+      "step": 5440
+    },
+    {
+      "epoch": 0.2551976025472935,
+      "grad_norm": 0.7528506517410278,
+      "learning_rate": 1.8571281795480632e-05,
+      "loss": 0.7121,
+      "step": 5450
+    },
+    {
+      "epoch": 0.25566585502903166,
+      "grad_norm": 0.4710090756416321,
+      "learning_rate": 1.8303470518362808e-05,
+      "loss": 0.7388,
+      "step": 5460
+    },
+    {
+      "epoch": 0.2561341075107698,
+      "grad_norm": 0.4562467634677887,
+      "learning_rate": 1.801523457828271e-05,
+      "loss": 0.7255,
+      "step": 5470
+    },
+    {
+      "epoch": 0.25660235999250797,
+      "grad_norm": 0.4709988236427307,
+      "learning_rate": 1.7707380047081796e-05,
+      "loss": 0.7584,
+      "step": 5480
+    },
+    {
+      "epoch": 0.2570706124742461,
+      "grad_norm": 0.5111814737319946,
+      "learning_rate": 1.7380767861355252e-05,
+      "loss": 0.7401,
+      "step": 5490
+    },
+    {
+      "epoch": 0.25753886495598427,
+      "grad_norm": 0.47118183970451355,
+      "learning_rate": 1.703631141478299e-05,
+      "loss": 0.7427,
+      "step": 5500
+    },
+    {
+      "epoch": 0.2580071174377224,
+      "grad_norm": 0.40499457716941833,
+      "learning_rate": 1.6674974003760705e-05,
+      "loss": 0.7324,
+      "step": 5510
+    },
+    {
+      "epoch": 0.2584753699194606,
+      "grad_norm": 0.46743279695510864,
+      "learning_rate": 1.6297766133474556e-05,
+      "loss": 0.727,
+      "step": 5520
+    },
+    {
+      "epoch": 0.2589436224011987,
+      "grad_norm": 0.44490283727645874,
+      "learning_rate": 1.5905742691953186e-05,
+      "loss": 0.747,
+      "step": 5530
+    },
+    {
+      "epoch": 0.2594118748829369,
+      "grad_norm": 0.5094086527824402,
+      "learning_rate": 1.55e-05,
+      "loss": 0.7585,
+      "step": 5540
+    },
+    {
+      "epoch": 0.25988012736467503,
+      "grad_norm": 0.6570689678192139,
+      "learning_rate": 1.5081672745255862e-05,
+      "loss": 0.7221,
+      "step": 5550
+    },
+    {
+      "epoch": 0.2603483798464132,
+      "grad_norm": 0.4389893114566803,
+      "learning_rate": 1.4651930808966206e-05,
+      "loss": 0.7258,
+      "step": 5560
+    },
+    {
+      "epoch": 0.26081663232815133,
+      "grad_norm": 0.6062896251678467,
+      "learning_rate": 1.4211975994326848e-05,
+      "loss": 0.7534,
+      "step": 5570
+    },
+    {
+      "epoch": 0.2612848848098895,
+      "grad_norm": 0.4593414068222046,
+      "learning_rate": 1.376303866555779e-05,
+      "loss": 0.6882,
+      "step": 5580
+    },
+    {
+      "epoch": 0.26175313729162764,
+      "grad_norm": 0.49872151017189026,
+      "learning_rate": 1.3306374307104164e-05,
+      "loss": 0.7131,
+      "step": 5590
+    },
+    {
+      "epoch": 0.2622213897733658,
+      "grad_norm": 0.42910924553871155,
+      "learning_rate": 1.2843260012586718e-05,
+      "loss": 0.7581,
+      "step": 5600
+    },
+    {
+      "epoch": 0.26268964225510394,
+      "grad_norm": 0.4514053761959076,
+      "learning_rate": 1.2374990913320726e-05,
+      "loss": 0.7269,
+      "step": 5610
+    },
+    {
+      "epoch": 0.2631578947368421,
+      "grad_norm": 0.49845200777053833,
+      "learning_rate": 1.1902876556391247e-05,
+      "loss": 0.742,
+      "step": 5620
+    },
+    {
+      "epoch": 0.26362614721858024,
+      "grad_norm": 0.4507065415382385,
+      "learning_rate": 1.1428237242413685e-05,
+      "loss": 0.7294,
+      "step": 5630
+    },
+    {
+      "epoch": 0.2640943997003184,
+      "grad_norm": 0.49284324049949646,
+      "learning_rate": 1.095240033322132e-05,
+      "loss": 0.7128,
+      "step": 5640
+    },
+    {
+      "epoch": 0.26456265218205655,
+      "grad_norm": 0.46432697772979736,
+      "learning_rate": 1.047669653980572e-05,
+      "loss": 0.7643,
+      "step": 5650
+    },
+    {
+      "epoch": 0.2650309046637947,
+      "grad_norm": 0.5275791883468628,
+      "learning_rate": 1.0002456200890902e-05,
+      "loss": 0.7324,
+      "step": 5660
+    },
+    {
+      "epoch": 0.26549915714553285,
+      "grad_norm": 0.4414176344871521,
+      "learning_rate": 9.531005562548668e-06,
+      "loss": 0.7486,
+      "step": 5670
+    },
+    {
+      "epoch": 0.265967409627271,
+      "grad_norm": 0.4890166223049164,
+      "learning_rate": 9.06366306925925e-06,
+      "loss": 0.7772,
+      "step": 5680
+    },
+    {
+      "epoch": 0.26643566210900915,
+      "grad_norm": 0.534798264503479,
+      "learning_rate": 8.601735676789686e-06,
+      "loss": 0.7124,
+      "step": 5690
+    },
+    {
+      "epoch": 0.2669039145907473,
+      "grad_norm": 0.514822244644165,
+      "learning_rate": 8.14651519720118e-06,
+      "loss": 0.7197,
+      "step": 5700
+    },
+    {
+      "epoch": 0.2673721670724855,
+      "grad_norm": 0.49526098370552063,
+      "learning_rate": 7.699274686206866e-06,
+      "loss": 0.7178,
+      "step": 5710
+    },
+    {
+      "epoch": 0.26784041955422366,
+      "grad_norm": 0.4386015832424164,
+      "learning_rate": 7.261264882983024e-06,
+      "loss": 0.6913,
+      "step": 5720
+    },
+    {
+      "epoch": 0.2683086720359618,
+      "grad_norm": 0.4785960912704468,
+      "learning_rate": 6.8337107123900386e-06,
+      "loss": 0.751,
+      "step": 5730
+    },
+    {
+      "epoch": 0.26877692451769997,
+      "grad_norm": 0.45442572236061096,
+      "learning_rate": 6.417807859384881e-06,
+      "loss": 0.7178,
+      "step": 5740
+    },
+    {
+      "epoch": 0.2692451769994381,
+      "grad_norm": 0.7225289940834045,
+      "learning_rate": 6.014719425205009e-06,
+      "loss": 0.7458,
+      "step": 5750
+    },
+    {
+      "epoch": 0.26971342948117627,
+      "grad_norm": 0.4711094796657562,
+      "learning_rate": 5.6255726746749275e-06,
+      "loss": 0.7363,
+      "step": 5760
+    },
+    {
+      "epoch": 0.2701816819629144,
+      "grad_norm": 0.47222524881362915,
+      "learning_rate": 5.251455883731728e-06,
+      "loss": 0.7541,
+      "step": 5770
+    },
+    {
+      "epoch": 0.2706499344446526,
+      "grad_norm": 0.5437127351760864,
+      "learning_rate": 4.893415295985812e-06,
+      "loss": 0.7053,
+      "step": 5780
+    },
+    {
+      "epoch": 0.2711181869263907,
+      "grad_norm": 0.45000630617141724,
+      "learning_rate": 4.552452196827885e-06,
+      "loss": 0.7405,
+      "step": 5790
+    },
+    {
+      "epoch": 0.2715864394081289,
+      "grad_norm": 0.4622010588645935,
+      "learning_rate": 4.229520113264785e-06,
+      "loss": 0.7319,
+      "step": 5800
+    },
+    {
+      "epoch": 0.27205469188986703,
+      "grad_norm": 0.4853818416595459,
+      "learning_rate": 3.925522147314915e-06,
+      "loss": 0.7301,
+      "step": 5810
+    },
+    {
+      "epoch": 0.2725229443716052,
+      "grad_norm": 0.5382267832756042,
+      "learning_rate": 3.641308450420675e-06,
+      "loss": 0.7154,
+      "step": 5820
+    },
+    {
+      "epoch": 0.27299119685334333,
+      "grad_norm": 0.46594077348709106,
+      "learning_rate": 3.37767384594087e-06,
+      "loss": 0.7075,
+      "step": 5830
+    },
+    {
+      "epoch": 0.2734594493350815,
+      "grad_norm": 0.45589321851730347,
+      "learning_rate": 3.1353556063719357e-06,
+      "loss": 0.7255,
+      "step": 5840
+    },
+    {
+      "epoch": 0.27392770181681964,
+      "grad_norm": 0.7770072221755981,
+      "learning_rate": 2.9150313915141474e-06,
+      "loss": 0.7316,
+      "step": 5850
+    },
+    {
+      "epoch": 0.2743959542985578,
+      "grad_norm": 0.5423017740249634,
+      "learning_rate": 2.717317353348891e-06,
+      "loss": 0.7269,
+      "step": 5860
+    },
+    {
+      "epoch": 0.27486420678029594,
+      "grad_norm": 0.4896930158138275,
+      "learning_rate": 2.5427664129268253e-06,
+      "loss": 0.7408,
+      "step": 5870
+    },
+    {
+      "epoch": 0.2753324592620341,
+      "grad_norm": 0.5106858015060425,
+      "learning_rate": 2.391866714085717e-06,
+      "loss": 0.7433,
+      "step": 5880
+    },
+    {
+      "epoch": 0.27580071174377224,
+      "grad_norm": 0.49272871017456055,
+      "learning_rate": 2.2650402583222753e-06,
+      "loss": 0.7108,
+      "step": 5890
+    },
+    {
+      "epoch": 0.2762689642255104,
+      "grad_norm": 0.49198707938194275,
+      "learning_rate": 2.1626417246356398e-06,
+      "loss": 0.7242,
+      "step": 5900
+    },
+    {
+      "epoch": 0.27673721670724855,
+      "grad_norm": 0.5110393166542053,
+      "learning_rate": 2.084957477642894e-06,
+      "loss": 0.7276,
+      "step": 5910
+    },
+    {
+      "epoch": 0.2772054691889867,
+      "grad_norm": 0.45616614818573,
+      "learning_rate": 2.0322047667405147e-06,
+      "loss": 0.6979,
+      "step": 5920
+    },
+    {
+      "epoch": 0.27767372167072485,
+      "grad_norm": 0.4469747841358185,
+      "learning_rate": 2.0045311185513342e-06,
+      "loss": 0.7245,
+      "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
+}