End of training

README.md

@@ -4,6 +4,7 @@ license: llama3.1
 base_model: meta-llama/Llama-3.1-8B
 tags:
 - llama-factory
+- full
 - generated_from_trainer
 model-index:
 - name: OH_DCFT_V3_wo_evol_instruct_140k
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # OH_DCFT_V3_wo_evol_instruct_140k
 
-This model is a fine-tuned version of [meta-llama/Llama-3.1-8B](https://huggingface.co/meta-llama/Llama-3.1-8B) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Llama-3.1-8B](https://huggingface.co/meta-llama/Llama-3.1-8B) on the mlfoundations-dev/OH_DCFT_V3_wo_evol_instruct_140k dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.6134
 

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 3.0,
-    "eval_loss": 0.6164913177490234,
-    "eval_runtime": 203.8494,
-    "eval_samples_per_second": 49.836,
+    "eval_loss": 0.6133805513381958,
+    "eval_runtime": 204.0253,
+    "eval_samples_per_second": 49.793,
     "eval_steps_per_second": 0.392,
     "total_flos": 1894048365281280.0,
-    "train_loss": 0.5957063028603713,
-    "train_runtime": 34171.7918,
-    "train_samples_per_second": 16.944,
+    "train_loss": 0.5970700682000076,
+    "train_runtime": 33949.4437,
+    "train_samples_per_second": 17.055,
     "train_steps_per_second": 0.033
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 3.0,
-    "eval_loss": 0.6164913177490234,
-    "eval_runtime": 203.8494,
-    "eval_samples_per_second": 49.836,
+    "eval_loss": 0.6133805513381958,
+    "eval_runtime": 204.0253,
+    "eval_samples_per_second": 49.793,
     "eval_steps_per_second": 0.392
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.0,
     "total_flos": 1894048365281280.0,
-    "train_loss": 0.5957063028603713,
-    "train_runtime": 34171.7918,
-    "train_samples_per_second": 16.944,
+    "train_loss": 0.5970700682000076,
+    "train_runtime": 33949.4437,
+    "train_samples_per_second": 17.055,
     "train_steps_per_second": 0.033
 }

trainer_state.json

@@ -10,826 +10,826 @@
   "log_history": [
     {
       "epoch": 0.026525198938992044,
-      "grad_norm": 2.329897814074136,
+      "grad_norm": 2.780399544451733,
       "learning_rate": 5e-06,
-      "loss": 0.8884,
+      "loss": 0.8608,
       "step": 10
     },
     {
       "epoch": 0.05305039787798409,
-      "grad_norm": 3.8595726803473838,
+      "grad_norm": 1.606321904164578,
       "learning_rate": 5e-06,
-      "loss": 0.7961,
+      "loss": 0.7691,
       "step": 20
     },
     {
       "epoch": 0.07957559681697612,
-      "grad_norm": 1.096843401953888,
+      "grad_norm": 1.4733236101897926,
       "learning_rate": 5e-06,
-      "loss": 0.7478,
+      "loss": 0.731,
       "step": 30
     },
     {
       "epoch": 0.10610079575596817,
-      "grad_norm": 1.1211395081081943,
+      "grad_norm": 1.0466130239794242,
       "learning_rate": 5e-06,
-      "loss": 0.719,
+      "loss": 0.7107,
       "step": 40
     },
     {
       "epoch": 0.13262599469496023,
-      "grad_norm": 1.4282528066175646,
+      "grad_norm": 0.9068095404759675,
       "learning_rate": 5e-06,
-      "loss": 0.7052,
+      "loss": 0.6977,
       "step": 50
     },
     {
       "epoch": 0.15915119363395225,
-      "grad_norm": 0.7596233206271922,
+      "grad_norm": 0.9537695534962456,
       "learning_rate": 5e-06,
-      "loss": 0.694,
+      "loss": 0.6882,
       "step": 60
     },
     {
       "epoch": 0.1856763925729443,
-      "grad_norm": 1.2777068139599712,
+      "grad_norm": 0.6703381757321946,
       "learning_rate": 5e-06,
-      "loss": 0.6882,
+      "loss": 0.6835,
       "step": 70
     },
     {
       "epoch": 0.21220159151193635,
-      "grad_norm": 1.1383877936666695,
+      "grad_norm": 1.00882019650449,
       "learning_rate": 5e-06,
-      "loss": 0.6771,
+      "loss": 0.6724,
       "step": 80
     },
     {
       "epoch": 0.23872679045092837,
-      "grad_norm": 1.0060703202699885,
+      "grad_norm": 1.0611084342807717,
       "learning_rate": 5e-06,
-      "loss": 0.6806,
+      "loss": 0.6764,
       "step": 90
     },
     {
       "epoch": 0.26525198938992045,
-      "grad_norm": 1.0179544972489034,
+      "grad_norm": 0.8654567865290824,
       "learning_rate": 5e-06,
-      "loss": 0.6665,
+      "loss": 0.663,
       "step": 100
     },
     {
       "epoch": 0.2917771883289125,
-      "grad_norm": 0.6720815026983992,
+      "grad_norm": 0.8547675860872707,
       "learning_rate": 5e-06,
-      "loss": 0.6637,
+      "loss": 0.6602,
       "step": 110
     },
     {
       "epoch": 0.3183023872679045,
-      "grad_norm": 0.9241587286794601,
+      "grad_norm": 0.6764470147171867,
       "learning_rate": 5e-06,
-      "loss": 0.6546,
+      "loss": 0.6511,
       "step": 120
     },
     {
       "epoch": 0.3448275862068966,
-      "grad_norm": 0.8824991950738043,
+      "grad_norm": 0.4996824578122701,
       "learning_rate": 5e-06,
-      "loss": 0.6522,
+      "loss": 0.649,
       "step": 130
     },
     {
       "epoch": 0.3713527851458886,
-      "grad_norm": 0.8302884563721783,
+      "grad_norm": 0.5555661729991889,
       "learning_rate": 5e-06,
-      "loss": 0.653,
+      "loss": 0.6499,
       "step": 140
     },
     {
       "epoch": 0.3978779840848806,
-      "grad_norm": 0.8327701504024553,
+      "grad_norm": 0.662354493585409,
       "learning_rate": 5e-06,
-      "loss": 0.6455,
+      "loss": 0.6428,
       "step": 150
     },
     {
       "epoch": 0.4244031830238727,
-      "grad_norm": 0.5492876289580202,
+      "grad_norm": 0.47582415706089126,
       "learning_rate": 5e-06,
-      "loss": 0.6553,
+      "loss": 0.6527,
       "step": 160
     },
     {
       "epoch": 0.4509283819628647,
-      "grad_norm": 0.7119323938911869,
+      "grad_norm": 1.3542813303379895,
       "learning_rate": 5e-06,
-      "loss": 0.6448,
+      "loss": 0.6423,
       "step": 170
     },
     {
       "epoch": 0.47745358090185674,
-      "grad_norm": 0.8042997280772343,
+      "grad_norm": 0.5202115494812097,
       "learning_rate": 5e-06,
-      "loss": 0.6403,
+      "loss": 0.6379,
       "step": 180
     },
     {
       "epoch": 0.5039787798408488,
-      "grad_norm": 0.7810674802506331,
+      "grad_norm": 0.5270125702790065,
       "learning_rate": 5e-06,
-      "loss": 0.655,
+      "loss": 0.6523,
       "step": 190
     },
     {
       "epoch": 0.5305039787798409,
-      "grad_norm": 0.5391416674289061,
+      "grad_norm": 0.4787233056340413,
       "learning_rate": 5e-06,
-      "loss": 0.6398,
+      "loss": 0.6375,
       "step": 200
     },
     {
       "epoch": 0.5570291777188329,
-      "grad_norm": 0.6885851666539189,
+      "grad_norm": 0.5279042698103554,
       "learning_rate": 5e-06,
-      "loss": 0.6452,
+      "loss": 0.643,
       "step": 210
     },
     {
       "epoch": 0.583554376657825,
-      "grad_norm": 0.6168676023297931,
+      "grad_norm": 0.7066879723483964,
       "learning_rate": 5e-06,
-      "loss": 0.6443,
+      "loss": 0.6424,
       "step": 220
     },
     {
       "epoch": 0.610079575596817,
-      "grad_norm": 0.5971017602698001,
+      "grad_norm": 0.7948472968894127,
       "learning_rate": 5e-06,
-      "loss": 0.636,
+      "loss": 0.6341,
       "step": 230
     },
     {
       "epoch": 0.636604774535809,
-      "grad_norm": 0.6022582014461996,
+      "grad_norm": 0.5390839840399401,
       "learning_rate": 5e-06,
-      "loss": 0.642,
+      "loss": 0.64,
       "step": 240
     },
     {
       "epoch": 0.6631299734748011,
-      "grad_norm": 0.500260575545034,
+      "grad_norm": 0.6650275203652714,
       "learning_rate": 5e-06,
-      "loss": 0.6344,
+      "loss": 0.6325,
       "step": 250
     },
     {
       "epoch": 0.6896551724137931,
-      "grad_norm": 0.537336125918856,
+      "grad_norm": 0.5941647978617659,
       "learning_rate": 5e-06,
-      "loss": 0.639,
+      "loss": 0.6373,
       "step": 260
     },
     {
       "epoch": 0.7161803713527851,
-      "grad_norm": 0.5365620197699547,
+      "grad_norm": 0.5582608177644663,
       "learning_rate": 5e-06,
-      "loss": 0.6403,
+      "loss": 0.6385,
       "step": 270
     },
     {
       "epoch": 0.7427055702917772,
-      "grad_norm": 0.7683273733374665,
+      "grad_norm": 0.5805453901667452,
       "learning_rate": 5e-06,
-      "loss": 0.6401,
+      "loss": 0.638,
       "step": 280
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 0.5276716709186424,
+      "grad_norm": 0.4514952002612208,
       "learning_rate": 5e-06,
-      "loss": 0.6338,
+      "loss": 0.6321,
       "step": 290
     },
     {
       "epoch": 0.7957559681697612,
-      "grad_norm": 0.5074059045052902,
+      "grad_norm": 0.5044213856224046,
       "learning_rate": 5e-06,
-      "loss": 0.6371,
+      "loss": 0.6355,
       "step": 300
     },
     {
       "epoch": 0.8222811671087533,
-      "grad_norm": 0.67204833216127,
+      "grad_norm": 0.5268464944312369,
       "learning_rate": 5e-06,
-      "loss": 0.639,
+      "loss": 0.6374,
       "step": 310
     },
     {
       "epoch": 0.8488063660477454,
-      "grad_norm": 0.5619464110309291,
+      "grad_norm": 0.47822319615489534,
       "learning_rate": 5e-06,
-      "loss": 0.6427,
+      "loss": 0.6409,
       "step": 320
     },
     {
       "epoch": 0.8753315649867374,
-      "grad_norm": 0.5724558992220249,
+      "grad_norm": 0.6797410044487332,
       "learning_rate": 5e-06,
-      "loss": 0.6294,
+      "loss": 0.6278,
       "step": 330
     },
     {
       "epoch": 0.9018567639257294,
-      "grad_norm": 0.619777380357849,
+      "grad_norm": 0.5170899802801728,
       "learning_rate": 5e-06,
-      "loss": 0.6334,
+      "loss": 0.6318,
       "step": 340
     },
     {
       "epoch": 0.9283819628647215,
-      "grad_norm": 0.7434000365286857,
+      "grad_norm": 0.6765922367059989,
       "learning_rate": 5e-06,
-      "loss": 0.6329,
+      "loss": 0.6314,
       "step": 350
     },
     {
       "epoch": 0.9549071618037135,
-      "grad_norm": 0.5490872236122448,
+      "grad_norm": 0.6459795123201603,
       "learning_rate": 5e-06,
-      "loss": 0.6286,
+      "loss": 0.6274,
       "step": 360
     },
     {
       "epoch": 0.9814323607427056,
-      "grad_norm": 0.584736306154939,
+      "grad_norm": 0.44017798959016546,
       "learning_rate": 5e-06,
-      "loss": 0.6296,
+      "loss": 0.6282,
       "step": 370
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.6237567663192749,
-      "eval_runtime": 202.8984,
-      "eval_samples_per_second": 50.069,
-      "eval_steps_per_second": 0.394,
+      "eval_loss": 0.622368574142456,
+      "eval_runtime": 203.4748,
+      "eval_samples_per_second": 49.928,
+      "eval_steps_per_second": 0.393,
       "step": 377
     },
     {
       "epoch": 1.0079575596816976,
-      "grad_norm": 1.012512560233075,
+      "grad_norm": 0.8095066775828493,
       "learning_rate": 5e-06,
-      "loss": 0.614,
+      "loss": 0.6139,
       "step": 380
     },
     {
       "epoch": 1.0344827586206897,
-      "grad_norm": 0.6423484790332888,
+      "grad_norm": 0.5482836222129767,
       "learning_rate": 5e-06,
-      "loss": 0.5872,
+      "loss": 0.5899,
       "step": 390
     },
     {
       "epoch": 1.0610079575596818,
-      "grad_norm": 0.6707809448038629,
+      "grad_norm": 0.5390710199271481,
       "learning_rate": 5e-06,
-      "loss": 0.5889,
+      "loss": 0.5918,
       "step": 400
     },
     {
       "epoch": 1.0875331564986737,
-      "grad_norm": 0.6650816829620442,
+      "grad_norm": 0.5745480856709776,
       "learning_rate": 5e-06,
-      "loss": 0.5807,
+      "loss": 0.5836,
       "step": 410
     },
     {
       "epoch": 1.1140583554376657,
-      "grad_norm": 0.6224384353734153,
+      "grad_norm": 0.5786473175674526,
       "learning_rate": 5e-06,
-      "loss": 0.5915,
+      "loss": 0.5942,
       "step": 420
     },
     {
       "epoch": 1.1405835543766578,
-      "grad_norm": 0.5684723509980193,
+      "grad_norm": 0.4754533900180701,
       "learning_rate": 5e-06,
-      "loss": 0.5868,
+      "loss": 0.5894,
       "step": 430
     },
     {
       "epoch": 1.16710875331565,
-      "grad_norm": 0.7059253953598508,
+      "grad_norm": 0.626662902645802,
       "learning_rate": 5e-06,
-      "loss": 0.5766,
+      "loss": 0.5792,
       "step": 440
     },
     {
       "epoch": 1.193633952254642,
-      "grad_norm": 0.523842539770435,
+      "grad_norm": 0.4520996771126915,
       "learning_rate": 5e-06,
-      "loss": 0.5848,
+      "loss": 0.5877,
       "step": 450
     },
     {
       "epoch": 1.2201591511936338,
-      "grad_norm": 0.6714226112627073,
+      "grad_norm": 0.48945343637681654,
       "learning_rate": 5e-06,
-      "loss": 0.5832,
+      "loss": 0.5859,
       "step": 460
     },
     {
       "epoch": 1.246684350132626,
-      "grad_norm": 0.7286525903858807,
+      "grad_norm": 0.464325446341802,
       "learning_rate": 5e-06,
-      "loss": 0.589,
+      "loss": 0.5915,
       "step": 470
     },
     {
       "epoch": 1.273209549071618,
-      "grad_norm": 0.52664344746201,
+      "grad_norm": 0.46045963914114574,
       "learning_rate": 5e-06,
-      "loss": 0.5781,
+      "loss": 0.5807,
       "step": 480
     },
     {
       "epoch": 1.29973474801061,
-      "grad_norm": 0.5394135765946012,
+      "grad_norm": 0.49722516653390847,
       "learning_rate": 5e-06,
-      "loss": 0.5793,
+      "loss": 0.5816,
       "step": 490
     },
     {
       "epoch": 1.3262599469496021,
-      "grad_norm": 0.5473942505095254,
+      "grad_norm": 0.5455802289844164,
       "learning_rate": 5e-06,
-      "loss": 0.5865,
+      "loss": 0.5887,
       "step": 500
     },
     {
       "epoch": 1.3527851458885942,
-      "grad_norm": 0.5239209443855906,
+      "grad_norm": 0.4427674181447264,
       "learning_rate": 5e-06,
-      "loss": 0.5851,
+      "loss": 0.5875,
       "step": 510
     },
     {
       "epoch": 1.3793103448275863,
-      "grad_norm": 0.638268567207473,
+      "grad_norm": 0.6045736193729062,
       "learning_rate": 5e-06,
-      "loss": 0.5868,
+      "loss": 0.5891,
       "step": 520
     },
     {
       "epoch": 1.4058355437665782,
-      "grad_norm": 0.5569448150591939,
+      "grad_norm": 0.45208550166771494,
       "learning_rate": 5e-06,
-      "loss": 0.5851,
+      "loss": 0.5876,
       "step": 530
     },
     {
       "epoch": 1.4323607427055702,
-      "grad_norm": 0.5739975276281734,
+      "grad_norm": 0.5217117204135557,
       "learning_rate": 5e-06,
-      "loss": 0.5807,
+      "loss": 0.5831,
       "step": 540
     },
     {
       "epoch": 1.4588859416445623,
-      "grad_norm": 0.5662395827613842,
+      "grad_norm": 0.5393363462032307,
       "learning_rate": 5e-06,
-      "loss": 0.5768,
+      "loss": 0.579,
       "step": 550
     },
     {
       "epoch": 1.4854111405835544,
-      "grad_norm": 0.7163603774804506,
+      "grad_norm": 0.6515785188868304,
       "learning_rate": 5e-06,
-      "loss": 0.5843,
+      "loss": 0.5867,
       "step": 560
     },
     {
       "epoch": 1.5119363395225465,
-      "grad_norm": 0.7163733609893056,
+      "grad_norm": 0.6394000137643647,
       "learning_rate": 5e-06,
-      "loss": 0.5917,
+      "loss": 0.5939,
       "step": 570
     },
     {
       "epoch": 1.5384615384615383,
-      "grad_norm": 0.6180542754631709,
+      "grad_norm": 0.5074512553300523,
       "learning_rate": 5e-06,
-      "loss": 0.5817,
+      "loss": 0.5839,
       "step": 580
     },
     {
       "epoch": 1.5649867374005306,
-      "grad_norm": 0.5638514267037327,
+      "grad_norm": 0.5172554765605915,
       "learning_rate": 5e-06,
-      "loss": 0.5862,
+      "loss": 0.5886,
       "step": 590
     },
     {
       "epoch": 1.5915119363395225,
-      "grad_norm": 0.561015542053421,
+      "grad_norm": 0.4946796812261883,
       "learning_rate": 5e-06,
-      "loss": 0.5815,
+      "loss": 0.5837,
       "step": 600
     },
     {
       "epoch": 1.6180371352785146,
-      "grad_norm": 0.5019925924846618,
+      "grad_norm": 0.45275057698879645,
       "learning_rate": 5e-06,
-      "loss": 0.5854,
+      "loss": 0.5875,
       "step": 610
     },
     {
       "epoch": 1.6445623342175066,
-      "grad_norm": 0.6456145062380878,
+      "grad_norm": 0.6141728231437198,
       "learning_rate": 5e-06,
-      "loss": 0.584,
+      "loss": 0.5863,
       "step": 620
     },
     {
       "epoch": 1.6710875331564987,
-      "grad_norm": 0.6273758065445275,
+      "grad_norm": 0.5764229361365383,
       "learning_rate": 5e-06,
-      "loss": 0.5794,
+      "loss": 0.5818,
       "step": 630
     },
     {
       "epoch": 1.6976127320954908,
-      "grad_norm": 0.47537875001219887,
+      "grad_norm": 0.43948836202023195,
       "learning_rate": 5e-06,
-      "loss": 0.5831,
+      "loss": 0.5854,
       "step": 640
     },
     {
       "epoch": 1.7241379310344827,
-      "grad_norm": 0.5799224686148554,
+      "grad_norm": 0.6212110766567985,
       "learning_rate": 5e-06,
-      "loss": 0.5876,
+      "loss": 0.5899,
       "step": 650
     },
     {
       "epoch": 1.750663129973475,
-      "grad_norm": 0.6477709788042249,
+      "grad_norm": 0.5230421236325133,
       "learning_rate": 5e-06,
-      "loss": 0.5786,
+      "loss": 0.5809,
       "step": 660
     },
     {
       "epoch": 1.7771883289124668,
-      "grad_norm": 0.5398722777232823,
+      "grad_norm": 0.42596513265710756,
       "learning_rate": 5e-06,
-      "loss": 0.5788,
+      "loss": 0.581,
       "step": 670
     },
     {
       "epoch": 1.8037135278514589,
-      "grad_norm": 0.503790352480121,
+      "grad_norm": 0.5019437503087736,
       "learning_rate": 5e-06,
-      "loss": 0.5789,
+      "loss": 0.5811,
       "step": 680
     },
     {
       "epoch": 1.830238726790451,
-      "grad_norm": 0.5980642131822352,
+      "grad_norm": 0.5695097062913548,
       "learning_rate": 5e-06,
-      "loss": 0.5836,
+      "loss": 0.5858,
       "step": 690
     },
     {
       "epoch": 1.8567639257294428,
-      "grad_norm": 0.5144731472377694,
+      "grad_norm": 0.48071601140752834,
       "learning_rate": 5e-06,
-      "loss": 0.5818,
+      "loss": 0.584,
       "step": 700
     },
     {
       "epoch": 1.8832891246684351,
-      "grad_norm": 0.748547689970371,
+      "grad_norm": 0.6322423860046313,
       "learning_rate": 5e-06,
-      "loss": 0.5882,
+      "loss": 0.5901,
       "step": 710
     },
     {
       "epoch": 1.909814323607427,
-      "grad_norm": 0.5454719348703825,
+      "grad_norm": 0.504386977138495,
       "learning_rate": 5e-06,
-      "loss": 0.5857,
+      "loss": 0.5878,
       "step": 720
     },
     {
       "epoch": 1.936339522546419,
-      "grad_norm": 0.537701148828446,
+      "grad_norm": 0.5132407169571725,
       "learning_rate": 5e-06,
-      "loss": 0.5919,
+      "loss": 0.5939,
       "step": 730
     },
     {
       "epoch": 1.9628647214854111,
-      "grad_norm": 0.5333812418899406,
+      "grad_norm": 0.49747549542012004,
       "learning_rate": 5e-06,
-      "loss": 0.5804,
+      "loss": 0.5824,
       "step": 740
     },
     {
       "epoch": 1.9893899204244032,
-      "grad_norm": 0.5945192673823688,
+      "grad_norm": 0.6053057729936807,
       "learning_rate": 5e-06,
-      "loss": 0.5875,
+      "loss": 0.5895,
       "step": 750
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.6138430833816528,
-      "eval_runtime": 204.1448,
-      "eval_samples_per_second": 49.764,
-      "eval_steps_per_second": 0.392,
+      "eval_loss": 0.6123443841934204,
+      "eval_runtime": 203.0015,
+      "eval_samples_per_second": 50.044,
+      "eval_steps_per_second": 0.394,
       "step": 754
     },
     {
       "epoch": 2.0159151193633953,
-      "grad_norm": 0.70666737481995,
+      "grad_norm": 0.6355273791432589,
       "learning_rate": 5e-06,
-      "loss": 0.5507,
+      "loss": 0.5558,
       "step": 760
     },
     {
       "epoch": 2.042440318302387,
-      "grad_norm": 0.6653710012047342,
+      "grad_norm": 0.5643951304254625,
       "learning_rate": 5e-06,
-      "loss": 0.5366,
+      "loss": 0.5439,
       "step": 770
     },
     {
       "epoch": 2.0689655172413794,
-      "grad_norm": 0.6392050869043902,
+      "grad_norm": 0.4960175423111283,
       "learning_rate": 5e-06,
-      "loss": 0.5362,
+      "loss": 0.5434,
       "step": 780
     },
     {
       "epoch": 2.0954907161803713,
-      "grad_norm": 0.552417268591106,
+      "grad_norm": 0.5284981337718996,
       "learning_rate": 5e-06,
-      "loss": 0.537,
+      "loss": 0.5441,
       "step": 790
     },
     {
       "epoch": 2.1220159151193636,
-      "grad_norm": 0.5237418246816455,
+      "grad_norm": 0.5187766913446101,
       "learning_rate": 5e-06,
-      "loss": 0.5387,
+      "loss": 0.5457,
       "step": 800
     },
     {
       "epoch": 2.1485411140583555,
-      "grad_norm": 0.5931401358357342,
+      "grad_norm": 0.4676393193911655,
       "learning_rate": 5e-06,
-      "loss": 0.5383,
+      "loss": 0.5453,
       "step": 810
     },
     {
       "epoch": 2.1750663129973473,
-      "grad_norm": 0.5859178511093684,
+      "grad_norm": 0.447245777123748,
       "learning_rate": 5e-06,
-      "loss": 0.5385,
+      "loss": 0.5454,
       "step": 820
     },
     {
       "epoch": 2.2015915119363396,
-      "grad_norm": 0.54174318871251,
+      "grad_norm": 0.4803013114072548,
       "learning_rate": 5e-06,
-      "loss": 0.536,
+      "loss": 0.5427,
       "step": 830
     },
     {
       "epoch": 2.2281167108753315,
-      "grad_norm": 0.6719837144625476,
+      "grad_norm": 0.6305967900882212,
       "learning_rate": 5e-06,
-      "loss": 0.5443,
+      "loss": 0.5512,
       "step": 840
     },
     {
       "epoch": 2.2546419098143238,
-      "grad_norm": 0.5913384023217347,
+      "grad_norm": 0.4660714087302693,
       "learning_rate": 5e-06,
-      "loss": 0.5275,
+      "loss": 0.5337,
       "step": 850
     },
     {
       "epoch": 2.2811671087533156,
-      "grad_norm": 0.5589992603235397,
+      "grad_norm": 0.5499183474925715,
       "learning_rate": 5e-06,
-      "loss": 0.5324,
+      "loss": 0.539,
       "step": 860
     },
     {
       "epoch": 2.3076923076923075,
-      "grad_norm": 0.7009013738100872,
+      "grad_norm": 0.6029739750306148,
       "learning_rate": 5e-06,
-      "loss": 0.5358,
+      "loss": 0.5423,
       "step": 870
     },
     {
       "epoch": 2.3342175066313,
-      "grad_norm": 0.5192518963562719,
+      "grad_norm": 0.5459822127048733,
       "learning_rate": 5e-06,
-      "loss": 0.5397,
+      "loss": 0.5462,
       "step": 880
     },
     {
       "epoch": 2.3607427055702916,
-      "grad_norm": 0.5673317232053819,
+      "grad_norm": 0.4843657467722299,
       "learning_rate": 5e-06,
-      "loss": 0.5441,
+      "loss": 0.5504,
       "step": 890
     },
     {
       "epoch": 2.387267904509284,
-      "grad_norm": 0.6565705800504016,
+      "grad_norm": 0.6497699445469315,
       "learning_rate": 5e-06,
-      "loss": 0.5376,
+      "loss": 0.5436,
       "step": 900
     },
     {
       "epoch": 2.413793103448276,
-      "grad_norm": 0.6497220724543952,
+      "grad_norm": 0.4716521124138746,
       "learning_rate": 5e-06,
-      "loss": 0.5309,
+      "loss": 0.537,
       "step": 910
     },
     {
       "epoch": 2.4403183023872677,
-      "grad_norm": 0.5907005582908587,
+      "grad_norm": 0.5180881653993036,
       "learning_rate": 5e-06,
-      "loss": 0.535,
+      "loss": 0.5412,
       "step": 920
     },
     {
       "epoch": 2.46684350132626,
-      "grad_norm": 0.6229830599060672,
+      "grad_norm": 0.6139310358311121,
       "learning_rate": 5e-06,
-      "loss": 0.5497,
+      "loss": 0.5563,
       "step": 930
     },
     {
       "epoch": 2.493368700265252,
-      "grad_norm": 0.5474753923784902,
+      "grad_norm": 0.5199070076942287,
       "learning_rate": 5e-06,
-      "loss": 0.5409,
+      "loss": 0.547,
       "step": 940
     },
     {
       "epoch": 2.519893899204244,
-      "grad_norm": 0.5472523872985887,
+      "grad_norm": 0.47708958463613943,
       "learning_rate": 5e-06,
-      "loss": 0.536,
+      "loss": 0.5419,
       "step": 950
     },
     {
       "epoch": 2.546419098143236,
-      "grad_norm": 0.9882268568658281,
+      "grad_norm": 0.6389465583810601,
       "learning_rate": 5e-06,
-      "loss": 0.5375,
+      "loss": 0.5433,
       "step": 960
     },
     {
       "epoch": 2.5729442970822283,
-      "grad_norm": 0.6076358216087694,
+      "grad_norm": 0.46642283536695006,
       "learning_rate": 5e-06,
-      "loss": 0.5417,
+      "loss": 0.5476,
       "step": 970
     },
     {
       "epoch": 2.59946949602122,
-      "grad_norm": 0.6233857713542503,
+      "grad_norm": 0.5177895417014805,
       "learning_rate": 5e-06,
-      "loss": 0.543,
+      "loss": 0.549,
       "step": 980
     },
     {
       "epoch": 2.6259946949602124,
-      "grad_norm": 0.662156931341361,
+      "grad_norm": 0.5889398648766363,
       "learning_rate": 5e-06,
-      "loss": 0.5437,
+      "loss": 0.5497,
       "step": 990
     },
     {
       "epoch": 2.6525198938992043,
-      "grad_norm": 0.520667113721229,
+      "grad_norm": 0.49922378313664745,
       "learning_rate": 5e-06,
-      "loss": 0.5392,
+      "loss": 0.5451,
       "step": 1000
     },
     {
       "epoch": 2.679045092838196,
-      "grad_norm": 0.5971607456495881,
+      "grad_norm": 0.5386720121866165,
       "learning_rate": 5e-06,
-      "loss": 0.5357,
+      "loss": 0.5415,
       "step": 1010
     },
     {
       "epoch": 2.7055702917771884,
-      "grad_norm": 0.5939890330376003,
+      "grad_norm": 0.500783203019607,
       "learning_rate": 5e-06,
-      "loss": 0.5371,
+      "loss": 0.5428,
       "step": 1020
     },
     {
       "epoch": 2.7320954907161803,
-      "grad_norm": 0.5773060764723967,
+      "grad_norm": 0.4931032038157372,
       "learning_rate": 5e-06,
-      "loss": 0.5359,
+      "loss": 0.5416,
       "step": 1030
     },
     {
       "epoch": 2.7586206896551726,
-      "grad_norm": 0.5827565949490094,
+      "grad_norm": 0.5391229756876829,
       "learning_rate": 5e-06,
-      "loss": 0.5431,
+      "loss": 0.5487,
       "step": 1040
     },
     {
       "epoch": 2.7851458885941645,
-      "grad_norm": 0.6314529885890559,
+      "grad_norm": 0.5489614966946841,
       "learning_rate": 5e-06,
-      "loss": 0.5444,
+      "loss": 0.5499,
       "step": 1050
     },
     {
       "epoch": 2.8116710875331563,
-      "grad_norm": 0.5422009039427094,
+      "grad_norm": 0.4779229991518947,
       "learning_rate": 5e-06,
-      "loss": 0.54,
+      "loss": 0.5457,
       "step": 1060
     },
     {
       "epoch": 2.8381962864721486,
-      "grad_norm": 0.6028981522236778,
+      "grad_norm": 0.52170064853256,
       "learning_rate": 5e-06,
-      "loss": 0.5433,
+      "loss": 0.549,
       "step": 1070
     },
     {
       "epoch": 2.8647214854111405,
-      "grad_norm": 0.5970004720986751,
+      "grad_norm": 0.5091309568444623,
       "learning_rate": 5e-06,
-      "loss": 0.5461,
+      "loss": 0.552,
       "step": 1080
     },
     {
       "epoch": 2.8912466843501328,
-      "grad_norm": 0.574798524733574,
+      "grad_norm": 0.45335922455307415,
       "learning_rate": 5e-06,
-      "loss": 0.5397,
+      "loss": 0.5453,
       "step": 1090
     },
     {
       "epoch": 2.9177718832891246,
-      "grad_norm": 0.5790776127397061,
+      "grad_norm": 0.565478737240957,
       "learning_rate": 5e-06,
-      "loss": 0.5404,
+      "loss": 0.546,
       "step": 1100
     },
     {
       "epoch": 2.9442970822281165,
-      "grad_norm": 0.5249016599382924,
+      "grad_norm": 0.5673285594805738,
       "learning_rate": 5e-06,
-      "loss": 0.5461,
+      "loss": 0.5517,
       "step": 1110
     },
     {
       "epoch": 2.970822281167109,
-      "grad_norm": 0.6578189789009287,
+      "grad_norm": 0.503611299440423,
       "learning_rate": 5e-06,
-      "loss": 0.5387,
+      "loss": 0.5444,
       "step": 1120
     },
     {
       "epoch": 2.9973474801061006,
-      "grad_norm": 0.5226173945421619,
+      "grad_norm": 0.5499097253524401,
       "learning_rate": 5e-06,
-      "loss": 0.5391,
+      "loss": 0.5447,
       "step": 1130
     },
     {
       "epoch": 3.0,
-      "eval_loss": 0.6164913177490234,
-      "eval_runtime": 204.5208,
-      "eval_samples_per_second": 49.672,
-      "eval_steps_per_second": 0.391,
+      "eval_loss": 0.6133805513381958,
+      "eval_runtime": 203.8433,
+      "eval_samples_per_second": 49.837,
+      "eval_steps_per_second": 0.392,
       "step": 1131
     },
     {
       "epoch": 3.0,
       "step": 1131,
       "total_flos": 1894048365281280.0,
-      "train_loss": 0.5957063028603713,
-      "train_runtime": 34171.7918,
-      "train_samples_per_second": 16.944,
+      "train_loss": 0.5970700682000076,
+      "train_runtime": 33949.4437,
+      "train_samples_per_second": 17.055,
       "train_steps_per_second": 0.033
     }
   ],