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_unnatural_instructions
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # OH_DCFT_V3_wo_unnatural_instructions
 
-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_unnatural_instructions dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.6454
 

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 2.9946777054997042,
-    "eval_loss": 0.648719847202301,
-    "eval_runtime": 224.9555,
-    "eval_samples_per_second": 50.619,
-    "eval_steps_per_second": 0.396,
+    "eval_loss": 0.6453979015350342,
+    "eval_runtime": 225.6683,
+    "eval_samples_per_second": 50.459,
+    "eval_steps_per_second": 0.394,
     "total_flos": 2120178393415680.0,
-    "train_loss": 0.6180764295478568,
-    "train_runtime": 37891.5513,
-    "train_samples_per_second": 17.129,
+    "train_loss": 0.6197805719164687,
+    "train_runtime": 38167.6556,
+    "train_samples_per_second": 17.005,
     "train_steps_per_second": 0.033
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 2.9946777054997042,
-    "eval_loss": 0.648719847202301,
-    "eval_runtime": 224.9555,
-    "eval_samples_per_second": 50.619,
-    "eval_steps_per_second": 0.396
+    "eval_loss": 0.6453979015350342,
+    "eval_runtime": 225.6683,
+    "eval_samples_per_second": 50.459,
+    "eval_steps_per_second": 0.394
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.9946777054997042,
     "total_flos": 2120178393415680.0,
-    "train_loss": 0.6180764295478568,
-    "train_runtime": 37891.5513,
-    "train_samples_per_second": 17.129,
+    "train_loss": 0.6197805719164687,
+    "train_runtime": 38167.6556,
+    "train_samples_per_second": 17.005,
     "train_steps_per_second": 0.033
 }

trainer_state.json

@@ -10,917 +10,917 @@
   "log_history": [
     {
       "epoch": 0.02365464222353637,
-      "grad_norm": 39.04922181795541,
+      "grad_norm": 4.438574633904732,
       "learning_rate": 5e-06,
-      "loss": 0.9233,
+      "loss": 0.8864,
       "step": 10
     },
     {
       "epoch": 0.04730928444707274,
-      "grad_norm": 8.275301054500245,
+      "grad_norm": 3.743436107422786,
       "learning_rate": 5e-06,
-      "loss": 0.8198,
+      "loss": 0.7865,
       "step": 20
     },
     {
       "epoch": 0.0709639266706091,
-      "grad_norm": 16.91349063149131,
+      "grad_norm": 1.3773266130731683,
       "learning_rate": 5e-06,
-      "loss": 0.8094,
+      "loss": 0.7718,
       "step": 30
     },
     {
       "epoch": 0.09461856889414548,
-      "grad_norm": 14.148988171179118,
+      "grad_norm": 1.0328014979269187,
       "learning_rate": 5e-06,
-      "loss": 0.7679,
+      "loss": 0.7321,
       "step": 40
     },
     {
       "epoch": 0.11827321111768184,
-      "grad_norm": 1.026404466612184,
+      "grad_norm": 0.9419322611015702,
       "learning_rate": 5e-06,
-      "loss": 0.7523,
+      "loss": 0.7263,
       "step": 50
     },
     {
       "epoch": 0.1419278533412182,
-      "grad_norm": 0.8542225183892468,
+      "grad_norm": 1.0108429467163185,
       "learning_rate": 5e-06,
-      "loss": 0.725,
+      "loss": 0.7095,
       "step": 60
     },
     {
       "epoch": 0.16558249556475457,
-      "grad_norm": 0.7426518873227826,
+      "grad_norm": 0.7116397296212089,
       "learning_rate": 5e-06,
-      "loss": 0.7185,
+      "loss": 0.7076,
       "step": 70
     },
     {
       "epoch": 0.18923713778829096,
-      "grad_norm": 0.6764382829611749,
+      "grad_norm": 0.6226004481870809,
       "learning_rate": 5e-06,
-      "loss": 0.6999,
+      "loss": 0.6923,
       "step": 80
     },
     {
       "epoch": 0.21289178001182732,
-      "grad_norm": 0.5663920137594394,
+      "grad_norm": 0.7788334547916798,
       "learning_rate": 5e-06,
-      "loss": 0.7,
+      "loss": 0.6939,
       "step": 90
     },
     {
       "epoch": 0.23654642223536368,
-      "grad_norm": 0.6218835377066562,
+      "grad_norm": 0.5579006889125326,
       "learning_rate": 5e-06,
-      "loss": 0.6968,
+      "loss": 0.6916,
       "step": 100
     },
     {
       "epoch": 0.26020106445890007,
-      "grad_norm": 0.601497886261039,
+      "grad_norm": 0.8348224630921526,
       "learning_rate": 5e-06,
-      "loss": 0.6885,
+      "loss": 0.6841,
       "step": 110
     },
     {
       "epoch": 0.2838557066824364,
-      "grad_norm": 0.8786039473525534,
+      "grad_norm": 1.0981466464621934,
       "learning_rate": 5e-06,
-      "loss": 0.6863,
+      "loss": 0.6822,
       "step": 120
     },
     {
       "epoch": 0.3075103489059728,
-      "grad_norm": 0.5373447312315734,
+      "grad_norm": 0.8917486118802607,
       "learning_rate": 5e-06,
-      "loss": 0.6792,
+      "loss": 0.6761,
       "step": 130
     },
     {
       "epoch": 0.33116499112950915,
-      "grad_norm": 0.5195473153997355,
+      "grad_norm": 0.6543556998725211,
       "learning_rate": 5e-06,
-      "loss": 0.6746,
+      "loss": 0.6715,
       "step": 140
     },
     {
       "epoch": 0.35481963335304556,
-      "grad_norm": 0.5999120946052041,
+      "grad_norm": 0.8125188565799322,
       "learning_rate": 5e-06,
-      "loss": 0.68,
+      "loss": 0.6773,
       "step": 150
     },
     {
       "epoch": 0.3784742755765819,
-      "grad_norm": 0.5351205582865509,
+      "grad_norm": 0.5688602721338011,
       "learning_rate": 5e-06,
-      "loss": 0.6766,
+      "loss": 0.674,
       "step": 160
     },
     {
       "epoch": 0.4021289178001183,
-      "grad_norm": 0.5425010920017291,
+      "grad_norm": 0.491756225050921,
       "learning_rate": 5e-06,
-      "loss": 0.6839,
+      "loss": 0.6815,
       "step": 170
     },
     {
       "epoch": 0.42578356002365464,
-      "grad_norm": 0.702582958286065,
+      "grad_norm": 0.6509377411475789,
       "learning_rate": 5e-06,
-      "loss": 0.6711,
+      "loss": 0.669,
       "step": 180
     },
     {
       "epoch": 0.449438202247191,
-      "grad_norm": 1.2564450479675764,
+      "grad_norm": 0.6207735885705108,
       "learning_rate": 5e-06,
-      "loss": 0.6776,
+      "loss": 0.6755,
       "step": 190
     },
     {
       "epoch": 0.47309284447072736,
-      "grad_norm": 0.6226292820244915,
+      "grad_norm": 0.528808960645126,
       "learning_rate": 5e-06,
-      "loss": 0.6695,
+      "loss": 0.6674,
       "step": 200
     },
     {
       "epoch": 0.4967474866942638,
-      "grad_norm": 0.6729451633589562,
+      "grad_norm": 0.8472585224068009,
       "learning_rate": 5e-06,
-      "loss": 0.6656,
+      "loss": 0.6637,
       "step": 210
     },
     {
       "epoch": 0.5204021289178001,
-      "grad_norm": 0.72360498041814,
+      "grad_norm": 0.5848774410670773,
       "learning_rate": 5e-06,
-      "loss": 0.6699,
+      "loss": 0.6681,
       "step": 220
     },
     {
       "epoch": 0.5440567711413364,
-      "grad_norm": 0.5730443009789895,
+      "grad_norm": 0.6344280914148243,
       "learning_rate": 5e-06,
-      "loss": 0.6693,
+      "loss": 0.6675,
       "step": 230
     },
     {
       "epoch": 0.5677114133648729,
-      "grad_norm": 0.6304204873609028,
+      "grad_norm": 0.7485250474806475,
       "learning_rate": 5e-06,
-      "loss": 0.6564,
+      "loss": 0.6549,
       "step": 240
     },
     {
       "epoch": 0.5913660555884093,
-      "grad_norm": 0.6327686473365169,
+      "grad_norm": 0.7332796111883003,
       "learning_rate": 5e-06,
-      "loss": 0.6707,
+      "loss": 0.6693,
       "step": 250
     },
     {
       "epoch": 0.6150206978119456,
-      "grad_norm": 0.5640335014419563,
+      "grad_norm": 0.5097881381268425,
       "learning_rate": 5e-06,
-      "loss": 0.6617,
+      "loss": 0.6603,
       "step": 260
     },
     {
       "epoch": 0.638675340035482,
-      "grad_norm": 0.6026872513783592,
+      "grad_norm": 0.49176638410796597,
       "learning_rate": 5e-06,
-      "loss": 0.6711,
+      "loss": 0.6697,
       "step": 270
     },
     {
       "epoch": 0.6623299822590183,
-      "grad_norm": 0.5194797310260643,
+      "grad_norm": 0.4404068786810332,
       "learning_rate": 5e-06,
-      "loss": 0.6561,
+      "loss": 0.6547,
       "step": 280
     },
     {
       "epoch": 0.6859846244825547,
-      "grad_norm": 0.6658270095766984,
+      "grad_norm": 0.45956579197536424,
       "learning_rate": 5e-06,
-      "loss": 0.663,
+      "loss": 0.6617,
       "step": 290
     },
     {
       "epoch": 0.7096392667060911,
-      "grad_norm": 0.5259962549449988,
+      "grad_norm": 0.5489471352518822,
       "learning_rate": 5e-06,
-      "loss": 0.6511,
+      "loss": 0.65,
       "step": 300
     },
     {
       "epoch": 0.7332939089296274,
-      "grad_norm": 0.5776532705758929,
+      "grad_norm": 0.5071925000559494,
       "learning_rate": 5e-06,
-      "loss": 0.6649,
+      "loss": 0.6639,
       "step": 310
     },
     {
       "epoch": 0.7569485511531638,
-      "grad_norm": 0.5249892835904177,
+      "grad_norm": 0.4479532221316009,
       "learning_rate": 5e-06,
-      "loss": 0.6648,
+      "loss": 0.6638,
       "step": 320
     },
     {
       "epoch": 0.7806031933767001,
-      "grad_norm": 0.5092145613062358,
+      "grad_norm": 0.5768991016278898,
       "learning_rate": 5e-06,
-      "loss": 0.6614,
+      "loss": 0.6601,
       "step": 330
     },
     {
       "epoch": 0.8042578356002366,
-      "grad_norm": 0.5273167065626364,
+      "grad_norm": 0.47440990509293773,
       "learning_rate": 5e-06,
-      "loss": 0.6468,
+      "loss": 0.6457,
       "step": 340
     },
     {
       "epoch": 0.8279124778237729,
-      "grad_norm": 0.5666036582386984,
+      "grad_norm": 0.5535872360742707,
       "learning_rate": 5e-06,
-      "loss": 0.6562,
+      "loss": 0.6552,
       "step": 350
     },
     {
       "epoch": 0.8515671200473093,
-      "grad_norm": 0.6164474600239763,
+      "grad_norm": 0.5692309710225549,
       "learning_rate": 5e-06,
-      "loss": 0.6544,
+      "loss": 0.6533,
       "step": 360
     },
     {
       "epoch": 0.8752217622708457,
-      "grad_norm": 0.5854672267431167,
+      "grad_norm": 0.477161966378483,
       "learning_rate": 5e-06,
-      "loss": 0.6547,
+      "loss": 0.6538,
       "step": 370
     },
     {
       "epoch": 0.898876404494382,
-      "grad_norm": 0.5313196039449892,
+      "grad_norm": 0.5742339288892304,
       "learning_rate": 5e-06,
-      "loss": 0.6584,
+      "loss": 0.6576,
       "step": 380
     },
     {
       "epoch": 0.9225310467179184,
-      "grad_norm": 0.5598019500152581,
+      "grad_norm": 0.6276883665595651,
       "learning_rate": 5e-06,
-      "loss": 0.6533,
+      "loss": 0.6524,
       "step": 390
     },
     {
       "epoch": 0.9461856889414547,
-      "grad_norm": 0.6003027055491813,
+      "grad_norm": 0.5739867998426434,
       "learning_rate": 5e-06,
-      "loss": 0.6458,
+      "loss": 0.645,
       "step": 400
     },
     {
       "epoch": 0.9698403311649911,
-      "grad_norm": 0.6341763962447327,
+      "grad_norm": 0.6714991896688636,
       "learning_rate": 5e-06,
-      "loss": 0.656,
+      "loss": 0.6552,
       "step": 410
     },
     {
       "epoch": 0.9934949733885275,
-      "grad_norm": 0.6083507877729101,
+      "grad_norm": 0.5532850958647405,
       "learning_rate": 5e-06,
-      "loss": 0.6538,
+      "loss": 0.6528,
       "step": 420
     },
     {
       "epoch": 0.9982259018332348,
-      "eval_loss": 0.6558582186698914,
-      "eval_runtime": 224.885,
-      "eval_samples_per_second": 50.635,
-      "eval_steps_per_second": 0.396,
+      "eval_loss": 0.6549943685531616,
+      "eval_runtime": 226.5831,
+      "eval_samples_per_second": 50.255,
+      "eval_steps_per_second": 0.393,
       "step": 422
     },
     {
       "epoch": 1.0171496156120639,
-      "grad_norm": 0.9361782463296332,
+      "grad_norm": 0.6577034498689445,
       "learning_rate": 5e-06,
-      "loss": 0.6157,
+      "loss": 0.6186,
       "step": 430
     },
     {
       "epoch": 1.0408042578356003,
-      "grad_norm": 0.6204113010596938,
+      "grad_norm": 0.6919211184212087,
       "learning_rate": 5e-06,
-      "loss": 0.6065,
+      "loss": 0.611,
       "step": 440
     },
     {
       "epoch": 1.0644589000591367,
-      "grad_norm": 0.6564336264095381,
+      "grad_norm": 0.49818432774881954,
       "learning_rate": 5e-06,
-      "loss": 0.6048,
+      "loss": 0.6094,
       "step": 450
     },
     {
       "epoch": 1.0881135422826729,
-      "grad_norm": 0.5555833545533679,
+      "grad_norm": 0.5304613381757841,
       "learning_rate": 5e-06,
-      "loss": 0.6072,
+      "loss": 0.6118,
       "step": 460
     },
     {
       "epoch": 1.1117681845062093,
-      "grad_norm": 0.568736434370096,
+      "grad_norm": 0.5661240677080396,
       "learning_rate": 5e-06,
-      "loss": 0.6026,
+      "loss": 0.6069,
       "step": 470
     },
     {
       "epoch": 1.1354228267297457,
-      "grad_norm": 0.5963174527245159,
+      "grad_norm": 0.4725657759678031,
       "learning_rate": 5e-06,
-      "loss": 0.6041,
+      "loss": 0.6081,
       "step": 480
     },
     {
       "epoch": 1.1590774689532821,
-      "grad_norm": 0.6296624775692966,
+      "grad_norm": 0.5348982555181953,
       "learning_rate": 5e-06,
-      "loss": 0.6036,
+      "loss": 0.608,
       "step": 490
     },
     {
       "epoch": 1.1827321111768185,
-      "grad_norm": 0.7667349314546962,
+      "grad_norm": 0.7009257467225577,
       "learning_rate": 5e-06,
-      "loss": 0.6068,
+      "loss": 0.6107,
       "step": 500
     },
     {
       "epoch": 1.2063867534003547,
-      "grad_norm": 0.6034621980970892,
+      "grad_norm": 0.516755234113577,
       "learning_rate": 5e-06,
-      "loss": 0.6104,
+      "loss": 0.6145,
       "step": 510
     },
     {
       "epoch": 1.2300413956238911,
-      "grad_norm": 0.5825117431703367,
+      "grad_norm": 0.5565870238553596,
       "learning_rate": 5e-06,
-      "loss": 0.6026,
+      "loss": 0.6064,
       "step": 520
     },
     {
       "epoch": 1.2536960378474276,
-      "grad_norm": 0.5671081402783421,
+      "grad_norm": 0.5176934237005286,
       "learning_rate": 5e-06,
-      "loss": 0.604,
+      "loss": 0.6078,
       "step": 530
     },
     {
       "epoch": 1.277350680070964,
-      "grad_norm": 0.5309591912112671,
+      "grad_norm": 0.5399011019791115,
       "learning_rate": 5e-06,
-      "loss": 0.6132,
+      "loss": 0.6173,
       "step": 540
     },
     {
       "epoch": 1.3010053222945004,
-      "grad_norm": 0.5636046858771947,
+      "grad_norm": 0.48804065232921706,
       "learning_rate": 5e-06,
-      "loss": 0.605,
+      "loss": 0.6089,
       "step": 550
     },
     {
       "epoch": 1.3246599645180366,
-      "grad_norm": 0.6623955102141809,
+      "grad_norm": 0.6247022748083035,
       "learning_rate": 5e-06,
-      "loss": 0.6082,
+      "loss": 0.612,
       "step": 560
     },
     {
       "epoch": 1.348314606741573,
-      "grad_norm": 0.5742305096790601,
+      "grad_norm": 0.5205181494692162,
       "learning_rate": 5e-06,
-      "loss": 0.605,
+      "loss": 0.6087,
       "step": 570
     },
     {
       "epoch": 1.3719692489651094,
-      "grad_norm": 0.5167065988140831,
+      "grad_norm": 0.4444906716754459,
       "learning_rate": 5e-06,
-      "loss": 0.6091,
+      "loss": 0.6129,
       "step": 580
     },
     {
       "epoch": 1.3956238911886458,
-      "grad_norm": 0.5112713876137833,
+      "grad_norm": 0.4699507974891951,
       "learning_rate": 5e-06,
-      "loss": 0.6049,
+      "loss": 0.6084,
       "step": 590
     },
     {
       "epoch": 1.4192785334121822,
-      "grad_norm": 0.515536375353522,
+      "grad_norm": 0.438759746705871,
       "learning_rate": 5e-06,
-      "loss": 0.6079,
+      "loss": 0.6118,
       "step": 600
     },
     {
       "epoch": 1.4429331756357184,
-      "grad_norm": 0.5943800369847494,
+      "grad_norm": 0.492167276336904,
       "learning_rate": 5e-06,
-      "loss": 0.602,
+      "loss": 0.6057,
       "step": 610
     },
     {
       "epoch": 1.4665878178592548,
-      "grad_norm": 0.5570413849081146,
+      "grad_norm": 0.508896134049524,
       "learning_rate": 5e-06,
-      "loss": 0.6074,
+      "loss": 0.6111,
       "step": 620
     },
     {
       "epoch": 1.4902424600827913,
-      "grad_norm": 0.5383074416990815,
+      "grad_norm": 0.5044935497801236,
       "learning_rate": 5e-06,
-      "loss": 0.6055,
+      "loss": 0.6091,
       "step": 630
     },
     {
       "epoch": 1.5138971023063275,
-      "grad_norm": 0.6221748842819845,
+      "grad_norm": 0.611513828523435,
       "learning_rate": 5e-06,
-      "loss": 0.6048,
+      "loss": 0.6084,
       "step": 640
     },
     {
       "epoch": 1.537551744529864,
-      "grad_norm": 0.623130737085543,
+      "grad_norm": 0.4284886945389684,
       "learning_rate": 5e-06,
-      "loss": 0.6124,
+      "loss": 0.6161,
       "step": 650
     },
     {
       "epoch": 1.5612063867534003,
-      "grad_norm": 0.7728758992657894,
+      "grad_norm": 0.7397737868115762,
       "learning_rate": 5e-06,
-      "loss": 0.6073,
+      "loss": 0.611,
       "step": 660
     },
     {
       "epoch": 1.5848610289769367,
-      "grad_norm": 0.5531126954202661,
+      "grad_norm": 0.45834776518516607,
       "learning_rate": 5e-06,
-      "loss": 0.6023,
+      "loss": 0.6054,
       "step": 670
     },
     {
       "epoch": 1.6085156712004731,
-      "grad_norm": 0.8207249388527519,
+      "grad_norm": 0.9830107568320281,
       "learning_rate": 5e-06,
-      "loss": 0.6041,
+      "loss": 0.6077,
       "step": 680
     },
     {
       "epoch": 1.6321703134240093,
-      "grad_norm": 0.7382668830128054,
+      "grad_norm": 0.7316301006660809,
       "learning_rate": 5e-06,
-      "loss": 0.6027,
+      "loss": 0.6066,
       "step": 690
     },
     {
       "epoch": 1.655824955647546,
-      "grad_norm": 0.8181634349883082,
+      "grad_norm": 0.8713979852654485,
       "learning_rate": 5e-06,
-      "loss": 0.6077,
+      "loss": 0.6112,
       "step": 700
     },
     {
       "epoch": 1.6794795978710821,
-      "grad_norm": 0.5715750112816181,
+      "grad_norm": 0.6609904307136948,
       "learning_rate": 5e-06,
-      "loss": 0.603,
+      "loss": 0.6063,
       "step": 710
     },
     {
       "epoch": 1.7031342400946186,
-      "grad_norm": 0.564060422032355,
+      "grad_norm": 0.5730385784454821,
       "learning_rate": 5e-06,
-      "loss": 0.6137,
+      "loss": 0.6174,
       "step": 720
     },
     {
       "epoch": 1.726788882318155,
-      "grad_norm": 0.5112934215435977,
+      "grad_norm": 0.428675321624077,
       "learning_rate": 5e-06,
-      "loss": 0.6041,
+      "loss": 0.6075,
       "step": 730
     },
     {
       "epoch": 1.7504435245416912,
-      "grad_norm": 0.6498040890743698,
+      "grad_norm": 0.5747845078803645,
       "learning_rate": 5e-06,
-      "loss": 0.6038,
+      "loss": 0.6073,
       "step": 740
     },
     {
       "epoch": 1.7740981667652278,
-      "grad_norm": 0.6625174306160165,
+      "grad_norm": 0.6019599666582006,
       "learning_rate": 5e-06,
-      "loss": 0.604,
+      "loss": 0.6074,
       "step": 750
     },
     {
       "epoch": 1.797752808988764,
-      "grad_norm": 0.5200239654238437,
+      "grad_norm": 0.484871680178572,
       "learning_rate": 5e-06,
-      "loss": 0.604,
+      "loss": 0.6076,
       "step": 760
     },
     {
       "epoch": 1.8214074512123004,
-      "grad_norm": 0.5056250667365105,
+      "grad_norm": 0.4801879662753807,
       "learning_rate": 5e-06,
-      "loss": 0.601,
+      "loss": 0.6046,
       "step": 770
     },
     {
       "epoch": 1.8450620934358368,
-      "grad_norm": 0.5465224841554837,
+      "grad_norm": 0.44588625446373603,
       "learning_rate": 5e-06,
-      "loss": 0.6043,
+      "loss": 0.6078,
       "step": 780
     },
     {
       "epoch": 1.868716735659373,
-      "grad_norm": 0.5173445168820222,
+      "grad_norm": 0.43921853360548113,
       "learning_rate": 5e-06,
-      "loss": 0.6051,
+      "loss": 0.6086,
       "step": 790
     },
     {
       "epoch": 1.8923713778829097,
-      "grad_norm": 0.5037163086029489,
+      "grad_norm": 0.4461189307976923,
       "learning_rate": 5e-06,
-      "loss": 0.6071,
+      "loss": 0.6105,
       "step": 800
     },
     {
       "epoch": 1.9160260201064458,
-      "grad_norm": 0.5032092904194995,
+      "grad_norm": 0.4949761836327779,
       "learning_rate": 5e-06,
-      "loss": 0.6097,
+      "loss": 0.613,
       "step": 810
     },
     {
       "epoch": 1.9396806623299823,
-      "grad_norm": 0.5153373413177225,
+      "grad_norm": 0.44062812260467765,
       "learning_rate": 5e-06,
-      "loss": 0.6017,
+      "loss": 0.6049,
       "step": 820
     },
     {
       "epoch": 1.9633353045535187,
-      "grad_norm": 0.48445434626456924,
+      "grad_norm": 0.4460160929815086,
       "learning_rate": 5e-06,
-      "loss": 0.6097,
+      "loss": 0.613,
       "step": 830
     },
     {
       "epoch": 1.9869899467770549,
-      "grad_norm": 0.5493690840998718,
+      "grad_norm": 0.5175110899521405,
       "learning_rate": 5e-06,
-      "loss": 0.5996,
+      "loss": 0.6028,
       "step": 840
     },
     {
       "epoch": 1.9988172678888232,
-      "eval_loss": 0.6458428502082825,
-      "eval_runtime": 226.2611,
-      "eval_samples_per_second": 50.327,
-      "eval_steps_per_second": 0.393,
+      "eval_loss": 0.6445377469062805,
+      "eval_runtime": 227.0244,
+      "eval_samples_per_second": 50.158,
+      "eval_steps_per_second": 0.392,
       "step": 845
     },
     {
       "epoch": 2.0106445890005915,
-      "grad_norm": 0.6225484532666892,
+      "grad_norm": 0.5892375412389526,
       "learning_rate": 5e-06,
-      "loss": 0.5795,
+      "loss": 0.5851,
       "step": 850
     },
     {
       "epoch": 2.0342992312241277,
-      "grad_norm": 0.5819857678964343,
+      "grad_norm": 0.5097516172118646,
       "learning_rate": 5e-06,
-      "loss": 0.5553,
+      "loss": 0.5643,
       "step": 860
     },
     {
       "epoch": 2.057953873447664,
-      "grad_norm": 0.7185360534865078,
+      "grad_norm": 0.6536746176311915,
       "learning_rate": 5e-06,
-      "loss": 0.5506,
+      "loss": 0.5592,
       "step": 870
     },
     {
       "epoch": 2.0816085156712005,
-      "grad_norm": 0.5134284842767335,
+      "grad_norm": 0.47983268810356666,
       "learning_rate": 5e-06,
-      "loss": 0.5539,
+      "loss": 0.5626,
       "step": 880
     },
     {
       "epoch": 2.1052631578947367,
-      "grad_norm": 0.5986326239884353,
+      "grad_norm": 0.6017282349204336,
       "learning_rate": 5e-06,
-      "loss": 0.5609,
+      "loss": 0.5697,
       "step": 890
     },
     {
       "epoch": 2.1289178001182734,
-      "grad_norm": 0.579714763513885,
+      "grad_norm": 0.5728407157654074,
       "learning_rate": 5e-06,
-      "loss": 0.5546,
+      "loss": 0.5632,
       "step": 900
     },
     {
       "epoch": 2.1525724423418096,
-      "grad_norm": 0.570292196409214,
+      "grad_norm": 0.5680779384221303,
       "learning_rate": 5e-06,
-      "loss": 0.5586,
+      "loss": 0.5672,
       "step": 910
     },
     {
       "epoch": 2.1762270845653457,
-      "grad_norm": 0.5811117001743673,
+      "grad_norm": 0.48858908601906337,
       "learning_rate": 5e-06,
-      "loss": 0.5585,
+      "loss": 0.567,
       "step": 920
     },
     {
       "epoch": 2.1998817267888824,
-      "grad_norm": 0.554144816987719,
+      "grad_norm": 0.5005707887249943,
       "learning_rate": 5e-06,
-      "loss": 0.5562,
+      "loss": 0.5646,
       "step": 930
     },
     {
       "epoch": 2.2235363690124186,
-      "grad_norm": 0.6493826388527278,
+      "grad_norm": 0.5829558904651037,
       "learning_rate": 5e-06,
-      "loss": 0.5565,
+      "loss": 0.5648,
       "step": 940
     },
     {
       "epoch": 2.247191011235955,
-      "grad_norm": 0.5220557856218626,
+      "grad_norm": 0.48798199303667406,
       "learning_rate": 5e-06,
-      "loss": 0.5694,
+      "loss": 0.578,
       "step": 950
     },
     {
       "epoch": 2.2708456534594914,
-      "grad_norm": 0.6378102281048501,
+      "grad_norm": 0.582446153234459,
       "learning_rate": 5e-06,
-      "loss": 0.5602,
+      "loss": 0.5682,
       "step": 960
     },
     {
       "epoch": 2.2945002956830276,
-      "grad_norm": 0.5497371876386185,
+      "grad_norm": 0.46970294592756995,
       "learning_rate": 5e-06,
-      "loss": 0.5628,
+      "loss": 0.571,
       "step": 970
     },
     {
       "epoch": 2.3181549379065642,
-      "grad_norm": 0.6521682175920844,
+      "grad_norm": 0.5759020549520256,
       "learning_rate": 5e-06,
-      "loss": 0.5565,
+      "loss": 0.5647,
       "step": 980
     },
     {
       "epoch": 2.3418095801301004,
-      "grad_norm": 0.5734936169662879,
+      "grad_norm": 0.549340588982862,
       "learning_rate": 5e-06,
-      "loss": 0.5674,
+      "loss": 0.5755,
       "step": 990
     },
     {
       "epoch": 2.365464222353637,
-      "grad_norm": 0.5394258314592499,
+      "grad_norm": 0.46429208051701265,
       "learning_rate": 5e-06,
-      "loss": 0.5569,
+      "loss": 0.5648,
       "step": 1000
     },
     {
       "epoch": 2.3891188645771733,
-      "grad_norm": 0.5306593171364488,
+      "grad_norm": 0.5160254392452897,
       "learning_rate": 5e-06,
-      "loss": 0.5502,
+      "loss": 0.558,
       "step": 1010
     },
     {
       "epoch": 2.4127735068007095,
-      "grad_norm": 0.5344199954837688,
+      "grad_norm": 0.4799281597192369,
       "learning_rate": 5e-06,
-      "loss": 0.5549,
+      "loss": 0.5627,
       "step": 1020
     },
     {
       "epoch": 2.436428149024246,
-      "grad_norm": 0.5892751227456119,
+      "grad_norm": 0.5121330286207769,
       "learning_rate": 5e-06,
-      "loss": 0.5533,
+      "loss": 0.5608,
       "step": 1030
     },
     {
       "epoch": 2.4600827912477823,
-      "grad_norm": 0.6529042003930223,
+      "grad_norm": 0.5841580086447481,
       "learning_rate": 5e-06,
-      "loss": 0.5613,
+      "loss": 0.5693,
       "step": 1040
     },
     {
       "epoch": 2.483737433471319,
-      "grad_norm": 0.5765438425321338,
+      "grad_norm": 0.557020183414569,
       "learning_rate": 5e-06,
-      "loss": 0.5646,
+      "loss": 0.5726,
       "step": 1050
     },
     {
       "epoch": 2.507392075694855,
-      "grad_norm": 0.6764490044193554,
+      "grad_norm": 0.6374112998842234,
       "learning_rate": 5e-06,
-      "loss": 0.567,
+      "loss": 0.5747,
       "step": 1060
     },
     {
       "epoch": 2.5310467179183913,
-      "grad_norm": 0.5365218264481744,
+      "grad_norm": 0.5343754869995426,
       "learning_rate": 5e-06,
-      "loss": 0.5532,
+      "loss": 0.5607,
       "step": 1070
     },
     {
       "epoch": 2.554701360141928,
-      "grad_norm": 0.6033785399498255,
+      "grad_norm": 0.5496554790900547,
       "learning_rate": 5e-06,
-      "loss": 0.5622,
+      "loss": 0.5698,
       "step": 1080
     },
     {
       "epoch": 2.578356002365464,
-      "grad_norm": 0.8004909937255467,
+      "grad_norm": 0.6822773581077988,
       "learning_rate": 5e-06,
-      "loss": 0.5661,
+      "loss": 0.5738,
       "step": 1090
     },
     {
       "epoch": 2.6020106445890008,
-      "grad_norm": 0.5819582134735406,
+      "grad_norm": 0.49632724788385346,
       "learning_rate": 5e-06,
-      "loss": 0.5616,
+      "loss": 0.5692,
       "step": 1100
     },
     {
       "epoch": 2.625665286812537,
-      "grad_norm": 0.5537773395049099,
+      "grad_norm": 0.4859614320386073,
       "learning_rate": 5e-06,
-      "loss": 0.5628,
+      "loss": 0.5704,
       "step": 1110
     },
     {
       "epoch": 2.649319929036073,
-      "grad_norm": 0.5539615560141525,
+      "grad_norm": 0.5005521245693028,
       "learning_rate": 5e-06,
-      "loss": 0.5648,
+      "loss": 0.5721,
       "step": 1120
     },
     {
       "epoch": 2.67297457125961,
-      "grad_norm": 0.6206027218523953,
+      "grad_norm": 0.5418331476470847,
       "learning_rate": 5e-06,
-      "loss": 0.5643,
+      "loss": 0.5719,
       "step": 1130
     },
     {
       "epoch": 2.696629213483146,
-      "grad_norm": 0.5108322877934205,
+      "grad_norm": 0.4518235189693759,
       "learning_rate": 5e-06,
-      "loss": 0.5586,
+      "loss": 0.5658,
       "step": 1140
     },
     {
       "epoch": 2.7202838557066826,
-      "grad_norm": 0.48797735494965916,
+      "grad_norm": 0.45229828628235735,
       "learning_rate": 5e-06,
-      "loss": 0.5563,
+      "loss": 0.5638,
       "step": 1150
     },
     {
       "epoch": 2.743938497930219,
-      "grad_norm": 0.5823974142352172,
+      "grad_norm": 0.5051148207876189,
       "learning_rate": 5e-06,
-      "loss": 0.5671,
+      "loss": 0.5748,
       "step": 1160
     },
     {
       "epoch": 2.767593140153755,
-      "grad_norm": 0.8599218035136146,
+      "grad_norm": 0.7455413514573421,
       "learning_rate": 5e-06,
-      "loss": 0.5723,
+      "loss": 0.5797,
       "step": 1170
     },
     {
       "epoch": 2.7912477823772917,
-      "grad_norm": 0.6555716714163583,
+      "grad_norm": 0.5362145068936747,
       "learning_rate": 5e-06,
-      "loss": 0.5633,
+      "loss": 0.5705,
       "step": 1180
     },
     {
       "epoch": 2.814902424600828,
-      "grad_norm": 0.49879910164951613,
+      "grad_norm": 0.46118669511344673,
       "learning_rate": 5e-06,
-      "loss": 0.5581,
+      "loss": 0.5653,
       "step": 1190
     },
     {
       "epoch": 2.8385570668243645,
-      "grad_norm": 0.5241725506783274,
+      "grad_norm": 0.5498761802579338,
       "learning_rate": 5e-06,
-      "loss": 0.5623,
+      "loss": 0.5694,
       "step": 1200
     },
     {
       "epoch": 2.8622117090479007,
-      "grad_norm": 0.6173811070502804,
+      "grad_norm": 0.5720658060375756,
       "learning_rate": 5e-06,
-      "loss": 0.569,
+      "loss": 0.5761,
       "step": 1210
     },
     {
       "epoch": 2.885866351271437,
-      "grad_norm": 0.5397292738316359,
+      "grad_norm": 0.4735883791639776,
       "learning_rate": 5e-06,
-      "loss": 0.5642,
+      "loss": 0.5714,
       "step": 1220
     },
     {
       "epoch": 2.9095209934949735,
-      "grad_norm": 0.7053290870019903,
+      "grad_norm": 0.6126626053091963,
       "learning_rate": 5e-06,
-      "loss": 0.5593,
+      "loss": 0.5665,
       "step": 1230
     },
     {
       "epoch": 2.9331756357185097,
-      "grad_norm": 0.5500348460578961,
+      "grad_norm": 0.5724885076669786,
       "learning_rate": 5e-06,
-      "loss": 0.5591,
+      "loss": 0.5666,
       "step": 1240
     },
     {
       "epoch": 2.9568302779420463,
-      "grad_norm": 0.5833114667049699,
+      "grad_norm": 0.5144727847784881,
       "learning_rate": 5e-06,
-      "loss": 0.5649,
+      "loss": 0.5721,
       "step": 1250
     },
     {
       "epoch": 2.9804849201655825,
-      "grad_norm": 0.569413301750619,
+      "grad_norm": 0.4637250585550989,
       "learning_rate": 5e-06,
-      "loss": 0.5577,
+      "loss": 0.5645,
       "step": 1260
     },
     {
       "epoch": 2.9946777054997042,
-      "eval_loss": 0.648719847202301,
-      "eval_runtime": 225.8886,
-      "eval_samples_per_second": 50.41,
-      "eval_steps_per_second": 0.394,
+      "eval_loss": 0.6453979015350342,
+      "eval_runtime": 226.895,
+      "eval_samples_per_second": 50.186,
+      "eval_steps_per_second": 0.392,
       "step": 1266
     },
     {
       "epoch": 2.9946777054997042,
       "step": 1266,
       "total_flos": 2120178393415680.0,
-      "train_loss": 0.6180764295478568,
-      "train_runtime": 37891.5513,
-      "train_samples_per_second": 17.129,
+      "train_loss": 0.6197805719164687,
+      "train_runtime": 38167.6556,
+      "train_samples_per_second": 17.005,
       "train_steps_per_second": 0.033
     }
   ],