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_gpt4_llm
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # OH_DCFT_V3_wo_gpt4_llm
 
-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_gpt4_llm dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.6373
 

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 2.996415770609319,
-    "eval_loss": 0.6404265761375427,
-    "eval_runtime": 225.386,
-    "eval_samples_per_second": 50.016,
-    "eval_steps_per_second": 0.395,
+    "eval_loss": 0.6373269557952881,
+    "eval_runtime": 226.1722,
+    "eval_samples_per_second": 49.843,
+    "eval_steps_per_second": 0.394,
     "total_flos": 2100077946470400.0,
-    "train_loss": 0.6158634758832162,
-    "train_runtime": 37532.7652,
-    "train_samples_per_second": 17.12,
+    "train_loss": 0.6173035094612523,
+    "train_runtime": 37751.1113,
+    "train_samples_per_second": 17.021,
     "train_steps_per_second": 0.033
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 2.996415770609319,
-    "eval_loss": 0.6404265761375427,
-    "eval_runtime": 225.386,
-    "eval_samples_per_second": 50.016,
-    "eval_steps_per_second": 0.395
+    "eval_loss": 0.6373269557952881,
+    "eval_runtime": 226.1722,
+    "eval_samples_per_second": 49.843,
+    "eval_steps_per_second": 0.394
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.996415770609319,
     "total_flos": 2100077946470400.0,
-    "train_loss": 0.6158634758832162,
-    "train_runtime": 37532.7652,
-    "train_samples_per_second": 17.12,
+    "train_loss": 0.6173035094612523,
+    "train_runtime": 37751.1113,
+    "train_samples_per_second": 17.021,
     "train_steps_per_second": 0.033
 }

trainer_state.json

@@ -10,910 +10,910 @@
   "log_history": [
     {
       "epoch": 0.023894862604540025,
-      "grad_norm": 6.081180104447192,
+      "grad_norm": 4.236361518666415,
       "learning_rate": 5e-06,
-      "loss": 0.9213,
+      "loss": 0.881,
       "step": 10
     },
     {
       "epoch": 0.04778972520908005,
-      "grad_norm": 2.5675420707703682,
+      "grad_norm": 5.741199257309677,
       "learning_rate": 5e-06,
-      "loss": 0.8064,
+      "loss": 0.7803,
       "step": 20
     },
     {
       "epoch": 0.07168458781362007,
-      "grad_norm": 1.116884956350582,
+      "grad_norm": 19.255957544263854,
       "learning_rate": 5e-06,
-      "loss": 0.764,
+      "loss": 0.7636,
       "step": 30
     },
     {
       "epoch": 0.0955794504181601,
-      "grad_norm": 0.9827896973879405,
+      "grad_norm": 1.4389740219584857,
       "learning_rate": 5e-06,
-      "loss": 0.7456,
+      "loss": 0.7601,
       "step": 40
     },
     {
       "epoch": 0.11947431302270012,
-      "grad_norm": 1.1716177346699552,
+      "grad_norm": 2.052033228969374,
       "learning_rate": 5e-06,
-      "loss": 0.7252,
+      "loss": 0.7307,
       "step": 50
     },
     {
       "epoch": 0.14336917562724014,
-      "grad_norm": 1.054189475268262,
+      "grad_norm": 1.4883942502737415,
       "learning_rate": 5e-06,
-      "loss": 0.7144,
+      "loss": 0.7194,
       "step": 60
     },
     {
       "epoch": 0.16726403823178015,
-      "grad_norm": 0.7494472231038515,
+      "grad_norm": 0.7721229562083435,
       "learning_rate": 5e-06,
-      "loss": 0.7153,
+      "loss": 0.7186,
       "step": 70
     },
     {
       "epoch": 0.1911589008363202,
-      "grad_norm": 0.5903583553666529,
+      "grad_norm": 0.5683129469939435,
       "learning_rate": 5e-06,
-      "loss": 0.6941,
+      "loss": 0.6965,
       "step": 80
     },
     {
       "epoch": 0.21505376344086022,
-      "grad_norm": 0.7850320924725688,
+      "grad_norm": 0.527733611127623,
       "learning_rate": 5e-06,
-      "loss": 0.6917,
+      "loss": 0.6922,
       "step": 90
     },
     {
       "epoch": 0.23894862604540024,
-      "grad_norm": 0.8082497951696401,
+      "grad_norm": 0.5540046789983225,
       "learning_rate": 5e-06,
-      "loss": 0.6931,
+      "loss": 0.693,
       "step": 100
     },
     {
       "epoch": 0.2628434886499403,
-      "grad_norm": 0.5831514385960807,
+      "grad_norm": 0.5451390307514128,
       "learning_rate": 5e-06,
-      "loss": 0.6808,
+      "loss": 0.6802,
       "step": 110
     },
     {
       "epoch": 0.2867383512544803,
-      "grad_norm": 0.7344040626713287,
+      "grad_norm": 0.5143838898116624,
       "learning_rate": 5e-06,
-      "loss": 0.689,
+      "loss": 0.688,
       "step": 120
     },
     {
       "epoch": 0.3106332138590203,
-      "grad_norm": 0.8291631762782786,
+      "grad_norm": 0.6502984472755421,
       "learning_rate": 5e-06,
-      "loss": 0.6852,
+      "loss": 0.6838,
       "step": 130
     },
     {
       "epoch": 0.3345280764635603,
-      "grad_norm": 0.555446198309624,
+      "grad_norm": 0.5635569077666838,
       "learning_rate": 5e-06,
-      "loss": 0.6746,
+      "loss": 0.6733,
       "step": 140
     },
     {
       "epoch": 0.35842293906810035,
-      "grad_norm": 0.8214482724693175,
+      "grad_norm": 0.6029469287016763,
       "learning_rate": 5e-06,
-      "loss": 0.6789,
+      "loss": 0.6776,
       "step": 150
     },
     {
       "epoch": 0.3823178016726404,
-      "grad_norm": 0.5332479678739207,
+      "grad_norm": 0.486292600711864,
       "learning_rate": 5e-06,
-      "loss": 0.6676,
+      "loss": 0.6661,
       "step": 160
     },
     {
       "epoch": 0.4062126642771804,
-      "grad_norm": 1.0446840993388027,
+      "grad_norm": 0.6615883711779132,
       "learning_rate": 5e-06,
-      "loss": 0.6669,
+      "loss": 0.6652,
       "step": 170
     },
     {
       "epoch": 0.43010752688172044,
-      "grad_norm": 0.5354819297917649,
+      "grad_norm": 0.4717863479299739,
       "learning_rate": 5e-06,
-      "loss": 0.6671,
+      "loss": 0.6655,
       "step": 180
     },
     {
       "epoch": 0.4540023894862604,
-      "grad_norm": 0.5790753975231967,
+      "grad_norm": 0.4888275284899482,
       "learning_rate": 5e-06,
-      "loss": 0.6637,
+      "loss": 0.662,
       "step": 190
     },
     {
       "epoch": 0.4778972520908005,
-      "grad_norm": 0.5611754139446838,
+      "grad_norm": 0.5394213188181476,
       "learning_rate": 5e-06,
-      "loss": 0.6706,
+      "loss": 0.6687,
       "step": 200
     },
     {
       "epoch": 0.5017921146953405,
-      "grad_norm": 0.541229197735182,
+      "grad_norm": 0.45576158948311507,
       "learning_rate": 5e-06,
-      "loss": 0.6678,
+      "loss": 0.6661,
       "step": 210
     },
     {
       "epoch": 0.5256869772998806,
-      "grad_norm": 0.6290230152316767,
+      "grad_norm": 0.45151984287636476,
       "learning_rate": 5e-06,
-      "loss": 0.6615,
+      "loss": 0.6597,
       "step": 220
     },
     {
       "epoch": 0.5495818399044206,
-      "grad_norm": 0.5103134958958712,
+      "grad_norm": 0.4799647706900106,
       "learning_rate": 5e-06,
-      "loss": 0.6621,
+      "loss": 0.6602,
       "step": 230
     },
     {
       "epoch": 0.5734767025089605,
-      "grad_norm": 0.4963914880777678,
+      "grad_norm": 0.47076133511342133,
       "learning_rate": 5e-06,
-      "loss": 0.6587,
+      "loss": 0.6571,
       "step": 240
     },
     {
       "epoch": 0.5973715651135006,
-      "grad_norm": 0.7047770736230026,
+      "grad_norm": 0.5919116297131423,
       "learning_rate": 5e-06,
-      "loss": 0.6633,
+      "loss": 0.6615,
       "step": 250
     },
     {
       "epoch": 0.6212664277180406,
-      "grad_norm": 0.6875405023947134,
+      "grad_norm": 0.5500231129527917,
       "learning_rate": 5e-06,
-      "loss": 0.6602,
+      "loss": 0.6585,
       "step": 260
     },
     {
       "epoch": 0.6451612903225806,
-      "grad_norm": 0.5469403807072362,
+      "grad_norm": 0.5242998976170237,
       "learning_rate": 5e-06,
-      "loss": 0.6649,
+      "loss": 0.6633,
       "step": 270
     },
     {
       "epoch": 0.6690561529271206,
-      "grad_norm": 0.6301316104025243,
+      "grad_norm": 0.44132900428051,
       "learning_rate": 5e-06,
-      "loss": 0.6605,
+      "loss": 0.6588,
       "step": 280
     },
     {
       "epoch": 0.6929510155316607,
-      "grad_norm": 0.7436109186331767,
+      "grad_norm": 0.6925054556015406,
       "learning_rate": 5e-06,
-      "loss": 0.657,
+      "loss": 0.6553,
       "step": 290
     },
     {
       "epoch": 0.7168458781362007,
-      "grad_norm": 0.6316182942840975,
+      "grad_norm": 0.4625241785333385,
       "learning_rate": 5e-06,
-      "loss": 0.659,
+      "loss": 0.6574,
       "step": 300
     },
     {
       "epoch": 0.7407407407407407,
-      "grad_norm": 0.5862276185836299,
+      "grad_norm": 0.4229402308269957,
       "learning_rate": 5e-06,
-      "loss": 0.6543,
+      "loss": 0.6527,
       "step": 310
     },
     {
       "epoch": 0.7646356033452808,
-      "grad_norm": 0.5727274506679324,
+      "grad_norm": 0.5130609463277542,
       "learning_rate": 5e-06,
-      "loss": 0.6578,
+      "loss": 0.6561,
       "step": 320
     },
     {
       "epoch": 0.7885304659498208,
-      "grad_norm": 0.8637237307062305,
+      "grad_norm": 0.6838274381409521,
       "learning_rate": 5e-06,
-      "loss": 0.6572,
+      "loss": 0.6555,
       "step": 330
     },
     {
       "epoch": 0.8124253285543608,
-      "grad_norm": 0.5716637099962937,
+      "grad_norm": 0.4426103821343896,
       "learning_rate": 5e-06,
-      "loss": 0.6544,
+      "loss": 0.6528,
       "step": 340
     },
     {
       "epoch": 0.8363201911589009,
-      "grad_norm": 0.5267829079741267,
+      "grad_norm": 0.4768048776745041,
       "learning_rate": 5e-06,
-      "loss": 0.6542,
+      "loss": 0.6526,
       "step": 350
     },
     {
       "epoch": 0.8602150537634409,
-      "grad_norm": 0.5819870889688735,
+      "grad_norm": 0.47979657505843953,
       "learning_rate": 5e-06,
-      "loss": 0.6524,
+      "loss": 0.6507,
       "step": 360
     },
     {
       "epoch": 0.8841099163679809,
-      "grad_norm": 0.5230379233220511,
+      "grad_norm": 0.43210991398577236,
       "learning_rate": 5e-06,
-      "loss": 0.6561,
+      "loss": 0.6545,
       "step": 370
     },
     {
       "epoch": 0.9080047789725209,
-      "grad_norm": 0.5405733925967506,
+      "grad_norm": 0.4219482631866451,
       "learning_rate": 5e-06,
-      "loss": 0.6576,
+      "loss": 0.6561,
       "step": 380
     },
     {
       "epoch": 0.931899641577061,
-      "grad_norm": 0.6252081499402665,
+      "grad_norm": 0.4889263682317913,
       "learning_rate": 5e-06,
-      "loss": 0.6431,
+      "loss": 0.6415,
       "step": 390
     },
     {
       "epoch": 0.955794504181601,
-      "grad_norm": 0.5241989931128205,
+      "grad_norm": 0.4994356501839893,
       "learning_rate": 5e-06,
-      "loss": 0.645,
+      "loss": 0.6434,
       "step": 400
     },
     {
       "epoch": 0.9796893667861409,
-      "grad_norm": 0.5917706193349264,
+      "grad_norm": 0.5756138907013993,
       "learning_rate": 5e-06,
-      "loss": 0.6548,
+      "loss": 0.6535,
       "step": 410
     },
     {
       "epoch": 0.998805256869773,
-      "eval_loss": 0.6481666564941406,
-      "eval_runtime": 225.2584,
-      "eval_samples_per_second": 50.045,
-      "eval_steps_per_second": 0.395,
+      "eval_loss": 0.6466529965400696,
+      "eval_runtime": 225.9354,
+      "eval_samples_per_second": 49.895,
+      "eval_steps_per_second": 0.394,
       "step": 418
     },
     {
       "epoch": 1.003584229390681,
-      "grad_norm": 0.8892508812440383,
+      "grad_norm": 0.7256416169536216,
       "learning_rate": 5e-06,
-      "loss": 0.645,
+      "loss": 0.6438,
       "step": 420
     },
     {
       "epoch": 1.027479091995221,
-      "grad_norm": 0.6856782905740205,
+      "grad_norm": 0.6564158902335233,
       "learning_rate": 5e-06,
-      "loss": 0.6088,
+      "loss": 0.6108,
       "step": 430
     },
     {
       "epoch": 1.0513739545997611,
-      "grad_norm": 0.5419930096362386,
+      "grad_norm": 0.4999679801637927,
       "learning_rate": 5e-06,
-      "loss": 0.6083,
+      "loss": 0.6106,
       "step": 440
     },
     {
       "epoch": 1.075268817204301,
-      "grad_norm": 0.6254120282769089,
+      "grad_norm": 0.5241048691611577,
       "learning_rate": 5e-06,
-      "loss": 0.6102,
+      "loss": 0.6124,
       "step": 450
     },
     {
       "epoch": 1.099163679808841,
-      "grad_norm": 0.7886162221301777,
+      "grad_norm": 0.5456228664692746,
       "learning_rate": 5e-06,
-      "loss": 0.6022,
+      "loss": 0.6042,
       "step": 460
     },
     {
       "epoch": 1.1230585424133812,
-      "grad_norm": 0.5575253634257799,
+      "grad_norm": 0.5456744152195628,
       "learning_rate": 5e-06,
-      "loss": 0.6008,
+      "loss": 0.6028,
       "step": 470
     },
     {
       "epoch": 1.146953405017921,
-      "grad_norm": 0.5457321558023005,
+      "grad_norm": 0.4664933079979728,
       "learning_rate": 5e-06,
-      "loss": 0.6076,
+      "loss": 0.6095,
       "step": 480
     },
     {
       "epoch": 1.1708482676224612,
-      "grad_norm": 0.4910125465988992,
+      "grad_norm": 0.4894583019401931,
       "learning_rate": 5e-06,
-      "loss": 0.5998,
+      "loss": 0.6019,
       "step": 490
     },
     {
       "epoch": 1.194743130227001,
-      "grad_norm": 0.5851438416753514,
+      "grad_norm": 0.4942642519947347,
       "learning_rate": 5e-06,
-      "loss": 0.6095,
+      "loss": 0.6114,
       "step": 500
     },
     {
       "epoch": 1.2186379928315412,
-      "grad_norm": 0.5544196352892787,
+      "grad_norm": 0.46554339302452813,
       "learning_rate": 5e-06,
-      "loss": 0.608,
+      "loss": 0.6099,
       "step": 510
     },
     {
       "epoch": 1.2425328554360813,
-      "grad_norm": 0.5835218258508235,
+      "grad_norm": 0.5215764597896382,
       "learning_rate": 5e-06,
-      "loss": 0.6053,
+      "loss": 0.6073,
       "step": 520
     },
     {
       "epoch": 1.2664277180406214,
-      "grad_norm": 0.6402155111012049,
+      "grad_norm": 0.5142341654295087,
       "learning_rate": 5e-06,
-      "loss": 0.6086,
+      "loss": 0.6105,
       "step": 530
     },
     {
       "epoch": 1.2903225806451613,
-      "grad_norm": 0.5274545363597922,
+      "grad_norm": 0.4429903840954624,
       "learning_rate": 5e-06,
-      "loss": 0.6082,
+      "loss": 0.61,
       "step": 540
     },
     {
       "epoch": 1.3142174432497014,
-      "grad_norm": 0.5023370306863523,
+      "grad_norm": 0.4244756990330428,
       "learning_rate": 5e-06,
-      "loss": 0.6097,
+      "loss": 0.6113,
       "step": 550
     },
     {
       "epoch": 1.3381123058542412,
-      "grad_norm": 0.5336199850801069,
+      "grad_norm": 0.4664930270424248,
       "learning_rate": 5e-06,
-      "loss": 0.6037,
+      "loss": 0.6057,
       "step": 560
     },
     {
       "epoch": 1.3620071684587813,
-      "grad_norm": 0.7734469958692578,
+      "grad_norm": 0.6747787167132405,
       "learning_rate": 5e-06,
-      "loss": 0.6066,
+      "loss": 0.6084,
       "step": 570
     },
     {
       "epoch": 1.3859020310633214,
-      "grad_norm": 0.8452375931165987,
+      "grad_norm": 0.8515989236641928,
       "learning_rate": 5e-06,
-      "loss": 0.604,
+      "loss": 0.6058,
       "step": 580
     },
     {
       "epoch": 1.4097968936678615,
-      "grad_norm": 0.5907657942794627,
+      "grad_norm": 0.634857639704424,
       "learning_rate": 5e-06,
-      "loss": 0.6042,
+      "loss": 0.6058,
       "step": 590
     },
     {
       "epoch": 1.4336917562724014,
-      "grad_norm": 0.6105214261701881,
+      "grad_norm": 0.5282115500074044,
       "learning_rate": 5e-06,
-      "loss": 0.6077,
+      "loss": 0.6096,
       "step": 600
     },
     {
       "epoch": 1.4575866188769415,
-      "grad_norm": 0.6803070474017702,
+      "grad_norm": 0.5576953727126037,
       "learning_rate": 5e-06,
-      "loss": 0.6145,
+      "loss": 0.616,
       "step": 610
     },
     {
       "epoch": 1.4814814814814814,
-      "grad_norm": 0.5137151572440128,
+      "grad_norm": 0.45965397939992636,
       "learning_rate": 5e-06,
-      "loss": 0.6068,
+      "loss": 0.6082,
       "step": 620
     },
     {
       "epoch": 1.5053763440860215,
-      "grad_norm": 0.5526611853120886,
+      "grad_norm": 0.5729607655893968,
       "learning_rate": 5e-06,
-      "loss": 0.6023,
+      "loss": 0.6037,
       "step": 630
     },
     {
       "epoch": 1.5292712066905616,
-      "grad_norm": 0.5276740317417068,
+      "grad_norm": 0.4420855639504453,
       "learning_rate": 5e-06,
-      "loss": 0.6066,
+      "loss": 0.608,
       "step": 640
     },
     {
       "epoch": 1.5531660692951017,
-      "grad_norm": 0.5266173782221237,
+      "grad_norm": 0.4815965030552482,
       "learning_rate": 5e-06,
-      "loss": 0.604,
+      "loss": 0.6053,
       "step": 650
     },
     {
       "epoch": 1.5770609318996416,
-      "grad_norm": 0.5879899658739348,
+      "grad_norm": 0.5446732967871324,
       "learning_rate": 5e-06,
-      "loss": 0.6062,
+      "loss": 0.6076,
       "step": 660
     },
     {
       "epoch": 1.6009557945041815,
-      "grad_norm": 0.6426153237314072,
+      "grad_norm": 0.5773921107864519,
       "learning_rate": 5e-06,
-      "loss": 0.6044,
+      "loss": 0.6058,
       "step": 670
     },
     {
       "epoch": 1.6248506571087216,
-      "grad_norm": 0.5633170243940351,
+      "grad_norm": 0.44904612161350127,
       "learning_rate": 5e-06,
-      "loss": 0.5975,
+      "loss": 0.5988,
       "step": 680
     },
     {
       "epoch": 1.6487455197132617,
-      "grad_norm": 0.5448798915341956,
+      "grad_norm": 0.4659803956684399,
       "learning_rate": 5e-06,
-      "loss": 0.5969,
+      "loss": 0.598,
       "step": 690
     },
     {
       "epoch": 1.6726403823178018,
-      "grad_norm": 0.484774901310647,
+      "grad_norm": 0.4361474003132107,
       "learning_rate": 5e-06,
-      "loss": 0.6069,
+      "loss": 0.6081,
       "step": 700
     },
     {
       "epoch": 1.6965352449223416,
-      "grad_norm": 0.5257867856081395,
+      "grad_norm": 0.4702827100539838,
       "learning_rate": 5e-06,
-      "loss": 0.5985,
+      "loss": 0.5997,
       "step": 710
     },
     {
       "epoch": 1.7204301075268817,
-      "grad_norm": 0.5602525718442715,
+      "grad_norm": 0.46962735672309736,
       "learning_rate": 5e-06,
-      "loss": 0.612,
+      "loss": 0.6135,
       "step": 720
     },
     {
       "epoch": 1.7443249701314216,
-      "grad_norm": 0.5040592279383703,
+      "grad_norm": 0.5064462322593579,
       "learning_rate": 5e-06,
-      "loss": 0.6021,
+      "loss": 0.6034,
       "step": 730
     },
     {
       "epoch": 1.7682198327359617,
-      "grad_norm": 0.6140508371910811,
+      "grad_norm": 0.6442892941899157,
       "learning_rate": 5e-06,
-      "loss": 0.6029,
+      "loss": 0.6044,
       "step": 740
     },
     {
       "epoch": 1.7921146953405018,
-      "grad_norm": 0.526723228546995,
+      "grad_norm": 0.449859458258856,
       "learning_rate": 5e-06,
-      "loss": 0.6051,
+      "loss": 0.6062,
       "step": 750
     },
     {
       "epoch": 1.816009557945042,
-      "grad_norm": 0.5485611376004595,
+      "grad_norm": 0.47467567108778363,
       "learning_rate": 5e-06,
-      "loss": 0.602,
+      "loss": 0.6035,
       "step": 760
     },
     {
       "epoch": 1.8399044205495818,
-      "grad_norm": 0.4925549773889819,
+      "grad_norm": 0.43550415026449085,
       "learning_rate": 5e-06,
-      "loss": 0.5976,
+      "loss": 0.5987,
       "step": 770
     },
     {
       "epoch": 1.863799283154122,
-      "grad_norm": 0.5127568984869073,
+      "grad_norm": 0.48913780227876247,
       "learning_rate": 5e-06,
-      "loss": 0.6015,
+      "loss": 0.6031,
       "step": 780
     },
     {
       "epoch": 1.8876941457586618,
-      "grad_norm": 0.7953660178238419,
+      "grad_norm": 0.5594004132295759,
       "learning_rate": 5e-06,
-      "loss": 0.5983,
+      "loss": 0.5995,
       "step": 790
     },
     {
       "epoch": 1.911589008363202,
-      "grad_norm": 0.4644171655680959,
+      "grad_norm": 0.4971730954697683,
       "learning_rate": 5e-06,
-      "loss": 0.6043,
+      "loss": 0.6056,
       "step": 800
     },
     {
       "epoch": 1.935483870967742,
-      "grad_norm": 0.5504656432596945,
+      "grad_norm": 0.46186692571258725,
       "learning_rate": 5e-06,
-      "loss": 0.5991,
+      "loss": 0.6004,
       "step": 810
     },
     {
       "epoch": 1.959378733572282,
-      "grad_norm": 0.5014583862445193,
+      "grad_norm": 0.4508830943663248,
       "learning_rate": 5e-06,
-      "loss": 0.6027,
+      "loss": 0.604,
       "step": 820
     },
     {
       "epoch": 1.983273596176822,
-      "grad_norm": 0.5169701196971361,
+      "grad_norm": 0.501207898912081,
       "learning_rate": 5e-06,
-      "loss": 0.6065,
+      "loss": 0.608,
       "step": 830
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.6383097767829895,
-      "eval_runtime": 225.1334,
-      "eval_samples_per_second": 50.073,
-      "eval_steps_per_second": 0.395,
+      "eval_loss": 0.6365451812744141,
+      "eval_runtime": 226.5918,
+      "eval_samples_per_second": 49.75,
+      "eval_steps_per_second": 0.393,
       "step": 837
     },
     {
       "epoch": 2.007168458781362,
-      "grad_norm": 0.9522384301889181,
+      "grad_norm": 0.85054207213812,
       "learning_rate": 5e-06,
-      "loss": 0.5892,
+      "loss": 0.592,
       "step": 840
     },
     {
       "epoch": 2.031063321385902,
-      "grad_norm": 0.5604355482368693,
+      "grad_norm": 0.5201507086108782,
       "learning_rate": 5e-06,
-      "loss": 0.555,
+      "loss": 0.5608,
       "step": 850
     },
     {
       "epoch": 2.054958183990442,
-      "grad_norm": 0.8454377137422782,
+      "grad_norm": 0.7188535226812537,
       "learning_rate": 5e-06,
-      "loss": 0.559,
+      "loss": 0.5647,
       "step": 860
     },
     {
       "epoch": 2.078853046594982,
-      "grad_norm": 0.6125609351007736,
+      "grad_norm": 0.4763195641282365,
       "learning_rate": 5e-06,
-      "loss": 0.5588,
+      "loss": 0.5644,
       "step": 870
     },
     {
       "epoch": 2.1027479091995223,
-      "grad_norm": 0.550684661961655,
+      "grad_norm": 0.5081415859208832,
       "learning_rate": 5e-06,
-      "loss": 0.5566,
+      "loss": 0.5622,
       "step": 880
     },
     {
       "epoch": 2.126642771804062,
-      "grad_norm": 0.5268020228547002,
+      "grad_norm": 0.5931792000293172,
       "learning_rate": 5e-06,
-      "loss": 0.5509,
+      "loss": 0.5563,
       "step": 890
     },
     {
       "epoch": 2.150537634408602,
-      "grad_norm": 0.5374686724727457,
+      "grad_norm": 0.49851033855755,
       "learning_rate": 5e-06,
-      "loss": 0.5558,
+      "loss": 0.5611,
       "step": 900
     },
     {
       "epoch": 2.174432497013142,
-      "grad_norm": 0.6591728468868026,
+      "grad_norm": 0.5379278365329638,
       "learning_rate": 5e-06,
-      "loss": 0.5523,
+      "loss": 0.5573,
       "step": 910
     },
     {
       "epoch": 2.198327359617682,
-      "grad_norm": 0.5758411262414109,
+      "grad_norm": 0.5350268044233742,
       "learning_rate": 5e-06,
-      "loss": 0.5626,
+      "loss": 0.5678,
       "step": 920
     },
     {
       "epoch": 2.2222222222222223,
-      "grad_norm": 0.5974631029863997,
+      "grad_norm": 0.5689123372030673,
       "learning_rate": 5e-06,
-      "loss": 0.5633,
+      "loss": 0.5686,
       "step": 930
     },
     {
       "epoch": 2.2461170848267624,
-      "grad_norm": 0.6693430289411901,
+      "grad_norm": 0.6671996296787344,
       "learning_rate": 5e-06,
-      "loss": 0.5569,
+      "loss": 0.5622,
       "step": 940
     },
     {
       "epoch": 2.270011947431302,
-      "grad_norm": 0.5235798291837075,
+      "grad_norm": 0.4506810352733908,
       "learning_rate": 5e-06,
-      "loss": 0.5592,
+      "loss": 0.5642,
       "step": 950
     },
     {
       "epoch": 2.293906810035842,
-      "grad_norm": 0.5951063252000355,
+      "grad_norm": 0.5358151280205125,
       "learning_rate": 5e-06,
-      "loss": 0.558,
+      "loss": 0.5628,
       "step": 960
     },
     {
       "epoch": 2.3178016726403823,
-      "grad_norm": 0.7007904171286725,
+      "grad_norm": 0.5566771627404731,
       "learning_rate": 5e-06,
-      "loss": 0.5585,
+      "loss": 0.5634,
       "step": 970
     },
     {
       "epoch": 2.3416965352449224,
-      "grad_norm": 0.6157005538099132,
+      "grad_norm": 0.49963936030628325,
       "learning_rate": 5e-06,
-      "loss": 0.5584,
+      "loss": 0.5632,
       "step": 980
     },
     {
       "epoch": 2.3655913978494625,
-      "grad_norm": 0.5560793638442904,
+      "grad_norm": 0.48679480824629434,
       "learning_rate": 5e-06,
-      "loss": 0.5536,
+      "loss": 0.5583,
       "step": 990
     },
     {
       "epoch": 2.389486260454002,
-      "grad_norm": 0.5994284875647052,
+      "grad_norm": 0.5074816823498985,
       "learning_rate": 5e-06,
-      "loss": 0.559,
+      "loss": 0.5636,
       "step": 1000
     },
     {
       "epoch": 2.4133811230585422,
-      "grad_norm": 0.6178447526276027,
+      "grad_norm": 0.5739148795335686,
       "learning_rate": 5e-06,
-      "loss": 0.5622,
+      "loss": 0.5671,
       "step": 1010
     },
     {
       "epoch": 2.4372759856630823,
-      "grad_norm": 0.598268249438012,
+      "grad_norm": 0.6501742104516552,
       "learning_rate": 5e-06,
-      "loss": 0.5617,
+      "loss": 0.5666,
       "step": 1020
     },
     {
       "epoch": 2.4611708482676224,
-      "grad_norm": 0.6456807780700198,
+      "grad_norm": 0.43406800220014613,
       "learning_rate": 5e-06,
-      "loss": 0.5598,
+      "loss": 0.5645,
       "step": 1030
     },
     {
       "epoch": 2.4850657108721625,
-      "grad_norm": 0.5838521623874042,
+      "grad_norm": 0.47946981158627366,
       "learning_rate": 5e-06,
-      "loss": 0.5565,
+      "loss": 0.5612,
       "step": 1040
     },
     {
       "epoch": 2.5089605734767026,
-      "grad_norm": 0.6160425605583054,
+      "grad_norm": 0.5508677225984592,
       "learning_rate": 5e-06,
-      "loss": 0.561,
+      "loss": 0.5658,
       "step": 1050
     },
     {
       "epoch": 2.5328554360812428,
-      "grad_norm": 0.6471012332418825,
+      "grad_norm": 0.6172108213167418,
       "learning_rate": 5e-06,
-      "loss": 0.5609,
+      "loss": 0.5656,
       "step": 1060
     },
     {
       "epoch": 2.5567502986857824,
-      "grad_norm": 0.7438495174826055,
+      "grad_norm": 0.6149816712572169,
       "learning_rate": 5e-06,
-      "loss": 0.559,
+      "loss": 0.5637,
       "step": 1070
     },
     {
       "epoch": 2.5806451612903225,
-      "grad_norm": 0.5965448247540126,
+      "grad_norm": 0.5494076230620691,
       "learning_rate": 5e-06,
-      "loss": 0.57,
+      "loss": 0.5748,
       "step": 1080
     },
     {
       "epoch": 2.6045400238948626,
-      "grad_norm": 0.6030850533182174,
+      "grad_norm": 0.5098015036653776,
       "learning_rate": 5e-06,
-      "loss": 0.5619,
+      "loss": 0.5665,
       "step": 1090
     },
     {
       "epoch": 2.6284348864994027,
-      "grad_norm": 0.5627845167167422,
+      "grad_norm": 0.4763003977246298,
       "learning_rate": 5e-06,
-      "loss": 0.5615,
+      "loss": 0.5659,
       "step": 1100
     },
     {
       "epoch": 2.652329749103943,
-      "grad_norm": 0.5586383353478368,
+      "grad_norm": 0.45015059391064355,
       "learning_rate": 5e-06,
-      "loss": 0.5567,
+      "loss": 0.5613,
       "step": 1110
     },
     {
       "epoch": 2.6762246117084825,
-      "grad_norm": 0.6169226364241693,
+      "grad_norm": 0.5195016081388676,
       "learning_rate": 5e-06,
-      "loss": 0.5614,
+      "loss": 0.5661,
       "step": 1120
     },
     {
       "epoch": 2.7001194743130226,
-      "grad_norm": 0.5655962840268283,
+      "grad_norm": 0.461979850463992,
       "learning_rate": 5e-06,
-      "loss": 0.566,
+      "loss": 0.5703,
       "step": 1130
     },
     {
       "epoch": 2.7240143369175627,
-      "grad_norm": 0.5379947967296507,
+      "grad_norm": 0.4611698536891998,
       "learning_rate": 5e-06,
-      "loss": 0.5585,
+      "loss": 0.5628,
       "step": 1140
     },
     {
       "epoch": 2.7479091995221028,
-      "grad_norm": 0.6022462071356245,
+      "grad_norm": 0.5474996121575114,
       "learning_rate": 5e-06,
-      "loss": 0.5647,
+      "loss": 0.5692,
       "step": 1150
     },
     {
       "epoch": 2.771804062126643,
-      "grad_norm": 0.5630793813942363,
+      "grad_norm": 0.477411852958178,
       "learning_rate": 5e-06,
-      "loss": 0.5701,
+      "loss": 0.5745,
       "step": 1160
     },
     {
       "epoch": 2.795698924731183,
-      "grad_norm": 0.7669976902930744,
+      "grad_norm": 0.48004817339516165,
       "learning_rate": 5e-06,
-      "loss": 0.5584,
+      "loss": 0.5625,
       "step": 1170
     },
     {
       "epoch": 2.819593787335723,
-      "grad_norm": 0.6081830710715922,
+      "grad_norm": 0.5043226922994581,
       "learning_rate": 5e-06,
-      "loss": 0.5619,
+      "loss": 0.5664,
       "step": 1180
     },
     {
       "epoch": 2.8434886499402627,
-      "grad_norm": 0.5471146843592755,
+      "grad_norm": 0.4988305698181874,
       "learning_rate": 5e-06,
-      "loss": 0.5607,
+      "loss": 0.5649,
       "step": 1190
     },
     {
       "epoch": 2.867383512544803,
-      "grad_norm": 0.500432815577127,
+      "grad_norm": 0.4569103859069353,
       "learning_rate": 5e-06,
-      "loss": 0.5605,
+      "loss": 0.5647,
       "step": 1200
     },
     {
       "epoch": 2.891278375149343,
-      "grad_norm": 0.5513789797790463,
+      "grad_norm": 0.46286445346886024,
       "learning_rate": 5e-06,
-      "loss": 0.558,
+      "loss": 0.5621,
       "step": 1210
     },
     {
       "epoch": 2.915173237753883,
-      "grad_norm": 0.5699534515350526,
+      "grad_norm": 0.5296890930558641,
       "learning_rate": 5e-06,
-      "loss": 0.5626,
+      "loss": 0.5668,
       "step": 1220
     },
     {
       "epoch": 2.9390681003584227,
-      "grad_norm": 0.6140827402540434,
+      "grad_norm": 0.5546209266748766,
       "learning_rate": 5e-06,
-      "loss": 0.5648,
+      "loss": 0.5688,
       "step": 1230
     },
     {
       "epoch": 2.962962962962963,
-      "grad_norm": 0.6486122276400309,
+      "grad_norm": 0.5910470543653078,
       "learning_rate": 5e-06,
-      "loss": 0.5627,
+      "loss": 0.5668,
       "step": 1240
     },
     {
       "epoch": 2.986857825567503,
-      "grad_norm": 0.5448840278284515,
+      "grad_norm": 0.5524701632367459,
       "learning_rate": 5e-06,
-      "loss": 0.5672,
+      "loss": 0.5714,
       "step": 1250
     },
     {
       "epoch": 2.996415770609319,
-      "eval_loss": 0.6404265761375427,
-      "eval_runtime": 226.6238,
-      "eval_samples_per_second": 49.743,
-      "eval_steps_per_second": 0.393,
+      "eval_loss": 0.6373269557952881,
+      "eval_runtime": 227.0283,
+      "eval_samples_per_second": 49.655,
+      "eval_steps_per_second": 0.392,
       "step": 1254
     },
     {
       "epoch": 2.996415770609319,
       "step": 1254,
       "total_flos": 2100077946470400.0,
-      "train_loss": 0.6158634758832162,
-      "train_runtime": 37532.7652,
-      "train_samples_per_second": 17.12,
+      "train_loss": 0.6173035094612523,
+      "train_runtime": 37751.1113,
+      "train_samples_per_second": 17.021,
       "train_steps_per_second": 0.033
     }
   ],