End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-km-10_v2
 metrics:
 - accuracy
 model-index:
 - name: bert_base_km_10_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-km-10_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-km-10_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.15457422161948353
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -14,10 +26,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # bert_base_km_10_v2
 
-This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
+This model is a fine-tuned version of [](https://huggingface.co/) on the Hartunka/processed_wikitext-103-raw-v1-km-10_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 6.3615
-- Accuracy: 0.1524
+- Loss: 6.2999
+- Accuracy: 0.1546
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15561959654178675,
-    "eval_loss": 6.325186252593994,
-    "eval_runtime": 1.0585,
+    "eval_accuracy": 0.15457422161948353,
+    "eval_loss": 6.299930572509766,
+    "eval_runtime": 1.0219,
     "eval_samples": 479,
-    "eval_samples_per_second": 452.547,
-    "eval_steps_per_second": 4.724,
-    "perplexity": 558.461818941483,
+    "eval_samples_per_second": 468.743,
+    "eval_steps_per_second": 4.893,
+    "perplexity": 544.5341031773928,
     "total_flos": 1.5149776477860864e+18,
-    "train_loss": 6.212542548319755,
-    "train_runtime": 24740.6956,
+    "train_loss": 6.208375089887207,
+    "train_runtime": 24672.4974,
     "train_samples": 228639,
-    "train_samples_per_second": 231.035,
-    "train_steps_per_second": 2.407
+    "train_samples_per_second": 231.674,
+    "train_steps_per_second": 2.414
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15561959654178675,
-    "eval_loss": 6.325186252593994,
-    "eval_runtime": 1.0585,
+    "eval_accuracy": 0.15457422161948353,
+    "eval_loss": 6.299930572509766,
+    "eval_runtime": 1.0219,
     "eval_samples": 479,
-    "eval_samples_per_second": 452.547,
-    "eval_steps_per_second": 4.724,
-    "perplexity": 558.461818941483
+    "eval_samples_per_second": 468.743,
+    "eval_steps_per_second": 4.893,
+    "perplexity": 544.5341031773928
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 1.5149776477860864e+18,
-    "train_loss": 6.212542548319755,
-    "train_runtime": 24740.6956,
+    "train_loss": 6.208375089887207,
+    "train_runtime": 24672.4974,
     "train_samples": 228639,
-    "train_samples_per_second": 231.035,
-    "train_steps_per_second": 2.407
+    "train_samples_per_second": 231.674,
+    "train_steps_per_second": 2.414
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_global_step": 40000,
-  "best_metric": 6.3062849044799805,
-  "best_model_checkpoint": "bert_base_km_10_v2/checkpoint-40000",
+  "best_global_step": 20000,
+  "best_metric": 6.302638053894043,
+  "best_model_checkpoint": "bert_base_km_10_v2/checkpoint-20000",
   "epoch": 25.0,
   "eval_steps": 10000,
   "global_step": 59550,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.9562325477600098,
+      "grad_norm": 3.58536434173584,
       "learning_rate": 5e-06,
-      "loss": 11.0235,
+      "loss": 11.4964,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 17.698453903198242,
+      "grad_norm": 10.171985626220703,
       "learning_rate": 1e-05,
-      "loss": 8.8552,
+      "loss": 9.3544,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 22.179046630859375,
+      "grad_norm": 18.19382667541504,
       "learning_rate": 1.5e-05,
-      "loss": 8.0013,
+      "loss": 8.2074,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 13.386920928955078,
+      "grad_norm": 15.068007469177246,
       "learning_rate": 2e-05,
-      "loss": 7.658,
+      "loss": 7.7786,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 4.086881160736084,
+      "grad_norm": 10.516756057739258,
       "learning_rate": 2.5e-05,
-      "loss": 7.4443,
+      "loss": 7.5311,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 18.590084075927734,
+      "grad_norm": 10.947854995727539,
       "learning_rate": 3e-05,
-      "loss": 7.2741,
+      "loss": 7.3186,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 5.107962608337402,
+      "grad_norm": 4.6012678146362305,
       "learning_rate": 3.5e-05,
-      "loss": 7.148,
+      "loss": 7.1826,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 5.024056911468506,
+      "grad_norm": 5.221875190734863,
       "learning_rate": 4e-05,
-      "loss": 7.0377,
+      "loss": 7.064,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 2.413734197616577,
+      "grad_norm": 3.059871196746826,
       "learning_rate": 4.5e-05,
-      "loss": 6.9474,
+      "loss": 6.9537,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 8.736750602722168,
+      "grad_norm": 3.354302167892456,
       "learning_rate": 5e-05,
-      "loss": 6.8616,
+      "loss": 6.853,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 2.9648070335388184,
+      "grad_norm": 3.8330137729644775,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 6.7957,
+      "loss": 6.7617,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 2.61800479888916,
+      "grad_norm": 3.2019448280334473,
       "learning_rate": 6e-05,
-      "loss": 6.7612,
+      "loss": 6.7295,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 1.6698569059371948,
+      "grad_norm": 2.814960479736328,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 6.7282,
+      "loss": 6.6955,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 3.9698450565338135,
+      "grad_norm": 2.536619186401367,
       "learning_rate": 7e-05,
-      "loss": 6.6944,
+      "loss": 6.6706,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 3.265909194946289,
+      "grad_norm": 1.7168173789978027,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 6.6338,
+      "loss": 6.5959,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 4.147052764892578,
+      "grad_norm": 3.416818141937256,
       "learning_rate": 8e-05,
-      "loss": 6.6197,
+      "loss": 6.5641,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 3.2067461013793945,
+      "grad_norm": 1.9000678062438965,
       "learning_rate": 8.5e-05,
-      "loss": 6.6051,
+      "loss": 6.5562,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 4.531407833099365,
+      "grad_norm": 2.7018251419067383,
       "learning_rate": 9e-05,
-      "loss": 6.5858,
+      "loss": 6.5505,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 3.9769198894500732,
+      "grad_norm": 2.684298276901245,
       "learning_rate": 9.5e-05,
-      "loss": 6.5653,
+      "loss": 6.5317,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 2.9494242668151855,
+      "grad_norm": 2.8553645610809326,
       "learning_rate": 0.0001,
-      "loss": 6.4969,
+      "loss": 6.4533,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.1499485238876986,
-      "eval_loss": 6.510183334350586,
-      "eval_runtime": 1.0354,
-      "eval_samples_per_second": 462.609,
-      "eval_steps_per_second": 4.829,
+      "eval_accuracy": 0.15014329836667686,
+      "eval_loss": 6.437900066375732,
+      "eval_runtime": 1.0338,
+      "eval_samples_per_second": 463.354,
+      "eval_steps_per_second": 4.837,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 1.840803623199463,
+      "grad_norm": 3.8193185329437256,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 6.4883,
+      "loss": 6.4487,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 1.7808061838150024,
+      "grad_norm": 1.8346351385116577,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 6.4772,
+      "loss": 6.4446,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 2.438516139984131,
+      "grad_norm": 1.7950433492660522,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 6.4639,
+      "loss": 6.44,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 1.6112890243530273,
+      "grad_norm": 1.7432068586349487,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 6.4355,
+      "loss": 6.4083,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 3.26106333732605,
+      "grad_norm": 2.3740103244781494,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 6.3789,
+      "loss": 6.3451,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 1.8721084594726562,
+      "grad_norm": 2.2546207904815674,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 6.3709,
+      "loss": 6.3388,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 1.87173330783844,
+      "grad_norm": 2.066467523574829,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 6.3687,
+      "loss": 6.3371,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 3.479503870010376,
+      "grad_norm": 2.4825870990753174,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 6.3636,
+      "loss": 6.3303,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 2.306701421737671,
+      "grad_norm": 2.1828935146331787,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 6.323,
+      "loss": 6.2988,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 1.920249581336975,
+      "grad_norm": 1.4010400772094727,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 6.2928,
+      "loss": 6.2573,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 1.3836824893951416,
+      "grad_norm": 2.2956342697143555,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 6.2833,
+      "loss": 6.259,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 1.9751460552215576,
+      "grad_norm": 1.6517410278320312,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 6.284,
+      "loss": 6.2507,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 1.621890902519226,
+      "grad_norm": 1.434820294380188,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 6.2808,
+      "loss": 6.2486,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 1.5578926801681519,
+      "grad_norm": 1.8767292499542236,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 6.2316,
+      "loss": 6.2043,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 1.4018528461456299,
+      "grad_norm": 2.1279685497283936,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 6.2145,
+      "loss": 6.1835,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 2.7806613445281982,
+      "grad_norm": 1.8389158248901367,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 6.2183,
+      "loss": 6.19,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 2.7784783840179443,
+      "grad_norm": 2.112571954727173,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 6.2219,
+      "loss": 6.1898,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 1.9644310474395752,
+      "grad_norm": 2.5010149478912354,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 6.2192,
+      "loss": 6.1894,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 2.034841537475586,
+      "grad_norm": 1.4285669326782227,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 6.1627,
+      "loss": 6.1437,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 3.0792479515075684,
+      "grad_norm": 1.5599991083145142,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 6.1617,
+      "loss": 6.1383,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15317489267375628,
-      "eval_loss": 6.349700927734375,
-      "eval_runtime": 1.015,
-      "eval_samples_per_second": 471.903,
-      "eval_steps_per_second": 4.926,
+      "eval_accuracy": 0.15224894076713713,
+      "eval_loss": 6.302638053894043,
+      "eval_runtime": 1.0065,
+      "eval_samples_per_second": 475.904,
+      "eval_steps_per_second": 4.968,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 1.869396686553955,
+      "grad_norm": 2.5455243587493896,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 6.1613,
+      "loss": 6.129,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 2.249077796936035,
+      "grad_norm": 2.5203423500061035,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 6.1601,
+      "loss": 6.1316,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 0.937393844127655,
+      "grad_norm": 1.9525269269943237,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 6.1498,
+      "loss": 6.1261,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 1.8309992551803589,
+      "grad_norm": 2.4618887901306152,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 6.1159,
+      "loss": 6.0892,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 2.2449676990509033,
+      "grad_norm": 1.0711874961853027,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 6.1121,
+      "loss": 6.0917,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 2.006056070327759,
+      "grad_norm": 1.9631586074829102,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 6.1132,
+      "loss": 6.0949,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 1.902909517288208,
+      "grad_norm": 2.284971237182617,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 6.1136,
+      "loss": 6.0862,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 1.2427181005477905,
+      "grad_norm": 2.3909666538238525,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 6.0892,
+      "loss": 6.069,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 2.3167381286621094,
+      "grad_norm": 0.7460120320320129,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 6.0766,
+      "loss": 6.0557,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 1.573493242263794,
+      "grad_norm": 2.5700876712799072,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 6.0668,
+      "loss": 6.0485,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 3.7507879734039307,
+      "grad_norm": 1.9020856618881226,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 6.0654,
+      "loss": 6.0461,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 2.302229166030884,
+      "grad_norm": 2.5678093433380127,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 6.07,
+      "loss": 6.0503,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 3.5238773822784424,
+      "grad_norm": 2.4644973278045654,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 6.0442,
+      "loss": 6.0298,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 2.082115411758423,
+      "grad_norm": 1.6001275777816772,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 6.0376,
+      "loss": 6.0205,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 2.6825358867645264,
+      "grad_norm": 1.3191250562667847,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 6.0319,
+      "loss": 6.0144,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 2.8914504051208496,
+      "grad_norm": 3.0827748775482178,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 6.0314,
+      "loss": 6.0097,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 2.513441801071167,
+      "grad_norm": 0.7663680911064148,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 6.0281,
+      "loss": 6.0088,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 2.4295971393585205,
+      "grad_norm": 1.7403725385665894,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 6.0116,
+      "loss": 5.993,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 1.3470953702926636,
+      "grad_norm": 1.2999112606048584,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 5.9952,
+      "loss": 5.9851,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 1.2119895219802856,
+      "grad_norm": 2.4106132984161377,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 6.0026,
+      "loss": 5.9882,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.154461762319165,
-      "eval_loss": 6.342120170593262,
-      "eval_runtime": 1.0069,
-      "eval_samples_per_second": 475.715,
-      "eval_steps_per_second": 4.966,
+      "eval_accuracy": 0.15465763774239583,
+      "eval_loss": 6.30724573135376,
+      "eval_runtime": 1.0385,
+      "eval_samples_per_second": 461.229,
+      "eval_steps_per_second": 4.814,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 2.608718156814575,
+      "grad_norm": 1.5281882286071777,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 5.996,
+      "loss": 5.9813,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 1.874100685119629,
+      "grad_norm": 1.9092435836791992,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 5.9987,
+      "loss": 5.9817,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 2.4008915424346924,
+      "grad_norm": 0.7660443186759949,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 5.9732,
+      "loss": 5.9619,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 0.8543368577957153,
+      "grad_norm": 2.1503288745880127,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 5.9705,
+      "loss": 5.9626,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 1.8783243894577026,
+      "grad_norm": 2.018294095993042,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 5.9761,
+      "loss": 5.9626,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 3.0358753204345703,
+      "grad_norm": 2.2128100395202637,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 5.9733,
+      "loss": 5.9588,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 2.7263076305389404,
+      "grad_norm": 0.7462359666824341,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 5.9591,
+      "loss": 5.9494,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 2.218528985977173,
+      "grad_norm": 0.8331591486930847,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 5.9475,
+      "loss": 5.9356,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 2.09617018699646,
+      "grad_norm": 2.5936596393585205,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 5.949,
+      "loss": 5.9355,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 0.8091275095939636,
+      "grad_norm": 1.2686055898666382,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 5.9477,
+      "loss": 5.934,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 0.9276658892631531,
+      "grad_norm": 2.0998692512512207,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 5.9454,
+      "loss": 5.9324,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 2.462109327316284,
+      "grad_norm": 1.187612771987915,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 5.9331,
+      "loss": 5.9227,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 2.254075288772583,
+      "grad_norm": 1.0955361127853394,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 5.9298,
+      "loss": 5.9181,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 1.9004478454589844,
+      "grad_norm": 0.9621930122375488,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 5.9271,
+      "loss": 5.9146,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 2.358386516571045,
+      "grad_norm": 2.776153087615967,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 5.9278,
+      "loss": 5.9147,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 2.4182186126708984,
+      "grad_norm": 2.1034648418426514,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 5.9239,
+      "loss": 5.9127,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 2.934300184249878,
+      "grad_norm": 2.1946804523468018,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 5.9158,
+      "loss": 5.9063,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 1.7431899309158325,
+      "grad_norm": 1.3870270252227783,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 5.9073,
+      "loss": 5.8973,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 1.3285027742385864,
+      "grad_norm": 0.937556266784668,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 5.908,
+      "loss": 5.8998,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 3.0293195247650146,
+      "grad_norm": 1.4358993768692017,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 5.8999,
+      "loss": 5.8876,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.15363911629471796,
-      "eval_loss": 6.3062849044799805,
-      "eval_runtime": 1.0052,
-      "eval_samples_per_second": 476.529,
-      "eval_steps_per_second": 4.974,
+      "eval_accuracy": 0.15305035325782212,
+      "eval_loss": 6.386124610900879,
+      "eval_runtime": 1.0209,
+      "eval_samples_per_second": 469.171,
+      "eval_steps_per_second": 4.897,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 1.3343366384506226,
+      "grad_norm": 1.1544150114059448,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 5.9068,
+      "loss": 5.8928,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 0.9144035577774048,
+      "grad_norm": 3.526372194290161,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 5.8911,
+      "loss": 5.8801,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 1.2867867946624756,
+      "grad_norm": 2.3305583000183105,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 5.8886,
+      "loss": 5.8785,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 0.9098787903785706,
+      "grad_norm": 2.5416979789733887,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 5.8906,
+      "loss": 5.8797,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 1.5115971565246582,
+      "grad_norm": 3.569450855255127,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 5.8865,
+      "loss": 5.876,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 0.9073482751846313,
+      "grad_norm": 0.7459242343902588,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 5.8811,
+      "loss": 5.8723,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 3.656424045562744,
+      "grad_norm": 2.551079750061035,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 5.8726,
+      "loss": 5.864,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 2.196565866470337,
+      "grad_norm": 2.6818277835845947,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 5.8744,
+      "loss": 5.8622,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 2.9759039878845215,
+      "grad_norm": 0.7526069283485413,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 5.8729,
+      "loss": 5.864,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 0.7581497430801392,
+      "grad_norm": 2.505458116531372,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 5.875,
+      "loss": 5.8651,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 2.455521583557129,
+      "grad_norm": 0.8404247760772705,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 5.8716,
+      "loss": 5.8593,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 3.327491283416748,
+      "grad_norm": 0.7641693353652954,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 5.858,
+      "loss": 5.8476,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 0.8205088973045349,
+      "grad_norm": 0.9754248857498169,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 5.8592,
+      "loss": 5.8501,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 0.7729224562644958,
+      "grad_norm": 2.362238883972168,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 5.8604,
+      "loss": 5.8494,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 3.09210467338562,
+      "grad_norm": 1.8124767541885376,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 5.8569,
+      "loss": 5.8456,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 0.7637450695037842,
+      "grad_norm": 2.56758189201355,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 5.8508,
+      "loss": 5.8399,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 0.7816163897514343,
+      "grad_norm": 2.143838405609131,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 5.8493,
+      "loss": 5.838,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 1.9569042921066284,
+      "grad_norm": 0.76775062084198,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 5.8435,
+      "loss": 5.834,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 1.819555401802063,
+      "grad_norm": 0.7878410220146179,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 5.8434,
+      "loss": 5.8331,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 3.0075411796569824,
+      "grad_norm": 1.194934368133545,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 5.8424,
+      "loss": 5.8344,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.1518210110185886,
-      "eval_loss": 6.399431228637695,
-      "eval_runtime": 1.0407,
-      "eval_samples_per_second": 460.263,
-      "eval_steps_per_second": 4.804,
+      "eval_accuracy": 0.152409790562705,
+      "eval_loss": 6.36154842376709,
+      "eval_runtime": 1.0113,
+      "eval_samples_per_second": 473.628,
+      "eval_steps_per_second": 4.944,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 1.2910844087600708,
+      "grad_norm": 1.0040415525436401,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 5.8371,
+      "loss": 5.8265,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 1.4958844184875488,
+      "grad_norm": 0.7996358275413513,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 5.8351,
+      "loss": 5.8269,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 3.3193728923797607,
+      "grad_norm": 1.9936509132385254,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 5.8341,
+      "loss": 5.827,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 0.7783209085464478,
+      "grad_norm": 1.050391435623169,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 5.8346,
+      "loss": 5.8246,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 4.355439186096191,
+      "grad_norm": 0.7597755193710327,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 5.8323,
+      "loss": 5.8239,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 2.6427314281463623,
+      "grad_norm": 0.762036144733429,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 5.8208,
+      "loss": 5.8129,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 2.9217841625213623,
+      "grad_norm": 3.3790647983551025,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 5.8258,
+      "loss": 5.8149,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 1.1128212213516235,
+      "grad_norm": 0.7693343162536621,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 5.8264,
+      "loss": 5.8187,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 0.7884045243263245,
+      "grad_norm": 4.682248115539551,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 5.8219,
+      "loss": 5.8147,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 0.7929973602294922,
+      "grad_norm": 1.8237544298171997,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 5.8197,
+      "loss": 5.8109,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 0.7606703042984009,
+      "grad_norm": 0.7912262082099915,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 5.8172,
+      "loss": 5.8103,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 1.0099503993988037,
+      "grad_norm": 0.7988426089286804,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 5.8169,
+      "loss": 5.8089,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 0.7755547165870667,
+      "grad_norm": 0.8773173093795776,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 5.8117,
+      "loss": 5.8048,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 0.7879419922828674,
+      "grad_norm": 0.7588059306144714,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 5.815,
+      "loss": 5.8056,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 0.7881122827529907,
+      "grad_norm": 0.7636212110519409,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 5.8103,
+      "loss": 5.8036,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 0.7837434411048889,
+      "grad_norm": 0.8246304392814636,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 5.8133,
+      "loss": 5.805,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 0.8001644611358643,
+      "grad_norm": 0.7530261874198914,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 5.8074,
+      "loss": 5.8004,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 0.7615817189216614,
+      "grad_norm": 0.7483027577400208,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 5.8077,
+      "loss": 5.8001,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 0.7823144793510437,
+      "grad_norm": 0.7466872334480286,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 5.8079,
+      "loss": 5.7992,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 1.5149776477860864e+18,
-      "train_loss": 6.212542548319755,
-      "train_runtime": 24740.6956,
-      "train_samples_per_second": 231.035,
-      "train_steps_per_second": 2.407
+      "train_loss": 6.208375089887207,
+      "train_runtime": 24672.4974,
+      "train_samples_per_second": 231.674,
+      "train_steps_per_second": 2.414
     }
   ],
   "logging_steps": 500,