End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-rand-100_v2
 metrics:
 - accuracy
 model-index:
 - name: distilbert_rand_100_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-rand-100_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-rand-100_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.1535288466971803
 ---
 
 <!-- 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. -->
 
 # distilbert_rand_100_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-rand-100_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 14.5562
-- Accuracy: 0.1517
+- Loss: 10.8332
+- Accuracy: 0.1535
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15299203254788948,
-    "eval_loss": 10.892149925231934,
-    "eval_runtime": 0.8428,
+    "eval_accuracy": 0.1535288466971803,
+    "eval_loss": 10.833195686340332,
+    "eval_runtime": 0.8447,
     "eval_samples": 479,
-    "eval_samples_per_second": 568.36,
-    "eval_steps_per_second": 5.933,
-    "perplexity": 53752.74020410498,
+    "eval_samples_per_second": 567.091,
+    "eval_steps_per_second": 5.92,
+    "perplexity": 50675.39096015328,
     "total_flos": 7.69437063436032e+17,
-    "train_loss": 8.341610258088004,
-    "train_runtime": 16911.6417,
+    "train_loss": 8.358330273888473,
+    "train_runtime": 17076.0076,
     "train_samples": 228639,
-    "train_samples_per_second": 337.991,
-    "train_steps_per_second": 3.521
+    "train_samples_per_second": 334.737,
+    "train_steps_per_second": 3.487
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15299203254788948,
-    "eval_loss": 10.892149925231934,
-    "eval_runtime": 0.8428,
+    "eval_accuracy": 0.1535288466971803,
+    "eval_loss": 10.833195686340332,
+    "eval_runtime": 0.8447,
     "eval_samples": 479,
-    "eval_samples_per_second": 568.36,
-    "eval_steps_per_second": 5.933,
-    "perplexity": 53752.74020410498
+    "eval_samples_per_second": 567.091,
+    "eval_steps_per_second": 5.92,
+    "perplexity": 50675.39096015328
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 7.69437063436032e+17,
-    "train_loss": 8.341610258088004,
-    "train_runtime": 16911.6417,
+    "train_loss": 8.358330273888473,
+    "train_runtime": 17076.0076,
     "train_samples": 228639,
-    "train_samples_per_second": 337.991,
-    "train_steps_per_second": 3.521
+    "train_samples_per_second": 334.737,
+    "train_steps_per_second": 3.487
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 10000,
-  "best_metric": 10.866716384887695,
+  "best_metric": 10.820178031921387,
   "best_model_checkpoint": "distilbert_rand_100_v2/checkpoint-10000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.623548746109009,
+      "grad_norm": 2.7810590267181396,
       "learning_rate": 5e-06,
-      "loss": 13.9349,
+      "loss": 13.9348,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 2.249107599258423,
+      "grad_norm": 2.1571836471557617,
       "learning_rate": 1e-05,
-      "loss": 12.2124,
+      "loss": 12.2131,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 2.007410764694214,
+      "grad_norm": 2.1669623851776123,
       "learning_rate": 1.5e-05,
-      "loss": 11.6792,
+      "loss": 11.6798,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 2.0401337146759033,
+      "grad_norm": 2.2685675621032715,
       "learning_rate": 2e-05,
-      "loss": 11.5311,
+      "loss": 11.5318,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 2.2878291606903076,
+      "grad_norm": 2.121546983718872,
       "learning_rate": 2.5e-05,
-      "loss": 11.3996,
+      "loss": 11.3986,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 2.2183690071105957,
+      "grad_norm": 2.088027238845825,
       "learning_rate": 3e-05,
-      "loss": 11.2858,
+      "loss": 11.2855,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 2.212686538696289,
+      "grad_norm": 2.490748882293701,
       "learning_rate": 3.5e-05,
-      "loss": 11.1939,
+      "loss": 11.1933,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 2.3838858604431152,
+      "grad_norm": 2.112895965576172,
       "learning_rate": 4e-05,
-      "loss": 11.1186,
+      "loss": 11.1191,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 2.437605857849121,
+      "grad_norm": 2.4714086055755615,
       "learning_rate": 4.5e-05,
-      "loss": 11.059,
+      "loss": 11.0607,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 2.260139226913452,
+      "grad_norm": 2.6922225952148438,
       "learning_rate": 5e-05,
-      "loss": 11.0079,
+      "loss": 11.0076,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 2.5615074634552,
+      "grad_norm": 2.3974456787109375,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 10.9692,
+      "loss": 10.9716,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 2.6356801986694336,
+      "grad_norm": 2.3665430545806885,
       "learning_rate": 6e-05,
-      "loss": 10.9397,
+      "loss": 10.9402,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 2.255794048309326,
+      "grad_norm": 2.40063738822937,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 10.9113,
+      "loss": 10.9131,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 2.3790087699890137,
+      "grad_norm": 2.4049057960510254,
       "learning_rate": 7e-05,
-      "loss": 10.8862,
+      "loss": 10.8886,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 2.4995689392089844,
+      "grad_norm": 2.5291497707366943,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 10.8125,
+      "loss": 10.8142,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 2.895751476287842,
+      "grad_norm": 2.7735931873321533,
       "learning_rate": 8e-05,
-      "loss": 10.7809,
+      "loss": 10.7816,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 2.476466655731201,
+      "grad_norm": 2.8998589515686035,
       "learning_rate": 8.5e-05,
-      "loss": 10.7733,
+      "loss": 10.7759,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 2.770347833633423,
+      "grad_norm": 2.9176485538482666,
       "learning_rate": 9e-05,
-      "loss": 10.759,
+      "loss": 10.7577,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 3.013545274734497,
+      "grad_norm": 2.7973697185516357,
       "learning_rate": 9.5e-05,
-      "loss": 10.748,
+      "loss": 10.7493,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 3.273892879486084,
+      "grad_norm": 3.1797733306884766,
       "learning_rate": 0.0001,
-      "loss": 10.5483,
+      "loss": 10.5487,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.15086674643145329,
-      "eval_loss": 10.866716384887695,
-      "eval_runtime": 0.8428,
-      "eval_samples_per_second": 568.365,
-      "eval_steps_per_second": 5.933,
+      "eval_accuracy": 0.15106152091043157,
+      "eval_loss": 10.820178031921387,
+      "eval_runtime": 0.8506,
+      "eval_samples_per_second": 563.115,
+      "eval_steps_per_second": 5.878,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 2.9765877723693848,
+      "grad_norm": 2.983929395675659,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 10.5578,
+      "loss": 10.5591,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 3.3972182273864746,
+      "grad_norm": 3.550022840499878,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 10.5622,
+      "loss": 10.5642,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 3.380134105682373,
+      "grad_norm": 3.246150493621826,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 10.5759,
+      "loss": 10.5806,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 3.4397242069244385,
+      "grad_norm": 3.4869985580444336,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 10.5203,
+      "loss": 10.5141,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 3.691770553588867,
+      "grad_norm": 3.589064836502075,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 10.2992,
+      "loss": 10.302,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 3.5637378692626953,
+      "grad_norm": 3.496105909347534,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 10.3358,
+      "loss": 10.3312,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 3.589146614074707,
+      "grad_norm": 3.550602674484253,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 10.3648,
+      "loss": 10.3664,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 3.59706711769104,
+      "grad_norm": 3.5406298637390137,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 10.3832,
+      "loss": 10.385,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 3.3952794075012207,
+      "grad_norm": 3.59983229637146,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 10.2463,
+      "loss": 10.248,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 3.605567693710327,
+      "grad_norm": 3.658291816711426,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 10.1037,
+      "loss": 10.1031,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 4.2224016189575195,
+      "grad_norm": 3.845203161239624,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 10.1442,
+      "loss": 10.1438,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 3.7331886291503906,
+      "grad_norm": 3.9203052520751953,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 10.1849,
+      "loss": 10.1775,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 3.8454816341400146,
+      "grad_norm": 3.997992992401123,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 10.207,
+      "loss": 10.21,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 4.344640731811523,
+      "grad_norm": 4.228085041046143,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 9.9534,
+      "loss": 9.9501,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 4.680375576019287,
+      "grad_norm": 4.66232442855835,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 9.8559,
+      "loss": 9.8469,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 4.534852504730225,
+      "grad_norm": 4.382004737854004,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 9.9008,
+      "loss": 9.8927,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 4.6581268310546875,
+      "grad_norm": 4.697808265686035,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 9.9419,
+      "loss": 9.9369,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 4.783538818359375,
+      "grad_norm": 4.500067710876465,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 9.9515,
+      "loss": 9.9465,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 5.095106601715088,
+      "grad_norm": 5.431141376495361,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 9.5185,
+      "loss": 9.5089,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 5.6867218017578125,
+      "grad_norm": 5.436167240142822,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 9.5051,
+      "loss": 9.5005,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15278206459216026,
-      "eval_loss": 11.361907005310059,
-      "eval_runtime": 0.8169,
-      "eval_samples_per_second": 586.381,
-      "eval_steps_per_second": 6.121,
+      "eval_accuracy": 0.15258565055136228,
+      "eval_loss": 11.374232292175293,
+      "eval_runtime": 0.8148,
+      "eval_samples_per_second": 587.842,
+      "eval_steps_per_second": 6.136,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 6.015316009521484,
+      "grad_norm": 5.705742359161377,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 9.5331,
+      "loss": 9.5201,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 5.788078784942627,
+      "grad_norm": 5.982388973236084,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 9.5873,
+      "loss": 9.5749,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 6.633572578430176,
+      "grad_norm": 6.552948474884033,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 9.5126,
+      "loss": 9.5082,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 7.035929203033447,
+      "grad_norm": 7.031944751739502,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 9.0144,
+      "loss": 9.0094,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 7.450264930725098,
+      "grad_norm": 7.557385444641113,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 9.0531,
+      "loss": 9.0536,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 6.961653232574463,
+      "grad_norm": 7.220298767089844,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 9.1108,
+      "loss": 9.1141,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 7.176884174346924,
+      "grad_norm": 7.19287109375,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 9.1259,
+      "loss": 9.1294,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 7.644462585449219,
+      "grad_norm": 7.745799541473389,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 8.9114,
+      "loss": 8.9289,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 7.759709358215332,
+      "grad_norm": 7.925483703613281,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 8.5453,
+      "loss": 8.557,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 7.842126369476318,
+      "grad_norm": 7.923360824584961,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 8.5853,
+      "loss": 8.6173,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 8.187042236328125,
+      "grad_norm": 8.141770362854004,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 8.6351,
+      "loss": 8.6523,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 8.118842124938965,
+      "grad_norm": 7.086618900299072,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 8.6694,
+      "loss": 8.7022,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 8.268385887145996,
+      "grad_norm": 8.559000015258789,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 8.3023,
+      "loss": 8.3396,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 9.762911796569824,
+      "grad_norm": 9.0249662399292,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 8.0963,
+      "loss": 8.1425,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 8.83743667602539,
+      "grad_norm": 8.788827896118164,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 8.1584,
+      "loss": 8.2002,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 9.637364387512207,
+      "grad_norm": 9.493494033813477,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 8.2039,
+      "loss": 8.2441,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 9.09965705871582,
+      "grad_norm": 9.14420223236084,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 8.2474,
+      "loss": 8.2831,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 8.384922981262207,
+      "grad_norm": 9.685812950134277,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 7.7417,
+      "loss": 7.7931,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 9.422770500183105,
+      "grad_norm": 10.053449630737305,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 7.6874,
+      "loss": 7.7363,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 10.507129669189453,
+      "grad_norm": 9.493916511535645,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 7.7561,
+      "loss": 7.8068,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.15415395808265944,
-      "eval_loss": 12.851670265197754,
-      "eval_runtime": 0.8151,
-      "eval_samples_per_second": 587.65,
-      "eval_steps_per_second": 6.134,
+      "eval_accuracy": 0.1529227411366371,
+      "eval_loss": 12.660033226013184,
+      "eval_runtime": 0.8184,
+      "eval_samples_per_second": 585.269,
+      "eval_steps_per_second": 6.109,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 9.831664085388184,
+      "grad_norm": 9.801485061645508,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 7.7899,
+      "loss": 7.8587,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 9.452991485595703,
+      "grad_norm": 9.46905517578125,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 7.7915,
+      "loss": 7.8407,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 9.460052490234375,
+      "grad_norm": 10.061419486999512,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 7.2869,
+      "loss": 7.3547,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 10.180335998535156,
+      "grad_norm": 10.121818542480469,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 7.3347,
+      "loss": 7.4044,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 10.93828010559082,
+      "grad_norm": 10.951651573181152,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 7.4064,
+      "loss": 7.4683,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 11.042454719543457,
+      "grad_norm": 10.483640670776367,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 7.4337,
+      "loss": 7.4945,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 10.585861206054688,
+      "grad_norm": 9.109230995178223,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 7.3246,
+      "loss": 7.3746,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 9.971748352050781,
+      "grad_norm": 9.702284812927246,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 7.0207,
+      "loss": 7.0751,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 10.239751815795898,
+      "grad_norm": 11.211170196533203,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 7.0628,
+      "loss": 7.1236,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 12.115407943725586,
+      "grad_norm": 10.092761993408203,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 7.11,
+      "loss": 7.1618,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 11.907709121704102,
+      "grad_norm": 11.896991729736328,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 7.1333,
+      "loss": 7.1961,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 9.875789642333984,
+      "grad_norm": 9.546497344970703,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.9586,
+      "loss": 7.0114,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 11.699231147766113,
+      "grad_norm": 11.717222213745117,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.8241,
+      "loss": 6.8654,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 10.097740173339844,
+      "grad_norm": 10.279557228088379,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.8637,
+      "loss": 6.8976,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 9.586366653442383,
+      "grad_norm": 10.96808910369873,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.894,
+      "loss": 6.9294,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 11.368609428405762,
+      "grad_norm": 11.235006332397461,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.9168,
+      "loss": 6.9638,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 9.660259246826172,
+      "grad_norm": 9.853087425231934,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.7177,
+      "loss": 6.7418,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 8.476889610290527,
+      "grad_norm": 8.750295639038086,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.698,
+      "loss": 6.7309,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 11.034769058227539,
+      "grad_norm": 10.288826942443848,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.6993,
+      "loss": 6.7339,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 9.719618797302246,
+      "grad_norm": 10.06374740600586,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.7204,
+      "loss": 6.765,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.1507794101155097,
-      "eval_loss": 14.135711669921875,
-      "eval_runtime": 0.8142,
-      "eval_samples_per_second": 588.291,
-      "eval_steps_per_second": 6.141,
+      "eval_accuracy": 0.1504149377593361,
+      "eval_loss": 14.03056526184082,
+      "eval_runtime": 0.8439,
+      "eval_samples_per_second": 567.613,
+      "eval_steps_per_second": 5.925,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 6.08411169052124,
+      "grad_norm": 8.312663078308105,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.7491,
+      "loss": 6.7817,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 7.530599117279053,
+      "grad_norm": 8.644062995910645,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.5746,
+      "loss": 6.5932,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 9.81454849243164,
+      "grad_norm": 10.031119346618652,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.5872,
+      "loss": 6.6121,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 8.780349731445312,
+      "grad_norm": 9.132280349731445,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.6003,
+      "loss": 6.6333,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 10.0014009475708,
+      "grad_norm": 9.706832885742188,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 6.6314,
+      "loss": 6.6536,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 7.868160724639893,
+      "grad_norm": 8.929744720458984,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.5975,
+      "loss": 6.6186,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 8.1547212600708,
+      "grad_norm": 9.22443962097168,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.5136,
+      "loss": 6.5293,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 8.351947784423828,
+      "grad_norm": 8.579099655151367,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.5266,
+      "loss": 6.5474,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 9.333196640014648,
+      "grad_norm": 9.799068450927734,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.5447,
+      "loss": 6.5603,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 7.543220520019531,
+      "grad_norm": 8.805617332458496,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.5615,
+      "loss": 6.5773,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 8.6489839553833,
+      "grad_norm": 8.691386222839355,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.4962,
+      "loss": 6.5167,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 8.978482246398926,
+      "grad_norm": 9.20458698272705,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.4631,
+      "loss": 6.4773,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 9.505804061889648,
+      "grad_norm": 9.205421447753906,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.4652,
+      "loss": 6.476,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 10.087528228759766,
+      "grad_norm": 8.747261047363281,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.4648,
+      "loss": 6.4856,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 8.372830390930176,
+      "grad_norm": 8.697514533996582,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.4747,
+      "loss": 6.4856,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 6.940515518188477,
+      "grad_norm": 6.696583271026611,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.4428,
+      "loss": 6.4534,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 9.426199913024902,
+      "grad_norm": 7.255544185638428,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.42,
+      "loss": 6.4353,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 7.125119209289551,
+      "grad_norm": 6.9102044105529785,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.4176,
+      "loss": 6.4322,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 8.605149269104004,
+      "grad_norm": 8.245722770690918,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.4178,
+      "loss": 6.4299,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 5.658854007720947,
+      "grad_norm": 7.834451198577881,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.4249,
+      "loss": 6.4433,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.15067148905150418,
-      "eval_loss": 14.746002197265625,
-      "eval_runtime": 0.8179,
-      "eval_samples_per_second": 585.658,
-      "eval_steps_per_second": 6.113,
+      "eval_accuracy": 0.15173689965514342,
+      "eval_loss": 14.556241035461426,
+      "eval_runtime": 0.847,
+      "eval_samples_per_second": 565.507,
+      "eval_steps_per_second": 5.903,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 9.741387367248535,
+      "grad_norm": 9.235966682434082,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.378,
+      "loss": 6.3917,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 7.579795837402344,
+      "grad_norm": 8.457624435424805,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.3897,
+      "loss": 6.3966,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 8.439435005187988,
+      "grad_norm": 6.855658531188965,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 6.374,
+      "loss": 6.3909,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 8.949285507202148,
+      "grad_norm": 6.519852638244629,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 6.3776,
+      "loss": 6.3931,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 4.752408981323242,
+      "grad_norm": 5.470546722412109,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 6.3705,
+      "loss": 6.3852,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 6.665569305419922,
+      "grad_norm": 6.591881275177002,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 6.3435,
+      "loss": 6.3505,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 8.843728065490723,
+      "grad_norm": 9.221796035766602,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 6.3398,
+      "loss": 6.3505,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 7.537617206573486,
+      "grad_norm": 8.409793853759766,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 6.3528,
+      "loss": 6.3681,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 7.035005569458008,
+      "grad_norm": 6.856619834899902,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 6.3397,
+      "loss": 6.3509,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 8.030730247497559,
+      "grad_norm": 6.267460823059082,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 6.3295,
+      "loss": 6.3424,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 6.538209438323975,
+      "grad_norm": 8.548900604248047,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 6.3122,
+      "loss": 6.3236,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 8.01657772064209,
+      "grad_norm": 6.857176303863525,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 6.3267,
+      "loss": 6.3391,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 7.038714408874512,
+      "grad_norm": 6.423684120178223,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 6.3101,
+      "loss": 6.317,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 9.071599960327148,
+      "grad_norm": 8.459613800048828,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 6.3173,
+      "loss": 6.3243,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 7.216913223266602,
+      "grad_norm": 8.112276077270508,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 6.3022,
+      "loss": 6.3075,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 8.274382591247559,
+      "grad_norm": 9.84228229522705,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 6.3014,
+      "loss": 6.3149,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 8.35374641418457,
+      "grad_norm": 6.50051212310791,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 6.3059,
+      "loss": 6.3166,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 6.7744927406311035,
+      "grad_norm": 6.993781566619873,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 6.2977,
+      "loss": 6.3091,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 7.471076488494873,
+      "grad_norm": 6.753544807434082,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 6.2865,
+      "loss": 6.2941,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 7.69437063436032e+17,
-      "train_loss": 8.341610258088004,
-      "train_runtime": 16911.6417,
-      "train_samples_per_second": 337.991,
-      "train_steps_per_second": 3.521
+      "train_loss": 8.358330273888473,
+      "train_runtime": 17076.0076,
+      "train_samples_per_second": 334.737,
+      "train_steps_per_second": 3.487
     }
   ],
   "logging_steps": 500,