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
 metrics:
 - accuracy
 model-index:
 - name: bert_base_rand_100_v1
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-rand-100
+      type: Hartunka/processed_wikitext-103-raw-v1-rand-100
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.15180539074419394
 ---
 
 <!-- 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_rand_100_v1
 
-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 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 14.2039
-- Accuracy: 0.1513
+- Loss: 10.7840
+- Accuracy: 0.1518
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15160761711024467,
-    "eval_loss": 10.7857027053833,
-    "eval_runtime": 1.4353,
+    "eval_accuracy": 0.15180539074419394,
+    "eval_loss": 10.783976554870605,
+    "eval_runtime": 1.0227,
     "eval_samples": 479,
-    "eval_samples_per_second": 333.721,
-    "eval_steps_per_second": 3.484,
-    "perplexity": 48324.92274321138,
+    "eval_samples_per_second": 468.363,
+    "eval_steps_per_second": 4.889,
+    "perplexity": 48241.57860599912,
     "total_flos": 1.5161929390646784e+18,
-    "train_loss": 8.29517004568771,
-    "train_runtime": 25269.2882,
+    "train_loss": 8.295239665226012,
+    "train_runtime": 24715.4683,
     "train_samples": 228639,
-    "train_samples_per_second": 226.202,
-    "train_steps_per_second": 2.357
+    "train_samples_per_second": 231.271,
+    "train_steps_per_second": 2.409
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15160761711024467,
-    "eval_loss": 10.7857027053833,
-    "eval_runtime": 1.4353,
+    "eval_accuracy": 0.15180539074419394,
+    "eval_loss": 10.783976554870605,
+    "eval_runtime": 1.0227,
     "eval_samples": 479,
-    "eval_samples_per_second": 333.721,
-    "eval_steps_per_second": 3.484,
-    "perplexity": 48324.92274321138
+    "eval_samples_per_second": 468.363,
+    "eval_steps_per_second": 4.889,
+    "perplexity": 48241.57860599912
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 1.5161929390646784e+18,
-    "train_loss": 8.29517004568771,
-    "train_runtime": 25269.2882,
+    "train_loss": 8.295239665226012,
+    "train_runtime": 24715.4683,
     "train_samples": 228639,
-    "train_samples_per_second": 226.202,
-    "train_steps_per_second": 2.357
+    "train_samples_per_second": 231.271,
+    "train_steps_per_second": 2.409
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 10000,
-  "best_metric": 10.769007682800293,
+  "best_metric": 10.76736068725586,
   "best_model_checkpoint": "bert_base_rand_100_v1/checkpoint-10000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.9042837619781494,
+      "grad_norm": 2.9043400287628174,
       "learning_rate": 5e-06,
       "loss": 13.8236,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 2.3711719512939453,
+      "grad_norm": 2.37027645111084,
       "learning_rate": 1e-05,
       "loss": 12.1691,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 2.1205294132232666,
+      "grad_norm": 2.120368480682373,
       "learning_rate": 1.5e-05,
       "loss": 11.6234,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 2.0399091243743896,
+      "grad_norm": 2.042137384414673,
       "learning_rate": 2e-05,
       "loss": 11.4719,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 2.3900954723358154,
+      "grad_norm": 2.3900394439697266,
       "learning_rate": 2.5e-05,
-      "loss": 11.3513,
+      "loss": 11.3512,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 2.2257678508758545,
+      "grad_norm": 2.2257821559906006,
       "learning_rate": 3e-05,
       "loss": 11.2406,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 2.260497570037842,
+      "grad_norm": 2.261098623275757,
       "learning_rate": 3.5e-05,
-      "loss": 11.1519,
+      "loss": 11.1518,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 2.370394468307495,
+      "grad_norm": 2.370227575302124,
       "learning_rate": 4e-05,
       "loss": 11.0846,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 2.4347972869873047,
+      "grad_norm": 2.435065269470215,
       "learning_rate": 4.5e-05,
       "loss": 11.0269,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 2.187213182449341,
+      "grad_norm": 2.1949784755706787,
       "learning_rate": 5e-05,
-      "loss": 10.9778,
+      "loss": 10.9777,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 2.5867912769317627,
+      "grad_norm": 2.5839924812316895,
       "learning_rate": 5.500000000000001e-05,
       "loss": 10.9436,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 2.634326696395874,
+      "grad_norm": 2.633960723876953,
       "learning_rate": 6e-05,
-      "loss": 10.9085,
+      "loss": 10.9083,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 2.207746982574463,
+      "grad_norm": 2.2050702571868896,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 10.8772,
+      "loss": 10.8769,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 2.3962109088897705,
+      "grad_norm": 2.393577814102173,
       "learning_rate": 7e-05,
-      "loss": 10.8501,
+      "loss": 10.85,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 2.4583606719970703,
+      "grad_norm": 2.4578983783721924,
       "learning_rate": 7.500000000000001e-05,
       "loss": 10.8094,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 2.708967924118042,
+      "grad_norm": 2.7051303386688232,
       "learning_rate": 8e-05,
       "loss": 10.7873,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 2.3622376918792725,
+      "grad_norm": 2.360792636871338,
       "learning_rate": 8.5e-05,
       "loss": 10.7784,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 2.6710660457611084,
+      "grad_norm": 2.6693968772888184,
       "learning_rate": 9e-05,
       "loss": 10.7604,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 2.704530715942383,
+      "grad_norm": 2.7021095752716064,
       "learning_rate": 9.5e-05,
-      "loss": 10.7486,
+      "loss": 10.7487,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 2.9432191848754883,
+      "grad_norm": 2.9364306926727295,
       "learning_rate": 0.0001,
-      "loss": 10.61,
+      "loss": 10.6101,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.14953115000417375,
-      "eval_loss": 10.769007682800293,
-      "eval_runtime": 1.0902,
-      "eval_samples_per_second": 439.361,
-      "eval_steps_per_second": 4.586,
+      "eval_accuracy": 0.14944767522746877,
+      "eval_loss": 10.76736068725586,
+      "eval_runtime": 1.0434,
+      "eval_samples_per_second": 459.067,
+      "eval_steps_per_second": 4.792,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 2.6833932399749756,
+      "grad_norm": 2.681375503540039,
       "learning_rate": 9.899091826437942e-05,
       "loss": 10.615,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 3.233334541320801,
+      "grad_norm": 3.2120511531829834,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 10.6127,
+      "loss": 10.6131,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 3.134445905685425,
+      "grad_norm": 3.1292941570281982,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 10.6178,
+      "loss": 10.6177,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 3.284961223602295,
+      "grad_norm": 3.279639482498169,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 10.5624,
+      "loss": 10.5628,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 3.565044641494751,
+      "grad_norm": 3.5632076263427734,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 10.355,
+      "loss": 10.356,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 3.2186875343322754,
+      "grad_norm": 3.187746047973633,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 10.3834,
+      "loss": 10.3828,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 3.482707977294922,
+      "grad_norm": 3.483473539352417,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 10.4053,
+      "loss": 10.4051,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 3.493900775909424,
+      "grad_norm": 3.5169174671173096,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 10.4204,
+      "loss": 10.4208,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 3.523930549621582,
+      "grad_norm": 3.506720781326294,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 10.2768,
+      "loss": 10.2771,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 3.6932802200317383,
+      "grad_norm": 3.7083451747894287,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 10.11,
+      "loss": 10.1113,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 4.2440032958984375,
+      "grad_norm": 4.175162315368652,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 10.1532,
+      "loss": 10.1539,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 4.022274494171143,
+      "grad_norm": 3.9958231449127197,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 10.1878,
+      "loss": 10.1875,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 3.917412281036377,
+      "grad_norm": 3.884005069732666,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 10.2067,
+      "loss": 10.2073,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 4.309359073638916,
+      "grad_norm": 4.298327922821045,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 9.9113,
+      "loss": 9.9116,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 4.902447700500488,
+      "grad_norm": 4.895790100097656,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 9.7779,
+      "loss": 9.7798,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 4.952358245849609,
+      "grad_norm": 4.92780065536499,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 9.8234,
+      "loss": 9.8235,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 5.0325422286987305,
+      "grad_norm": 4.988821506500244,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 9.8649,
+      "loss": 9.8654,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 5.1185078620910645,
+      "grad_norm": 5.0547990798950195,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 9.8866,
+      "loss": 9.888,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 5.815095901489258,
+      "grad_norm": 5.894690990447998,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 9.351,
+      "loss": 9.3523,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 5.886953830718994,
+      "grad_norm": 5.780293941497803,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 9.3352,
+      "loss": 9.3327,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15205252672633912,
-      "eval_loss": 11.293512344360352,
-      "eval_runtime": 1.159,
-      "eval_samples_per_second": 413.29,
-      "eval_steps_per_second": 4.314,
+      "eval_accuracy": 0.15278206459216026,
+      "eval_loss": 11.305651664733887,
+      "eval_runtime": 1.0273,
+      "eval_samples_per_second": 466.258,
+      "eval_steps_per_second": 4.867,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 6.195263385772705,
+      "grad_norm": 5.931501865386963,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 9.3556,
+      "loss": 9.3552,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 6.156721115112305,
+      "grad_norm": 6.061377048492432,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 9.4093,
+      "loss": 9.4139,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 6.612517833709717,
+      "grad_norm": 6.685647010803223,
       "learning_rate": 7.679112008072655e-05,
       "loss": 9.345,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 6.567813873291016,
+      "grad_norm": 6.919853210449219,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 8.7891,
+      "loss": 8.7858,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 7.372893333435059,
+      "grad_norm": 7.349966049194336,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 8.8319,
+      "loss": 8.8277,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 7.385047912597656,
+      "grad_norm": 7.589054107666016,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 8.894,
+      "loss": 8.8889,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 7.407289981842041,
+      "grad_norm": 7.083017349243164,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 8.9068,
+      "loss": 8.9119,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 7.530327796936035,
+      "grad_norm": 7.565244197845459,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 8.7138,
+      "loss": 8.7134,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 7.097758769989014,
+      "grad_norm": 6.967763900756836,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 8.3451,
+      "loss": 8.3524,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 6.849892616271973,
+      "grad_norm": 6.766768455505371,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 8.3961,
+      "loss": 8.3869,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 8.5947847366333,
+      "grad_norm": 8.808615684509277,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 8.4496,
+      "loss": 8.4474,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 8.466645240783691,
+      "grad_norm": 8.201273918151855,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 8.4811,
+      "loss": 8.4821,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 7.488640308380127,
+      "grad_norm": 6.808009147644043,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 8.1578,
+      "loss": 8.1586,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 9.680927276611328,
+      "grad_norm": 10.347038269042969,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 7.9774,
+      "loss": 7.978,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 8.196619987487793,
+      "grad_norm": 7.1785407066345215,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 8.0335,
+      "loss": 8.0327,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 9.016983032226562,
+      "grad_norm": 9.36381721496582,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 8.063,
+      "loss": 8.0651,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 8.409159660339355,
+      "grad_norm": 8.557469367980957,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 8.1125,
+      "loss": 8.1101,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 7.864636421203613,
+      "grad_norm": 7.42149019241333,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 7.6838,
+      "loss": 7.6789,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 8.791885375976562,
+      "grad_norm": 8.780143737792969,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 7.6226,
+      "loss": 7.6264,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 7.979760646820068,
+      "grad_norm": 7.676377296447754,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 7.6969,
+      "loss": 7.6978,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.15289475893331841,
-      "eval_loss": 12.737491607666016,
-      "eval_runtime": 1.0463,
-      "eval_samples_per_second": 457.821,
-      "eval_steps_per_second": 4.779,
+      "eval_accuracy": 0.15334247418641744,
+      "eval_loss": 12.596437454223633,
+      "eval_runtime": 1.0175,
+      "eval_samples_per_second": 470.742,
+      "eval_steps_per_second": 4.914,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 8.597046852111816,
+      "grad_norm": 8.385842323303223,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 7.7242,
+      "loss": 7.731,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 10.116854667663574,
+      "grad_norm": 10.642855644226074,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 7.7177,
+      "loss": 7.7093,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 9.874378204345703,
+      "grad_norm": 8.902974128723145,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 7.2822,
+      "loss": 7.2842,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 9.96634292602539,
+      "grad_norm": 8.315555572509766,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 7.3366,
+      "loss": 7.3393,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 7.915868759155273,
+      "grad_norm": 7.8403239250183105,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 7.3993,
+      "loss": 7.3964,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 9.921277046203613,
+      "grad_norm": 9.209378242492676,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 7.4048,
+      "loss": 7.3977,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 8.88127326965332,
+      "grad_norm": 8.331807136535645,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 7.3091,
+      "loss": 7.3057,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 9.53819465637207,
+      "grad_norm": 10.185376167297363,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 7.0438,
+      "loss": 7.0405,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 8.33125114440918,
+      "grad_norm": 8.794229507446289,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 7.0943,
+      "loss": 7.0942,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 10.220344543457031,
+      "grad_norm": 10.817024230957031,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 7.1146,
+      "loss": 7.1171,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 8.603779792785645,
+      "grad_norm": 8.237445831298828,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 7.1284,
+      "loss": 7.132,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 9.361529350280762,
+      "grad_norm": 10.144477844238281,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.9751,
+      "loss": 6.9735,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 9.789393424987793,
+      "grad_norm": 10.091374397277832,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.8353,
+      "loss": 6.8384,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 10.226776123046875,
+      "grad_norm": 8.900592803955078,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.8657,
+      "loss": 6.8597,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 8.201433181762695,
+      "grad_norm": 9.65026569366455,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.9006,
+      "loss": 6.9007,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 12.848469734191895,
+      "grad_norm": 10.96306037902832,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.9246,
+      "loss": 6.9206,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 7.868854999542236,
+      "grad_norm": 8.602423667907715,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.7071,
+      "loss": 6.7132,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 8.136842727661133,
+      "grad_norm": 11.067571640014648,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.6972,
+      "loss": 6.6932,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 7.379518032073975,
+      "grad_norm": 8.451064109802246,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.6954,
+      "loss": 6.6978,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 9.728989601135254,
+      "grad_norm": 9.018234252929688,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.7114,
+      "loss": 6.711,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.1509195917909611,
-      "eval_loss": 13.75811767578125,
-      "eval_runtime": 1.0504,
-      "eval_samples_per_second": 456.008,
-      "eval_steps_per_second": 4.76,
+      "eval_accuracy": 0.15125602781204442,
+      "eval_loss": 13.537100791931152,
+      "eval_runtime": 1.0787,
+      "eval_samples_per_second": 444.034,
+      "eval_steps_per_second": 4.635,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 6.4882378578186035,
+      "grad_norm": 8.95382022857666,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.7356,
+      "loss": 6.7369,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 7.2141008377075195,
+      "grad_norm": 9.570463180541992,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.5515,
+      "loss": 6.5492,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 9.768448829650879,
+      "grad_norm": 11.023043632507324,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.5638,
+      "loss": 6.5668,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 9.917952537536621,
+      "grad_norm": 6.771119117736816,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.5735,
+      "loss": 6.5749,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 9.767972946166992,
+      "grad_norm": 6.928664207458496,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 6.5996,
+      "loss": 6.5968,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 7.926119804382324,
+      "grad_norm": 6.557650089263916,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.5648,
+      "loss": 6.5665,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 6.734870910644531,
+      "grad_norm": 9.319487571716309,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.4787,
+      "loss": 6.4821,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 8.4148530960083,
+      "grad_norm": 6.7069807052612305,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.4963,
+      "loss": 6.495,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 8.035300254821777,
+      "grad_norm": 8.839655876159668,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.5038,
+      "loss": 6.5034,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 8.386587142944336,
+      "grad_norm": 8.352717399597168,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.5292,
+      "loss": 6.5231,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 8.425530433654785,
+      "grad_norm": 10.152676582336426,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.4663,
+      "loss": 6.4684,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 7.744596004486084,
+      "grad_norm": 7.716465950012207,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.4273,
+      "loss": 6.4346,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 8.2754545211792,
+      "grad_norm": 7.97578239440918,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.4312,
+      "loss": 6.4292,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 8.379171371459961,
+      "grad_norm": 9.827136993408203,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.4364,
+      "loss": 6.4386,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 9.298855781555176,
+      "grad_norm": 8.448431968688965,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.4337,
+      "loss": 6.4378,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 6.819767951965332,
+      "grad_norm": 4.799870014190674,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.4097,
+      "loss": 6.4102,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 11.679108619689941,
+      "grad_norm": 10.621322631835938,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.3829,
+      "loss": 6.3849,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 9.999126434326172,
+      "grad_norm": 7.9846320152282715,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.3806,
+      "loss": 6.3831,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 8.14251708984375,
+      "grad_norm": 9.265085220336914,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.3757,
+      "loss": 6.3771,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 5.68617582321167,
+      "grad_norm": 8.088281631469727,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.3913,
+      "loss": 6.3914,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.15221353071466623,
-      "eval_loss": 14.300688743591309,
-      "eval_runtime": 1.0393,
-      "eval_samples_per_second": 460.887,
-      "eval_steps_per_second": 4.811,
+      "eval_accuracy": 0.1513163428379174,
+      "eval_loss": 14.20391845703125,
+      "eval_runtime": 1.0199,
+      "eval_samples_per_second": 469.674,
+      "eval_steps_per_second": 4.903,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 10.17446231842041,
+      "grad_norm": 7.882266521453857,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.341,
+      "loss": 6.3395,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 5.787772178649902,
+      "grad_norm": 6.02915096282959,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.3499,
+      "loss": 6.3551,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 8.405139923095703,
+      "grad_norm": 6.716942310333252,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 6.3451,
+      "loss": 6.3396,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 7.770719051361084,
+      "grad_norm": 6.620936393737793,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 6.3394,
+      "loss": 6.3377,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 5.889435768127441,
+      "grad_norm": 7.303170204162598,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 6.3379,
+      "loss": 6.3348,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 9.526209831237793,
+      "grad_norm": 5.3609113693237305,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 6.3086,
+      "loss": 6.3096,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 8.701118469238281,
+      "grad_norm": 8.874212265014648,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 6.2999,
+      "loss": 6.3005,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 6.540258407592773,
+      "grad_norm": 7.095221042633057,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 6.3145,
+      "loss": 6.3148,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 5.922007083892822,
+      "grad_norm": 6.500632286071777,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 6.3016,
+      "loss": 6.2985,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 5.774903297424316,
+      "grad_norm": 8.113584518432617,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 6.2921,
+      "loss": 6.2984,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 7.3452534675598145,
+      "grad_norm": 7.188576698303223,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 6.272,
+      "loss": 6.2736,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 7.179046630859375,
+      "grad_norm": 5.888152599334717,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 6.2902,
+      "loss": 6.2891,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 8.841833114624023,
+      "grad_norm": 6.741386890411377,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 6.2644,
+      "loss": 6.2661,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 6.666844844818115,
+      "grad_norm": 6.2131428718566895,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 6.2729,
+      "loss": 6.2724,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 5.185845851898193,
+      "grad_norm": 6.424174785614014,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 6.264,
+      "loss": 6.2612,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 7.542902946472168,
+      "grad_norm": 8.169211387634277,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 6.2608,
+      "loss": 6.264,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 6.536818027496338,
+      "grad_norm": 6.250510215759277,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 6.2605,
+      "loss": 6.2625,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 5.594988822937012,
+      "grad_norm": 6.013635158538818,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 6.2513,
+      "loss": 6.2497,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 6.380642890930176,
+      "grad_norm": 6.833993434906006,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 6.2466,
+      "loss": 6.2482,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 1.5161929390646784e+18,
-      "train_loss": 8.29517004568771,
-      "train_runtime": 25269.2882,
-      "train_samples_per_second": 226.202,
-      "train_steps_per_second": 2.357
+      "train_loss": 8.295239665226012,
+      "train_runtime": 24715.4683,
+      "train_samples_per_second": 231.271,
+      "train_steps_per_second": 2.409
     }
   ],
   "logging_steps": 500,