End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-km-100_v2
 metrics:
 - accuracy
 model-index:
 - name: bert_base_km_100_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-km-100_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-km-100_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.1534440865683449
 ---
 
 <!-- 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_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-km-100_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 7.3908
-- Accuracy: 0.1511
+- Loss: 7.0530
+- Accuracy: 0.1534
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15330281968695258,
-    "eval_loss": 7.078766345977783,
-    "eval_runtime": 1.0374,
+    "eval_accuracy": 0.1534440865683449,
+    "eval_loss": 7.0529632568359375,
+    "eval_runtime": 1.0371,
     "eval_samples": 479,
-    "eval_samples_per_second": 461.719,
-    "eval_steps_per_second": 4.82,
-    "perplexity": 1186.5038799823253,
+    "eval_samples_per_second": 461.877,
+    "eval_steps_per_second": 4.821,
+    "perplexity": 1156.2800259006108,
     "total_flos": 1.5161929390646784e+18,
-    "train_loss": 6.7262844943439735,
-    "train_runtime": 24835.344,
+    "train_loss": 6.650803347186418,
+    "train_runtime": 24926.7454,
     "train_samples": 228639,
-    "train_samples_per_second": 230.155,
-    "train_steps_per_second": 2.398
+    "train_samples_per_second": 229.311,
+    "train_steps_per_second": 2.389
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15330281968695258,
-    "eval_loss": 7.078766345977783,
-    "eval_runtime": 1.0374,
+    "eval_accuracy": 0.1534440865683449,
+    "eval_loss": 7.0529632568359375,
+    "eval_runtime": 1.0371,
     "eval_samples": 479,
-    "eval_samples_per_second": 461.719,
-    "eval_steps_per_second": 4.82,
-    "perplexity": 1186.5038799823253
+    "eval_samples_per_second": 461.877,
+    "eval_steps_per_second": 4.821,
+    "perplexity": 1156.2800259006108
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 1.5161929390646784e+18,
-    "train_loss": 6.7262844943439735,
-    "train_runtime": 24835.344,
+    "train_loss": 6.650803347186418,
+    "train_runtime": 24926.7454,
     "train_samples": 228639,
-    "train_samples_per_second": 230.155,
-    "train_steps_per_second": 2.398
+    "train_samples_per_second": 229.311,
+    "train_steps_per_second": 2.389
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 20000,
-  "best_metric": 7.085968971252441,
+  "best_metric": 7.036741733551025,
   "best_model_checkpoint": "bert_base_km_100_v2/checkpoint-20000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 3.0025107860565186,
+      "grad_norm": 3.0325520038604736,
       "learning_rate": 5e-06,
-      "loss": 13.7908,
+      "loss": 13.7954,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 5.112063884735107,
+      "grad_norm": 2.8106133937835693,
       "learning_rate": 1e-05,
-      "loss": 11.9879,
+      "loss": 12.0341,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 16.70848274230957,
+      "grad_norm": 11.055712699890137,
       "learning_rate": 1.5e-05,
-      "loss": 10.5182,
+      "loss": 10.5831,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 17.31781578063965,
+      "grad_norm": 5.347803115844727,
       "learning_rate": 2e-05,
-      "loss": 9.5552,
+      "loss": 9.7271,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 10.2286958694458,
+      "grad_norm": 15.038311004638672,
       "learning_rate": 2.5e-05,
-      "loss": 8.8963,
+      "loss": 9.1027,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 26.42514419555664,
+      "grad_norm": 17.481904983520508,
       "learning_rate": 3e-05,
-      "loss": 8.4575,
+      "loss": 8.6071,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 18.870014190673828,
+      "grad_norm": 14.806625366210938,
       "learning_rate": 3.5e-05,
-      "loss": 8.225,
+      "loss": 8.2317,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 12.687928199768066,
+      "grad_norm": 15.759472846984863,
       "learning_rate": 4e-05,
-      "loss": 8.0818,
+      "loss": 8.0202,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 10.143560409545898,
+      "grad_norm": 7.746140480041504,
       "learning_rate": 4.5e-05,
-      "loss": 7.9677,
+      "loss": 7.8677,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 14.060294151306152,
+      "grad_norm": 9.09663200378418,
       "learning_rate": 5e-05,
-      "loss": 7.8224,
+      "loss": 7.699,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 11.197796821594238,
+      "grad_norm": 8.326930046081543,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 7.7363,
+      "loss": 7.5839,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 6.759899616241455,
+      "grad_norm": 4.7854905128479,
       "learning_rate": 6e-05,
-      "loss": 7.7026,
+      "loss": 7.5285,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 13.113651275634766,
+      "grad_norm": 5.352976322174072,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 7.6635,
+      "loss": 7.5069,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 9.93619441986084,
+      "grad_norm": 6.005526065826416,
       "learning_rate": 7e-05,
-      "loss": 7.6104,
+      "loss": 7.4692,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 11.52259635925293,
+      "grad_norm": 5.010321617126465,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 7.5022,
+      "loss": 7.3121,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 7.724097728729248,
+      "grad_norm": 5.591887950897217,
       "learning_rate": 8e-05,
-      "loss": 7.4652,
+      "loss": 7.2788,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 8.379256248474121,
+      "grad_norm": 4.3518385887146,
       "learning_rate": 8.5e-05,
-      "loss": 7.4749,
+      "loss": 7.2869,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 8.36221694946289,
+      "grad_norm": 4.468409061431885,
       "learning_rate": 9e-05,
-      "loss": 7.4383,
+      "loss": 7.2643,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 5.691290855407715,
+      "grad_norm": 4.664194107055664,
       "learning_rate": 9.5e-05,
-      "loss": 7.4046,
+      "loss": 7.2599,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 5.282100677490234,
+      "grad_norm": 4.949222564697266,
       "learning_rate": 0.0001,
-      "loss": 7.2949,
+      "loss": 7.0906,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.14900247641837558,
-      "eval_loss": 7.556832790374756,
-      "eval_runtime": 1.043,
-      "eval_samples_per_second": 459.244,
-      "eval_steps_per_second": 4.794,
+      "eval_accuracy": 0.14933637552519546,
+      "eval_loss": 7.23159122467041,
+      "eval_runtime": 1.0708,
+      "eval_samples_per_second": 447.345,
+      "eval_steps_per_second": 4.67,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 4.287475109100342,
+      "grad_norm": 4.300989627838135,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 7.304,
+      "loss": 7.1063,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 5.505510330200195,
+      "grad_norm": 4.733574390411377,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 7.2835,
+      "loss": 7.1007,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 8.971573829650879,
+      "grad_norm": 5.090231418609619,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 7.2731,
+      "loss": 7.0946,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 4.525511741638184,
+      "grad_norm": 3.6812405586242676,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 7.2014,
+      "loss": 7.0535,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 4.402318477630615,
+      "grad_norm": 3.7199642658233643,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 7.0982,
+      "loss": 6.9155,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 4.837042808532715,
+      "grad_norm": 3.8014750480651855,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 7.1127,
+      "loss": 6.9244,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 4.827608585357666,
+      "grad_norm": 4.38668155670166,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 7.1018,
+      "loss": 6.9336,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 5.197642803192139,
+      "grad_norm": 4.114847183227539,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 7.0843,
+      "loss": 6.9379,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 6.849023818969727,
+      "grad_norm": 3.991894006729126,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 7.0147,
+      "loss": 6.8481,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 4.558162689208984,
+      "grad_norm": 3.796557903289795,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 6.96,
+      "loss": 6.7738,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 5.003061771392822,
+      "grad_norm": 4.997390270233154,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 6.9587,
+      "loss": 6.7876,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 6.175533771514893,
+      "grad_norm": 3.105367422103882,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 6.9574,
+      "loss": 6.8055,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 3.8840558528900146,
+      "grad_norm": 3.6604092121124268,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 6.9631,
+      "loss": 6.811,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 5.194244861602783,
+      "grad_norm": 4.096405029296875,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 6.8594,
+      "loss": 6.6924,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 4.484853267669678,
+      "grad_norm": 3.752100706100464,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 6.8151,
+      "loss": 6.6527,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 4.555990219116211,
+      "grad_norm": 3.462996244430542,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 6.8386,
+      "loss": 6.6678,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 5.909701347351074,
+      "grad_norm": 3.705519437789917,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 6.8453,
+      "loss": 6.6887,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 4.352225303649902,
+      "grad_norm": 4.445788860321045,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 6.8414,
+      "loss": 6.6944,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 3.9375240802764893,
+      "grad_norm": 3.868389368057251,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 6.7079,
+      "loss": 6.5522,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 3.901026487350464,
+      "grad_norm": 3.729473829269409,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 6.7147,
+      "loss": 6.5625,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15179999438817027,
-      "eval_loss": 7.085968971252441,
-      "eval_runtime": 1.0227,
-      "eval_samples_per_second": 468.386,
-      "eval_steps_per_second": 4.889,
+      "eval_accuracy": 0.1526137097000477,
+      "eval_loss": 7.036741733551025,
+      "eval_runtime": 1.0589,
+      "eval_samples_per_second": 452.355,
+      "eval_steps_per_second": 4.722,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 6.245331287384033,
+      "grad_norm": 3.6588330268859863,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 6.7238,
+      "loss": 6.5614,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 4.60683536529541,
+      "grad_norm": 4.638293743133545,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 6.7341,
+      "loss": 6.5764,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 3.868889808654785,
+      "grad_norm": 4.317777633666992,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 6.7211,
+      "loss": 6.5686,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 4.567445278167725,
+      "grad_norm": 4.681490898132324,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 6.5951,
+      "loss": 6.4478,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 3.638380527496338,
+      "grad_norm": 4.315625190734863,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 6.6143,
+      "loss": 6.4594,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 5.055649280548096,
+      "grad_norm": 3.847973346710205,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 6.631,
+      "loss": 6.4783,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 6.595772743225098,
+      "grad_norm": 6.255940914154053,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 6.6273,
+      "loss": 6.4777,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 6.4645843505859375,
+      "grad_norm": 3.730644941329956,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 6.5857,
+      "loss": 6.4379,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 7.419833183288574,
+      "grad_norm": 3.6163721084594727,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 6.4984,
+      "loss": 6.3757,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 5.692862510681152,
+      "grad_norm": 6.299481391906738,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 6.52,
+      "loss": 6.3724,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 7.923045635223389,
+      "grad_norm": 4.287810802459717,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 6.5261,
+      "loss": 6.3921,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 4.724234580993652,
+      "grad_norm": 4.100649356842041,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 6.5381,
+      "loss": 6.402,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 5.390316486358643,
+      "grad_norm": 4.9406819343566895,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 6.459,
+      "loss": 6.3409,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 4.8012590408325195,
+      "grad_norm": 3.7887113094329834,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 6.4199,
+      "loss": 6.3122,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 4.7395405769348145,
+      "grad_norm": 2.6583824157714844,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 6.4265,
+      "loss": 6.3141,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 4.3479509353637695,
+      "grad_norm": 4.447488784790039,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 6.4346,
+      "loss": 6.3233,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 4.003292560577393,
+      "grad_norm": 3.666630983352661,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 6.4396,
+      "loss": 6.3288,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 4.667975902557373,
+      "grad_norm": 5.452030658721924,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 6.3512,
+      "loss": 6.2633,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 4.609392166137695,
+      "grad_norm": 4.218036651611328,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 6.3368,
+      "loss": 6.2574,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 5.208188056945801,
+      "grad_norm": 2.8640193939208984,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 6.355,
+      "loss": 6.2634,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.15384615384615385,
-      "eval_loss": 7.2959394454956055,
-      "eval_runtime": 1.0205,
-      "eval_samples_per_second": 469.376,
-      "eval_steps_per_second": 4.9,
+      "eval_accuracy": 0.15451772672580238,
+      "eval_loss": 7.229600429534912,
+      "eval_runtime": 1.0176,
+      "eval_samples_per_second": 470.714,
+      "eval_steps_per_second": 4.914,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 5.530768394470215,
+      "grad_norm": 5.212186813354492,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 6.3615,
+      "loss": 6.2586,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 4.560688495635986,
+      "grad_norm": 3.6361076831817627,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 6.3601,
+      "loss": 6.2629,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 5.984344959259033,
+      "grad_norm": 5.413948059082031,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 6.2667,
+      "loss": 6.206,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 4.298348903656006,
+      "grad_norm": 4.079280853271484,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 6.2842,
+      "loss": 6.213,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 4.859953880310059,
+      "grad_norm": 4.215555667877197,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 6.2894,
+      "loss": 6.2147,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 4.065077304840088,
+      "grad_norm": 3.849573850631714,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 6.2944,
+      "loss": 6.2189,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 4.847146034240723,
+      "grad_norm": 4.60481071472168,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 6.2613,
+      "loss": 6.1962,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 5.9763994216918945,
+      "grad_norm": 6.386091232299805,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 6.2229,
+      "loss": 6.1677,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 5.000112533569336,
+      "grad_norm": 4.499414443969727,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 6.224,
+      "loss": 6.1705,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 5.4157915115356445,
+      "grad_norm": 3.7441251277923584,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 6.2344,
+      "loss": 6.1722,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 5.365735054016113,
+      "grad_norm": 3.144296646118164,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 6.2383,
+      "loss": 6.1723,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 4.138433933258057,
+      "grad_norm": 3.7621850967407227,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.201,
+      "loss": 6.1493,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 5.998276233673096,
+      "grad_norm": 3.3722171783447266,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.1842,
+      "loss": 6.1382,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 5.508979320526123,
+      "grad_norm": 4.0898237228393555,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.1843,
+      "loss": 6.1344,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 4.854427337646484,
+      "grad_norm": 2.528909206390381,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.185,
+      "loss": 6.1404,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 6.459695339202881,
+      "grad_norm": 3.6384353637695312,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.1853,
+      "loss": 6.1333,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 5.194928169250488,
+      "grad_norm": 4.224349021911621,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.1512,
+      "loss": 6.1154,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 4.583359241485596,
+      "grad_norm": 2.7164251804351807,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.1383,
+      "loss": 6.1076,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 3.236578941345215,
+      "grad_norm": 3.8217499256134033,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.1455,
+      "loss": 6.109,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 3.4994120597839355,
+      "grad_norm": 4.337832927703857,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.1362,
+      "loss": 6.0994,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.15254569922619715,
-      "eval_loss": 7.29498291015625,
-      "eval_runtime": 1.0185,
-      "eval_samples_per_second": 470.295,
-      "eval_steps_per_second": 4.909,
+      "eval_accuracy": 0.15229337221038466,
+      "eval_loss": 7.491790294647217,
+      "eval_runtime": 1.0187,
+      "eval_samples_per_second": 470.228,
+      "eval_steps_per_second": 4.908,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 5.33200216293335,
+      "grad_norm": 3.712265729904175,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.1483,
+      "loss": 6.1037,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 3.6958775520324707,
+      "grad_norm": 3.09163761138916,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.1059,
+      "loss": 6.0796,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 5.909574508666992,
+      "grad_norm": 3.670301675796509,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.1038,
+      "loss": 6.0753,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 4.738137245178223,
+      "grad_norm": 3.4407198429107666,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.1097,
+      "loss": 6.0812,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 5.095664024353027,
+      "grad_norm": 2.8078510761260986,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 6.1102,
+      "loss": 6.0788,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 4.215599536895752,
+      "grad_norm": 2.582552194595337,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.0973,
+      "loss": 6.0706,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 4.486133575439453,
+      "grad_norm": 3.352553606033325,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.0722,
+      "loss": 6.0535,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 7.16460657119751,
+      "grad_norm": 1.4775241613388062,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.0772,
+      "loss": 6.0499,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 4.880932331085205,
+      "grad_norm": 2.1691818237304688,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.0752,
+      "loss": 6.0492,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 3.0645861625671387,
+      "grad_norm": 3.906991720199585,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.0766,
+      "loss": 6.0534,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 3.3358895778656006,
+      "grad_norm": 4.238534927368164,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.0637,
+      "loss": 6.0443,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 2.9742538928985596,
+      "grad_norm": 3.73866868019104,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.0414,
+      "loss": 6.0267,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 4.6407246589660645,
+      "grad_norm": 4.061879634857178,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.0482,
+      "loss": 6.029,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 4.145533561706543,
+      "grad_norm": 1.6275923252105713,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.0502,
+      "loss": 6.0312,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 2.4257829189300537,
+      "grad_norm": 2.3194937705993652,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.0411,
+      "loss": 6.024,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 2.911485195159912,
+      "grad_norm": 2.4898221492767334,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.0322,
+      "loss": 6.0184,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 4.442984104156494,
+      "grad_norm": 1.5689467191696167,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.0257,
+      "loss": 6.0113,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 2.145681381225586,
+      "grad_norm": 2.0924572944641113,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.0252,
+      "loss": 6.0047,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 3.470374584197998,
+      "grad_norm": 3.0981178283691406,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.0198,
+      "loss": 6.0064,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 4.202281475067139,
+      "grad_norm": 4.074100494384766,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.019,
+      "loss": 6.0062,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.1516527882916982,
-      "eval_loss": 7.419981479644775,
-      "eval_runtime": 1.0224,
-      "eval_samples_per_second": 468.524,
-      "eval_steps_per_second": 4.891,
+      "eval_accuracy": 0.1510920458687302,
+      "eval_loss": 7.390818119049072,
+      "eval_runtime": 1.0237,
+      "eval_samples_per_second": 467.917,
+      "eval_steps_per_second": 4.884,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 3.843820810317993,
+      "grad_norm": 2.681124687194824,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.0042,
+      "loss": 5.9937,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 3.3001763820648193,
+      "grad_norm": 2.0725343227386475,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.0042,
+      "loss": 5.9944,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 2.0066754817962646,
+      "grad_norm": 5.385054111480713,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 5.9985,
+      "loss": 5.9913,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 5.390365123748779,
+      "grad_norm": 1.695888876914978,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 5.999,
+      "loss": 5.9879,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 3.7821195125579834,
+      "grad_norm": 2.4581727981567383,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 5.9948,
+      "loss": 5.9876,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 2.3854730129241943,
+      "grad_norm": 0.8789335489273071,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 5.9828,
+      "loss": 5.9758,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 4.705515384674072,
+      "grad_norm": 6.385268211364746,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 5.9815,
+      "loss": 5.9773,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 5.391833305358887,
+      "grad_norm": 3.9702112674713135,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 5.9886,
+      "loss": 5.9772,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 4.696737289428711,
+      "grad_norm": 3.841106653213501,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 5.9775,
+      "loss": 5.9728,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 1.0100961923599243,
+      "grad_norm": 2.1624410152435303,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 5.9738,
+      "loss": 5.9677,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 3.317594528198242,
+      "grad_norm": 1.8635454177856445,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 5.9696,
+      "loss": 5.9647,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 3.988764524459839,
+      "grad_norm": 1.5375641584396362,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 5.9699,
+      "loss": 5.9643,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 1.8712412118911743,
+      "grad_norm": 2.0004587173461914,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 5.9626,
+      "loss": 5.9589,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 2.5608625411987305,
+      "grad_norm": 4.578677654266357,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 5.9643,
+      "loss": 5.9608,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 3.8009350299835205,
+      "grad_norm": 2.199143171310425,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 5.9597,
+      "loss": 5.9534,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 2.031345844268799,
+      "grad_norm": 0.8375799655914307,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 5.9567,
+      "loss": 5.9541,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 3.481703042984009,
+      "grad_norm": 3.7426412105560303,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 5.9533,
+      "loss": 5.9519,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 1.0147701501846313,
+      "grad_norm": 0.7772536277770996,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 5.9533,
+      "loss": 5.9495,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 2.468587636947632,
+      "grad_norm": 0.813712477684021,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 5.9517,
+      "loss": 5.9498,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 1.5161929390646784e+18,
-      "train_loss": 6.7262844943439735,
-      "train_runtime": 24835.344,
-      "train_samples_per_second": 230.155,
-      "train_steps_per_second": 2.398
+      "train_loss": 6.650803347186418,
+      "train_runtime": 24926.7454,
+      "train_samples_per_second": 229.311,
+      "train_steps_per_second": 2.389
     }
   ],
   "logging_steps": 500,