End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-rand-10_v2
 metrics:
 - accuracy
 model-index:
 - name: bert_base_rand_10_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-rand-10_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-rand-10_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.15282251229021868
 ---
 
 <!-- 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_10_v2
 
-This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
+This model is a fine-tuned version of [](https://huggingface.co/) on the Hartunka/processed_wikitext-103-raw-v1-rand-10_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 11.1017
-- Accuracy: 0.1508
+- Loss: 8.3712
+- Accuracy: 0.1528
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15268124540882635,
-    "eval_loss": 8.355876922607422,
-    "eval_runtime": 1.0307,
+    "eval_accuracy": 0.15282251229021868,
+    "eval_loss": 8.3711576461792,
+    "eval_runtime": 1.0257,
     "eval_samples": 479,
-    "eval_samples_per_second": 464.734,
-    "eval_steps_per_second": 4.851,
-    "perplexity": 4255.114382745165,
+    "eval_samples_per_second": 466.98,
+    "eval_steps_per_second": 4.875,
+    "perplexity": 4320.634935200771,
     "total_flos": 1.5149776477860864e+18,
-    "train_loss": 7.332706870744451,
-    "train_runtime": 24796.9574,
+    "train_loss": 7.298674304619444,
+    "train_runtime": 24656.8275,
     "train_samples": 228639,
-    "train_samples_per_second": 230.511,
-    "train_steps_per_second": 2.402
+    "train_samples_per_second": 231.821,
+    "train_steps_per_second": 2.415
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15268124540882635,
-    "eval_loss": 8.355876922607422,
-    "eval_runtime": 1.0307,
+    "eval_accuracy": 0.15282251229021868,
+    "eval_loss": 8.3711576461792,
+    "eval_runtime": 1.0257,
     "eval_samples": 479,
-    "eval_samples_per_second": 464.734,
-    "eval_steps_per_second": 4.851,
-    "perplexity": 4255.114382745165
+    "eval_samples_per_second": 466.98,
+    "eval_steps_per_second": 4.875,
+    "perplexity": 4320.634935200771
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 1.5149776477860864e+18,
-    "train_loss": 7.332706870744451,
-    "train_runtime": 24796.9574,
+    "train_loss": 7.298674304619444,
+    "train_runtime": 24656.8275,
     "train_samples": 228639,
-    "train_samples_per_second": 230.511,
-    "train_steps_per_second": 2.402
+    "train_samples_per_second": 231.821,
+    "train_steps_per_second": 2.415
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 10000,
-  "best_metric": 8.353975296020508,
+  "best_metric": 8.372286796569824,
   "best_model_checkpoint": "bert_base_rand_10_v2/checkpoint-10000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.7386207580566406,
+      "grad_norm": 3.6852850914001465,
       "learning_rate": 5e-06,
-      "loss": 11.5311,
+      "loss": 11.5321,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 2.343797206878662,
+      "grad_norm": 2.2791285514831543,
       "learning_rate": 1e-05,
-      "loss": 9.8497,
+      "loss": 9.8516,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 2.4195940494537354,
+      "grad_norm": 2.6390469074249268,
       "learning_rate": 1.5e-05,
-      "loss": 9.3046,
+      "loss": 9.3008,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 2.4725236892700195,
+      "grad_norm": 1.7550042867660522,
       "learning_rate": 2e-05,
-      "loss": 9.147,
+      "loss": 9.1443,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 2.195281505584717,
+      "grad_norm": 2.3174149990081787,
       "learning_rate": 2.5e-05,
-      "loss": 9.0266,
+      "loss": 9.0251,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 2.7833189964294434,
+      "grad_norm": 2.786781072616577,
       "learning_rate": 3e-05,
-      "loss": 8.9158,
+      "loss": 8.9154,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 2.8701202869415283,
+      "grad_norm": 2.0159904956817627,
       "learning_rate": 3.5e-05,
-      "loss": 8.8255,
+      "loss": 8.8249,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 2.5824337005615234,
+      "grad_norm": 2.1325860023498535,
       "learning_rate": 4e-05,
-      "loss": 8.7586,
+      "loss": 8.7579,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 1.4923522472381592,
+      "grad_norm": 2.5831844806671143,
       "learning_rate": 4.5e-05,
-      "loss": 8.7001,
+      "loss": 8.6993,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 2.264029026031494,
+      "grad_norm": 2.185391426086426,
       "learning_rate": 5e-05,
-      "loss": 8.65,
+      "loss": 8.6487,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 2.5837624073028564,
+      "grad_norm": 2.713773012161255,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 8.6083,
+      "loss": 8.6079,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 3.833742618560791,
+      "grad_norm": 2.877927780151367,
       "learning_rate": 6e-05,
-      "loss": 8.5711,
+      "loss": 8.5705,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 2.1671464443206787,
+      "grad_norm": 3.3467283248901367,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 8.5378,
+      "loss": 8.5368,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 2.455141544342041,
+      "grad_norm": 2.0169737339019775,
       "learning_rate": 7e-05,
-      "loss": 8.5089,
+      "loss": 8.5103,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 2.34185528755188,
+      "grad_norm": 3.169982433319092,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 8.4854,
+      "loss": 8.4839,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 2.6939852237701416,
+      "grad_norm": 2.6595113277435303,
       "learning_rate": 8e-05,
-      "loss": 8.4633,
+      "loss": 8.462,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 2.867020845413208,
+      "grad_norm": 3.434765338897705,
       "learning_rate": 8.5e-05,
-      "loss": 8.4461,
+      "loss": 8.4464,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 2.205739736557007,
+      "grad_norm": 3.2256903648376465,
       "learning_rate": 9e-05,
-      "loss": 8.4208,
+      "loss": 8.4202,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 2.926563262939453,
+      "grad_norm": 2.2853989601135254,
       "learning_rate": 9.5e-05,
       "loss": 8.4024,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 3.00920033454895,
+      "grad_norm": 3.075472354888916,
       "learning_rate": 0.0001,
-      "loss": 8.37,
+      "loss": 8.374,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.15067197195247503,
-      "eval_loss": 8.353975296020508,
-      "eval_runtime": 1.0372,
-      "eval_samples_per_second": 461.815,
-      "eval_steps_per_second": 4.821,
+      "eval_accuracy": 0.1504215476223601,
+      "eval_loss": 8.372286796569824,
+      "eval_runtime": 1.02,
+      "eval_samples_per_second": 469.617,
+      "eval_steps_per_second": 4.902,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 3.1034252643585205,
+      "grad_norm": 2.2937734127044678,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 8.3583,
+      "loss": 8.3611,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 2.8090872764587402,
+      "grad_norm": 2.6895759105682373,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 8.3412,
+      "loss": 8.3415,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 3.098296880722046,
+      "grad_norm": 3.8590962886810303,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 8.3346,
+      "loss": 8.3364,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 3.6177151203155518,
+      "grad_norm": 2.2545762062072754,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 8.3095,
+      "loss": 8.3105,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 2.9146227836608887,
+      "grad_norm": 3.4054758548736572,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 8.2474,
+      "loss": 8.2518,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 2.585206985473633,
+      "grad_norm": 3.0823278427124023,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 8.249,
+      "loss": 8.2505,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 3.3508150577545166,
+      "grad_norm": 3.253948211669922,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 8.2504,
+      "loss": 8.2507,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 2.227829933166504,
+      "grad_norm": 3.6160099506378174,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 8.2453,
+      "loss": 8.2465,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 2.9316141605377197,
+      "grad_norm": 3.6914443969726562,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 8.1971,
+      "loss": 8.2012,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 3.4122517108917236,
+      "grad_norm": 2.717073440551758,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 8.158,
+      "loss": 8.154,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 3.834278106689453,
+      "grad_norm": 3.506162166595459,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 8.159,
+      "loss": 8.1615,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 3.4940927028656006,
+      "grad_norm": 2.6031546592712402,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 8.1764,
+      "loss": 8.1672,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 3.1475300788879395,
+      "grad_norm": 2.740102767944336,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 8.1736,
+      "loss": 8.1701,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 2.9396891593933105,
+      "grad_norm": 3.0991599559783936,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 8.094,
+      "loss": 8.0863,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 3.0760741233825684,
+      "grad_norm": 3.861220598220825,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 8.0627,
+      "loss": 8.0581,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 2.9063827991485596,
+      "grad_norm": 3.77158522605896,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 8.0774,
+      "loss": 8.0758,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 3.3095574378967285,
+      "grad_norm": 2.998100757598877,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 8.0906,
+      "loss": 8.0836,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 3.671017646789551,
+      "grad_norm": 2.62511944770813,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 8.0912,
+      "loss": 8.085,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 3.6357405185699463,
+      "grad_norm": 4.364990711212158,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 7.9367,
+      "loss": 7.9198,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 3.9633190631866455,
+      "grad_norm": 2.857267379760742,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 7.9402,
+      "loss": 7.9188,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15191223098291198,
-      "eval_loss": 8.597428321838379,
-      "eval_runtime": 1.0203,
-      "eval_samples_per_second": 469.451,
-      "eval_steps_per_second": 4.9,
+      "eval_accuracy": 0.15168775779342855,
+      "eval_loss": 8.573280334472656,
+      "eval_runtime": 1.0058,
+      "eval_samples_per_second": 476.223,
+      "eval_steps_per_second": 4.971,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 3.9242844581604004,
+      "grad_norm": 4.015566825866699,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 7.948,
+      "loss": 7.9285,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 3.322613477706909,
+      "grad_norm": 3.7750561237335205,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 7.9675,
+      "loss": 7.9474,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 4.762775897979736,
+      "grad_norm": 5.549905776977539,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 7.9288,
+      "loss": 7.9053,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 4.790997505187988,
+      "grad_norm": 5.526270389556885,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 7.696,
+      "loss": 7.6521,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 5.435602188110352,
+      "grad_norm": 5.86563777923584,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 7.7155,
+      "loss": 7.6691,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 4.448131084442139,
+      "grad_norm": 4.7667436599731445,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 7.7444,
+      "loss": 7.6996,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 5.873594284057617,
+      "grad_norm": 5.395050048828125,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 7.7417,
+      "loss": 7.707,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 6.450915813446045,
+      "grad_norm": 5.892194747924805,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 7.6323,
+      "loss": 7.5903,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 5.398285388946533,
+      "grad_norm": 5.045933246612549,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 7.4505,
+      "loss": 7.4006,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 6.161104679107666,
+      "grad_norm": 5.722672462463379,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 7.463,
+      "loss": 7.42,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 6.576310157775879,
+      "grad_norm": 6.793386459350586,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 7.4863,
+      "loss": 7.4472,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 5.577183723449707,
+      "grad_norm": 5.019600868225098,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 7.5076,
+      "loss": 7.4676,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 5.695350170135498,
+      "grad_norm": 7.059942245483398,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 7.3367,
+      "loss": 7.2898,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 7.413161754608154,
+      "grad_norm": 6.75661039352417,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 7.2485,
+      "loss": 7.2027,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 7.09622859954834,
+      "grad_norm": 7.141287326812744,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 7.2704,
+      "loss": 7.2284,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 6.137447834014893,
+      "grad_norm": 5.021152973175049,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 7.287,
+      "loss": 7.2403,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 6.190756797790527,
+      "grad_norm": 6.937757968902588,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 7.3065,
+      "loss": 7.2654,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 5.271512031555176,
+      "grad_norm": 6.400770664215088,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 7.101,
+      "loss": 7.048,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 7.171426296234131,
+      "grad_norm": 9.594337463378906,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 7.0708,
+      "loss": 7.0271,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 6.148562431335449,
+      "grad_norm": 7.834346294403076,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 7.1082,
+      "loss": 7.0642,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.15309063435654924,
-      "eval_loss": 9.577547073364258,
-      "eval_runtime": 1.1962,
-      "eval_samples_per_second": 400.438,
-      "eval_steps_per_second": 4.18,
+      "eval_accuracy": 0.15339843859305483,
+      "eval_loss": 9.611098289489746,
+      "eval_runtime": 1.0061,
+      "eval_samples_per_second": 476.103,
+      "eval_steps_per_second": 4.97,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 6.524195194244385,
+      "grad_norm": 6.474629878997803,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 7.1284,
+      "loss": 7.081,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 5.461683750152588,
+      "grad_norm": 7.800127029418945,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 7.1167,
+      "loss": 7.072,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 5.116941452026367,
+      "grad_norm": 5.215531826019287,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 6.9099,
+      "loss": 6.8515,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 7.876025676727295,
+      "grad_norm": 7.740579128265381,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 6.9343,
+      "loss": 6.8846,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 6.082109451293945,
+      "grad_norm": 8.729143142700195,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 6.9668,
+      "loss": 6.9149,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 6.903444290161133,
+      "grad_norm": 7.193752288818359,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 6.9685,
+      "loss": 6.9133,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 5.676742076873779,
+      "grad_norm": 7.278873443603516,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 6.9083,
+      "loss": 6.86,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 8.090641021728516,
+      "grad_norm": 8.18996810913086,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 6.7851,
+      "loss": 6.7287,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 7.927946090698242,
+      "grad_norm": 8.34183406829834,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 6.808,
+      "loss": 6.7532,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 5.399305820465088,
+      "grad_norm": 6.523552894592285,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 6.8176,
+      "loss": 6.7692,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 7.270740032196045,
+      "grad_norm": 7.099226951599121,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 6.8321,
+      "loss": 6.7841,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 6.8463358879089355,
+      "grad_norm": 6.869073867797852,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.739,
+      "loss": 6.689,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 8.434353828430176,
+      "grad_norm": 8.294517517089844,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.6743,
+      "loss": 6.6235,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 5.886655807495117,
+      "grad_norm": 9.18161678314209,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.6907,
+      "loss": 6.6395,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 9.10944938659668,
+      "grad_norm": 6.705373764038086,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.7091,
+      "loss": 6.6594,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 8.006986618041992,
+      "grad_norm": 8.675580024719238,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.7187,
+      "loss": 6.661,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 6.354709625244141,
+      "grad_norm": 5.100327014923096,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.6063,
+      "loss": 6.5488,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 8.193419456481934,
+      "grad_norm": 6.246947288513184,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.5956,
+      "loss": 6.5364,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 9.004114151000977,
+      "grad_norm": 8.054546356201172,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.5984,
+      "loss": 6.5339,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 6.446766376495361,
+      "grad_norm": 8.97150707244873,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.6025,
+      "loss": 6.5459,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.15192889985421107,
-      "eval_loss": 10.42084789276123,
-      "eval_runtime": 1.0275,
-      "eval_samples_per_second": 466.195,
-      "eval_steps_per_second": 4.866,
+      "eval_accuracy": 0.1510037007962319,
+      "eval_loss": 10.364506721496582,
+      "eval_runtime": 1.026,
+      "eval_samples_per_second": 466.866,
+      "eval_steps_per_second": 4.873,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 6.60426664352417,
+      "grad_norm": 6.644847393035889,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.6194,
+      "loss": 6.5685,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 7.228277206420898,
+      "grad_norm": 8.234219551086426,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.5066,
+      "loss": 6.4355,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 8.09839916229248,
+      "grad_norm": 7.898758411407471,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.5231,
+      "loss": 6.4477,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 6.092846393585205,
+      "grad_norm": 5.532283782958984,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.5291,
+      "loss": 6.4561,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 7.129655838012695,
+      "grad_norm": 9.313681602478027,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 6.5341,
+      "loss": 6.4727,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 4.712491035461426,
+      "grad_norm": 9.502847671508789,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.5046,
+      "loss": 6.4386,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 5.977617263793945,
+      "grad_norm": 6.0692057609558105,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.4385,
+      "loss": 6.3679,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 8.676977157592773,
+      "grad_norm": 6.463051795959473,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.445,
+      "loss": 6.372,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 6.589122295379639,
+      "grad_norm": 9.71617317199707,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.4603,
+      "loss": 6.3865,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 5.998814582824707,
+      "grad_norm": 8.971209526062012,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.4607,
+      "loss": 6.3962,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 5.517434120178223,
+      "grad_norm": 8.518393516540527,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.4095,
+      "loss": 6.3379,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 6.569741725921631,
+      "grad_norm": 7.180614471435547,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.3718,
+      "loss": 6.2966,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 5.036622047424316,
+      "grad_norm": 5.732406139373779,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.3688,
+      "loss": 6.3011,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 7.818873405456543,
+      "grad_norm": 4.699340343475342,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.3726,
+      "loss": 6.3127,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 5.3081955909729,
+      "grad_norm": 4.854285717010498,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.3741,
+      "loss": 6.3086,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 5.733414173126221,
+      "grad_norm": 7.51551628112793,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.3229,
+      "loss": 6.2612,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 9.06678295135498,
+      "grad_norm": 8.607324600219727,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.2961,
+      "loss": 6.2382,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 4.187677383422852,
+      "grad_norm": 7.082845211029053,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.2937,
+      "loss": 6.2377,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 8.01919174194336,
+      "grad_norm": 6.963397026062012,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.2997,
+      "loss": 6.2378,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 4.963134288787842,
+      "grad_norm": 4.931850433349609,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.2995,
+      "loss": 6.2464,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.1519331595031822,
-      "eval_loss": 11.06142520904541,
-      "eval_runtime": 1.0136,
-      "eval_samples_per_second": 472.594,
-      "eval_steps_per_second": 4.933,
+      "eval_accuracy": 0.15081167465724618,
+      "eval_loss": 11.101739883422852,
+      "eval_runtime": 1.0118,
+      "eval_samples_per_second": 473.392,
+      "eval_steps_per_second": 4.941,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 10.060037612915039,
+      "grad_norm": 7.39634895324707,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.2277,
+      "loss": 6.179,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 3.576784372329712,
+      "grad_norm": 4.351124286651611,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.2332,
+      "loss": 6.1939,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 7.847463607788086,
+      "grad_norm": 5.487509727478027,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 6.2321,
+      "loss": 6.1821,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 5.3019232749938965,
+      "grad_norm": 4.813266277313232,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 6.232,
+      "loss": 6.1822,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 3.935030937194824,
+      "grad_norm": 3.7798547744750977,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 6.2224,
+      "loss": 6.1733,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 3.691894054412842,
+      "grad_norm": 7.521478652954102,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 6.1747,
+      "loss": 6.1302,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 7.271762847900391,
+      "grad_norm": 6.492825984954834,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 6.1786,
+      "loss": 6.1308,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 3.5396854877471924,
+      "grad_norm": 3.990598678588867,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 6.182,
+      "loss": 6.1427,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 2.217651128768921,
+      "grad_norm": 4.194901466369629,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 6.1654,
+      "loss": 6.1251,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 7.711702346801758,
+      "grad_norm": 3.876901149749756,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 6.1587,
+      "loss": 6.1162,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 6.7426981925964355,
+      "grad_norm": 6.772987365722656,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 6.1373,
+      "loss": 6.1002,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 3.855668544769287,
+      "grad_norm": 3.5543112754821777,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 6.1439,
+      "loss": 6.1027,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 3.8093292713165283,
+      "grad_norm": 4.911917209625244,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 6.1323,
+      "loss": 6.0919,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 3.554218292236328,
+      "grad_norm": 4.851410865783691,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 6.1344,
+      "loss": 6.1013,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 4.135804653167725,
+      "grad_norm": 6.0461039543151855,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 6.1233,
+      "loss": 6.0854,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 7.920344352722168,
+      "grad_norm": 6.592803001403809,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 6.1253,
+      "loss": 6.0857,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 8.885543823242188,
+      "grad_norm": 5.261019706726074,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 6.1242,
+      "loss": 6.0856,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 3.082768201828003,
+      "grad_norm": 3.316622734069824,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 6.115,
+      "loss": 6.0747,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 5.572300434112549,
+      "grad_norm": 4.893493175506592,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 6.1039,
+      "loss": 6.0665,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 1.5149776477860864e+18,
-      "train_loss": 7.332706870744451,
-      "train_runtime": 24796.9574,
-      "train_samples_per_second": 230.511,
-      "train_steps_per_second": 2.402
+      "train_loss": 7.298674304619444,
+      "train_runtime": 24656.8275,
+      "train_samples_per_second": 231.821,
+      "train_steps_per_second": 2.415
     }
   ],
   "logging_steps": 500,