End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-rand-50_v2
 metrics:
 - accuracy
 model-index:
 - name: bert_base_rand_50_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-rand-50_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-rand-50_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.151748883991637
 ---
 
 <!-- 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_50_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-50_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 13.7891
-- Accuracy: 0.1520
+- Loss: 10.1382
+- Accuracy: 0.1517
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15330281968695258,
-    "eval_loss": 10.085304260253906,
-    "eval_runtime": 1.0293,
+    "eval_accuracy": 0.151748883991637,
+    "eval_loss": 10.138225555419922,
+    "eval_runtime": 1.0511,
     "eval_samples": 479,
-    "eval_samples_per_second": 465.348,
-    "eval_steps_per_second": 4.857,
-    "perplexity": 23987.88668012822,
+    "eval_samples_per_second": 455.694,
+    "eval_steps_per_second": 4.757,
+    "perplexity": 25291.548193853436,
     "total_flos": 1.5155177772432384e+18,
-    "train_loss": 8.017368304479232,
-    "train_runtime": 24861.7911,
+    "train_loss": 8.027886620048312,
+    "train_runtime": 24986.4669,
     "train_samples": 228639,
-    "train_samples_per_second": 229.91,
-    "train_steps_per_second": 2.395
+    "train_samples_per_second": 228.763,
+    "train_steps_per_second": 2.383
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15330281968695258,
-    "eval_loss": 10.085304260253906,
-    "eval_runtime": 1.0293,
+    "eval_accuracy": 0.151748883991637,
+    "eval_loss": 10.138225555419922,
+    "eval_runtime": 1.0511,
     "eval_samples": 479,
-    "eval_samples_per_second": 465.348,
-    "eval_steps_per_second": 4.857,
-    "perplexity": 23987.88668012822
+    "eval_samples_per_second": 455.694,
+    "eval_steps_per_second": 4.757,
+    "perplexity": 25291.548193853436
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 1.5155177772432384e+18,
-    "train_loss": 8.017368304479232,
-    "train_runtime": 24861.7911,
+    "train_loss": 8.027886620048312,
+    "train_runtime": 24986.4669,
     "train_samples": 228639,
-    "train_samples_per_second": 229.91,
-    "train_steps_per_second": 2.395
+    "train_samples_per_second": 228.763,
+    "train_steps_per_second": 2.383
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 10000,
-  "best_metric": 10.05613899230957,
+  "best_metric": 10.111589431762695,
   "best_model_checkpoint": "bert_base_rand_50_v2/checkpoint-10000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.9099268913269043,
+      "grad_norm": 2.9707529544830322,
       "learning_rate": 5e-06,
-      "loss": 13.1533,
+      "loss": 13.1554,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 1.9890450239181519,
+      "grad_norm": 1.9404082298278809,
       "learning_rate": 1e-05,
-      "loss": 11.4748,
+      "loss": 11.4762,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 2.1525957584381104,
+      "grad_norm": 1.878383994102478,
       "learning_rate": 1.5e-05,
-      "loss": 10.9273,
+      "loss": 10.9292,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 2.329235792160034,
+      "grad_norm": 2.2431883811950684,
       "learning_rate": 2e-05,
-      "loss": 10.7736,
+      "loss": 10.7757,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 2.406121015548706,
+      "grad_norm": 2.226682424545288,
       "learning_rate": 2.5e-05,
-      "loss": 10.6557,
+      "loss": 10.6539,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 2.3026375770568848,
+      "grad_norm": 2.4292383193969727,
       "learning_rate": 3e-05,
-      "loss": 10.5438,
+      "loss": 10.5441,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 2.6435930728912354,
+      "grad_norm": 2.472700357437134,
       "learning_rate": 3.5e-05,
-      "loss": 10.4519,
+      "loss": 10.4527,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 2.505197286605835,
+      "grad_norm": 2.2452244758605957,
       "learning_rate": 4e-05,
-      "loss": 10.3824,
+      "loss": 10.3843,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 2.0937581062316895,
+      "grad_norm": 2.4805452823638916,
       "learning_rate": 4.5e-05,
-      "loss": 10.3245,
+      "loss": 10.3271,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 2.4586992263793945,
+      "grad_norm": 2.4201602935791016,
       "learning_rate": 5e-05,
-      "loss": 10.2748,
+      "loss": 10.2781,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 2.683626413345337,
+      "grad_norm": 2.1865737438201904,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 10.2351,
+      "loss": 10.2407,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 2.38889217376709,
+      "grad_norm": 2.409081220626831,
       "learning_rate": 6e-05,
-      "loss": 10.1982,
+      "loss": 10.2053,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 2.4529201984405518,
+      "grad_norm": 2.762845039367676,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 10.1657,
+      "loss": 10.1719,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 3.1758840084075928,
+      "grad_norm": 2.8632805347442627,
       "learning_rate": 7e-05,
-      "loss": 10.1384,
+      "loss": 10.1446,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 2.4082701206207275,
+      "grad_norm": 2.655979633331299,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 10.1111,
+      "loss": 10.115,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 2.531470775604248,
+      "grad_norm": 2.5348496437072754,
       "learning_rate": 8e-05,
-      "loss": 10.0913,
+      "loss": 10.0935,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 2.6216888427734375,
+      "grad_norm": 2.47623610496521,
       "learning_rate": 8.5e-05,
-      "loss": 10.0767,
+      "loss": 10.0793,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 2.823248863220215,
+      "grad_norm": 2.29648756980896,
       "learning_rate": 9e-05,
-      "loss": 10.0562,
+      "loss": 10.0585,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 2.4972293376922607,
+      "grad_norm": 2.7251288890838623,
       "learning_rate": 9.5e-05,
-      "loss": 10.0398,
+      "loss": 10.0433,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 3.1853861808776855,
+      "grad_norm": 3.1358323097229004,
       "learning_rate": 0.0001,
-      "loss": 9.9561,
+      "loss": 9.9441,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.15047719747349675,
-      "eval_loss": 10.05613899230957,
-      "eval_runtime": 1.0851,
-      "eval_samples_per_second": 441.45,
-      "eval_steps_per_second": 4.608,
+      "eval_accuracy": 0.1491416010462172,
+      "eval_loss": 10.111589431762695,
+      "eval_runtime": 1.0518,
+      "eval_samples_per_second": 455.415,
+      "eval_steps_per_second": 4.754,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 2.944951295852661,
+      "grad_norm": 3.0342698097229004,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 9.9524,
+      "loss": 9.9454,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 3.06426739692688,
+      "grad_norm": 3.2399396896362305,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 9.9408,
+      "loss": 9.9392,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 3.042531728744507,
+      "grad_norm": 2.9815125465393066,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 9.9423,
+      "loss": 9.9401,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 3.07763671875,
+      "grad_norm": 3.674541473388672,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 9.8982,
+      "loss": 9.8878,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 3.4107093811035156,
+      "grad_norm": 3.4258456230163574,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 9.738,
+      "loss": 9.731,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 3.5907392501831055,
+      "grad_norm": 3.4225902557373047,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 9.7568,
+      "loss": 9.7512,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 3.213953733444214,
+      "grad_norm": 3.278662919998169,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 9.7777,
+      "loss": 9.7728,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 3.4074673652648926,
+      "grad_norm": 3.3490500450134277,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 9.7864,
+      "loss": 9.7853,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 3.344696521759033,
+      "grad_norm": 3.52455472946167,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 9.6659,
+      "loss": 9.6649,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 3.5423378944396973,
+      "grad_norm": 3.455338478088379,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 9.5485,
+      "loss": 9.5491,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 3.4122226238250732,
+      "grad_norm": 3.7562873363494873,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 9.5791,
+      "loss": 9.5784,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 3.7815067768096924,
+      "grad_norm": 3.659771203994751,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 9.6023,
+      "loss": 9.6044,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 3.9482810497283936,
+      "grad_norm": 3.8113808631896973,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 9.622,
+      "loss": 9.6219,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 4.10134744644165,
+      "grad_norm": 3.8798913955688477,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 9.3796,
+      "loss": 9.3837,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 4.931428909301758,
+      "grad_norm": 4.825026988983154,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 9.2804,
+      "loss": 9.2877,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 4.4619035720825195,
+      "grad_norm": 4.486569404602051,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 9.3171,
+      "loss": 9.327,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 4.992231369018555,
+      "grad_norm": 5.084190845489502,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 9.3526,
+      "loss": 9.3632,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 4.491001605987549,
+      "grad_norm": 4.671340465545654,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 9.3705,
+      "loss": 9.3704,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 5.572983264923096,
+      "grad_norm": 5.141868591308594,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 8.9272,
+      "loss": 8.944,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 5.715620517730713,
+      "grad_norm": 5.74922513961792,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 8.9037,
+      "loss": 8.9223,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15247341395662056,
-      "eval_loss": 10.681720733642578,
-      "eval_runtime": 1.0393,
-      "eval_samples_per_second": 460.876,
-      "eval_steps_per_second": 4.811,
+      "eval_accuracy": 0.1528101237408457,
+      "eval_loss": 10.787955284118652,
+      "eval_runtime": 1.0309,
+      "eval_samples_per_second": 464.647,
+      "eval_steps_per_second": 4.85,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 5.897007942199707,
+      "grad_norm": 6.01029109954834,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 8.9314,
+      "loss": 8.9445,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 6.214521408081055,
+      "grad_norm": 5.30519437789917,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 8.9805,
+      "loss": 8.9922,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 6.8039069175720215,
+      "grad_norm": 6.141433238983154,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 8.9169,
+      "loss": 8.9304,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 6.507361888885498,
+      "grad_norm": 6.596657752990723,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 8.4538,
+      "loss": 8.4829,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 7.411440372467041,
+      "grad_norm": 6.786694526672363,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 8.4945,
+      "loss": 8.5094,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 6.491020679473877,
+      "grad_norm": 7.219344615936279,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 8.5462,
+      "loss": 8.5676,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 7.425654888153076,
+      "grad_norm": 7.966107368469238,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 8.5592,
+      "loss": 8.5788,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 7.183740139007568,
+      "grad_norm": 7.448211193084717,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 8.4007,
+      "loss": 8.4328,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 7.079336643218994,
+      "grad_norm": 7.143707275390625,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 8.1158,
+      "loss": 8.1465,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 7.690479278564453,
+      "grad_norm": 6.648916721343994,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 8.1468,
+      "loss": 8.1743,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 8.05373764038086,
+      "grad_norm": 7.4369683265686035,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 8.1876,
+      "loss": 8.2125,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 7.30242395401001,
+      "grad_norm": 6.361053466796875,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 8.2264,
+      "loss": 8.249,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 7.552883625030518,
+      "grad_norm": 7.151641368865967,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 7.9535,
+      "loss": 7.9823,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 8.210590362548828,
+      "grad_norm": 8.406608581542969,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 7.8062,
+      "loss": 7.8567,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 8.393407821655273,
+      "grad_norm": 7.313540458679199,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 7.8662,
+      "loss": 7.8898,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 7.620030403137207,
+      "grad_norm": 8.477072715759277,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 7.8813,
+      "loss": 7.9167,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 9.025675773620605,
+      "grad_norm": 7.650353908538818,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 7.9285,
+      "loss": 7.9557,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 6.877237796783447,
+      "grad_norm": 9.022331237792969,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 7.5567,
+      "loss": 7.594,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 7.735791206359863,
+      "grad_norm": 7.6493353843688965,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 7.5234,
+      "loss": 7.5633,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 8.022677421569824,
+      "grad_norm": 8.859118461608887,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 7.5812,
+      "loss": 7.6087,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.15390211825279124,
-      "eval_loss": 12.069398880004883,
-      "eval_runtime": 1.0189,
-      "eval_samples_per_second": 470.133,
-      "eval_steps_per_second": 4.907,
+      "eval_accuracy": 0.1539860648627473,
+      "eval_loss": 11.965897560119629,
+      "eval_runtime": 1.05,
+      "eval_samples_per_second": 456.203,
+      "eval_steps_per_second": 4.762,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 7.011293888092041,
+      "grad_norm": 8.735087394714355,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 7.6079,
+      "loss": 7.6486,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 9.066295623779297,
+      "grad_norm": 8.684382438659668,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 7.5975,
+      "loss": 7.6268,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 6.338802814483643,
+      "grad_norm": 8.555488586425781,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 7.2223,
+      "loss": 7.2555,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 9.85566234588623,
+      "grad_norm": 9.307188987731934,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 7.273,
+      "loss": 7.296,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 7.851714134216309,
+      "grad_norm": 8.361678123474121,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 7.3248,
+      "loss": 7.3526,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 8.250563621520996,
+      "grad_norm": 9.484582901000977,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 7.3395,
+      "loss": 7.3558,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 6.940363883972168,
+      "grad_norm": 7.470253944396973,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 7.2425,
+      "loss": 7.2673,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 8.973872184753418,
+      "grad_norm": 10.630374908447266,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 7.0056,
+      "loss": 7.0194,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 10.261417388916016,
+      "grad_norm": 9.159460067749023,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 7.0529,
+      "loss": 7.0692,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 10.443400382995605,
+      "grad_norm": 9.374323844909668,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 7.0808,
+      "loss": 7.0937,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 8.094125747680664,
+      "grad_norm": 10.391216278076172,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 7.0903,
+      "loss": 7.1094,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 8.093913078308105,
+      "grad_norm": 7.5976386070251465,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.9361,
+      "loss": 6.9448,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 8.476778984069824,
+      "grad_norm": 8.927652359008789,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.8211,
+      "loss": 6.8178,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 7.970340728759766,
+      "grad_norm": 9.162418365478516,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.8499,
+      "loss": 6.8473,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 10.228806495666504,
+      "grad_norm": 10.707791328430176,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.879,
+      "loss": 6.8819,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 8.830512046813965,
+      "grad_norm": 10.748319625854492,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.89,
+      "loss": 6.9071,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 6.972827434539795,
+      "grad_norm": 8.3500394821167,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.6836,
+      "loss": 6.6874,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 9.140987396240234,
+      "grad_norm": 7.616583347320557,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.6726,
+      "loss": 6.6696,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 8.59704875946045,
+      "grad_norm": 9.38701343536377,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.6704,
+      "loss": 6.6748,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 8.396159172058105,
+      "grad_norm": 8.05835247039795,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.698,
+      "loss": 6.6937,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.151115846136593,
-      "eval_loss": 13.08890151977539,
-      "eval_runtime": 1.0379,
-      "eval_samples_per_second": 461.522,
-      "eval_steps_per_second": 4.818,
+      "eval_accuracy": 0.15061119210496804,
+      "eval_loss": 13.027770042419434,
+      "eval_runtime": 1.0264,
+      "eval_samples_per_second": 466.677,
+      "eval_steps_per_second": 4.871,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 8.734539031982422,
+      "grad_norm": 9.350462913513184,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.7099,
+      "loss": 6.717,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 10.298134803771973,
+      "grad_norm": 9.866216659545898,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.5213,
+      "loss": 6.5246,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 8.436574935913086,
+      "grad_norm": 9.163368225097656,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.5382,
+      "loss": 6.5392,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 8.568227767944336,
+      "grad_norm": 8.24027156829834,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.5512,
+      "loss": 6.5577,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 10.303642272949219,
+      "grad_norm": 8.120686531066895,
       "learning_rate": 3.440968718466196e-05,
       "loss": 6.5747,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 6.319511890411377,
+      "grad_norm": 6.9892473220825195,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.5355,
+      "loss": 6.5415,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 7.791539669036865,
+      "grad_norm": 8.302289962768555,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.4412,
+      "loss": 6.4506,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 8.565069198608398,
+      "grad_norm": 7.935204029083252,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.4591,
+      "loss": 6.4641,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 8.081923484802246,
+      "grad_norm": 6.349064350128174,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.4688,
+      "loss": 6.4759,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 8.338988304138184,
+      "grad_norm": 8.392001152038574,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.4904,
+      "loss": 6.493,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 7.188676834106445,
+      "grad_norm": 7.667578220367432,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.4305,
+      "loss": 6.4343,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 8.027253150939941,
+      "grad_norm": 7.433592319488525,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.3892,
+      "loss": 6.3996,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 10.849565505981445,
+      "grad_norm": 9.781779289245605,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.3917,
+      "loss": 6.4004,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 7.715574741363525,
+      "grad_norm": 7.436337471008301,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.398,
+      "loss": 6.3999,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 7.849226474761963,
+      "grad_norm": 7.300240993499756,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.3903,
+      "loss": 6.403,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 3.8096721172332764,
+      "grad_norm": 9.506343841552734,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.3642,
+      "loss": 6.3776,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 7.980307579040527,
+      "grad_norm": 8.23635482788086,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.3426,
+      "loss": 6.3528,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 11.38594913482666,
+      "grad_norm": 9.698938369750977,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.3359,
+      "loss": 6.3486,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 8.007773399353027,
+      "grad_norm": 8.498291969299316,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.333,
+      "loss": 6.3499,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 9.579621315002441,
+      "grad_norm": 8.12326717376709,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.3527,
+      "loss": 6.355,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.1511200829898786,
-      "eval_loss": 13.698497772216797,
-      "eval_runtime": 1.0288,
-      "eval_samples_per_second": 465.6,
-      "eval_steps_per_second": 4.86,
+      "eval_accuracy": 0.1519611966243306,
+      "eval_loss": 13.789127349853516,
+      "eval_runtime": 1.0298,
+      "eval_samples_per_second": 465.133,
+      "eval_steps_per_second": 4.855,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 7.129371166229248,
+      "grad_norm": 6.933436870574951,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.3026,
+      "loss": 6.3149,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 7.323594093322754,
+      "grad_norm": 7.0417399406433105,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.3087,
+      "loss": 6.3262,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 5.939312934875488,
+      "grad_norm": 6.051033020019531,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 6.3067,
+      "loss": 6.3138,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 7.441634178161621,
+      "grad_norm": 6.116024971008301,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 6.2996,
+      "loss": 6.3152,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 3.924901008605957,
+      "grad_norm": 7.532071113586426,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 6.3001,
+      "loss": 6.304,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 5.1285200119018555,
+      "grad_norm": 5.917665958404541,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 6.2659,
+      "loss": 6.2765,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 7.868868350982666,
+      "grad_norm": 8.547967910766602,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 6.2639,
+      "loss": 6.2766,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 6.70086669921875,
+      "grad_norm": 9.161884307861328,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 6.2764,
+      "loss": 6.2854,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 5.3119635581970215,
+      "grad_norm": 4.192229747772217,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 6.2614,
+      "loss": 6.2708,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 4.03257417678833,
+      "grad_norm": 5.182990550994873,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 6.2512,
+      "loss": 6.2605,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 11.998774528503418,
+      "grad_norm": 6.774938106536865,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 6.2369,
+      "loss": 6.2476,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 7.52518367767334,
+      "grad_norm": 7.716025352478027,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 6.2438,
+      "loss": 6.2571,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 7.338294982910156,
+      "grad_norm": 7.3651652336120605,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 6.2247,
+      "loss": 6.2405,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 4.920290946960449,
+      "grad_norm": 6.686182022094727,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 6.2355,
+      "loss": 6.2474,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 4.500492095947266,
+      "grad_norm": 5.635970592498779,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 6.2176,
+      "loss": 6.233,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 7.811495304107666,
+      "grad_norm": 6.446502685546875,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 6.2283,
+      "loss": 6.2342,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 6.0045623779296875,
+      "grad_norm": 6.000938415527344,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 6.217,
+      "loss": 6.2376,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 4.935161113739014,
+      "grad_norm": 6.459875106811523,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 6.2151,
+      "loss": 6.2275,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 6.402288436889648,
+      "grad_norm": 5.6358489990234375,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 6.2068,
+      "loss": 6.2202,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 1.5155177772432384e+18,
-      "train_loss": 8.017368304479232,
-      "train_runtime": 24861.7911,
-      "train_samples_per_second": 229.91,
-      "train_steps_per_second": 2.395
+      "train_loss": 8.027886620048312,
+      "train_runtime": 24986.4669,
+      "train_samples_per_second": 228.763,
+      "train_steps_per_second": 2.383
     }
   ],
   "logging_steps": 500,