End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-km-10_v2
 metrics:
 - accuracy
 model-index:
 - name: distilbert_km_10_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-km-10_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-km-10_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.15502627563993898
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -14,10 +26,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # distilbert_km_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-km-10_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 6.3391
-- Accuracy: 0.1524
+- Loss: 6.2864
+- Accuracy: 0.1550
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15423518110414194,
-    "eval_loss": 6.305652141571045,
-    "eval_runtime": 0.8528,
+    "eval_accuracy": 0.15502627563993898,
+    "eval_loss": 6.286385536193848,
+    "eval_runtime": 0.8213,
     "eval_samples": 479,
-    "eval_samples_per_second": 561.669,
-    "eval_steps_per_second": 5.863,
-    "perplexity": 547.6586227083073,
+    "eval_samples_per_second": 583.214,
+    "eval_steps_per_second": 6.088,
+    "perplexity": 537.208096494861,
     "total_flos": 7.6822177215744e+17,
-    "train_loss": 6.218012153119425,
-    "train_runtime": 17148.6644,
+    "train_loss": 6.223999219200974,
+    "train_runtime": 16965.8185,
     "train_samples": 228639,
-    "train_samples_per_second": 333.319,
-    "train_steps_per_second": 3.473
+    "train_samples_per_second": 336.911,
+    "train_steps_per_second": 3.51
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15423518110414194,
-    "eval_loss": 6.305652141571045,
-    "eval_runtime": 0.8528,
+    "eval_accuracy": 0.15502627563993898,
+    "eval_loss": 6.286385536193848,
+    "eval_runtime": 0.8213,
     "eval_samples": 479,
-    "eval_samples_per_second": 561.669,
-    "eval_steps_per_second": 5.863,
-    "perplexity": 547.6586227083073
+    "eval_samples_per_second": 583.214,
+    "eval_steps_per_second": 6.088,
+    "perplexity": 537.208096494861
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 7.6822177215744e+17,
-    "train_loss": 6.218012153119425,
-    "train_runtime": 17148.6644,
+    "train_loss": 6.223999219200974,
+    "train_runtime": 16965.8185,
     "train_samples": 228639,
-    "train_samples_per_second": 333.319,
-    "train_steps_per_second": 3.473
+    "train_samples_per_second": 336.911,
+    "train_steps_per_second": 3.51
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_global_step": 20000,
-  "best_metric": 6.307790279388428,
-  "best_model_checkpoint": "distilbert_km_10_v2/checkpoint-20000",
+  "best_global_step": 30000,
+  "best_metric": 6.312332630157471,
+  "best_model_checkpoint": "distilbert_km_10_v2/checkpoint-30000",
   "epoch": 25.0,
   "eval_steps": 10000,
   "global_step": 59550,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 3.150407075881958,
+      "grad_norm": 3.0913989543914795,
       "learning_rate": 5e-06,
-      "loss": 11.0763,
+      "loss": 11.5507,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 10.050383567810059,
+      "grad_norm": 3.9964706897735596,
       "learning_rate": 1e-05,
-      "loss": 8.927,
+      "loss": 9.5896,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 8.682597160339355,
+      "grad_norm": 8.248088836669922,
       "learning_rate": 1.5e-05,
-      "loss": 7.8866,
+      "loss": 8.2229,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 20.868480682373047,
+      "grad_norm": 9.571820259094238,
       "learning_rate": 2e-05,
-      "loss": 7.5837,
+      "loss": 7.7459,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 11.644207954406738,
+      "grad_norm": 9.813409805297852,
       "learning_rate": 2.5e-05,
-      "loss": 7.4123,
+      "loss": 7.4976,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 9.56916618347168,
+      "grad_norm": 6.8137078285217285,
       "learning_rate": 3e-05,
-      "loss": 7.2678,
+      "loss": 7.3035,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 3.5344231128692627,
+      "grad_norm": 4.454509258270264,
       "learning_rate": 3.5e-05,
-      "loss": 7.1395,
+      "loss": 7.1659,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 6.1385273933410645,
+      "grad_norm": 4.408298969268799,
       "learning_rate": 4e-05,
-      "loss": 7.0361,
+      "loss": 7.035,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 5.029597759246826,
+      "grad_norm": 3.7217628955841064,
       "learning_rate": 4.5e-05,
-      "loss": 6.9433,
+      "loss": 6.9278,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 8.310822486877441,
+      "grad_norm": 3.723860740661621,
       "learning_rate": 5e-05,
-      "loss": 6.8394,
+      "loss": 6.8158,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 1.8969370126724243,
+      "grad_norm": 2.9869563579559326,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 6.7782,
+      "loss": 6.7349,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 2.1686832904815674,
+      "grad_norm": 3.580021381378174,
       "learning_rate": 6e-05,
-      "loss": 6.7481,
+      "loss": 6.7012,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 2.3065264225006104,
+      "grad_norm": 2.0765135288238525,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 6.7081,
+      "loss": 6.6703,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 3.0020182132720947,
+      "grad_norm": 2.655747890472412,
       "learning_rate": 7e-05,
-      "loss": 6.6844,
+      "loss": 6.6418,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 2.257652521133423,
+      "grad_norm": 1.6839361190795898,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 6.6217,
+      "loss": 6.5716,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 2.9006402492523193,
+      "grad_norm": 3.693770408630371,
       "learning_rate": 8e-05,
-      "loss": 6.6082,
+      "loss": 6.5504,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 1.9896626472473145,
+      "grad_norm": 2.004279613494873,
       "learning_rate": 8.5e-05,
-      "loss": 6.6034,
+      "loss": 6.5431,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 3.605994701385498,
+      "grad_norm": 2.780397415161133,
       "learning_rate": 9e-05,
-      "loss": 6.5784,
+      "loss": 6.5403,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 3.555478572845459,
+      "grad_norm": 2.6447219848632812,
       "learning_rate": 9.5e-05,
-      "loss": 6.5616,
+      "loss": 6.5275,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 3.820685625076294,
+      "grad_norm": 2.2194011211395264,
       "learning_rate": 0.0001,
-      "loss": 6.4951,
+      "loss": 6.4486,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.15047719747349675,
-      "eval_loss": 6.483672142028809,
-      "eval_runtime": 0.8699,
-      "eval_samples_per_second": 550.665,
-      "eval_steps_per_second": 5.748,
+      "eval_accuracy": 0.15159019449622974,
+      "eval_loss": 6.427244186401367,
+      "eval_runtime": 0.8571,
+      "eval_samples_per_second": 558.841,
+      "eval_steps_per_second": 5.833,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 2.839904308319092,
+      "grad_norm": 4.444994926452637,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 6.4842,
+      "loss": 6.4438,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 1.8162016868591309,
+      "grad_norm": 2.5240862369537354,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 6.4732,
+      "loss": 6.4422,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 2.417090892791748,
+      "grad_norm": 1.9277300834655762,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 6.4607,
+      "loss": 6.4359,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 2.758234977722168,
+      "grad_norm": 2.0342838764190674,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 6.4308,
+      "loss": 6.4098,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 2.7664124965667725,
+      "grad_norm": 1.9326578378677368,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 6.3696,
+      "loss": 6.3468,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 1.8280402421951294,
+      "grad_norm": 2.6669442653656006,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 6.3635,
+      "loss": 6.3378,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 1.9037777185440063,
+      "grad_norm": 1.374561071395874,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 6.3637,
+      "loss": 6.3387,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 3.311438798904419,
+      "grad_norm": 1.913304090499878,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 6.3549,
+      "loss": 6.3309,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 2.81231689453125,
+      "grad_norm": 2.308854579925537,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 6.3206,
+      "loss": 6.3009,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 1.9139678478240967,
+      "grad_norm": 1.655401587486267,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 6.2828,
+      "loss": 6.2679,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 1.5594526529312134,
+      "grad_norm": 3.0039784908294678,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 6.2772,
+      "loss": 6.2672,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 1.8456449508666992,
+      "grad_norm": 1.6358284950256348,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 6.2806,
+      "loss": 6.2583,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 1.3920214176177979,
+      "grad_norm": 1.525787591934204,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 6.2785,
+      "loss": 6.2591,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 1.2545069456100464,
+      "grad_norm": 2.4053471088409424,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 6.2294,
+      "loss": 6.2163,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 1.872115135192871,
+      "grad_norm": 1.8042075634002686,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 6.2094,
+      "loss": 6.1939,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 1.420103907585144,
+      "grad_norm": 2.0802719593048096,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 6.2161,
+      "loss": 6.2032,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 2.163827419281006,
+      "grad_norm": 1.846529245376587,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 6.219,
+      "loss": 6.2008,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 1.9957118034362793,
+      "grad_norm": 2.998645782470703,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 6.215,
+      "loss": 6.2019,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 2.5708162784576416,
+      "grad_norm": 1.5601189136505127,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 6.161,
+      "loss": 6.1568,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 2.9247000217437744,
+      "grad_norm": 1.497053861618042,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 6.1618,
+      "loss": 6.1571,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15292236033558743,
-      "eval_loss": 6.307790279388428,
-      "eval_runtime": 0.8459,
-      "eval_samples_per_second": 566.252,
-      "eval_steps_per_second": 5.911,
+      "eval_accuracy": 0.15250147310530598,
+      "eval_loss": 6.384017467498779,
+      "eval_runtime": 0.8049,
+      "eval_samples_per_second": 595.137,
+      "eval_steps_per_second": 6.212,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 2.03761887550354,
+      "grad_norm": 2.2093071937561035,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 6.1605,
+      "loss": 6.1456,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 1.3724454641342163,
+      "grad_norm": 1.9814311265945435,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 6.1581,
+      "loss": 6.1494,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 1.9964754581451416,
+      "grad_norm": 2.023616313934326,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 6.1502,
+      "loss": 6.1416,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 3.6079020500183105,
+      "grad_norm": 1.8390247821807861,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 6.1122,
+      "loss": 6.1101,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 1.9001446962356567,
+      "grad_norm": 1.6448228359222412,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 6.115,
+      "loss": 6.1093,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 1.6073627471923828,
+      "grad_norm": 2.1586291790008545,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 6.1134,
+      "loss": 6.1088,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 2.9083070755004883,
+      "grad_norm": 0.8653324246406555,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 6.1108,
+      "loss": 6.1018,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 1.5109046697616577,
+      "grad_norm": 2.379240036010742,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 6.0905,
+      "loss": 6.0861,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 2.261303186416626,
+      "grad_norm": 2.1743650436401367,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 6.0762,
+      "loss": 6.075,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 1.638889193534851,
+      "grad_norm": 0.8681607246398926,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 6.0723,
+      "loss": 6.0686,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 2.529879093170166,
+      "grad_norm": 2.3597679138183594,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 6.07,
+      "loss": 6.0628,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 1.039648175239563,
+      "grad_norm": 3.0710947513580322,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 6.0695,
+      "loss": 6.0724,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 2.0828940868377686,
+      "grad_norm": 3.005756378173828,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 6.0511,
+      "loss": 6.0492,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 2.210174798965454,
+      "grad_norm": 2.287787914276123,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 6.0412,
+      "loss": 6.0445,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 1.7135660648345947,
+      "grad_norm": 0.7432211637496948,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 6.0342,
+      "loss": 6.0339,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 2.6034951210021973,
+      "grad_norm": 2.2258312702178955,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 6.03,
+      "loss": 6.0316,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 1.5736366510391235,
+      "grad_norm": 0.8729032278060913,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 6.0286,
+      "loss": 6.0313,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 2.4166982173919678,
+      "grad_norm": 3.3516526222229004,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 6.0145,
+      "loss": 6.0164,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 3.0722899436950684,
+      "grad_norm": 0.7179839015007019,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 6.0049,
+      "loss": 6.0063,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 1.7896665334701538,
+      "grad_norm": 1.9476817846298218,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 6.0097,
+      "loss": 6.0111,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.1539860648627473,
-      "eval_loss": 6.370279788970947,
-      "eval_runtime": 0.8545,
-      "eval_samples_per_second": 560.535,
-      "eval_steps_per_second": 5.851,
+      "eval_accuracy": 0.15409799367602206,
+      "eval_loss": 6.312332630157471,
+      "eval_runtime": 0.8094,
+      "eval_samples_per_second": 591.806,
+      "eval_steps_per_second": 6.178,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 0.9673942923545837,
+      "grad_norm": 1.2007274627685547,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 6.003,
+      "loss": 6.0035,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 2.033679485321045,
+      "grad_norm": 1.2920252084732056,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 6.0039,
+      "loss": 6.0048,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 1.8327769041061401,
+      "grad_norm": 2.3055388927459717,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 5.9835,
+      "loss": 5.985,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 0.9647663235664368,
+      "grad_norm": 1.011922001838684,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 5.9799,
+      "loss": 5.9838,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 2.6440300941467285,
+      "grad_norm": 2.5464136600494385,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 5.9848,
+      "loss": 5.9836,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 2.5690128803253174,
+      "grad_norm": 1.663964867591858,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 5.9815,
+      "loss": 5.9798,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 1.104326844215393,
+      "grad_norm": 4.078064441680908,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 5.9677,
+      "loss": 5.968,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 1.111171007156372,
+      "grad_norm": 1.4199191331863403,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 5.9592,
+      "loss": 5.9619,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 1.3937731981277466,
+      "grad_norm": 1.5001009702682495,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 5.9594,
+      "loss": 5.9581,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 1.7694659233093262,
+      "grad_norm": 1.059971809387207,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 5.9568,
+      "loss": 5.9553,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 0.8559344410896301,
+      "grad_norm": 2.0065133571624756,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 5.9557,
+      "loss": 5.9536,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 1.2789491415023804,
+      "grad_norm": 3.530527353286743,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 5.9455,
+      "loss": 5.9432,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 2.1253957748413086,
+      "grad_norm": 1.7634003162384033,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 5.9398,
+      "loss": 5.9416,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 2.34883451461792,
+      "grad_norm": 1.8264235258102417,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 5.9364,
+      "loss": 5.9357,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 1.949161171913147,
+      "grad_norm": 2.3028690814971924,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 5.9397,
+      "loss": 5.9381,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 2.70025372505188,
+      "grad_norm": 2.6295433044433594,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 5.9363,
+      "loss": 5.9345,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 3.276752471923828,
+      "grad_norm": 3.4183132648468018,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 5.9274,
+      "loss": 5.9277,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 2.2213597297668457,
+      "grad_norm": 1.4393947124481201,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 5.921,
+      "loss": 5.9196,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 0.9024696946144104,
+      "grad_norm": 3.317281723022461,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 5.9232,
+      "loss": 5.9229,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 2.0587399005889893,
+      "grad_norm": 3.812405586242676,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 5.9114,
+      "loss": 5.9119,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.15209711786475272,
-      "eval_loss": 6.38532829284668,
-      "eval_runtime": 0.8502,
-      "eval_samples_per_second": 563.395,
-      "eval_steps_per_second": 5.881,
+      "eval_accuracy": 0.15240551755074577,
+      "eval_loss": 6.347248554229736,
+      "eval_runtime": 0.83,
+      "eval_samples_per_second": 577.092,
+      "eval_steps_per_second": 6.024,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 0.7935212850570679,
+      "grad_norm": 0.9209339618682861,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 5.9192,
+      "loss": 5.9151,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 2.7210114002227783,
+      "grad_norm": 1.6605467796325684,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 5.9058,
+      "loss": 5.9045,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 3.2669694423675537,
+      "grad_norm": 2.7795817852020264,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 5.9026,
+      "loss": 5.9006,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 1.0916407108306885,
+      "grad_norm": 4.178030967712402,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 5.903,
+      "loss": 5.9023,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 2.3075475692749023,
+      "grad_norm": 1.5764979124069214,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 5.9022,
+      "loss": 5.8993,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 2.378257989883423,
+      "grad_norm": 2.2686071395874023,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 5.8981,
+      "loss": 5.895,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 2.6638600826263428,
+      "grad_norm": 0.8200526833534241,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 5.8905,
+      "loss": 5.8859,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 2.0818095207214355,
+      "grad_norm": 2.398494005203247,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 5.8895,
+      "loss": 5.8873,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 1.4235777854919434,
+      "grad_norm": 0.7526465654373169,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 5.8888,
+      "loss": 5.8857,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 1.954344391822815,
+      "grad_norm": 1.0299161672592163,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 5.8908,
+      "loss": 5.8885,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 3.4054806232452393,
+      "grad_norm": 1.3554900884628296,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 5.8871,
+      "loss": 5.8824,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 0.8294370174407959,
+      "grad_norm": 1.76670503616333,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 5.8728,
+      "loss": 5.8708,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 1.288112998008728,
+      "grad_norm": 0.7913095951080322,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 5.8753,
+      "loss": 5.8745,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 0.8261142373085022,
+      "grad_norm": 1.3672707080841064,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 5.8778,
+      "loss": 5.8733,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 0.8744669556617737,
+      "grad_norm": 4.57985782623291,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 5.875,
+      "loss": 5.8701,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 1.528863549232483,
+      "grad_norm": 2.7127416133880615,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 5.8687,
+      "loss": 5.8653,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 1.1342629194259644,
+      "grad_norm": 2.380730628967285,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 5.8654,
+      "loss": 5.8623,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 1.0352425575256348,
+      "grad_norm": 0.7706246376037598,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 5.8615,
+      "loss": 5.8587,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 1.1459453105926514,
+      "grad_norm": 0.7926819920539856,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 5.8617,
+      "loss": 5.8562,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 2.6860058307647705,
+      "grad_norm": 0.7799748778343201,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 5.8616,
+      "loss": 5.8597,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.15170886253399501,
-      "eval_loss": 6.470077037811279,
-      "eval_runtime": 0.8507,
-      "eval_samples_per_second": 563.071,
-      "eval_steps_per_second": 5.878,
+      "eval_accuracy": 0.152409790562705,
+      "eval_loss": 6.339051723480225,
+      "eval_runtime": 0.8379,
+      "eval_samples_per_second": 571.658,
+      "eval_steps_per_second": 5.967,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 0.7595142126083374,
+      "grad_norm": 0.74815434217453,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 5.854,
+      "loss": 5.8511,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 3.7565836906433105,
+      "grad_norm": 0.8203549981117249,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 5.8555,
+      "loss": 5.8536,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 0.786226749420166,
+      "grad_norm": 0.7631138563156128,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 5.8545,
+      "loss": 5.8513,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 1.6939553022384644,
+      "grad_norm": 1.3732026815414429,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 5.8533,
+      "loss": 5.8479,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 1.5326541662216187,
+      "grad_norm": 0.7802448868751526,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 5.8524,
+      "loss": 5.8493,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 2.2361979484558105,
+      "grad_norm": 0.7773926854133606,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 5.8428,
+      "loss": 5.8393,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 0.8266626000404358,
+      "grad_norm": 3.208472967147827,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 5.8471,
+      "loss": 5.8429,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 1.09735906124115,
+      "grad_norm": 3.8855886459350586,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 5.8474,
+      "loss": 5.8431,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 0.7828765511512756,
+      "grad_norm": 0.7867493629455566,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 5.8436,
+      "loss": 5.8394,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 0.7960859537124634,
+      "grad_norm": 0.7768656611442566,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 5.84,
+      "loss": 5.8363,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 2.636533260345459,
+      "grad_norm": 0.7813059687614441,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 5.84,
+      "loss": 5.8362,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 1.344537615776062,
+      "grad_norm": 0.7924099564552307,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 5.8391,
+      "loss": 5.836,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 0.7723442316055298,
+      "grad_norm": 2.308229923248291,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 5.8338,
+      "loss": 5.8307,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 0.7971383333206177,
+      "grad_norm": 0.7893543839454651,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 5.8373,
+      "loss": 5.8332,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 1.1021674871444702,
+      "grad_norm": 1.8878270387649536,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 5.8339,
+      "loss": 5.8305,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 0.784916877746582,
+      "grad_norm": 0.7944318056106567,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 5.8349,
+      "loss": 5.8324,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 0.8498377203941345,
+      "grad_norm": 0.9040911197662354,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 5.832,
+      "loss": 5.8281,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 0.7591010332107544,
+      "grad_norm": 1.0953303575515747,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 5.8316,
+      "loss": 5.8273,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 3.36128568649292,
+      "grad_norm": 1.0026205778121948,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 5.8313,
+      "loss": 5.8271,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 7.6822177215744e+17,
-      "train_loss": 6.218012153119425,
-      "train_runtime": 17148.6644,
-      "train_samples_per_second": 333.319,
-      "train_steps_per_second": 3.473
+      "train_loss": 6.223999219200974,
+      "train_runtime": 16965.8185,
+      "train_samples_per_second": 336.911,
+      "train_steps_per_second": 3.51
     }
   ],
   "logging_steps": 500,