End of training

README.md

@@ -2,11 +2,23 @@
 library_name: transformers
 tags:
 - generated_from_trainer
+datasets:
+- Hartunka/processed_wikitext-103-raw-v1-km-100_v2
 metrics:
 - accuracy
 model-index:
 - name: distilbert_km_100_v2
-  results: []
+  results:
+  - task:
+      name: Masked Language Modeling
+      type: fill-mask
+    dataset:
+      name: Hartunka/processed_wikitext-103-raw-v1-km-100_v2
+      type: Hartunka/processed_wikitext-103-raw-v1-km-100_v2
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.15431994123297735
 ---
 
 <!-- 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_100_v2
 
-This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
+This model is a fine-tuned version of [](https://huggingface.co/) on the Hartunka/processed_wikitext-103-raw-v1-km-100_v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 7.3543
-- Accuracy: 0.1515
+- Loss: 7.0274
+- Accuracy: 0.1543
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15400915409391422,
-    "eval_loss": 7.071955680847168,
-    "eval_runtime": 0.858,
+    "eval_accuracy": 0.15431994123297735,
+    "eval_loss": 7.027431488037109,
+    "eval_runtime": 0.8343,
     "eval_samples": 479,
-    "eval_samples_per_second": 558.288,
-    "eval_steps_per_second": 5.828,
-    "perplexity": 1178.4504550995314,
+    "eval_samples_per_second": 574.149,
+    "eval_steps_per_second": 5.993,
+    "perplexity": 1127.1318374056182,
     "total_flos": 7.69437063436032e+17,
-    "train_loss": 6.748414883681649,
-    "train_runtime": 16891.5237,
+    "train_loss": 6.664753860742679,
+    "train_runtime": 16825.3697,
     "train_samples": 228639,
-    "train_samples_per_second": 338.393,
-    "train_steps_per_second": 3.525
+    "train_samples_per_second": 339.724,
+    "train_steps_per_second": 3.539
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15400915409391422,
-    "eval_loss": 7.071955680847168,
-    "eval_runtime": 0.858,
+    "eval_accuracy": 0.15431994123297735,
+    "eval_loss": 7.027431488037109,
+    "eval_runtime": 0.8343,
     "eval_samples": 479,
-    "eval_samples_per_second": 558.288,
-    "eval_steps_per_second": 5.828,
-    "perplexity": 1178.4504550995314
+    "eval_samples_per_second": 574.149,
+    "eval_steps_per_second": 5.993,
+    "perplexity": 1127.1318374056182
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 7.69437063436032e+17,
-    "train_loss": 6.748414883681649,
-    "train_runtime": 16891.5237,
+    "train_loss": 6.664753860742679,
+    "train_runtime": 16825.3697,
     "train_samples": 228639,
-    "train_samples_per_second": 338.393,
-    "train_steps_per_second": 3.525
+    "train_samples_per_second": 339.724,
+    "train_steps_per_second": 3.539
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 20000,
-  "best_metric": 7.057971954345703,
+  "best_metric": 7.044628143310547,
   "best_model_checkpoint": "distilbert_km_100_v2/checkpoint-20000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.7113592624664307,
+      "grad_norm": 2.784573554992676,
       "learning_rate": 5e-06,
-      "loss": 13.9058,
+      "loss": 13.9113,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 2.375203847885132,
+      "grad_norm": 2.5845043659210205,
       "learning_rate": 1e-05,
-      "loss": 12.1512,
+      "loss": 12.1495,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 4.600935459136963,
+      "grad_norm": 8.000569343566895,
       "learning_rate": 1.5e-05,
-      "loss": 11.0564,
+      "loss": 11.1005,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 8.284174919128418,
+      "grad_norm": 6.192606449127197,
       "learning_rate": 2e-05,
-      "loss": 9.8409,
+      "loss": 10.0606,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 7.562943935394287,
+      "grad_norm": 4.060250282287598,
       "learning_rate": 2.5e-05,
-      "loss": 8.9968,
+      "loss": 9.2999,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 5.516976833343506,
+      "grad_norm": 9.090065002441406,
       "learning_rate": 3e-05,
-      "loss": 8.4646,
+      "loss": 8.8083,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 18.835968017578125,
+      "grad_norm": 7.904158115386963,
       "learning_rate": 3.5e-05,
-      "loss": 8.1656,
+      "loss": 8.4579,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 8.071735382080078,
+      "grad_norm": 10.219331741333008,
       "learning_rate": 4e-05,
-      "loss": 7.9951,
+      "loss": 8.1012,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 14.845528602600098,
+      "grad_norm": 7.001996994018555,
       "learning_rate": 4.5e-05,
-      "loss": 7.8686,
+      "loss": 7.8566,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 5.2275238037109375,
+      "grad_norm": 8.5523681640625,
       "learning_rate": 5e-05,
-      "loss": 7.7211,
+      "loss": 7.6603,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 11.481599807739258,
+      "grad_norm": 4.500087738037109,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 7.6294,
+      "loss": 7.5138,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 16.489280700683594,
+      "grad_norm": 6.297169208526611,
       "learning_rate": 6e-05,
-      "loss": 7.6034,
+      "loss": 7.4642,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 9.843246459960938,
+      "grad_norm": 5.106413841247559,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 7.5598,
+      "loss": 7.4318,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 6.220517158508301,
+      "grad_norm": 5.254331588745117,
       "learning_rate": 7e-05,
-      "loss": 7.5364,
+      "loss": 7.3941,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 10.564053535461426,
+      "grad_norm": 7.277562141418457,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 7.4227,
+      "loss": 7.2441,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 8.94958209991455,
+      "grad_norm": 4.704337120056152,
       "learning_rate": 8e-05,
-      "loss": 7.397,
+      "loss": 7.2148,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 8.731792449951172,
+      "grad_norm": 5.036457538604736,
       "learning_rate": 8.5e-05,
-      "loss": 7.4138,
+      "loss": 7.2291,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 6.825326442718506,
+      "grad_norm": 3.824119806289673,
       "learning_rate": 9e-05,
-      "loss": 7.3904,
+      "loss": 7.2217,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 7.09285306930542,
+      "grad_norm": 4.354274272918701,
       "learning_rate": 9.5e-05,
-      "loss": 7.3691,
+      "loss": 7.2089,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 4.00818395614624,
+      "grad_norm": 4.518379211425781,
       "learning_rate": 0.0001,
-      "loss": 7.2575,
+      "loss": 7.039,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.14894682656723893,
-      "eval_loss": 7.438680648803711,
-      "eval_runtime": 0.8302,
-      "eval_samples_per_second": 577.004,
-      "eval_steps_per_second": 6.023,
+      "eval_accuracy": 0.14967027463201535,
+      "eval_loss": 7.228392124176025,
+      "eval_runtime": 0.8584,
+      "eval_samples_per_second": 558.005,
+      "eval_steps_per_second": 5.825,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 4.358462333679199,
+      "grad_norm": 4.492236614227295,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 7.2655,
+      "loss": 7.0718,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 5.749093532562256,
+      "grad_norm": 4.8100762367248535,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 7.2469,
+      "loss": 7.0643,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 8.624403953552246,
+      "grad_norm": 5.540683746337891,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 7.2402,
+      "loss": 7.0658,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 6.518628120422363,
+      "grad_norm": 4.311294078826904,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 7.1819,
+      "loss": 7.0231,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 4.440794944763184,
+      "grad_norm": 4.135281562805176,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 7.0817,
+      "loss": 6.8828,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 4.726202964782715,
+      "grad_norm": 3.9243323802948,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 7.0918,
+      "loss": 6.8903,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 4.670093536376953,
+      "grad_norm": 3.4225540161132812,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 7.0922,
+      "loss": 6.9057,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 5.7864861488342285,
+      "grad_norm": 4.613020420074463,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 7.0712,
+      "loss": 6.9142,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 5.725551128387451,
+      "grad_norm": 3.643721580505371,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 7.0123,
+      "loss": 6.826,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 5.02437686920166,
+      "grad_norm": 4.394964694976807,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 6.9501,
+      "loss": 6.7422,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 3.9292702674865723,
+      "grad_norm": 4.120483875274658,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 6.9476,
+      "loss": 6.7615,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 5.404318809509277,
+      "grad_norm": 3.474606990814209,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 6.9523,
+      "loss": 6.7768,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 4.438605785369873,
+      "grad_norm": 4.457559585571289,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 6.9565,
+      "loss": 6.7839,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 6.045186519622803,
+      "grad_norm": 4.320551872253418,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 6.8538,
+      "loss": 6.6655,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 3.2410764694213867,
+      "grad_norm": 2.919321298599243,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 6.8083,
+      "loss": 6.6271,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 8.613337516784668,
+      "grad_norm": 3.527904510498047,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 6.8266,
+      "loss": 6.6405,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 5.131848335266113,
+      "grad_norm": 4.277390480041504,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 6.8391,
+      "loss": 6.6603,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 5.321475982666016,
+      "grad_norm": 5.155904769897461,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 6.843,
+      "loss": 6.6706,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 4.39497184753418,
+      "grad_norm": 3.8268680572509766,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 6.7044,
+      "loss": 6.5286,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 4.414360523223877,
+      "grad_norm": 4.085644245147705,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 6.7034,
+      "loss": 6.5358,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15247341395662056,
-      "eval_loss": 7.057971954345703,
-      "eval_runtime": 0.8055,
-      "eval_samples_per_second": 594.669,
-      "eval_steps_per_second": 6.207,
+      "eval_accuracy": 0.1528101237408457,
+      "eval_loss": 7.044628143310547,
+      "eval_runtime": 0.8179,
+      "eval_samples_per_second": 585.662,
+      "eval_steps_per_second": 6.113,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 4.554222583770752,
+      "grad_norm": 3.932213544845581,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 6.7252,
+      "loss": 6.5419,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 4.7603278160095215,
+      "grad_norm": 4.101009368896484,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 6.7325,
+      "loss": 6.5599,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 5.447115898132324,
+      "grad_norm": 4.759427070617676,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 6.7182,
+      "loss": 6.546,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 3.981916904449463,
+      "grad_norm": 5.555412292480469,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 6.5932,
+      "loss": 6.4372,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 4.683881759643555,
+      "grad_norm": 3.8104710578918457,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 6.6058,
+      "loss": 6.4445,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 5.057840824127197,
+      "grad_norm": 3.9375782012939453,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 6.6287,
+      "loss": 6.4583,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 4.702626705169678,
+      "grad_norm": 4.717337131500244,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 6.6288,
+      "loss": 6.4633,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 5.8243889808654785,
+      "grad_norm": 4.42396354675293,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 6.5846,
+      "loss": 6.4254,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 5.50222110748291,
+      "grad_norm": 3.7944576740264893,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 6.511,
+      "loss": 6.3711,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 5.341757774353027,
+      "grad_norm": 4.983055114746094,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 6.5179,
+      "loss": 6.3726,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 4.866272926330566,
+      "grad_norm": 3.8398661613464355,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 6.534,
+      "loss": 6.3855,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 4.987467288970947,
+      "grad_norm": 4.5126633644104,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 6.5462,
+      "loss": 6.3993,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 4.852264404296875,
+      "grad_norm": 4.435245037078857,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 6.467,
+      "loss": 6.3383,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 4.646698474884033,
+      "grad_norm": 4.072259902954102,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 6.4296,
+      "loss": 6.3137,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 4.733177661895752,
+      "grad_norm": 2.745482921600342,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 6.4389,
+      "loss": 6.3158,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 4.592160224914551,
+      "grad_norm": 5.201105117797852,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 6.4479,
+      "loss": 6.321,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 3.777204751968384,
+      "grad_norm": 4.3690996170043945,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 6.4617,
+      "loss": 6.3309,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 7.123958110809326,
+      "grad_norm": 5.218190670013428,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 6.373,
+      "loss": 6.2662,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 4.165009498596191,
+      "grad_norm": 3.5860462188720703,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 6.362,
+      "loss": 6.2577,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 3.9091787338256836,
+      "grad_norm": 3.0204660892486572,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 6.3786,
+      "loss": 6.2663,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.15432185130257156,
-      "eval_loss": 7.214364528656006,
-      "eval_runtime": 0.8062,
-      "eval_samples_per_second": 594.166,
-      "eval_steps_per_second": 6.202,
+      "eval_accuracy": 0.15412597587934074,
+      "eval_loss": 7.245998859405518,
+      "eval_runtime": 0.8149,
+      "eval_samples_per_second": 587.837,
+      "eval_steps_per_second": 6.136,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 6.216394424438477,
+      "grad_norm": 4.847262382507324,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 6.3755,
+      "loss": 6.2671,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 3.730398416519165,
+      "grad_norm": 4.043307781219482,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 6.3806,
+      "loss": 6.2704,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 4.789046764373779,
+      "grad_norm": 2.4291205406188965,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 6.306,
+      "loss": 6.2131,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 5.080028057098389,
+      "grad_norm": 4.489384174346924,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 6.304,
+      "loss": 6.2167,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 4.223598003387451,
+      "grad_norm": 4.288168430328369,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 6.3138,
+      "loss": 6.2246,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 6.034927845001221,
+      "grad_norm": 4.168376445770264,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 6.3206,
+      "loss": 6.2261,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 4.911815643310547,
+      "grad_norm": 3.092984437942505,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 6.2908,
+      "loss": 6.2068,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 3.741004467010498,
+      "grad_norm": 5.309474468231201,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 6.2574,
+      "loss": 6.1782,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 4.319329261779785,
+      "grad_norm": 2.4930341243743896,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 6.2622,
+      "loss": 6.1817,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 5.501757621765137,
+      "grad_norm": 3.5880661010742188,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 6.263,
+      "loss": 6.1836,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 6.313974857330322,
+      "grad_norm": 3.5491909980773926,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 6.2679,
+      "loss": 6.1855,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 3.7311179637908936,
+      "grad_norm": 2.905510425567627,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.2426,
+      "loss": 6.164,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 5.0952253341674805,
+      "grad_norm": 2.551401138305664,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.2246,
+      "loss": 6.1533,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 3.5889272689819336,
+      "grad_norm": 3.9075331687927246,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.2186,
+      "loss": 6.1517,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 3.486255407333374,
+      "grad_norm": 3.4250755310058594,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.2212,
+      "loss": 6.1518,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 7.651794910430908,
+      "grad_norm": 4.045588970184326,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.2207,
+      "loss": 6.1506,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 3.6229660511016846,
+      "grad_norm": 3.8768672943115234,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.1922,
+      "loss": 6.133,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 4.261775970458984,
+      "grad_norm": 3.469980001449585,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.1811,
+      "loss": 6.1212,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 7.182936191558838,
+      "grad_norm": 4.312319755554199,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.1862,
+      "loss": 6.1236,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 5.000805377960205,
+      "grad_norm": 4.129187107086182,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.1779,
+      "loss": 6.1164,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
-      "eval_accuracy": 0.15229337221038466,
-      "eval_loss": 7.225529670715332,
-      "eval_runtime": 0.8199,
-      "eval_samples_per_second": 584.196,
-      "eval_steps_per_second": 6.098,
+      "eval_accuracy": 0.15209711786475272,
+      "eval_loss": 7.413624286651611,
+      "eval_runtime": 0.8127,
+      "eval_samples_per_second": 589.429,
+      "eval_steps_per_second": 6.153,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 4.597228050231934,
+      "grad_norm": 2.001995325088501,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.1861,
+      "loss": 6.1222,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 3.7905569076538086,
+      "grad_norm": 2.0768425464630127,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.1464,
+      "loss": 6.0973,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 4.74383020401001,
+      "grad_norm": 4.364762783050537,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.1477,
+      "loss": 6.092,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 5.506079196929932,
+      "grad_norm": 5.304425239562988,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.1522,
+      "loss": 6.099,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 4.9415459632873535,
+      "grad_norm": 3.797673463821411,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 6.149,
+      "loss": 6.0987,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 4.241745948791504,
+      "grad_norm": 2.320584535598755,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.1407,
+      "loss": 6.0891,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 4.646262168884277,
+      "grad_norm": 3.300384044647217,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.1175,
+      "loss": 6.0756,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 5.132898807525635,
+      "grad_norm": 2.7061753273010254,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.1197,
+      "loss": 6.0726,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 5.060595512390137,
+      "grad_norm": 2.1149325370788574,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.1198,
+      "loss": 6.0748,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 4.880249977111816,
+      "grad_norm": 1.9222161769866943,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.124,
+      "loss": 6.077,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 2.574049711227417,
+      "grad_norm": 2.988523006439209,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.1091,
+      "loss": 6.0632,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 3.6874139308929443,
+      "grad_norm": 2.1261062622070312,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.0918,
+      "loss": 6.0508,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 8.639307975769043,
+      "grad_norm": 4.243496894836426,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.0945,
+      "loss": 6.0515,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 4.701109409332275,
+      "grad_norm": 1.7424923181533813,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.0949,
+      "loss": 6.0543,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 3.732200860977173,
+      "grad_norm": 1.0676500797271729,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.0898,
+      "loss": 6.0494,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 3.8800060749053955,
+      "grad_norm": 5.008332252502441,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.0794,
+      "loss": 6.0417,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 4.853990077972412,
+      "grad_norm": 0.8818482160568237,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.0726,
+      "loss": 6.0369,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 2.7448856830596924,
+      "grad_norm": 1.6551254987716675,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.0735,
+      "loss": 6.033,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 4.7044830322265625,
+      "grad_norm": 3.1230814456939697,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.0695,
+      "loss": 6.033,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 2.968531370162964,
+      "grad_norm": 2.8738510608673096,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.0698,
+      "loss": 6.0315,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.1510640087475818,
-      "eval_loss": 7.457143306732178,
-      "eval_runtime": 0.8096,
-      "eval_samples_per_second": 591.648,
-      "eval_steps_per_second": 6.176,
+      "eval_accuracy": 0.1515406398071046,
+      "eval_loss": 7.354280948638916,
+      "eval_runtime": 0.8362,
+      "eval_samples_per_second": 572.813,
+      "eval_steps_per_second": 5.979,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 4.587925910949707,
+      "grad_norm": 2.0873351097106934,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.0554,
+      "loss": 6.0214,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 4.3473029136657715,
+      "grad_norm": 3.0156331062316895,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.0574,
+      "loss": 6.0204,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 1.9689273834228516,
+      "grad_norm": 1.9316205978393555,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 6.0549,
+      "loss": 6.0164,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 4.508573055267334,
+      "grad_norm": 1.7503081560134888,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 6.0537,
+      "loss": 6.0154,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 4.022706508636475,
+      "grad_norm": 2.066725254058838,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 6.0499,
+      "loss": 6.0162,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 2.755496025085449,
+      "grad_norm": 1.7203088998794556,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 6.0322,
+      "loss": 6.0027,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 3.5307862758636475,
+      "grad_norm": 2.9079740047454834,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 6.0369,
+      "loss": 6.0046,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 1.913621425628662,
+      "grad_norm": 2.9654033184051514,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 6.0388,
+      "loss": 6.008,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 1.4337750673294067,
+      "grad_norm": 1.6890202760696411,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 6.0292,
+      "loss": 6.0001,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 0.9701590538024902,
+      "grad_norm": 1.3355780839920044,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 6.0287,
+      "loss": 5.9987,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 3.845411777496338,
+      "grad_norm": 4.112831115722656,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 6.0263,
+      "loss": 5.994,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 2.6532528400421143,
+      "grad_norm": 4.485779762268066,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 6.0247,
+      "loss": 5.9947,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 3.6103475093841553,
+      "grad_norm": 1.0883957147598267,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 6.0183,
+      "loss": 5.9888,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 3.059880018234253,
+      "grad_norm": 1.0840790271759033,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 6.0192,
+      "loss": 5.99,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 3.420591115951538,
+      "grad_norm": 1.092151403427124,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 6.0133,
+      "loss": 5.9857,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 3.1370186805725098,
+      "grad_norm": 1.402838110923767,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 6.0144,
+      "loss": 5.9872,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 2.226641893386841,
+      "grad_norm": 1.4342875480651855,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 6.01,
+      "loss": 5.984,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 1.4273358583450317,
+      "grad_norm": 0.9019073247909546,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 6.0098,
+      "loss": 5.9823,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 2.529050827026367,
+      "grad_norm": 0.8056707382202148,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 6.0087,
+      "loss": 5.9812,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 7.69437063436032e+17,
-      "train_loss": 6.748414883681649,
-      "train_runtime": 16891.5237,
-      "train_samples_per_second": 338.393,
-      "train_steps_per_second": 3.525
+      "train_loss": 6.664753860742679,
+      "train_runtime": 16825.3697,
+      "train_samples_per_second": 339.724,
+      "train_steps_per_second": 3.539
     }
   ],
   "logging_steps": 500,