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: distilbert_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: 0.15299203254788948
 ---
 
 <!-- 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_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: 14.1213
-- Accuracy: 0.1517
+- Loss: 10.1505
+- Accuracy: 0.1530
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15186189749675086,
-    "eval_loss": 10.148606300354004,
-    "eval_runtime": 0.8211,
+    "eval_accuracy": 0.15299203254788948,
+    "eval_loss": 10.150479316711426,
+    "eval_runtime": 0.8497,
     "eval_samples": 479,
-    "eval_samples_per_second": 583.349,
-    "eval_steps_per_second": 6.089,
-    "perplexity": 25555.460739123107,
+    "eval_samples_per_second": 563.725,
+    "eval_steps_per_second": 5.884,
+    "perplexity": 25603.37138981852,
     "total_flos": 7.68761901614592e+17,
-    "train_loss": 8.060933875813392,
-    "train_runtime": 16751.5356,
+    "train_loss": 8.045040057203932,
+    "train_runtime": 16960.8581,
     "train_samples": 228639,
-    "train_samples_per_second": 341.221,
-    "train_steps_per_second": 3.555
+    "train_samples_per_second": 337.01,
+    "train_steps_per_second": 3.511
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 25.0,
-    "eval_accuracy": 0.15186189749675086,
-    "eval_loss": 10.148606300354004,
-    "eval_runtime": 0.8211,
+    "eval_accuracy": 0.15299203254788948,
+    "eval_loss": 10.150479316711426,
+    "eval_runtime": 0.8497,
     "eval_samples": 479,
-    "eval_samples_per_second": 583.349,
-    "eval_steps_per_second": 6.089,
-    "perplexity": 25555.460739123107
+    "eval_samples_per_second": 563.725,
+    "eval_steps_per_second": 5.884,
+    "perplexity": 25603.37138981852
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 25.0,
     "total_flos": 7.68761901614592e+17,
-    "train_loss": 8.060933875813392,
-    "train_runtime": 16751.5356,
+    "train_loss": 8.045040057203932,
+    "train_runtime": 16960.8581,
     "train_samples": 228639,
-    "train_samples_per_second": 341.221,
-    "train_steps_per_second": 3.555
+    "train_samples_per_second": 337.01,
+    "train_steps_per_second": 3.511
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 10000,
-  "best_metric": 10.124896049499512,
+  "best_metric": 10.129691123962402,
   "best_model_checkpoint": "distilbert_rand_50_v2/checkpoint-10000",
   "epoch": 25.0,
   "eval_steps": 10000,
@@ -11,890 +11,890 @@
   "log_history": [
     {
       "epoch": 0.20990764063811923,
-      "grad_norm": 2.635725736618042,
+      "grad_norm": 2.654482841491699,
       "learning_rate": 5e-06,
-      "loss": 13.2342,
+      "loss": 13.2356,
       "step": 500
     },
     {
       "epoch": 0.41981528127623846,
-      "grad_norm": 1.89760422706604,
+      "grad_norm": 1.8086973428726196,
       "learning_rate": 1e-05,
-      "loss": 11.5227,
+      "loss": 11.5224,
       "step": 1000
     },
     {
       "epoch": 0.6297229219143576,
-      "grad_norm": 2.041102647781372,
+      "grad_norm": 1.734184741973877,
       "learning_rate": 1.5e-05,
-      "loss": 10.9869,
+      "loss": 10.9868,
       "step": 1500
     },
     {
       "epoch": 0.8396305625524769,
-      "grad_norm": 2.3095645904541016,
+      "grad_norm": 2.2412476539611816,
       "learning_rate": 2e-05,
-      "loss": 10.8348,
+      "loss": 10.8357,
       "step": 2000
     },
     {
       "epoch": 1.0495382031905962,
-      "grad_norm": 2.2821314334869385,
+      "grad_norm": 2.0357539653778076,
       "learning_rate": 2.5e-05,
-      "loss": 10.703,
+      "loss": 10.7021,
       "step": 2500
     },
     {
       "epoch": 1.2594458438287153,
-      "grad_norm": 2.2018096446990967,
+      "grad_norm": 2.3354532718658447,
       "learning_rate": 3e-05,
-      "loss": 10.5867,
+      "loss": 10.5888,
       "step": 3000
     },
     {
       "epoch": 1.4693534844668346,
-      "grad_norm": 2.673328399658203,
+      "grad_norm": 2.455918550491333,
       "learning_rate": 3.5e-05,
-      "loss": 10.4911,
+      "loss": 10.4955,
       "step": 3500
     },
     {
       "epoch": 1.6792611251049538,
-      "grad_norm": 2.457890748977661,
+      "grad_norm": 2.2312426567077637,
       "learning_rate": 4e-05,
-      "loss": 10.4145,
+      "loss": 10.4175,
       "step": 4000
     },
     {
       "epoch": 1.8891687657430731,
-      "grad_norm": 2.067857027053833,
+      "grad_norm": 2.462714910507202,
       "learning_rate": 4.5e-05,
-      "loss": 10.3571,
+      "loss": 10.3589,
       "step": 4500
     },
     {
       "epoch": 2.0990764063811924,
-      "grad_norm": 2.4149136543273926,
+      "grad_norm": 2.400935649871826,
       "learning_rate": 5e-05,
-      "loss": 10.308,
+      "loss": 10.3095,
       "step": 5000
     },
     {
       "epoch": 2.3089840470193117,
-      "grad_norm": 2.607347011566162,
+      "grad_norm": 2.179032564163208,
       "learning_rate": 5.500000000000001e-05,
-      "loss": 10.2673,
+      "loss": 10.2702,
       "step": 5500
     },
     {
       "epoch": 2.5188916876574305,
-      "grad_norm": 2.360692024230957,
+      "grad_norm": 2.4195713996887207,
       "learning_rate": 6e-05,
-      "loss": 10.2338,
+      "loss": 10.2365,
       "step": 6000
     },
     {
       "epoch": 2.72879932829555,
-      "grad_norm": 2.4523544311523438,
+      "grad_norm": 2.730342149734497,
       "learning_rate": 6.500000000000001e-05,
-      "loss": 10.2053,
+      "loss": 10.2047,
       "step": 6500
     },
     {
       "epoch": 2.938706968933669,
-      "grad_norm": 3.0250744819641113,
+      "grad_norm": 2.8091626167297363,
       "learning_rate": 7e-05,
-      "loss": 10.1799,
+      "loss": 10.1773,
       "step": 7000
     },
     {
       "epoch": 3.1486146095717884,
-      "grad_norm": 2.492704391479492,
+      "grad_norm": 2.63844895362854,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 10.1207,
+      "loss": 10.1273,
       "step": 7500
     },
     {
       "epoch": 3.3585222502099077,
-      "grad_norm": 2.93556809425354,
+      "grad_norm": 2.6782772541046143,
       "learning_rate": 8e-05,
-      "loss": 10.0885,
+      "loss": 10.0974,
       "step": 8000
     },
     {
       "epoch": 3.568429890848027,
-      "grad_norm": 2.849184513092041,
+      "grad_norm": 2.6178882122039795,
       "learning_rate": 8.5e-05,
-      "loss": 10.0792,
+      "loss": 10.0834,
       "step": 8500
     },
     {
       "epoch": 3.7783375314861463,
-      "grad_norm": 3.0643296241760254,
+      "grad_norm": 2.458190679550171,
       "learning_rate": 9e-05,
-      "loss": 10.0607,
+      "loss": 10.0643,
       "step": 9000
     },
     {
       "epoch": 3.988245172124265,
-      "grad_norm": 2.7466018199920654,
+      "grad_norm": 2.828200340270996,
       "learning_rate": 9.5e-05,
-      "loss": 10.0499,
+      "loss": 10.0514,
       "step": 9500
     },
     {
       "epoch": 4.198152812762385,
-      "grad_norm": 3.5570523738861084,
+      "grad_norm": 3.316427707672119,
       "learning_rate": 0.0001,
-      "loss": 9.8831,
+      "loss": 9.8946,
       "step": 10000
     },
     {
       "epoch": 4.198152812762385,
-      "eval_accuracy": 0.15106152091043157,
-      "eval_loss": 10.124896049499512,
-      "eval_runtime": 0.8474,
-      "eval_samples_per_second": 565.247,
-      "eval_steps_per_second": 5.9,
+      "eval_accuracy": 0.1506441470269067,
+      "eval_loss": 10.129691123962402,
+      "eval_runtime": 0.8628,
+      "eval_samples_per_second": 555.15,
+      "eval_steps_per_second": 5.795,
       "step": 10000
     },
     {
       "epoch": 4.408060453400504,
-      "grad_norm": 3.0391769409179688,
+      "grad_norm": 3.1053104400634766,
       "learning_rate": 9.899091826437942e-05,
-      "loss": 9.8913,
+      "loss": 9.9001,
       "step": 10500
     },
     {
       "epoch": 4.617968094038623,
-      "grad_norm": 3.57462739944458,
+      "grad_norm": 3.395521402359009,
       "learning_rate": 9.798183652875883e-05,
-      "loss": 9.8939,
+      "loss": 9.903,
       "step": 11000
     },
     {
       "epoch": 4.827875734676742,
-      "grad_norm": 3.1066319942474365,
+      "grad_norm": 3.172823190689087,
       "learning_rate": 9.697275479313825e-05,
-      "loss": 9.9055,
+      "loss": 9.9128,
       "step": 11500
     },
     {
       "epoch": 5.037783375314861,
-      "grad_norm": 3.334506034851074,
+      "grad_norm": 3.802762269973755,
       "learning_rate": 9.596367305751767e-05,
-      "loss": 9.854,
+      "loss": 9.8554,
       "step": 12000
     },
     {
       "epoch": 5.247691015952981,
-      "grad_norm": 3.3224949836730957,
+      "grad_norm": 3.419677972793579,
       "learning_rate": 9.495459132189707e-05,
-      "loss": 9.677,
+      "loss": 9.6843,
       "step": 12500
     },
     {
       "epoch": 5.4575986565911,
-      "grad_norm": 3.5245678424835205,
+      "grad_norm": 3.1203525066375732,
       "learning_rate": 9.394550958627649e-05,
-      "loss": 9.7049,
+      "loss": 9.7106,
       "step": 13000
     },
     {
       "epoch": 5.667506297229219,
-      "grad_norm": 3.4072303771972656,
+      "grad_norm": 3.380894899368286,
       "learning_rate": 9.293642785065591e-05,
-      "loss": 9.7342,
+      "loss": 9.7358,
       "step": 13500
     },
     {
       "epoch": 5.877413937867338,
-      "grad_norm": 3.344393014907837,
+      "grad_norm": 3.52885103225708,
       "learning_rate": 9.192734611503532e-05,
-      "loss": 9.7428,
+      "loss": 9.7491,
       "step": 14000
     },
     {
       "epoch": 6.087321578505458,
-      "grad_norm": 3.4316670894622803,
+      "grad_norm": 3.5351815223693848,
       "learning_rate": 9.091826437941474e-05,
-      "loss": 9.6339,
+      "loss": 9.6395,
       "step": 14500
     },
     {
       "epoch": 6.297229219143577,
-      "grad_norm": 3.4701027870178223,
+      "grad_norm": 3.486654281616211,
       "learning_rate": 8.990918264379415e-05,
-      "loss": 9.5216,
+      "loss": 9.528,
       "step": 15000
     },
     {
       "epoch": 6.507136859781696,
-      "grad_norm": 3.419870376586914,
+      "grad_norm": 3.5609726905822754,
       "learning_rate": 8.890010090817356e-05,
-      "loss": 9.556,
+      "loss": 9.5576,
       "step": 15500
     },
     {
       "epoch": 6.717044500419815,
-      "grad_norm": 3.671058416366577,
+      "grad_norm": 3.4209423065185547,
       "learning_rate": 8.789101917255298e-05,
-      "loss": 9.5815,
+      "loss": 9.5892,
       "step": 16000
     },
     {
       "epoch": 6.926952141057934,
-      "grad_norm": 3.951993703842163,
+      "grad_norm": 3.8070693016052246,
       "learning_rate": 8.68819374369324e-05,
-      "loss": 9.6018,
+      "loss": 9.6046,
       "step": 16500
     },
     {
       "epoch": 7.136859781696054,
-      "grad_norm": 3.774444580078125,
+      "grad_norm": 3.740973711013794,
       "learning_rate": 8.58728557013118e-05,
-      "loss": 9.3935,
+      "loss": 9.4002,
       "step": 17000
     },
     {
       "epoch": 7.346767422334173,
-      "grad_norm": 4.343321323394775,
+      "grad_norm": 4.353114604949951,
       "learning_rate": 8.486377396569122e-05,
-      "loss": 9.3104,
+      "loss": 9.3255,
       "step": 17500
     },
     {
       "epoch": 7.5566750629722925,
-      "grad_norm": 4.312654972076416,
+      "grad_norm": 3.9671778678894043,
       "learning_rate": 8.385469223007064e-05,
-      "loss": 9.344,
+      "loss": 9.3593,
       "step": 18000
     },
     {
       "epoch": 7.766582703610411,
-      "grad_norm": 4.411766052246094,
+      "grad_norm": 4.088095664978027,
       "learning_rate": 8.284561049445006e-05,
-      "loss": 9.3802,
+      "loss": 9.3918,
       "step": 18500
     },
     {
       "epoch": 7.97649034424853,
-      "grad_norm": 4.005197525024414,
+      "grad_norm": 4.308635234832764,
       "learning_rate": 8.183652875882947e-05,
-      "loss": 9.3985,
+      "loss": 9.4039,
       "step": 19000
     },
     {
       "epoch": 8.18639798488665,
-      "grad_norm": 4.736388683319092,
+      "grad_norm": 4.793691635131836,
       "learning_rate": 8.082744702320889e-05,
-      "loss": 9.0333,
+      "loss": 9.0403,
       "step": 19500
     },
     {
       "epoch": 8.39630562552477,
-      "grad_norm": 4.861172199249268,
+      "grad_norm": 5.178264617919922,
       "learning_rate": 7.98183652875883e-05,
-      "loss": 9.0198,
+      "loss": 9.0265,
       "step": 20000
     },
     {
       "epoch": 8.39630562552477,
-      "eval_accuracy": 0.15210864502371,
-      "eval_loss": 10.614727020263672,
-      "eval_runtime": 0.8067,
-      "eval_samples_per_second": 593.8,
-      "eval_steps_per_second": 6.198,
+      "eval_accuracy": 0.1519402901315974,
+      "eval_loss": 10.667503356933594,
+      "eval_runtime": 0.818,
+      "eval_samples_per_second": 585.587,
+      "eval_steps_per_second": 6.113,
       "step": 20000
     },
     {
       "epoch": 8.606213266162888,
-      "grad_norm": 5.332311153411865,
+      "grad_norm": 5.22618293762207,
       "learning_rate": 7.880928355196771e-05,
-      "loss": 9.0463,
+      "loss": 9.0511,
       "step": 20500
     },
     {
       "epoch": 8.816120906801007,
-      "grad_norm": 5.440824031829834,
+      "grad_norm": 5.140614986419678,
       "learning_rate": 7.780020181634713e-05,
-      "loss": 9.087,
+      "loss": 9.0922,
       "step": 21000
     },
     {
       "epoch": 9.026028547439127,
-      "grad_norm": 6.202512741088867,
+      "grad_norm": 6.019027233123779,
       "learning_rate": 7.679112008072655e-05,
-      "loss": 9.0257,
+      "loss": 9.0378,
       "step": 21500
     },
     {
       "epoch": 9.235936188077247,
-      "grad_norm": 5.817151069641113,
+      "grad_norm": 5.943739891052246,
       "learning_rate": 7.578203834510595e-05,
-      "loss": 8.6126,
+      "loss": 8.6197,
       "step": 22000
     },
     {
       "epoch": 9.445843828715365,
-      "grad_norm": 6.3854451179504395,
+      "grad_norm": 6.256453037261963,
       "learning_rate": 7.477295660948537e-05,
-      "loss": 8.6487,
+      "loss": 8.644,
       "step": 22500
     },
     {
       "epoch": 9.655751469353484,
-      "grad_norm": 6.242320537567139,
+      "grad_norm": 6.752525806427002,
       "learning_rate": 7.376387487386479e-05,
-      "loss": 8.6954,
+      "loss": 8.7002,
       "step": 23000
     },
     {
       "epoch": 9.865659109991604,
-      "grad_norm": 7.406656265258789,
+      "grad_norm": 6.760085105895996,
       "learning_rate": 7.27547931382442e-05,
-      "loss": 8.7105,
+      "loss": 8.7088,
       "step": 23500
     },
     {
       "epoch": 10.075566750629722,
-      "grad_norm": 7.50529146194458,
+      "grad_norm": 6.823685646057129,
       "learning_rate": 7.174571140262362e-05,
-      "loss": 8.5476,
+      "loss": 8.5395,
       "step": 24000
     },
     {
       "epoch": 10.285474391267842,
-      "grad_norm": 6.765026092529297,
+      "grad_norm": 6.6949028968811035,
       "learning_rate": 7.073662966700304e-05,
-      "loss": 8.2554,
+      "loss": 8.2398,
       "step": 24500
     },
     {
       "epoch": 10.495382031905962,
-      "grad_norm": 7.266622543334961,
+      "grad_norm": 6.399642467498779,
       "learning_rate": 6.972754793138244e-05,
-      "loss": 8.2974,
+      "loss": 8.2619,
       "step": 25000
     },
     {
       "epoch": 10.705289672544081,
-      "grad_norm": 7.264074802398682,
+      "grad_norm": 7.6814494132995605,
       "learning_rate": 6.871846619576186e-05,
-      "loss": 8.3286,
+      "loss": 8.3142,
       "step": 25500
     },
     {
       "epoch": 10.9151973131822,
-      "grad_norm": 7.223959922790527,
+      "grad_norm": 7.36984920501709,
       "learning_rate": 6.770938446014128e-05,
-      "loss": 8.3643,
+      "loss": 8.3451,
       "step": 26000
     },
     {
       "epoch": 11.125104953820319,
-      "grad_norm": 7.5548014640808105,
+      "grad_norm": 7.151028156280518,
       "learning_rate": 6.670030272452068e-05,
-      "loss": 8.0772,
+      "loss": 8.0434,
       "step": 26500
     },
     {
       "epoch": 11.335012594458439,
-      "grad_norm": 7.976690292358398,
+      "grad_norm": 8.322630882263184,
       "learning_rate": 6.56912209889001e-05,
-      "loss": 7.9346,
+      "loss": 7.8841,
       "step": 27000
     },
     {
       "epoch": 11.544920235096557,
-      "grad_norm": 6.7330098152160645,
+      "grad_norm": 7.856650352478027,
       "learning_rate": 6.468213925327952e-05,
-      "loss": 7.9736,
+      "loss": 7.9312,
       "step": 27500
     },
     {
       "epoch": 11.754827875734676,
-      "grad_norm": 8.504254341125488,
+      "grad_norm": 8.230381965637207,
       "learning_rate": 6.367305751765893e-05,
-      "loss": 7.9997,
+      "loss": 7.9579,
       "step": 28000
     },
     {
       "epoch": 11.964735516372796,
-      "grad_norm": 8.328359603881836,
+      "grad_norm": 7.677419185638428,
       "learning_rate": 6.266397578203835e-05,
-      "loss": 8.0404,
+      "loss": 7.9988,
       "step": 28500
     },
     {
       "epoch": 12.174643157010916,
-      "grad_norm": 8.02249526977539,
+      "grad_norm": 8.350260734558105,
       "learning_rate": 6.165489404641777e-05,
-      "loss": 7.6489,
+      "loss": 7.5902,
       "step": 29000
     },
     {
       "epoch": 12.384550797649034,
-      "grad_norm": 9.46779727935791,
+      "grad_norm": 9.383851051330566,
       "learning_rate": 6.064581231079718e-05,
-      "loss": 7.5991,
+      "loss": 7.5395,
       "step": 29500
     },
     {
       "epoch": 12.594458438287154,
-      "grad_norm": 8.518192291259766,
+      "grad_norm": 9.167280197143555,
       "learning_rate": 5.963673057517659e-05,
-      "loss": 7.6593,
+      "loss": 7.5938,
       "step": 30000
     },
     {
       "epoch": 12.594458438287154,
-      "eval_accuracy": 0.1537062428295604,
-      "eval_loss": 12.006357192993164,
-      "eval_runtime": 0.8089,
-      "eval_samples_per_second": 592.159,
-      "eval_steps_per_second": 6.181,
+      "eval_accuracy": 0.1528667767299997,
+      "eval_loss": 12.15234088897705,
+      "eval_runtime": 0.8363,
+      "eval_samples_per_second": 572.732,
+      "eval_steps_per_second": 5.978,
       "step": 30000
     },
     {
       "epoch": 12.804366078925273,
-      "grad_norm": 9.009134292602539,
+      "grad_norm": 9.172540664672852,
       "learning_rate": 5.8627648839556e-05,
-      "loss": 7.7002,
+      "loss": 7.6402,
       "step": 30500
     },
     {
       "epoch": 13.014273719563391,
-      "grad_norm": 9.298676490783691,
+      "grad_norm": 9.858193397521973,
       "learning_rate": 5.7618567103935416e-05,
-      "loss": 7.6841,
+      "loss": 7.6299,
       "step": 31000
     },
     {
       "epoch": 13.224181360201511,
-      "grad_norm": 8.665403366088867,
+      "grad_norm": 8.385052680969238,
       "learning_rate": 5.660948536831484e-05,
-      "loss": 7.2722,
+      "loss": 7.1953,
       "step": 31500
     },
     {
       "epoch": 13.43408900083963,
-      "grad_norm": 9.538726806640625,
+      "grad_norm": 10.600648880004883,
       "learning_rate": 5.5600403632694254e-05,
-      "loss": 7.3275,
+      "loss": 7.2477,
       "step": 32000
     },
     {
       "epoch": 13.64399664147775,
-      "grad_norm": 8.789461135864258,
+      "grad_norm": 9.292642593383789,
       "learning_rate": 5.4591321897073666e-05,
-      "loss": 7.3716,
+      "loss": 7.3049,
       "step": 32500
     },
     {
       "epoch": 13.853904282115868,
-      "grad_norm": 8.967900276184082,
+      "grad_norm": 9.895312309265137,
       "learning_rate": 5.358224016145308e-05,
-      "loss": 7.3807,
+      "loss": 7.3147,
       "step": 33000
     },
     {
       "epoch": 14.063811922753988,
-      "grad_norm": 8.663387298583984,
+      "grad_norm": 9.32699203491211,
       "learning_rate": 5.257315842583249e-05,
-      "loss": 7.2787,
+      "loss": 7.2173,
       "step": 33500
     },
     {
       "epoch": 14.273719563392108,
-      "grad_norm": 9.423514366149902,
+      "grad_norm": 8.1248779296875,
       "learning_rate": 5.15640766902119e-05,
-      "loss": 7.011,
+      "loss": 6.9594,
       "step": 34000
     },
     {
       "epoch": 14.483627204030226,
-      "grad_norm": 9.283555030822754,
+      "grad_norm": 9.73206615447998,
       "learning_rate": 5.055499495459133e-05,
-      "loss": 7.0583,
+      "loss": 6.9976,
       "step": 34500
     },
     {
       "epoch": 14.693534844668346,
-      "grad_norm": 10.366921424865723,
+      "grad_norm": 10.707878112792969,
       "learning_rate": 4.954591321897074e-05,
-      "loss": 7.0982,
+      "loss": 7.0311,
       "step": 35000
     },
     {
       "epoch": 14.903442485306465,
-      "grad_norm": 9.820926666259766,
+      "grad_norm": 10.610901832580566,
       "learning_rate": 4.853683148335015e-05,
-      "loss": 7.1108,
+      "loss": 7.0584,
       "step": 35500
     },
     {
       "epoch": 15.113350125944585,
-      "grad_norm": 9.178672790527344,
+      "grad_norm": 8.190956115722656,
       "learning_rate": 4.7527749747729566e-05,
-      "loss": 6.9524,
+      "loss": 6.8945,
       "step": 36000
     },
     {
       "epoch": 15.323257766582703,
-      "grad_norm": 8.867226600646973,
+      "grad_norm": 11.0116605758667,
       "learning_rate": 4.6518668012108985e-05,
-      "loss": 6.8192,
+      "loss": 6.77,
       "step": 36500
     },
     {
       "epoch": 15.533165407220823,
-      "grad_norm": 9.827154159545898,
+      "grad_norm": 10.029074668884277,
       "learning_rate": 4.55095862764884e-05,
-      "loss": 6.858,
+      "loss": 6.8023,
       "step": 37000
     },
     {
       "epoch": 15.743073047858942,
-      "grad_norm": 11.180596351623535,
+      "grad_norm": 9.985063552856445,
       "learning_rate": 4.450050454086781e-05,
-      "loss": 6.8874,
+      "loss": 6.8255,
       "step": 37500
     },
     {
       "epoch": 15.95298068849706,
-      "grad_norm": 11.212045669555664,
+      "grad_norm": 12.623684883117676,
       "learning_rate": 4.349142280524723e-05,
-      "loss": 6.9024,
+      "loss": 6.8486,
       "step": 38000
     },
     {
       "epoch": 16.162888329135182,
-      "grad_norm": 9.407392501831055,
+      "grad_norm": 9.287301063537598,
       "learning_rate": 4.248234106962664e-05,
-      "loss": 6.7053,
+      "loss": 6.6617,
       "step": 38500
     },
     {
       "epoch": 16.3727959697733,
-      "grad_norm": 8.527677536010742,
+      "grad_norm": 10.953415870666504,
       "learning_rate": 4.147325933400605e-05,
-      "loss": 6.6778,
+      "loss": 6.6503,
       "step": 39000
     },
     {
       "epoch": 16.582703610411418,
-      "grad_norm": 8.391467094421387,
+      "grad_norm": 10.664518356323242,
       "learning_rate": 4.046417759838547e-05,
-      "loss": 6.6931,
+      "loss": 6.6505,
       "step": 39500
     },
     {
       "epoch": 16.79261125104954,
-      "grad_norm": 10.770227432250977,
+      "grad_norm": 9.114075660705566,
       "learning_rate": 3.9455095862764884e-05,
-      "loss": 6.7206,
+      "loss": 6.6753,
       "step": 40000
     },
     {
       "epoch": 16.79261125104954,
       "eval_accuracy": 0.15156442749803745,
-      "eval_loss": 13.526458740234375,
-      "eval_runtime": 0.8055,
-      "eval_samples_per_second": 594.69,
-      "eval_steps_per_second": 6.208,
+      "eval_loss": 13.50600814819336,
+      "eval_runtime": 0.8439,
+      "eval_samples_per_second": 567.59,
+      "eval_steps_per_second": 5.925,
       "step": 40000
     },
     {
       "epoch": 17.002518891687657,
-      "grad_norm": 7.6711578369140625,
+      "grad_norm": 7.26846981048584,
       "learning_rate": 3.84460141271443e-05,
-      "loss": 6.736,
+      "loss": 6.6952,
       "step": 40500
     },
     {
       "epoch": 17.212426532325775,
-      "grad_norm": 6.2812676429748535,
+      "grad_norm": 6.937128067016602,
       "learning_rate": 3.7436932391523715e-05,
-      "loss": 6.5574,
+      "loss": 6.5264,
       "step": 41000
     },
     {
       "epoch": 17.422334172963897,
-      "grad_norm": 9.812127113342285,
+      "grad_norm": 9.404059410095215,
       "learning_rate": 3.642785065590313e-05,
-      "loss": 6.5741,
+      "loss": 6.5351,
       "step": 41500
     },
     {
       "epoch": 17.632241813602015,
-      "grad_norm": 8.351988792419434,
+      "grad_norm": 9.156888008117676,
       "learning_rate": 3.541876892028255e-05,
-      "loss": 6.5801,
+      "loss": 6.558,
       "step": 42000
     },
     {
       "epoch": 17.842149454240133,
-      "grad_norm": 9.75585651397705,
+      "grad_norm": 7.921253681182861,
       "learning_rate": 3.440968718466196e-05,
-      "loss": 6.6068,
+      "loss": 6.5764,
       "step": 42500
     },
     {
       "epoch": 18.052057094878254,
-      "grad_norm": 7.7363457679748535,
+      "grad_norm": 8.91440200805664,
       "learning_rate": 3.340060544904137e-05,
-      "loss": 6.58,
+      "loss": 6.5503,
       "step": 43000
     },
     {
       "epoch": 18.261964735516372,
-      "grad_norm": 7.955825328826904,
+      "grad_norm": 9.434792518615723,
       "learning_rate": 3.239152371342079e-05,
-      "loss": 6.4826,
+      "loss": 6.4687,
       "step": 43500
     },
     {
       "epoch": 18.471872376154494,
-      "grad_norm": 7.37641716003418,
+      "grad_norm": 7.329489707946777,
       "learning_rate": 3.13824419778002e-05,
-      "loss": 6.4953,
+      "loss": 6.4783,
       "step": 44000
     },
     {
       "epoch": 18.68178001679261,
-      "grad_norm": 8.995388984680176,
+      "grad_norm": 9.097280502319336,
       "learning_rate": 3.0373360242179615e-05,
-      "loss": 6.5129,
+      "loss": 6.4916,
       "step": 44500
     },
     {
       "epoch": 18.89168765743073,
-      "grad_norm": 10.685982704162598,
+      "grad_norm": 7.823692321777344,
       "learning_rate": 2.9364278506559034e-05,
-      "loss": 6.528,
+      "loss": 6.5121,
       "step": 45000
     },
     {
       "epoch": 19.10159529806885,
-      "grad_norm": 9.09445858001709,
+      "grad_norm": 9.118895530700684,
       "learning_rate": 2.8355196770938446e-05,
-      "loss": 6.4693,
+      "loss": 6.4479,
       "step": 45500
     },
     {
       "epoch": 19.31150293870697,
-      "grad_norm": 8.21455192565918,
+      "grad_norm": 8.967138290405273,
       "learning_rate": 2.734611503531786e-05,
-      "loss": 6.4292,
+      "loss": 6.418,
       "step": 46000
     },
     {
       "epoch": 19.521410579345087,
-      "grad_norm": 11.054938316345215,
+      "grad_norm": 9.329202651977539,
       "learning_rate": 2.6337033299697278e-05,
-      "loss": 6.4354,
+      "loss": 6.4214,
       "step": 46500
     },
     {
       "epoch": 19.73131821998321,
-      "grad_norm": 9.078868865966797,
+      "grad_norm": 9.571353912353516,
       "learning_rate": 2.532795156407669e-05,
-      "loss": 6.4375,
+      "loss": 6.4222,
       "step": 47000
     },
     {
       "epoch": 19.941225860621326,
-      "grad_norm": 9.143745422363281,
+      "grad_norm": 6.103086948394775,
       "learning_rate": 2.4318869828456106e-05,
-      "loss": 6.4416,
+      "loss": 6.426,
       "step": 47500
     },
     {
       "epoch": 20.151133501259444,
-      "grad_norm": 7.7672576904296875,
+      "grad_norm": 7.004849433898926,
       "learning_rate": 2.330978809283552e-05,
-      "loss": 6.4051,
+      "loss": 6.3975,
       "step": 48000
     },
     {
       "epoch": 20.361041141897566,
-      "grad_norm": 7.813689708709717,
+      "grad_norm": 9.450573921203613,
       "learning_rate": 2.2300706357214934e-05,
-      "loss": 6.3898,
+      "loss": 6.3794,
       "step": 48500
     },
     {
       "epoch": 20.570948782535684,
-      "grad_norm": 6.870074272155762,
+      "grad_norm": 5.927501678466797,
       "learning_rate": 2.129162462159435e-05,
-      "loss": 6.3815,
+      "loss": 6.3728,
       "step": 49000
     },
     {
       "epoch": 20.7808564231738,
-      "grad_norm": 6.914427757263184,
+      "grad_norm": 7.504936695098877,
       "learning_rate": 2.0282542885973765e-05,
-      "loss": 6.38,
+      "loss": 6.375,
       "step": 49500
     },
     {
       "epoch": 20.990764063811923,
-      "grad_norm": 6.275176525115967,
+      "grad_norm": 5.263129711151123,
       "learning_rate": 1.927346115035318e-05,
-      "loss": 6.3962,
+      "loss": 6.3865,
       "step": 50000
     },
     {
       "epoch": 20.990764063811923,
-      "eval_accuracy": 0.150867748899543,
-      "eval_loss": 14.363707542419434,
-      "eval_runtime": 0.8215,
-      "eval_samples_per_second": 583.072,
-      "eval_steps_per_second": 6.086,
+      "eval_accuracy": 0.1516527882916982,
+      "eval_loss": 14.12132453918457,
+      "eval_runtime": 0.8368,
+      "eval_samples_per_second": 572.413,
+      "eval_steps_per_second": 5.975,
       "step": 50000
     },
     {
       "epoch": 21.20067170445004,
-      "grad_norm": 7.347343921661377,
+      "grad_norm": 5.663426876068115,
       "learning_rate": 1.8264379414732593e-05,
-      "loss": 6.3426,
+      "loss": 6.3381,
       "step": 50500
     },
     {
       "epoch": 21.410579345088163,
-      "grad_norm": 8.971973419189453,
+      "grad_norm": 7.3178181648254395,
       "learning_rate": 1.725529767911201e-05,
-      "loss": 6.3554,
+      "loss": 6.3551,
       "step": 51000
     },
     {
       "epoch": 21.62048698572628,
-      "grad_norm": 6.905628681182861,
+      "grad_norm": 8.326308250427246,
       "learning_rate": 1.6246215943491424e-05,
-      "loss": 6.3459,
+      "loss": 6.3379,
       "step": 51500
     },
     {
       "epoch": 21.8303946263644,
-      "grad_norm": 5.673792839050293,
+      "grad_norm": 8.359925270080566,
       "learning_rate": 1.5237134207870837e-05,
-      "loss": 6.345,
+      "loss": 6.3383,
       "step": 52000
     },
     {
       "epoch": 22.04030226700252,
-      "grad_norm": 5.387075424194336,
+      "grad_norm": 3.8704984188079834,
       "learning_rate": 1.4228052472250252e-05,
-      "loss": 6.345,
+      "loss": 6.3344,
       "step": 52500
     },
     {
       "epoch": 22.250209907640638,
-      "grad_norm": 5.177424907684326,
+      "grad_norm": 5.124746799468994,
       "learning_rate": 1.3218970736629668e-05,
-      "loss": 6.3094,
+      "loss": 6.3043,
       "step": 53000
     },
     {
       "epoch": 22.460117548278756,
-      "grad_norm": 7.82404088973999,
+      "grad_norm": 8.9225492477417,
       "learning_rate": 1.2209889001009082e-05,
-      "loss": 6.3139,
+      "loss": 6.3044,
       "step": 53500
     },
     {
       "epoch": 22.670025188916878,
-      "grad_norm": 8.11977481842041,
+      "grad_norm": 6.467714786529541,
       "learning_rate": 1.1200807265388496e-05,
-      "loss": 6.3222,
+      "loss": 6.3148,
       "step": 54000
     },
     {
       "epoch": 22.879932829554996,
-      "grad_norm": 6.464744567871094,
+      "grad_norm": 5.8907012939453125,
       "learning_rate": 1.0191725529767912e-05,
-      "loss": 6.3058,
+      "loss": 6.3029,
       "step": 54500
     },
     {
       "epoch": 23.089840470193113,
-      "grad_norm": 6.440338611602783,
+      "grad_norm": 5.726820468902588,
       "learning_rate": 9.182643794147326e-06,
-      "loss": 6.3025,
+      "loss": 6.2945,
       "step": 55000
     },
     {
       "epoch": 23.299748110831235,
-      "grad_norm": 7.113975524902344,
+      "grad_norm": 8.676447868347168,
       "learning_rate": 8.173562058526741e-06,
-      "loss": 6.2824,
+      "loss": 6.2752,
       "step": 55500
     },
     {
       "epoch": 23.509655751469353,
-      "grad_norm": 7.145349979400635,
+      "grad_norm": 6.037963390350342,
       "learning_rate": 7.164480322906156e-06,
-      "loss": 6.2934,
+      "loss": 6.2896,
       "step": 56000
     },
     {
       "epoch": 23.719563392107474,
-      "grad_norm": 8.096223831176758,
+      "grad_norm": 6.0700883865356445,
       "learning_rate": 6.155398587285571e-06,
-      "loss": 6.2722,
+      "loss": 6.2668,
       "step": 56500
     },
     {
       "epoch": 23.929471032745592,
-      "grad_norm": 7.301811218261719,
+      "grad_norm": 6.540452003479004,
       "learning_rate": 5.146316851664986e-06,
-      "loss": 6.2863,
+      "loss": 6.2783,
       "step": 57000
     },
     {
       "epoch": 24.13937867338371,
-      "grad_norm": 5.255412578582764,
+      "grad_norm": 4.93905782699585,
       "learning_rate": 4.1372351160444e-06,
-      "loss": 6.2732,
+      "loss": 6.2632,
       "step": 57500
     },
     {
       "epoch": 24.349286314021832,
-      "grad_norm": 7.897463798522949,
+      "grad_norm": 8.611550331115723,
       "learning_rate": 3.1281533804238145e-06,
-      "loss": 6.2744,
+      "loss": 6.2699,
       "step": 58000
     },
     {
       "epoch": 24.55919395465995,
-      "grad_norm": 5.157082557678223,
+      "grad_norm": 5.015688896179199,
       "learning_rate": 2.1190716448032293e-06,
-      "loss": 6.2767,
+      "loss": 6.2675,
       "step": 58500
     },
     {
       "epoch": 24.769101595298068,
-      "grad_norm": 7.014009952545166,
+      "grad_norm": 6.46795129776001,
       "learning_rate": 1.109989909182644e-06,
-      "loss": 6.2635,
+      "loss": 6.2634,
       "step": 59000
     },
     {
       "epoch": 24.97900923593619,
-      "grad_norm": 5.050642490386963,
+      "grad_norm": 6.047255992889404,
       "learning_rate": 1.0090817356205853e-07,
-      "loss": 6.255,
+      "loss": 6.2544,
       "step": 59500
     },
     {
       "epoch": 25.0,
       "step": 59550,
       "total_flos": 7.68761901614592e+17,
-      "train_loss": 8.060933875813392,
-      "train_runtime": 16751.5356,
-      "train_samples_per_second": 341.221,
-      "train_steps_per_second": 3.555
+      "train_loss": 8.045040057203932,
+      "train_runtime": 16960.8581,
+      "train_samples_per_second": 337.01,
+      "train_steps_per_second": 3.511
     }
   ],
   "logging_steps": 500,