Model save

README.md

@@ -2,13 +2,12 @@
 license: gemma
 library_name: peft
 tags:
-- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
 base_model: google/gemma-2b
 datasets:
-- llama-duo/synth_summarize_dataset_dedup
+- generator
 model-index:
 - name: gemma2b-summarize-claude3sonnet-32k
   results: []
@@ -19,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma2b-summarize-claude3sonnet-32k
 
-This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on the llama-duo/synth_summarize_dataset_dedup dataset.
+This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.5238
+- Loss: 2.5627
 
 ## Model description
 
@@ -58,22 +57,22 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| 1.2361        | 0.9951 | 101  | 2.5186          |
-| 1.0968        | 2.0    | 203  | 2.4845          |
-| 1.0436        | 2.9951 | 304  | 2.4796          |
-| 1.0084        | 4.0    | 406  | 2.4944          |
-| 0.9913        | 4.9951 | 507  | 2.5010          |
-| 0.9588        | 6.0    | 609  | 2.5066          |
-| 0.9459        | 6.9951 | 710  | 2.5164          |
-| 0.943         | 8.0    | 812  | 2.5233          |
-| 0.9169        | 8.9951 | 913  | 2.5240          |
-| 0.925         | 9.9507 | 1010 | 2.5238          |
+| 1.187         | 0.9951 | 101  | 2.4971          |
+| 1.0678        | 2.0    | 203  | 2.4806          |
+| 1.0113        | 2.9951 | 304  | 2.4777          |
+| 0.9692        | 4.0    | 406  | 2.5005          |
+| 0.9448        | 4.9951 | 507  | 2.5205          |
+| 0.906         | 6.0    | 609  | 2.5325          |
+| 0.8867        | 6.9951 | 710  | 2.5459          |
+| 0.8783        | 8.0    | 812  | 2.5595          |
+| 0.8501        | 8.9951 | 913  | 2.5624          |
+| 0.8569        | 9.9507 | 1010 | 2.5627          |
 
 
 ### Framework versions
 
-- PEFT 0.10.0
-- Transformers 4.40.0
+- PEFT 0.11.1
+- Transformers 4.41.2
 - Pytorch 2.2.2+cu121
 - Datasets 2.19.1
 - Tokenizers 0.19.1

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f6a4a724deaee053e9ab878aa4801e880a2403b1e28870f2ff291bb90974727e
+oid sha256:6aa408029a9ce8f05f7d178a0dee5dcc155dce04189450d98353a468dbb0edcb
 size 39256960

all_results.json

@@ -1,14 +1,9 @@
 {
     "epoch": 9.950738916256158,
-    "eval_loss": 2.523752212524414,
-    "eval_runtime": 0.4919,
-    "eval_samples": 25,
-    "eval_samples_per_second": 20.328,
-    "eval_steps_per_second": 2.033,
-    "total_flos": 5.932470720905871e+17,
-    "train_loss": 1.07745361446154,
-    "train_runtime": 3468.402,
+    "total_flos": 5.9616784580975e+17,
+    "train_loss": 1.0187734235631358,
+    "train_runtime": 5372.8296,
     "train_samples": 31747,
-    "train_samples_per_second": 13.998,
-    "train_steps_per_second": 0.291
+    "train_samples_per_second": 9.036,
+    "train_steps_per_second": 0.188
 }

runs/Jun10_10-58-48_user-HP-Z8-Fury-G5-Workstation-Desktop-PC/events.out.tfevents.1717984754.user-HP-Z8-Fury-G5-Workstation-Desktop-PC.2790.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1568e21285fd2cb679c10aa968cf5b0d3c7b690081e55bfd5dea03b82c929f84
-size 50370
+oid sha256:908b45a02e9755d074c639719685982a517d7beee45c66fe36a4d40c26cf802c
+size 51417

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 9.950738916256158,
-    "total_flos": 5.932470720905871e+17,
-    "train_loss": 1.07745361446154,
-    "train_runtime": 3468.402,
+    "total_flos": 5.9616784580975e+17,
+    "train_loss": 1.0187734235631358,
+    "train_runtime": 5372.8296,
     "train_samples": 31747,
-    "train_samples_per_second": 13.998,
-    "train_steps_per_second": 0.291
+    "train_samples_per_second": 9.036,
+    "train_steps_per_second": 0.188
 }

trainer_state.json

@@ -10,1513 +10,1513 @@
   "log_history": [
     {
       "epoch": 0.009852216748768473,
-      "grad_norm": 1.765625,
+      "grad_norm": 2.5,
       "learning_rate": 1.9801980198019803e-06,
-      "loss": 2.8604,
+      "loss": 2.8603,
       "step": 1
     },
     {
       "epoch": 0.04926108374384237,
-      "grad_norm": 1.4921875,
+      "grad_norm": 2.125,
       "learning_rate": 9.900990099009901e-06,
-      "loss": 2.8185,
+      "loss": 2.8169,
       "step": 5
     },
     {
       "epoch": 0.09852216748768473,
-      "grad_norm": 2.1875,
+      "grad_norm": 5.84375,
       "learning_rate": 1.9801980198019803e-05,
-      "loss": 2.8289,
+      "loss": 2.818,
       "step": 10
     },
     {
       "epoch": 0.1477832512315271,
-      "grad_norm": 2.0625,
+      "grad_norm": 1.828125,
       "learning_rate": 2.9702970297029702e-05,
-      "loss": 2.7485,
+      "loss": 2.691,
       "step": 15
     },
     {
       "epoch": 0.19704433497536947,
-      "grad_norm": 3.3125,
+      "grad_norm": 2.390625,
       "learning_rate": 3.9603960396039605e-05,
-      "loss": 2.6773,
+      "loss": 2.5462,
       "step": 20
     },
     {
       "epoch": 0.24630541871921183,
-      "grad_norm": 2.3125,
+      "grad_norm": 3.109375,
       "learning_rate": 4.950495049504951e-05,
-      "loss": 2.5348,
+      "loss": 2.3501,
       "step": 25
     },
     {
       "epoch": 0.2955665024630542,
-      "grad_norm": 1.3359375,
+      "grad_norm": 1.0546875,
       "learning_rate": 5.9405940594059404e-05,
-      "loss": 2.375,
+      "loss": 2.1929,
       "step": 30
     },
     {
       "epoch": 0.3448275862068966,
-      "grad_norm": 1.5,
+      "grad_norm": 1.265625,
       "learning_rate": 6.93069306930693e-05,
-      "loss": 2.225,
+      "loss": 2.02,
       "step": 35
     },
     {
       "epoch": 0.39408866995073893,
-      "grad_norm": 1.1796875,
+      "grad_norm": 0.98828125,
       "learning_rate": 7.920792079207921e-05,
-      "loss": 2.0577,
+      "loss": 1.8628,
       "step": 40
     },
     {
       "epoch": 0.4433497536945813,
-      "grad_norm": 2.3125,
+      "grad_norm": 0.98046875,
       "learning_rate": 8.910891089108912e-05,
-      "loss": 1.9332,
+      "loss": 1.7401,
       "step": 45
     },
     {
       "epoch": 0.49261083743842365,
-      "grad_norm": 0.890625,
+      "grad_norm": 1.046875,
       "learning_rate": 9.900990099009902e-05,
-      "loss": 1.8054,
+      "loss": 1.6161,
       "step": 50
     },
     {
       "epoch": 0.541871921182266,
-      "grad_norm": 0.82421875,
+      "grad_norm": 0.5546875,
       "learning_rate": 0.00010891089108910893,
-      "loss": 1.7066,
+      "loss": 1.5233,
       "step": 55
     },
     {
       "epoch": 0.5911330049261084,
-      "grad_norm": 0.55078125,
+      "grad_norm": 0.47265625,
       "learning_rate": 0.00011881188118811881,
-      "loss": 1.6055,
+      "loss": 1.4435,
       "step": 60
     },
     {
       "epoch": 0.6403940886699507,
-      "grad_norm": 0.44921875,
+      "grad_norm": 0.45703125,
       "learning_rate": 0.00012871287128712872,
-      "loss": 1.5102,
+      "loss": 1.3694,
       "step": 65
     },
     {
       "epoch": 0.6896551724137931,
-      "grad_norm": 0.8125,
+      "grad_norm": 0.58984375,
       "learning_rate": 0.0001386138613861386,
-      "loss": 1.4332,
+      "loss": 1.3129,
       "step": 70
     },
     {
       "epoch": 0.7389162561576355,
-      "grad_norm": 0.48046875,
+      "grad_norm": 0.58203125,
       "learning_rate": 0.0001485148514851485,
-      "loss": 1.3703,
+      "loss": 1.2671,
       "step": 75
     },
     {
       "epoch": 0.7881773399014779,
-      "grad_norm": 0.62109375,
+      "grad_norm": 0.6796875,
       "learning_rate": 0.00015841584158415842,
-      "loss": 1.3483,
+      "loss": 1.2618,
       "step": 80
     },
     {
       "epoch": 0.8374384236453202,
-      "grad_norm": 0.5625,
+      "grad_norm": 0.3828125,
       "learning_rate": 0.00016831683168316833,
-      "loss": 1.2989,
+      "loss": 1.2273,
       "step": 85
     },
     {
       "epoch": 0.8866995073891626,
-      "grad_norm": 0.55078125,
+      "grad_norm": 1.4609375,
       "learning_rate": 0.00017821782178217824,
-      "loss": 1.2627,
+      "loss": 1.2031,
       "step": 90
     },
     {
       "epoch": 0.9359605911330049,
-      "grad_norm": 0.71875,
+      "grad_norm": 0.515625,
       "learning_rate": 0.00018811881188118812,
-      "loss": 1.2456,
+      "loss": 1.1932,
       "step": 95
     },
     {
       "epoch": 0.9852216748768473,
-      "grad_norm": 0.5,
+      "grad_norm": 0.431640625,
       "learning_rate": 0.00019801980198019803,
-      "loss": 1.2361,
+      "loss": 1.187,
       "step": 100
     },
     {
       "epoch": 0.9950738916256158,
-      "eval_loss": 2.518617868423462,
-      "eval_runtime": 0.6245,
-      "eval_samples_per_second": 16.014,
-      "eval_steps_per_second": 1.601,
+      "eval_loss": 2.497124195098877,
+      "eval_runtime": 0.9532,
+      "eval_samples_per_second": 10.491,
+      "eval_steps_per_second": 1.049,
       "step": 101
     },
     {
       "epoch": 1.0344827586206897,
-      "grad_norm": 0.609375,
+      "grad_norm": 0.384765625,
       "learning_rate": 0.000199990444464082,
-      "loss": 1.2023,
+      "loss": 1.1539,
       "step": 105
     },
     {
       "epoch": 1.083743842364532,
-      "grad_norm": 0.5546875,
+      "grad_norm": 0.37890625,
       "learning_rate": 0.00019995162822919883,
-      "loss": 1.1906,
+      "loss": 1.146,
       "step": 110
     },
     {
       "epoch": 1.1330049261083743,
-      "grad_norm": 0.4375,
+      "grad_norm": 0.40234375,
       "learning_rate": 0.00019988296565626987,
-      "loss": 1.1786,
+      "loss": 1.1351,
       "step": 115
     },
     {
       "epoch": 1.1822660098522166,
-      "grad_norm": 0.62109375,
+      "grad_norm": 0.65234375,
       "learning_rate": 0.00019978447724847652,
-      "loss": 1.1654,
+      "loss": 1.1261,
       "step": 120
     },
     {
       "epoch": 1.2315270935960592,
-      "grad_norm": 0.466796875,
+      "grad_norm": 0.458984375,
       "learning_rate": 0.0001996561924152278,
-      "loss": 1.1533,
+      "loss": 1.1158,
       "step": 125
     },
     {
       "epoch": 1.2807881773399015,
-      "grad_norm": 0.466796875,
+      "grad_norm": 0.5703125,
       "learning_rate": 0.00019949814946337838,
-      "loss": 1.149,
+      "loss": 1.1137,
       "step": 130
     },
     {
       "epoch": 1.3300492610837438,
-      "grad_norm": 0.5234375,
+      "grad_norm": 0.84765625,
       "learning_rate": 0.00019931039558578997,
-      "loss": 1.1363,
+      "loss": 1.1012,
       "step": 135
     },
     {
       "epoch": 1.3793103448275863,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.5234375,
       "learning_rate": 0.00019909298684723904,
-      "loss": 1.1292,
+      "loss": 1.0942,
       "step": 140
     },
     {
       "epoch": 1.4285714285714286,
-      "grad_norm": 0.7109375,
+      "grad_norm": 0.51953125,
       "learning_rate": 0.00019884598816767563,
-      "loss": 1.1319,
+      "loss": 1.0984,
       "step": 145
     },
     {
       "epoch": 1.477832512315271,
-      "grad_norm": 0.384765625,
+      "grad_norm": 0.392578125,
       "learning_rate": 0.00019856947330283752,
-      "loss": 1.1248,
+      "loss": 1.0926,
       "step": 150
     },
     {
       "epoch": 1.5270935960591134,
-      "grad_norm": 0.578125,
+      "grad_norm": 0.384765625,
       "learning_rate": 0.00019826352482222638,
-      "loss": 1.1241,
+      "loss": 1.0912,
       "step": 155
     },
     {
       "epoch": 1.5763546798029555,
-      "grad_norm": 0.47265625,
+      "grad_norm": 0.451171875,
       "learning_rate": 0.00019792823408445174,
-      "loss": 1.1179,
+      "loss": 1.0846,
       "step": 160
     },
     {
       "epoch": 1.625615763546798,
-      "grad_norm": 0.55859375,
+      "grad_norm": 0.58203125,
       "learning_rate": 0.00019756370120995066,
-      "loss": 1.1112,
+      "loss": 1.0794,
       "step": 165
     },
     {
       "epoch": 1.6748768472906403,
-      "grad_norm": 0.462890625,
+      "grad_norm": 0.50390625,
       "learning_rate": 0.00019717003505109095,
-      "loss": 1.1021,
+      "loss": 1.0735,
       "step": 170
     },
     {
       "epoch": 1.7241379310344827,
-      "grad_norm": 0.93359375,
+      "grad_norm": 0.8828125,
       "learning_rate": 0.0001967473531596671,
-      "loss": 1.1162,
+      "loss": 1.0918,
       "step": 175
     },
     {
       "epoch": 1.7733990147783252,
-      "grad_norm": 0.46875,
+      "grad_norm": 0.47265625,
       "learning_rate": 0.0001962957817517982,
-      "loss": 1.0989,
+      "loss": 1.0742,
       "step": 180
     },
     {
       "epoch": 1.8226600985221675,
-      "grad_norm": 0.482421875,
+      "grad_norm": 0.5390625,
       "learning_rate": 0.000195815455670239,
-      "loss": 1.1004,
+      "loss": 1.0722,
       "step": 185
     },
     {
       "epoch": 1.8719211822660098,
-      "grad_norm": 0.55859375,
+      "grad_norm": 0.57421875,
       "learning_rate": 0.00019530651834411474,
-      "loss": 1.0986,
+      "loss": 1.0693,
       "step": 190
     },
     {
       "epoch": 1.9211822660098523,
-      "grad_norm": 0.40625,
+      "grad_norm": 0.4140625,
       "learning_rate": 0.0001947691217460921,
-      "loss": 1.0921,
+      "loss": 1.0638,
       "step": 195
     },
     {
       "epoch": 1.9704433497536946,
-      "grad_norm": 0.59375,
+      "grad_norm": 0.5390625,
       "learning_rate": 0.0001942034263469989,
-      "loss": 1.0968,
+      "loss": 1.0678,
       "step": 200
     },
     {
       "epoch": 2.0,
-      "eval_loss": 2.484497547149658,
-      "eval_runtime": 0.5386,
-      "eval_samples_per_second": 18.567,
-      "eval_steps_per_second": 1.857,
+      "eval_loss": 2.4805524349212646,
+      "eval_runtime": 0.5456,
+      "eval_samples_per_second": 18.329,
+      "eval_steps_per_second": 1.833,
       "step": 203
     },
     {
       "epoch": 2.019704433497537,
-      "grad_norm": 0.75390625,
+      "grad_norm": 0.67578125,
       "learning_rate": 0.00019360960106790643,
-      "loss": 1.0792,
+      "loss": 1.049,
       "step": 205
     },
     {
       "epoch": 2.0689655172413794,
-      "grad_norm": 0.455078125,
+      "grad_norm": 0.48046875,
       "learning_rate": 0.00019298782322968815,
-      "loss": 1.0645,
+      "loss": 1.0286,
       "step": 210
     },
     {
       "epoch": 2.1182266009852215,
-      "grad_norm": 0.6015625,
+      "grad_norm": 0.56640625,
       "learning_rate": 0.00019233827850007027,
-      "loss": 1.069,
+      "loss": 1.0332,
       "step": 215
     },
     {
       "epoch": 2.167487684729064,
-      "grad_norm": 0.48046875,
+      "grad_norm": 0.435546875,
       "learning_rate": 0.00019166116083819002,
-      "loss": 1.0589,
+      "loss": 1.0231,
       "step": 220
     },
     {
       "epoch": 2.2167487684729066,
-      "grad_norm": 0.53515625,
+      "grad_norm": 0.625,
       "learning_rate": 0.0001909566724366779,
-      "loss": 1.0646,
+      "loss": 1.029,
       "step": 225
     },
     {
       "epoch": 2.2660098522167487,
-      "grad_norm": 0.5546875,
+      "grad_norm": 0.39453125,
       "learning_rate": 0.00019022502366128135,
-      "loss": 1.0631,
+      "loss": 1.0278,
       "step": 230
     },
     {
       "epoch": 2.315270935960591,
-      "grad_norm": 0.55078125,
+      "grad_norm": 0.49609375,
       "learning_rate": 0.00018946643298804793,
-      "loss": 1.0559,
+      "loss": 1.0202,
       "step": 235
     },
     {
       "epoch": 2.3645320197044333,
-      "grad_norm": 0.609375,
+      "grad_norm": 0.408203125,
       "learning_rate": 0.00018868112693808665,
-      "loss": 1.0469,
+      "loss": 1.0123,
       "step": 240
     },
     {
       "epoch": 2.413793103448276,
-      "grad_norm": 0.6953125,
+      "grad_norm": 0.55078125,
       "learning_rate": 0.00018786934000992688,
-      "loss": 1.0502,
+      "loss": 1.0166,
       "step": 245
     },
     {
       "epoch": 2.4630541871921183,
-      "grad_norm": 0.53515625,
+      "grad_norm": 0.443359375,
       "learning_rate": 0.00018703131460949554,
-      "loss": 1.058,
+      "loss": 1.0241,
       "step": 250
     },
     {
       "epoch": 2.512315270935961,
-      "grad_norm": 0.53515625,
+      "grad_norm": 0.515625,
       "learning_rate": 0.0001861673009777325,
-      "loss": 1.0501,
+      "loss": 1.0164,
       "step": 255
     },
     {
       "epoch": 2.561576354679803,
-      "grad_norm": 0.5546875,
+      "grad_norm": 0.482421875,
       "learning_rate": 0.00018527755711586678,
-      "loss": 1.0516,
+      "loss": 1.0179,
       "step": 260
     },
     {
       "epoch": 2.6108374384236455,
-      "grad_norm": 0.6484375,
+      "grad_norm": 0.52734375,
       "learning_rate": 0.00018436234870837547,
-      "loss": 1.0503,
+      "loss": 1.0175,
       "step": 265
     },
     {
       "epoch": 2.6600985221674875,
-      "grad_norm": 0.52734375,
+      "grad_norm": 0.546875,
       "learning_rate": 0.00018342194904364813,
-      "loss": 1.0539,
+      "loss": 1.0204,
       "step": 270
     },
     {
       "epoch": 2.70935960591133,
-      "grad_norm": 0.5078125,
+      "grad_norm": 0.447265625,
       "learning_rate": 0.00018245663893238075,
-      "loss": 1.0407,
+      "loss": 1.0079,
       "step": 275
     },
     {
       "epoch": 2.7586206896551726,
-      "grad_norm": 0.466796875,
+      "grad_norm": 0.494140625,
       "learning_rate": 0.00018146670662372354,
-      "loss": 1.0412,
+      "loss": 1.0087,
       "step": 280
     },
     {
       "epoch": 2.8078817733990147,
-      "grad_norm": 0.72265625,
+      "grad_norm": 0.64453125,
       "learning_rate": 0.0001804524477192075,
-      "loss": 1.0412,
+      "loss": 1.007,
       "step": 285
     },
     {
       "epoch": 2.857142857142857,
-      "grad_norm": 0.4921875,
+      "grad_norm": 0.51171875,
       "learning_rate": 0.00017941416508447536,
-      "loss": 1.0286,
+      "loss": 0.9952,
       "step": 290
     },
     {
       "epoch": 2.9064039408866993,
-      "grad_norm": 0.63671875,
+      "grad_norm": 0.65234375,
       "learning_rate": 0.00017835216875884368,
-      "loss": 1.0476,
+      "loss": 1.014,
       "step": 295
     },
     {
       "epoch": 2.955665024630542,
-      "grad_norm": 0.5390625,
+      "grad_norm": 0.494140625,
       "learning_rate": 0.00017726677586272263,
-      "loss": 1.0436,
+      "loss": 1.0113,
       "step": 300
     },
     {
       "epoch": 2.9950738916256157,
-      "eval_loss": 2.479555606842041,
-      "eval_runtime": 0.6669,
-      "eval_samples_per_second": 14.995,
-      "eval_steps_per_second": 1.5,
+      "eval_loss": 2.4776642322540283,
+      "eval_runtime": 0.9583,
+      "eval_samples_per_second": 10.435,
+      "eval_steps_per_second": 1.043,
       "step": 304
     },
     {
       "epoch": 3.0049261083743843,
-      "grad_norm": 0.455078125,
+      "grad_norm": 0.435546875,
       "learning_rate": 0.0001761583105029213,
-      "loss": 1.0275,
+      "loss": 0.9932,
       "step": 305
     },
     {
       "epoch": 3.0541871921182264,
-      "grad_norm": 0.6640625,
+      "grad_norm": 0.46875,
       "learning_rate": 0.00017502710367586687,
-      "loss": 1.0062,
+      "loss": 0.9648,
       "step": 310
     },
     {
       "epoch": 3.103448275862069,
-      "grad_norm": 0.6484375,
+      "grad_norm": 0.578125,
       "learning_rate": 0.00017387349316876666,
-      "loss": 1.018,
+      "loss": 0.9771,
       "step": 315
     },
     {
       "epoch": 3.1527093596059115,
-      "grad_norm": 0.68359375,
+      "grad_norm": 0.546875,
       "learning_rate": 0.00017269782345874203,
-      "loss": 1.005,
+      "loss": 0.9638,
       "step": 320
     },
     {
       "epoch": 3.2019704433497536,
-      "grad_norm": 0.66015625,
+      "grad_norm": 0.5625,
       "learning_rate": 0.00017150044560996488,
-      "loss": 1.0104,
+      "loss": 0.9706,
       "step": 325
     },
     {
       "epoch": 3.251231527093596,
-      "grad_norm": 0.609375,
+      "grad_norm": 0.56640625,
       "learning_rate": 0.00017028171716882714,
-      "loss": 1.0039,
+      "loss": 0.965,
       "step": 330
     },
     {
       "epoch": 3.3004926108374386,
-      "grad_norm": 0.494140625,
+      "grad_norm": 0.3984375,
       "learning_rate": 0.0001690420020571747,
-      "loss": 1.0177,
+      "loss": 0.9805,
       "step": 335
     },
     {
       "epoch": 3.3497536945812807,
-      "grad_norm": 0.76171875,
+      "grad_norm": 0.609375,
       "learning_rate": 0.00016778167046363734,
-      "loss": 1.0066,
+      "loss": 0.9666,
       "step": 340
     },
     {
       "epoch": 3.399014778325123,
-      "grad_norm": 0.5859375,
+      "grad_norm": 0.515625,
       "learning_rate": 0.00016650109873308765,
-      "loss": 1.0187,
+      "loss": 0.9795,
       "step": 345
     },
     {
       "epoch": 3.4482758620689653,
-      "grad_norm": 0.498046875,
+      "grad_norm": 0.390625,
       "learning_rate": 0.00016520066925426144,
-      "loss": 1.0157,
+      "loss": 0.9765,
       "step": 350
     },
     {
       "epoch": 3.497536945812808,
-      "grad_norm": 0.734375,
+      "grad_norm": 0.462890625,
       "learning_rate": 0.00016388077034557355,
-      "loss": 1.0104,
+      "loss": 0.9721,
       "step": 355
     },
     {
       "epoch": 3.5467980295566504,
-      "grad_norm": 0.5546875,
+      "grad_norm": 0.44921875,
       "learning_rate": 0.00016254179613916278,
-      "loss": 1.0177,
+      "loss": 0.9784,
       "step": 360
     },
     {
       "epoch": 3.596059113300493,
-      "grad_norm": 0.4765625,
+      "grad_norm": 0.384765625,
       "learning_rate": 0.0001611841464632011,
-      "loss": 1.0193,
+      "loss": 0.9792,
       "step": 365
     },
     {
       "epoch": 3.645320197044335,
-      "grad_norm": 0.52734375,
+      "grad_norm": 0.38671875,
       "learning_rate": 0.0001598082267225018,
-      "loss": 1.0096,
+      "loss": 0.9692,
       "step": 370
     },
     {
       "epoch": 3.6945812807881775,
-      "grad_norm": 0.6640625,
+      "grad_norm": 0.54296875,
       "learning_rate": 0.0001584144477774623,
-      "loss": 1.025,
+      "loss": 0.9838,
       "step": 375
     },
     {
       "epoch": 3.7438423645320196,
-      "grad_norm": 0.53515625,
+      "grad_norm": 0.51171875,
       "learning_rate": 0.00015700322582137827,
-      "loss": 1.0125,
+      "loss": 0.9721,
       "step": 380
     },
     {
       "epoch": 3.793103448275862,
-      "grad_norm": 0.50390625,
+      "grad_norm": 0.470703125,
       "learning_rate": 0.00015557498225616487,
-      "loss": 1.0022,
+      "loss": 0.9632,
       "step": 385
     },
     {
       "epoch": 3.8423645320197046,
-      "grad_norm": 0.578125,
+      "grad_norm": 0.458984375,
       "learning_rate": 0.00015413014356652286,
-      "loss": 1.007,
+      "loss": 0.9676,
       "step": 390
     },
     {
       "epoch": 3.8916256157635467,
-      "grad_norm": 0.5078125,
+      "grad_norm": 0.447265625,
       "learning_rate": 0.000152669141192587,
-      "loss": 1.0013,
+      "loss": 0.9617,
       "step": 395
     },
     {
       "epoch": 3.9408866995073892,
-      "grad_norm": 0.50390625,
+      "grad_norm": 0.515625,
       "learning_rate": 0.00015119241140109467,
-      "loss": 1.0009,
+      "loss": 0.962,
       "step": 400
     },
     {
       "epoch": 3.9901477832512313,
-      "grad_norm": 0.5625,
+      "grad_norm": 0.486328125,
       "learning_rate": 0.00014970039515511304,
-      "loss": 1.0084,
+      "loss": 0.9692,
       "step": 405
     },
     {
       "epoch": 4.0,
-      "eval_loss": 2.494363307952881,
-      "eval_runtime": 0.5386,
-      "eval_samples_per_second": 18.567,
-      "eval_steps_per_second": 1.857,
+      "eval_loss": 2.500500202178955,
+      "eval_runtime": 0.543,
+      "eval_samples_per_second": 18.417,
+      "eval_steps_per_second": 1.842,
       "step": 406
     },
     {
       "epoch": 4.039408866995074,
-      "grad_norm": 0.50390625,
+      "grad_norm": 0.439453125,
       "learning_rate": 0.00014819353798236427,
-      "loss": 0.9801,
+      "loss": 0.9328,
       "step": 410
     },
     {
       "epoch": 4.088669950738916,
-      "grad_norm": 0.60546875,
+      "grad_norm": 0.478515625,
       "learning_rate": 0.0001466722898421873,
-      "loss": 0.9817,
+      "loss": 0.9336,
       "step": 415
     },
     {
       "epoch": 4.137931034482759,
-      "grad_norm": 0.5390625,
+      "grad_norm": 0.435546875,
       "learning_rate": 0.00014513710499117647,
-      "loss": 0.988,
+      "loss": 0.9385,
       "step": 420
     },
     {
       "epoch": 4.187192118226601,
-      "grad_norm": 0.51171875,
+      "grad_norm": 0.4375,
       "learning_rate": 0.00014358844184753712,
-      "loss": 0.9782,
+      "loss": 0.93,
       "step": 425
     },
     {
       "epoch": 4.236453201970443,
-      "grad_norm": 0.546875,
+      "grad_norm": 0.474609375,
       "learning_rate": 0.00014202676285419812,
-      "loss": 0.9812,
+      "loss": 0.9353,
       "step": 430
     },
     {
       "epoch": 4.285714285714286,
-      "grad_norm": 0.7421875,
+      "grad_norm": 0.546875,
       "learning_rate": 0.0001404525343407228,
-      "loss": 0.9897,
+      "loss": 0.9412,
       "step": 435
     },
     {
       "epoch": 4.334975369458128,
-      "grad_norm": 0.859375,
+      "grad_norm": 0.71875,
       "learning_rate": 0.00013886622638405952,
-      "loss": 0.992,
+      "loss": 0.9438,
       "step": 440
     },
     {
       "epoch": 4.384236453201971,
-      "grad_norm": 0.5390625,
+      "grad_norm": 0.458984375,
       "learning_rate": 0.00013726831266817278,
-      "loss": 0.9933,
+      "loss": 0.9453,
       "step": 445
     },
     {
       "epoch": 4.433497536945813,
-      "grad_norm": 0.54296875,
+      "grad_norm": 0.47265625,
       "learning_rate": 0.0001356592703425976,
-      "loss": 0.9742,
+      "loss": 0.9278,
       "step": 450
     },
     {
       "epoch": 4.482758620689655,
-      "grad_norm": 0.6953125,
+      "grad_norm": 0.55078125,
       "learning_rate": 0.00013403957987995882,
-      "loss": 0.9777,
+      "loss": 0.9314,
       "step": 455
     },
     {
       "epoch": 4.532019704433497,
-      "grad_norm": 0.63671875,
+      "grad_norm": 0.5390625,
       "learning_rate": 0.00013240972493249847,
-      "loss": 0.9853,
+      "loss": 0.9375,
       "step": 460
     },
     {
       "epoch": 4.58128078817734,
-      "grad_norm": 0.474609375,
+      "grad_norm": 0.6171875,
       "learning_rate": 0.00013077019218765305,
-      "loss": 0.9791,
+      "loss": 0.9333,
       "step": 465
     },
     {
       "epoch": 4.630541871921182,
-      "grad_norm": 0.53125,
+      "grad_norm": 0.4375,
       "learning_rate": 0.00012912147122272523,
-      "loss": 0.9857,
+      "loss": 0.9388,
       "step": 470
     },
     {
       "epoch": 4.679802955665025,
-      "grad_norm": 0.48046875,
+      "grad_norm": 0.419921875,
       "learning_rate": 0.00012746405435869198,
-      "loss": 0.9854,
+      "loss": 0.9405,
       "step": 475
     },
     {
       "epoch": 4.7290640394088665,
-      "grad_norm": 0.43359375,
+      "grad_norm": 0.400390625,
       "learning_rate": 0.0001257984365131938,
-      "loss": 0.9836,
+      "loss": 0.9374,
       "step": 480
     },
     {
       "epoch": 4.778325123152709,
-      "grad_norm": 0.478515625,
+      "grad_norm": 0.48046875,
       "learning_rate": 0.00012412511505274844,
-      "loss": 0.9939,
+      "loss": 0.9474,
       "step": 485
     },
     {
       "epoch": 4.827586206896552,
-      "grad_norm": 0.494140625,
+      "grad_norm": 0.44921875,
       "learning_rate": 0.00012244458964423327,
-      "loss": 0.9685,
+      "loss": 0.9217,
       "step": 490
     },
     {
       "epoch": 4.876847290640394,
-      "grad_norm": 0.55859375,
+      "grad_norm": 0.48046875,
       "learning_rate": 0.0001207573621056809,
-      "loss": 0.9806,
+      "loss": 0.9344,
       "step": 495
     },
     {
       "epoch": 4.926108374384237,
-      "grad_norm": 0.494140625,
+      "grad_norm": 0.390625,
       "learning_rate": 0.00011906393625643244,
-      "loss": 0.979,
+      "loss": 0.9325,
       "step": 500
     },
     {
       "epoch": 4.975369458128079,
-      "grad_norm": 0.6484375,
+      "grad_norm": 0.515625,
       "learning_rate": 0.00011736481776669306,
-      "loss": 0.9913,
+      "loss": 0.9448,
       "step": 505
     },
     {
       "epoch": 4.995073891625616,
-      "eval_loss": 2.50097918510437,
-      "eval_runtime": 0.6794,
-      "eval_samples_per_second": 14.718,
-      "eval_steps_per_second": 1.472,
+      "eval_loss": 2.520510196685791,
+      "eval_runtime": 0.9492,
+      "eval_samples_per_second": 10.535,
+      "eval_steps_per_second": 1.054,
       "step": 507
     },
     {
       "epoch": 5.024630541871921,
-      "grad_norm": 0.57421875,
+      "grad_norm": 0.5234375,
       "learning_rate": 0.00011566051400653486,
-      "loss": 0.9714,
+      "loss": 0.9194,
       "step": 510
     },
     {
       "epoch": 5.073891625615763,
-      "grad_norm": 0.6171875,
+      "grad_norm": 0.53515625,
       "learning_rate": 0.00011395153389439233,
-      "loss": 0.9602,
+      "loss": 0.9051,
       "step": 515
     },
     {
       "epoch": 5.123152709359606,
-      "grad_norm": 0.5234375,
+      "grad_norm": 0.45703125,
       "learning_rate": 0.00011223838774509514,
-      "loss": 0.9657,
+      "loss": 0.9109,
       "step": 520
     },
     {
       "epoch": 5.172413793103448,
-      "grad_norm": 0.490234375,
+      "grad_norm": 0.45703125,
       "learning_rate": 0.00011052158711748434,
-      "loss": 0.9526,
+      "loss": 0.8987,
       "step": 525
     },
     {
       "epoch": 5.221674876847291,
-      "grad_norm": 0.5234375,
+      "grad_norm": 0.455078125,
       "learning_rate": 0.00010880164466165674,
-      "loss": 0.958,
+      "loss": 0.9022,
       "step": 530
     },
     {
       "epoch": 5.2709359605911335,
-      "grad_norm": 0.498046875,
+      "grad_norm": 0.447265625,
       "learning_rate": 0.00010707907396588361,
-      "loss": 0.9666,
+      "loss": 0.9103,
       "step": 535
     },
     {
       "epoch": 5.320197044334975,
-      "grad_norm": 0.52734375,
+      "grad_norm": 0.453125,
       "learning_rate": 0.0001053543894032493,
-      "loss": 0.9625,
+      "loss": 0.9085,
       "step": 540
     },
     {
       "epoch": 5.369458128078818,
-      "grad_norm": 0.515625,
+      "grad_norm": 0.419921875,
       "learning_rate": 0.00010362810597805526,
-      "loss": 0.9657,
+      "loss": 0.9127,
       "step": 545
     },
     {
       "epoch": 5.41871921182266,
-      "grad_norm": 0.482421875,
+      "grad_norm": 0.44140625,
       "learning_rate": 0.00010190073917203589,
-      "loss": 0.9655,
+      "loss": 0.9116,
       "step": 550
     },
     {
       "epoch": 5.467980295566503,
-      "grad_norm": 0.515625,
+      "grad_norm": 0.423828125,
       "learning_rate": 0.00010017280479043147,
-      "loss": 0.9665,
+      "loss": 0.9122,
       "step": 555
     },
     {
       "epoch": 5.517241379310345,
-      "grad_norm": 0.498046875,
+      "grad_norm": 0.4296875,
       "learning_rate": 9.844481880796491e-05,
-      "loss": 0.9587,
+      "loss": 0.9065,
       "step": 560
     },
     {
       "epoch": 5.566502463054187,
-      "grad_norm": 0.609375,
+      "grad_norm": 0.451171875,
       "learning_rate": 9.671729721476746e-05,
-      "loss": 0.9665,
+      "loss": 0.9107,
       "step": 565
     },
     {
       "epoch": 5.615763546798029,
-      "grad_norm": 0.5390625,
+      "grad_norm": 0.443359375,
       "learning_rate": 9.499075586230013e-05,
-      "loss": 0.9554,
+      "loss": 0.9025,
       "step": 570
     },
     {
       "epoch": 5.665024630541872,
-      "grad_norm": 0.625,
+      "grad_norm": 0.462890625,
       "learning_rate": 9.326571030931637e-05,
-      "loss": 0.9607,
+      "loss": 0.9086,
       "step": 575
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 0.68359375,
+      "grad_norm": 0.51953125,
       "learning_rate": 9.154267566791223e-05,
-      "loss": 0.9669,
+      "loss": 0.9123,
       "step": 580
     },
     {
       "epoch": 5.763546798029557,
-      "grad_norm": 0.61328125,
+      "grad_norm": 0.498046875,
       "learning_rate": 8.982216644970979e-05,
-      "loss": 0.9628,
+      "loss": 0.9085,
       "step": 585
     },
     {
       "epoch": 5.812807881773399,
-      "grad_norm": 0.6171875,
+      "grad_norm": 0.435546875,
       "learning_rate": 8.810469641222001e-05,
-      "loss": 0.9511,
+      "loss": 0.8975,
       "step": 590
     },
     {
       "epoch": 5.862068965517241,
-      "grad_norm": 0.51171875,
+      "grad_norm": 0.5,
       "learning_rate": 8.639077840543077e-05,
-      "loss": 0.9684,
+      "loss": 0.9139,
       "step": 595
     },
     {
       "epoch": 5.911330049261084,
-      "grad_norm": 0.671875,
+      "grad_norm": 0.5390625,
       "learning_rate": 8.468092421866573e-05,
-      "loss": 0.9678,
+      "loss": 0.9145,
       "step": 600
     },
     {
       "epoch": 5.960591133004926,
-      "grad_norm": 0.65625,
+      "grad_norm": 0.59375,
       "learning_rate": 8.297564442776014e-05,
-      "loss": 0.9588,
+      "loss": 0.906,
       "step": 605
     },
     {
       "epoch": 6.0,
-      "eval_loss": 2.5066332817077637,
-      "eval_runtime": 0.5391,
-      "eval_samples_per_second": 18.548,
-      "eval_steps_per_second": 1.855,
+      "eval_loss": 2.5324759483337402,
+      "eval_runtime": 0.5446,
+      "eval_samples_per_second": 18.362,
+      "eval_steps_per_second": 1.836,
       "step": 609
     },
     {
       "epoch": 6.009852216748769,
-      "grad_norm": 0.68359375,
+      "grad_norm": 0.5546875,
       "learning_rate": 8.127544824259889e-05,
-      "loss": 0.953,
+      "loss": 0.8978,
       "step": 610
     },
     {
       "epoch": 6.059113300492611,
-      "grad_norm": 0.734375,
+      "grad_norm": 0.5703125,
       "learning_rate": 7.958084335506239e-05,
-      "loss": 0.9536,
+      "loss": 0.8914,
       "step": 615
     },
     {
       "epoch": 6.108374384236453,
-      "grad_norm": 0.50390625,
+      "grad_norm": 0.52734375,
       "learning_rate": 7.789233578742582e-05,
-      "loss": 0.9446,
+      "loss": 0.8826,
       "step": 620
     },
     {
       "epoch": 6.157635467980295,
-      "grad_norm": 0.498046875,
+      "grad_norm": 0.4609375,
       "learning_rate": 7.6210429741257e-05,
-      "loss": 0.9353,
+      "loss": 0.8733,
       "step": 625
     },
     {
       "epoch": 6.206896551724138,
-      "grad_norm": 0.5078125,
+      "grad_norm": 0.431640625,
       "learning_rate": 7.453562744685778e-05,
-      "loss": 0.9449,
+      "loss": 0.8837,
       "step": 630
     },
     {
       "epoch": 6.25615763546798,
-      "grad_norm": 0.5625,
+      "grad_norm": 0.498046875,
       "learning_rate": 7.286842901329412e-05,
-      "loss": 0.9535,
+      "loss": 0.8932,
       "step": 635
     },
     {
       "epoch": 6.305418719211823,
-      "grad_norm": 0.498046875,
+      "grad_norm": 0.4765625,
       "learning_rate": 7.12093322790597e-05,
-      "loss": 0.9376,
+      "loss": 0.8791,
       "step": 640
     },
     {
       "epoch": 6.3546798029556655,
-      "grad_norm": 0.51953125,
+      "grad_norm": 0.578125,
       "learning_rate": 6.955883266341741e-05,
-      "loss": 0.9461,
+      "loss": 0.8853,
       "step": 645
     },
     {
       "epoch": 6.403940886699507,
-      "grad_norm": 0.5,
+      "grad_norm": 0.5078125,
       "learning_rate": 6.791742301846326e-05,
-      "loss": 0.9503,
+      "loss": 0.8909,
       "step": 650
     },
     {
       "epoch": 6.45320197044335,
-      "grad_norm": 0.46875,
+      "grad_norm": 0.41796875,
       "learning_rate": 6.62855934819569e-05,
-      "loss": 0.9447,
+      "loss": 0.886,
       "step": 655
     },
     {
       "epoch": 6.502463054187192,
-      "grad_norm": 0.6328125,
+      "grad_norm": 0.46484375,
       "learning_rate": 6.466383133096267e-05,
-      "loss": 0.9453,
+      "loss": 0.8862,
       "step": 660
     },
     {
       "epoch": 6.551724137931035,
-      "grad_norm": 0.53125,
+      "grad_norm": 0.44921875,
       "learning_rate": 6.305262083634488e-05,
-      "loss": 0.944,
+      "loss": 0.8844,
       "step": 665
     },
     {
       "epoch": 6.600985221674877,
-      "grad_norm": 0.53125,
+      "grad_norm": 0.44921875,
       "learning_rate": 6.145244311816063e-05,
-      "loss": 0.9366,
+      "loss": 0.8788,
       "step": 670
     },
     {
       "epoch": 6.650246305418719,
-      "grad_norm": 0.5234375,
+      "grad_norm": 0.466796875,
       "learning_rate": 5.986377600199371e-05,
-      "loss": 0.9405,
+      "loss": 0.8799,
       "step": 675
     },
     {
       "epoch": 6.699507389162561,
-      "grad_norm": 0.482421875,
+      "grad_norm": 0.416015625,
       "learning_rate": 5.828709387627218e-05,
-      "loss": 0.9521,
+      "loss": 0.8931,
       "step": 680
     },
     {
       "epoch": 6.748768472906404,
-      "grad_norm": 0.4453125,
+      "grad_norm": 0.46875,
       "learning_rate": 5.6722867550612116e-05,
-      "loss": 0.9509,
+      "loss": 0.8907,
       "step": 685
     },
     {
       "epoch": 6.798029556650246,
-      "grad_norm": 0.59765625,
+      "grad_norm": 0.451171875,
       "learning_rate": 5.5171564115230254e-05,
-      "loss": 0.9625,
+      "loss": 0.9035,
       "step": 690
     },
     {
       "epoch": 6.847290640394089,
-      "grad_norm": 0.447265625,
+      "grad_norm": 0.416015625,
       "learning_rate": 5.363364680146725e-05,
-      "loss": 0.9496,
+      "loss": 0.8889,
       "step": 695
     },
     {
       "epoch": 6.896551724137931,
-      "grad_norm": 0.439453125,
+      "grad_norm": 0.421875,
       "learning_rate": 5.210957484346314e-05,
-      "loss": 0.9457,
+      "loss": 0.8851,
       "step": 700
     },
     {
       "epoch": 6.945812807881773,
-      "grad_norm": 0.482421875,
+      "grad_norm": 0.41796875,
       "learning_rate": 5.059980334102637e-05,
-      "loss": 0.9377,
+      "loss": 0.8793,
       "step": 705
     },
     {
       "epoch": 6.995073891625616,
-      "grad_norm": 0.453125,
+      "grad_norm": 0.4375,
       "learning_rate": 4.9104783123737566e-05,
-      "loss": 0.9459,
+      "loss": 0.8867,
       "step": 710
     },
     {
       "epoch": 6.995073891625616,
-      "eval_loss": 2.516418933868408,
-      "eval_runtime": 0.674,
-      "eval_samples_per_second": 14.837,
-      "eval_steps_per_second": 1.484,
+      "eval_loss": 2.54585599899292,
+      "eval_runtime": 0.9527,
+      "eval_samples_per_second": 10.497,
+      "eval_steps_per_second": 1.05,
       "step": 710
     },
     {
       "epoch": 7.044334975369458,
-      "grad_norm": 0.48828125,
+      "grad_norm": 0.408203125,
       "learning_rate": 4.762496061632814e-05,
-      "loss": 0.9341,
+      "loss": 0.867,
       "step": 715
     },
     {
       "epoch": 7.093596059113301,
-      "grad_norm": 0.51953125,
+      "grad_norm": 0.40234375,
       "learning_rate": 4.6160777705374524e-05,
-      "loss": 0.938,
+      "loss": 0.8716,
       "step": 720
     },
     {
       "epoch": 7.142857142857143,
-      "grad_norm": 0.50390625,
+      "grad_norm": 0.451171875,
       "learning_rate": 4.471267160734731e-05,
-      "loss": 0.9366,
+      "loss": 0.8719,
       "step": 725
     },
     {
       "epoch": 7.192118226600985,
-      "grad_norm": 0.53125,
+      "grad_norm": 0.48046875,
       "learning_rate": 4.328107473805487e-05,
-      "loss": 0.9403,
+      "loss": 0.8745,
       "step": 730
     },
     {
       "epoch": 7.241379310344827,
-      "grad_norm": 0.45703125,
+      "grad_norm": 0.41015625,
       "learning_rate": 4.1866414583520877e-05,
-      "loss": 0.9387,
+      "loss": 0.8744,
       "step": 735
     },
     {
       "epoch": 7.29064039408867,
-      "grad_norm": 0.71484375,
+      "grad_norm": 0.42578125,
       "learning_rate": 4.046911357233343e-05,
-      "loss": 0.9334,
+      "loss": 0.8686,
       "step": 740
     },
     {
       "epoch": 7.3399014778325125,
-      "grad_norm": 0.498046875,
+      "grad_norm": 0.40234375,
       "learning_rate": 3.9089588949504655e-05,
-      "loss": 0.93,
+      "loss": 0.8646,
       "step": 745
     },
     {
       "epoch": 7.389162561576355,
-      "grad_norm": 0.453125,
+      "grad_norm": 0.412109375,
       "learning_rate": 3.772825265187802e-05,
-      "loss": 0.9298,
+      "loss": 0.8653,
       "step": 750
     },
     {
       "epoch": 7.4384236453201975,
-      "grad_norm": 0.474609375,
+      "grad_norm": 0.423828125,
       "learning_rate": 3.638551118512089e-05,
-      "loss": 0.9486,
+      "loss": 0.8843,
       "step": 755
     },
     {
       "epoch": 7.487684729064039,
-      "grad_norm": 0.447265625,
+      "grad_norm": 0.41796875,
       "learning_rate": 3.506176550233863e-05,
-      "loss": 0.9373,
+      "loss": 0.871,
       "step": 760
     },
     {
       "epoch": 7.536945812807882,
-      "grad_norm": 0.423828125,
+      "grad_norm": 0.41015625,
       "learning_rate": 3.3757410884346894e-05,
-      "loss": 0.939,
+      "loss": 0.8743,
       "step": 765
     },
     {
       "epoch": 7.586206896551724,
-      "grad_norm": 0.42578125,
+      "grad_norm": 0.3984375,
       "learning_rate": 3.2472836821637744e-05,
-      "loss": 0.9325,
+      "loss": 0.8672,
       "step": 770
     },
     {
       "epoch": 7.635467980295567,
-      "grad_norm": 0.458984375,
+      "grad_norm": 0.40234375,
       "learning_rate": 3.120842689807468e-05,
-      "loss": 0.932,
+      "loss": 0.8675,
       "step": 775
     },
     {
       "epoch": 7.684729064039409,
-      "grad_norm": 0.51171875,
+      "grad_norm": 0.439453125,
       "learning_rate": 2.996455867635155e-05,
-      "loss": 0.9259,
+      "loss": 0.8623,
       "step": 780
     },
     {
       "epoch": 7.733990147783251,
-      "grad_norm": 0.44921875,
+      "grad_norm": 0.41796875,
       "learning_rate": 2.874160358524931e-05,
-      "loss": 0.9277,
+      "loss": 0.8638,
       "step": 785
     },
     {
       "epoch": 7.783251231527093,
-      "grad_norm": 0.458984375,
+      "grad_norm": 0.390625,
       "learning_rate": 2.753992680872457e-05,
-      "loss": 0.9259,
+      "loss": 0.8606,
       "step": 790
     },
     {
       "epoch": 7.832512315270936,
-      "grad_norm": 0.447265625,
+      "grad_norm": 0.39453125,
       "learning_rate": 2.6359887176862718e-05,
-      "loss": 0.9431,
+      "loss": 0.8782,
       "step": 795
     },
     {
       "epoch": 7.8817733990147785,
-      "grad_norm": 0.46875,
+      "grad_norm": 0.40234375,
       "learning_rate": 2.5201837058728505e-05,
-      "loss": 0.9323,
+      "loss": 0.868,
       "step": 800
     },
     {
       "epoch": 7.931034482758621,
-      "grad_norm": 0.4296875,
+      "grad_norm": 0.421875,
       "learning_rate": 2.4066122257145894e-05,
-      "loss": 0.9294,
+      "loss": 0.8652,
       "step": 805
     },
     {
       "epoch": 7.980295566502463,
-      "grad_norm": 0.453125,
+      "grad_norm": 0.412109375,
       "learning_rate": 2.295308190543859e-05,
-      "loss": 0.943,
+      "loss": 0.8783,
       "step": 810
     },
     {
       "epoch": 8.0,
-      "eval_loss": 2.523322582244873,
-      "eval_runtime": 0.5404,
-      "eval_samples_per_second": 18.504,
-      "eval_steps_per_second": 1.85,
+      "eval_loss": 2.55953311920166,
+      "eval_runtime": 0.5444,
+      "eval_samples_per_second": 18.369,
+      "eval_steps_per_second": 1.837,
       "step": 812
     },
     {
       "epoch": 8.029556650246306,
-      "grad_norm": 0.419921875,
+      "grad_norm": 0.392578125,
       "learning_rate": 2.1863048366162208e-05,
-      "loss": 0.9428,
+      "loss": 0.875,
       "step": 815
     },
     {
       "epoch": 8.078817733990148,
-      "grad_norm": 0.44140625,
+      "grad_norm": 0.43359375,
       "learning_rate": 2.0796347131858186e-05,
-      "loss": 0.931,
+      "loss": 0.8631,
       "step": 820
     },
     {
       "epoch": 8.12807881773399,
-      "grad_norm": 0.431640625,
+      "grad_norm": 0.39453125,
       "learning_rate": 1.9753296727859195e-05,
-      "loss": 0.9263,
+      "loss": 0.8587,
       "step": 825
     },
     {
       "epoch": 8.177339901477833,
-      "grad_norm": 0.455078125,
+      "grad_norm": 0.40234375,
       "learning_rate": 1.8734208617174988e-05,
-      "loss": 0.9426,
+      "loss": 0.8745,
       "step": 830
     },
     {
       "epoch": 8.226600985221674,
-      "grad_norm": 0.453125,
+      "grad_norm": 0.4140625,
       "learning_rate": 1.773938710748706e-05,
-      "loss": 0.9267,
+      "loss": 0.8604,
       "step": 835
     },
     {
       "epoch": 8.275862068965518,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.404296875,
       "learning_rate": 1.676912926028007e-05,
-      "loss": 0.9269,
+      "loss": 0.8604,
       "step": 840
     },
     {
       "epoch": 8.32512315270936,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.392578125,
       "learning_rate": 1.5823724802136865e-05,
-      "loss": 0.9294,
+      "loss": 0.8618,
       "step": 845
     },
     {
       "epoch": 8.374384236453203,
-      "grad_norm": 0.419921875,
+      "grad_norm": 0.3828125,
       "learning_rate": 1.4903456038223939e-05,
-      "loss": 0.9222,
+      "loss": 0.8546,
       "step": 850
     },
     {
       "epoch": 8.423645320197044,
-      "grad_norm": 0.412109375,
+      "grad_norm": 0.37890625,
       "learning_rate": 1.4008597767992871e-05,
-      "loss": 0.929,
+      "loss": 0.861,
       "step": 855
     },
     {
       "epoch": 8.472906403940886,
-      "grad_norm": 0.42578125,
+      "grad_norm": 0.421875,
       "learning_rate": 1.3139417203123027e-05,
-      "loss": 0.9285,
+      "loss": 0.8593,
       "step": 860
     },
     {
       "epoch": 8.52216748768473,
-      "grad_norm": 0.421875,
+      "grad_norm": 0.3828125,
       "learning_rate": 1.2296173887730123e-05,
-      "loss": 0.9346,
+      "loss": 0.8665,
       "step": 865
     },
     {
       "epoch": 8.571428571428571,
-      "grad_norm": 0.42578125,
+      "grad_norm": 0.41015625,
       "learning_rate": 1.1479119620864276e-05,
-      "loss": 0.9252,
+      "loss": 0.8565,
       "step": 870
     },
     {
       "epoch": 8.620689655172415,
-      "grad_norm": 0.4375,
+      "grad_norm": 0.40234375,
       "learning_rate": 1.0688498381320855e-05,
-      "loss": 0.929,
+      "loss": 0.8611,
       "step": 875
     },
     {
       "epoch": 8.669950738916256,
-      "grad_norm": 0.66796875,
+      "grad_norm": 0.384765625,
       "learning_rate": 9.924546254786493e-06,
-      "loss": 0.9381,
+      "loss": 0.8703,
       "step": 880
     },
     {
       "epoch": 8.719211822660098,
-      "grad_norm": 0.453125,
+      "grad_norm": 0.376953125,
       "learning_rate": 9.187491363342093e-06,
-      "loss": 0.9374,
+      "loss": 0.8697,
       "step": 885
     },
     {
       "epoch": 8.768472906403941,
-      "grad_norm": 0.458984375,
+      "grad_norm": 0.396484375,
       "learning_rate": 8.47755379734373e-06,
-      "loss": 0.9267,
+      "loss": 0.8597,
       "step": 890
     },
     {
       "epoch": 8.817733990147783,
-      "grad_norm": 0.447265625,
+      "grad_norm": 0.419921875,
       "learning_rate": 7.794945549701993e-06,
-      "loss": 0.9369,
+      "loss": 0.8694,
       "step": 895
     },
     {
       "epoch": 8.866995073891626,
-      "grad_norm": 0.43359375,
+      "grad_norm": 0.390625,
       "learning_rate": 7.1398704525792e-06,
-      "loss": 0.9371,
+      "loss": 0.8685,
       "step": 900
     },
     {
       "epoch": 8.916256157635468,
-      "grad_norm": 0.4375,
+      "grad_norm": 0.40625,
       "learning_rate": 6.512524116523633e-06,
-      "loss": 0.9348,
+      "loss": 0.8675,
       "step": 905
     },
     {
       "epoch": 8.96551724137931,
-      "grad_norm": 0.44140625,
+      "grad_norm": 0.3828125,
       "learning_rate": 5.913093872058528e-06,
-      "loss": 0.9169,
+      "loss": 0.8501,
       "step": 910
     },
     {
       "epoch": 8.995073891625616,
-      "eval_loss": 2.523988723754883,
-      "eval_runtime": 0.6711,
-      "eval_samples_per_second": 14.9,
-      "eval_steps_per_second": 1.49,
+      "eval_loss": 2.562363624572754,
+      "eval_runtime": 0.9485,
+      "eval_samples_per_second": 10.543,
+      "eval_steps_per_second": 1.054,
       "step": 913
     },
     {
       "epoch": 9.014778325123153,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.400390625,
       "learning_rate": 5.341758713743828e-06,
-      "loss": 0.9254,
+      "loss": 0.8582,
       "step": 915
     },
     {
       "epoch": 9.064039408866995,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.40625,
       "learning_rate": 4.798689246727006e-06,
-      "loss": 0.929,
+      "loss": 0.8616,
       "step": 920
     },
     {
       "epoch": 9.113300492610838,
-      "grad_norm": 0.421875,
+      "grad_norm": 0.388671875,
       "learning_rate": 4.2840476357989825e-06,
-      "loss": 0.9346,
+      "loss": 0.8664,
       "step": 925
     },
     {
       "epoch": 9.16256157635468,
-      "grad_norm": 0.4140625,
+      "grad_norm": 0.3828125,
       "learning_rate": 3.797987556970495e-06,
-      "loss": 0.936,
+      "loss": 0.8662,
       "step": 930
     },
     {
       "epoch": 9.211822660098521,
-      "grad_norm": 0.419921875,
+      "grad_norm": 0.37890625,
       "learning_rate": 3.3406541515832003e-06,
-      "loss": 0.9283,
+      "loss": 0.8596,
       "step": 935
     },
     {
       "epoch": 9.261083743842365,
-      "grad_norm": 0.41015625,
+      "grad_norm": 0.37890625,
       "learning_rate": 2.912183982969385e-06,
-      "loss": 0.9255,
+      "loss": 0.8566,
       "step": 940
     },
     {
       "epoch": 9.310344827586206,
-      "grad_norm": 0.421875,
+      "grad_norm": 0.392578125,
       "learning_rate": 2.5127049956730207e-06,
-      "loss": 0.9252,
+      "loss": 0.8562,
       "step": 945
     },
     {
       "epoch": 9.35960591133005,
-      "grad_norm": 0.431640625,
+      "grad_norm": 0.3984375,
       "learning_rate": 2.1423364772445887e-06,
-      "loss": 0.9276,
+      "loss": 0.8597,
       "step": 950
     },
     {
       "epoch": 9.408866995073891,
-      "grad_norm": 0.423828125,
+      "grad_norm": 0.390625,
       "learning_rate": 1.8011890226208527e-06,
-      "loss": 0.9251,
+      "loss": 0.8565,
       "step": 955
     },
     {
       "epoch": 9.458128078817733,
-      "grad_norm": 0.4140625,
+      "grad_norm": 0.412109375,
       "learning_rate": 1.489364501100332e-06,
-      "loss": 0.9291,
+      "loss": 0.8608,
       "step": 960
     },
     {
       "epoch": 9.507389162561577,
-      "grad_norm": 0.423828125,
+      "grad_norm": 0.390625,
       "learning_rate": 1.2069560259243328e-06,
-      "loss": 0.9333,
+      "loss": 0.8639,
       "step": 965
     },
     {
       "epoch": 9.556650246305418,
-      "grad_norm": 0.4375,
+      "grad_norm": 0.392578125,
       "learning_rate": 9.540479264726676e-07,
-      "loss": 0.9332,
+      "loss": 0.8654,
       "step": 970
     },
     {
       "epoch": 9.605911330049262,
-      "grad_norm": 0.4140625,
+      "grad_norm": 0.380859375,
       "learning_rate": 7.307157230821426e-07,
-      "loss": 0.9297,
+      "loss": 0.8604,
       "step": 975
     },
     {
       "epoch": 9.655172413793103,
-      "grad_norm": 0.40234375,
+      "grad_norm": 0.44921875,
       "learning_rate": 5.370261044956971e-07,
-      "loss": 0.9222,
+      "loss": 0.8551,
       "step": 980
     },
     {
       "epoch": 9.704433497536947,
-      "grad_norm": 0.421875,
+      "grad_norm": 0.38671875,
       "learning_rate": 3.73036907948543e-07,
-      "loss": 0.9396,
+      "loss": 0.8721,
       "step": 985
     },
     {
       "epoch": 9.753694581280788,
-      "grad_norm": 0.416015625,
+      "grad_norm": 0.376953125,
       "learning_rate": 2.3879710189753656e-07,
-      "loss": 0.931,
+      "loss": 0.8636,
       "step": 990
     },
     {
       "epoch": 9.80295566502463,
-      "grad_norm": 0.43359375,
+      "grad_norm": 0.3828125,
       "learning_rate": 1.3434677139885222e-07,
-      "loss": 0.9309,
+      "loss": 0.8607,
       "step": 995
     },
     {
       "epoch": 9.852216748768473,
-      "grad_norm": 0.4140625,
+      "grad_norm": 0.384765625,
       "learning_rate": 5.971710613821291e-08,
-      "loss": 0.9223,
+      "loss": 0.8563,
       "step": 1000
     },
     {
       "epoch": 9.901477832512315,
-      "grad_norm": 0.470703125,
+      "grad_norm": 0.388671875,
       "learning_rate": 1.4930391117451426e-08,
-      "loss": 0.9251,
+      "loss": 0.8566,
       "step": 1005
     },
     {
       "epoch": 9.950738916256158,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.38671875,
       "learning_rate": 0.0,
-      "loss": 0.925,
+      "loss": 0.8569,
       "step": 1010
     },
     {
       "epoch": 9.950738916256158,
-      "eval_loss": 2.523752212524414,
-      "eval_runtime": 0.5416,
-      "eval_samples_per_second": 18.464,
-      "eval_steps_per_second": 1.846,
+      "eval_loss": 2.562716007232666,
+      "eval_runtime": 0.5429,
+      "eval_samples_per_second": 18.418,
+      "eval_steps_per_second": 1.842,
       "step": 1010
     },
     {
       "epoch": 9.950738916256158,
       "step": 1010,
-      "total_flos": 5.932470720905871e+17,
-      "train_loss": 1.07745361446154,
-      "train_runtime": 3468.402,
-      "train_samples_per_second": 13.998,
-      "train_steps_per_second": 0.291
+      "total_flos": 5.9616784580975e+17,
+      "train_loss": 1.0187734235631358,
+      "train_runtime": 5372.8296,
+      "train_samples_per_second": 9.036,
+      "train_steps_per_second": 0.188
     }
   ],
   "logging_steps": 5,
@@ -1524,7 +1524,19 @@
   "num_input_tokens_seen": 0,
   "num_train_epochs": 10,
   "save_steps": 100,
-  "total_flos": 5.932470720905871e+17,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.9616784580975e+17,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null