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-gpt4o-32k
   results: []
@@ -19,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma2b-summarize-gpt4o-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: 2.5618
+- Loss: 2.6191
 
 ## Model description
 
@@ -58,16 +57,16 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 1.4358        | 1.0   | 73   | 2.5470          |
-| 1.2092        | 2.0   | 146  | 2.5064          |
-| 1.1501        | 3.0   | 219  | 2.5000          |
-| 1.0995        | 4.0   | 292  | 2.5064          |
-| 1.0809        | 5.0   | 365  | 2.5177          |
-| 1.0583        | 6.0   | 438  | 2.5419          |
-| 1.04          | 7.0   | 511  | 2.5488          |
-| 1.0248        | 8.0   | 584  | 2.5574          |
-| 1.021         | 9.0   | 657  | 2.5614          |
-| 1.0179        | 10.0  | 730  | 2.5618          |
+| 1.3438        | 1.0   | 73   | 2.5419          |
+| 1.1767        | 2.0   | 146  | 2.4979          |
+| 1.1163        | 3.0   | 219  | 2.4967          |
+| 1.0605        | 4.0   | 292  | 2.5121          |
+| 1.0362        | 5.0   | 365  | 2.5401          |
+| 1.0052        | 6.0   | 438  | 2.5711          |
+| 0.9813        | 7.0   | 511  | 2.5912          |
+| 0.9593        | 8.0   | 584  | 2.6132          |
+| 0.953         | 9.0   | 657  | 2.6180          |
+| 0.9482        | 10.0  | 730  | 2.6191          |
 
 
 ### Framework versions

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:179f616eb8d50d298e1a8b6c146aaa0ab92f1f0294298ee925107af1df7d67c7
+oid sha256:983177e9ca1e77913edba3ab7a25c04930b8160a7de82b8326b9dc06b2e283eb
 size 39256960

all_results.json

@@ -1,14 +1,9 @@
 {
     "epoch": 10.0,
-    "eval_loss": 2.5617620944976807,
-    "eval_runtime": 0.5264,
-    "eval_samples": 25,
-    "eval_samples_per_second": 18.996,
-    "eval_steps_per_second": 1.9,
-    "total_flos": 4.287825372721971e+17,
-    "train_loss": 1.1969801510850044,
-    "train_runtime": 3891.9077,
+    "total_flos": 4.3089359154446336e+17,
+    "train_loss": 1.133717959547696,
+    "train_runtime": 3901.5754,
     "train_samples": 32305,
-    "train_samples_per_second": 9.003,
-    "train_steps_per_second": 0.188
+    "train_samples_per_second": 8.981,
+    "train_steps_per_second": 0.187
 }

runs/Jun10_10-37-10_user-HP-Z8-Fury-G5-Workstation-Desktop-PC/events.out.tfevents.1717983445.user-HP-Z8-Fury-G5-Workstation-Desktop-PC.72146.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:34d1f47e94e47f70f05aaef14a8c0e391eb9f489ab77ba18130ff0e4c31ae3bd
-size 37678
+oid sha256:24f51c5bdf6fa07509c44d96de8fbc37e1907387921a4023118235ebf96e2e62
+size 39569

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 10.0,
-    "total_flos": 4.287825372721971e+17,
-    "train_loss": 1.1969801510850044,
-    "train_runtime": 3891.9077,
+    "total_flos": 4.3089359154446336e+17,
+    "train_loss": 1.133717959547696,
+    "train_runtime": 3901.5754,
     "train_samples": 32305,
-    "train_samples_per_second": 9.003,
-    "train_steps_per_second": 0.188
+    "train_samples_per_second": 8.981,
+    "train_steps_per_second": 0.187
 }

trainer_state.json

@@ -10,1121 +10,1121 @@
   "log_history": [
     {
       "epoch": 0.0136986301369863,
-      "grad_norm": 3.5625,
+      "grad_norm": 5.0625,
       "learning_rate": 2.7397260273972604e-06,
       "loss": 3.0538,
       "step": 1
     },
     {
       "epoch": 0.0684931506849315,
-      "grad_norm": 2.296875,
+      "grad_norm": 2.875,
       "learning_rate": 1.3698630136986302e-05,
-      "loss": 3.0804,
+      "loss": 3.0796,
       "step": 5
     },
     {
       "epoch": 0.136986301369863,
-      "grad_norm": 2.59375,
+      "grad_norm": 2.515625,
       "learning_rate": 2.7397260273972603e-05,
-      "loss": 3.036,
+      "loss": 3.0061,
       "step": 10
     },
     {
       "epoch": 0.2054794520547945,
-      "grad_norm": 2.1875,
+      "grad_norm": 6.15625,
       "learning_rate": 4.1095890410958905e-05,
-      "loss": 3.0119,
+      "loss": 2.8959,
       "step": 15
     },
     {
       "epoch": 0.273972602739726,
-      "grad_norm": 1.84375,
+      "grad_norm": 1.921875,
       "learning_rate": 5.479452054794521e-05,
-      "loss": 2.7939,
+      "loss": 2.636,
       "step": 20
     },
     {
       "epoch": 0.3424657534246575,
-      "grad_norm": 4.4375,
+      "grad_norm": 1.5546875,
       "learning_rate": 6.84931506849315e-05,
-      "loss": 2.5661,
+      "loss": 2.4025,
       "step": 25
     },
     {
       "epoch": 0.410958904109589,
-      "grad_norm": 7.25,
+      "grad_norm": 1.3671875,
       "learning_rate": 8.219178082191781e-05,
-      "loss": 2.3942,
+      "loss": 2.1878,
       "step": 30
     },
     {
       "epoch": 0.4794520547945205,
-      "grad_norm": 2.28125,
+      "grad_norm": 1.59375,
       "learning_rate": 9.58904109589041e-05,
-      "loss": 2.2555,
+      "loss": 1.9943,
       "step": 35
     },
     {
       "epoch": 0.547945205479452,
-      "grad_norm": 0.94140625,
+      "grad_norm": 1.796875,
       "learning_rate": 0.00010958904109589041,
-      "loss": 2.1131,
+      "loss": 1.8382,
       "step": 40
     },
     {
       "epoch": 0.6164383561643836,
-      "grad_norm": 0.6875,
+      "grad_norm": 1.015625,
       "learning_rate": 0.0001232876712328767,
-      "loss": 1.9336,
+      "loss": 1.7025,
       "step": 45
     },
     {
       "epoch": 0.684931506849315,
-      "grad_norm": 0.462890625,
+      "grad_norm": 1.0859375,
       "learning_rate": 0.000136986301369863,
-      "loss": 1.7804,
+      "loss": 1.5949,
       "step": 50
     },
     {
       "epoch": 0.7534246575342466,
-      "grad_norm": 0.39453125,
+      "grad_norm": 0.33984375,
       "learning_rate": 0.00015068493150684933,
-      "loss": 1.6514,
+      "loss": 1.5023,
       "step": 55
     },
     {
       "epoch": 0.821917808219178,
-      "grad_norm": 0.4453125,
+      "grad_norm": 0.328125,
       "learning_rate": 0.00016438356164383562,
-      "loss": 1.5682,
+      "loss": 1.4428,
       "step": 60
     },
     {
       "epoch": 0.8904109589041096,
-      "grad_norm": 0.39453125,
+      "grad_norm": 0.65234375,
       "learning_rate": 0.00017808219178082192,
-      "loss": 1.5013,
+      "loss": 1.3924,
       "step": 65
     },
     {
       "epoch": 0.958904109589041,
-      "grad_norm": 0.4375,
+      "grad_norm": 0.40234375,
       "learning_rate": 0.0001917808219178082,
-      "loss": 1.4358,
+      "loss": 1.3438,
       "step": 70
     },
     {
       "epoch": 1.0,
-      "eval_loss": 2.546966075897217,
-      "eval_runtime": 0.552,
-      "eval_samples_per_second": 18.116,
-      "eval_steps_per_second": 1.812,
+      "eval_loss": 2.541907787322998,
+      "eval_runtime": 0.5592,
+      "eval_samples_per_second": 17.884,
+      "eval_steps_per_second": 1.788,
       "step": 73
     },
     {
       "epoch": 1.0273972602739727,
-      "grad_norm": 0.2412109375,
+      "grad_norm": 0.4296875,
       "learning_rate": 0.00019999542705801296,
-      "loss": 1.3872,
+      "loss": 1.3109,
       "step": 75
     },
     {
       "epoch": 1.095890410958904,
-      "grad_norm": 0.326171875,
+      "grad_norm": 0.478515625,
       "learning_rate": 0.00019994398626371643,
-      "loss": 1.3485,
+      "loss": 1.2812,
       "step": 80
     },
     {
       "epoch": 1.1643835616438356,
-      "grad_norm": 0.478515625,
+      "grad_norm": 0.515625,
       "learning_rate": 0.0001998354179989585,
-      "loss": 1.3316,
+      "loss": 1.2712,
       "step": 85
     },
     {
       "epoch": 1.2328767123287672,
-      "grad_norm": 1.1015625,
+      "grad_norm": 0.50390625,
       "learning_rate": 0.00019966978432080316,
-      "loss": 1.3049,
+      "loss": 1.2483,
       "step": 90
     },
     {
       "epoch": 1.3013698630136985,
-      "grad_norm": 0.458984375,
+      "grad_norm": 0.37109375,
       "learning_rate": 0.00019944717990461207,
-      "loss": 1.2894,
+      "loss": 1.2383,
       "step": 95
     },
     {
       "epoch": 1.36986301369863,
-      "grad_norm": 0.427734375,
+      "grad_norm": 0.65234375,
       "learning_rate": 0.000199167731989929,
-      "loss": 1.2703,
+      "loss": 1.2227,
       "step": 100
     },
     {
       "epoch": 1.4383561643835616,
-      "grad_norm": 0.2890625,
+      "grad_norm": 0.404296875,
       "learning_rate": 0.00019883160030775016,
-      "loss": 1.2547,
+      "loss": 1.2099,
       "step": 105
     },
     {
       "epoch": 1.5068493150684932,
-      "grad_norm": 0.291015625,
+      "grad_norm": 0.318359375,
       "learning_rate": 0.00019843897698922284,
-      "loss": 1.2613,
+      "loss": 1.2186,
       "step": 110
     },
     {
       "epoch": 1.5753424657534247,
-      "grad_norm": 0.2734375,
+      "grad_norm": 0.443359375,
       "learning_rate": 0.0001979900864558242,
-      "loss": 1.2388,
+      "loss": 1.1989,
       "step": 115
     },
     {
       "epoch": 1.643835616438356,
-      "grad_norm": 0.275390625,
+      "grad_norm": 0.32421875,
       "learning_rate": 0.00019748518529108316,
-      "loss": 1.2528,
+      "loss": 1.2146,
       "step": 120
     },
     {
       "epoch": 1.7123287671232876,
-      "grad_norm": 0.419921875,
+      "grad_norm": 0.435546875,
       "learning_rate": 0.00019692456209391846,
-      "loss": 1.2373,
+      "loss": 1.2015,
       "step": 125
     },
     {
       "epoch": 1.7808219178082192,
-      "grad_norm": 0.8046875,
+      "grad_norm": 0.671875,
       "learning_rate": 0.00019630853731367713,
-      "loss": 1.2241,
+      "loss": 1.1884,
       "step": 130
     },
     {
       "epoch": 1.8493150684931505,
-      "grad_norm": 0.33984375,
+      "grad_norm": 0.259765625,
       "learning_rate": 0.0001956374630669672,
-      "loss": 1.2263,
+      "loss": 1.1902,
       "step": 135
     },
     {
       "epoch": 1.9178082191780823,
-      "grad_norm": 0.287109375,
+      "grad_norm": 0.5234375,
       "learning_rate": 0.00019491172293638968,
-      "loss": 1.212,
+      "loss": 1.1781,
       "step": 140
     },
     {
       "epoch": 1.9863013698630136,
-      "grad_norm": 0.2734375,
+      "grad_norm": 0.396484375,
       "learning_rate": 0.00019413173175128473,
-      "loss": 1.2092,
+      "loss": 1.1767,
       "step": 145
     },
     {
       "epoch": 2.0,
-      "eval_loss": 2.5063788890838623,
-      "eval_runtime": 0.5512,
-      "eval_samples_per_second": 18.144,
-      "eval_steps_per_second": 1.814,
+      "eval_loss": 2.497932195663452,
+      "eval_runtime": 0.5527,
+      "eval_samples_per_second": 18.093,
+      "eval_steps_per_second": 1.809,
       "step": 146
     },
     {
       "epoch": 2.0547945205479454,
-      "grad_norm": 0.380859375,
+      "grad_norm": 0.494140625,
       "learning_rate": 0.00019329793535061723,
-      "loss": 1.1791,
+      "loss": 1.1406,
       "step": 150
     },
     {
       "epoch": 2.1232876712328768,
-      "grad_norm": 0.458984375,
+      "grad_norm": 0.65234375,
       "learning_rate": 0.00019241081032813772,
-      "loss": 1.1745,
+      "loss": 1.1353,
       "step": 155
     },
     {
       "epoch": 2.191780821917808,
-      "grad_norm": 0.380859375,
+      "grad_norm": 0.296875,
       "learning_rate": 0.0001914708637599636,
-      "loss": 1.1783,
+      "loss": 1.1387,
       "step": 160
     },
     {
       "epoch": 2.26027397260274,
-      "grad_norm": 0.2734375,
+      "grad_norm": 0.328125,
       "learning_rate": 0.00019047863291473717,
-      "loss": 1.156,
+      "loss": 1.1186,
       "step": 165
     },
     {
       "epoch": 2.328767123287671,
-      "grad_norm": 0.5625,
+      "grad_norm": 0.349609375,
       "learning_rate": 0.0001894346849465257,
-      "loss": 1.1642,
+      "loss": 1.1269,
       "step": 170
     },
     {
       "epoch": 2.3972602739726026,
-      "grad_norm": 0.65625,
+      "grad_norm": 0.435546875,
       "learning_rate": 0.00018833961657063885,
-      "loss": 1.176,
+      "loss": 1.1377,
       "step": 175
     },
     {
       "epoch": 2.4657534246575343,
-      "grad_norm": 0.640625,
+      "grad_norm": 0.400390625,
       "learning_rate": 0.00018719405372254948,
-      "loss": 1.16,
+      "loss": 1.1241,
       "step": 180
     },
     {
       "epoch": 2.5342465753424657,
-      "grad_norm": 0.400390625,
+      "grad_norm": 0.470703125,
       "learning_rate": 0.00018599865120011192,
-      "loss": 1.1653,
+      "loss": 1.1301,
       "step": 185
     },
     {
       "epoch": 2.602739726027397,
-      "grad_norm": 0.3125,
+      "grad_norm": 0.40625,
       "learning_rate": 0.00018475409228928312,
-      "loss": 1.1616,
+      "loss": 1.1252,
       "step": 190
     },
     {
       "epoch": 2.671232876712329,
-      "grad_norm": 0.6796875,
+      "grad_norm": 0.53515625,
       "learning_rate": 0.00018346108837355972,
-      "loss": 1.1549,
+      "loss": 1.1193,
       "step": 195
     },
     {
       "epoch": 2.73972602739726,
-      "grad_norm": 0.54296875,
+      "grad_norm": 0.494140625,
       "learning_rate": 0.00018212037852735486,
-      "loss": 1.1518,
+      "loss": 1.1163,
       "step": 200
     },
     {
       "epoch": 2.808219178082192,
-      "grad_norm": 0.52734375,
+      "grad_norm": 0.478515625,
       "learning_rate": 0.00018073272909354727,
-      "loss": 1.1516,
+      "loss": 1.1165,
       "step": 205
     },
     {
       "epoch": 2.8767123287671232,
-      "grad_norm": 0.375,
+      "grad_norm": 0.314453125,
       "learning_rate": 0.00017929893324544332,
-      "loss": 1.154,
+      "loss": 1.1205,
       "step": 210
     },
     {
       "epoch": 2.9452054794520546,
-      "grad_norm": 0.44921875,
+      "grad_norm": 0.703125,
       "learning_rate": 0.00017781981053340337,
-      "loss": 1.1501,
+      "loss": 1.1163,
       "step": 215
     },
     {
       "epoch": 3.0,
-      "eval_loss": 2.4999959468841553,
-      "eval_runtime": 0.5406,
-      "eval_samples_per_second": 18.499,
-      "eval_steps_per_second": 1.85,
+      "eval_loss": 2.496716022491455,
+      "eval_runtime": 0.5457,
+      "eval_samples_per_second": 18.324,
+      "eval_steps_per_second": 1.832,
       "step": 219
     },
     {
       "epoch": 3.0136986301369864,
-      "grad_norm": 0.353515625,
+      "grad_norm": 0.310546875,
       "learning_rate": 0.00017629620641639103,
-      "loss": 1.1368,
+      "loss": 1.1017,
       "step": 220
     },
     {
       "epoch": 3.0821917808219177,
-      "grad_norm": 0.337890625,
+      "grad_norm": 0.357421875,
       "learning_rate": 0.00017472899177871297,
-      "loss": 1.118,
+      "loss": 1.0756,
       "step": 225
     },
     {
       "epoch": 3.1506849315068495,
-      "grad_norm": 0.353515625,
+      "grad_norm": 0.36328125,
       "learning_rate": 0.00017311906243222614,
-      "loss": 1.1122,
+      "loss": 1.0712,
       "step": 230
     },
     {
       "epoch": 3.219178082191781,
-      "grad_norm": 0.34765625,
+      "grad_norm": 0.341796875,
       "learning_rate": 0.00017146733860429612,
-      "loss": 1.1205,
+      "loss": 1.0791,
       "step": 235
     },
     {
       "epoch": 3.287671232876712,
-      "grad_norm": 0.5625,
+      "grad_norm": 0.396484375,
       "learning_rate": 0.00016977476441179992,
-      "loss": 1.1197,
+      "loss": 1.078,
       "step": 240
     },
     {
       "epoch": 3.356164383561644,
-      "grad_norm": 0.796875,
+      "grad_norm": 0.51171875,
       "learning_rate": 0.0001680423073214737,
-      "loss": 1.1155,
+      "loss": 1.0753,
       "step": 245
     },
     {
       "epoch": 3.4246575342465753,
-      "grad_norm": 0.47265625,
+      "grad_norm": 0.359375,
       "learning_rate": 0.00016627095759691362,
-      "loss": 1.1255,
+      "loss": 1.0861,
       "step": 250
     },
     {
       "epoch": 3.493150684931507,
-      "grad_norm": 0.451171875,
+      "grad_norm": 0.7578125,
       "learning_rate": 0.00016446172773254629,
-      "loss": 1.1113,
+      "loss": 1.0716,
       "step": 255
     },
     {
       "epoch": 3.5616438356164384,
-      "grad_norm": 0.486328125,
+      "grad_norm": 0.453125,
       "learning_rate": 0.0001626156518748922,
-      "loss": 1.1073,
+      "loss": 1.0669,
       "step": 260
     },
     {
       "epoch": 3.6301369863013697,
-      "grad_norm": 1.046875,
+      "grad_norm": 0.71875,
       "learning_rate": 0.0001607337852314527,
-      "loss": 1.1155,
+      "loss": 1.0759,
       "step": 265
     },
     {
       "epoch": 3.6986301369863015,
-      "grad_norm": 0.6171875,
+      "grad_norm": 0.412109375,
       "learning_rate": 0.00015881720346755905,
-      "loss": 1.1239,
+      "loss": 1.0831,
       "step": 270
     },
     {
       "epoch": 3.767123287671233,
-      "grad_norm": 0.416015625,
+      "grad_norm": 0.34765625,
       "learning_rate": 0.00015686700209152738,
-      "loss": 1.1185,
+      "loss": 1.08,
       "step": 275
     },
     {
       "epoch": 3.8356164383561646,
-      "grad_norm": 0.380859375,
+      "grad_norm": 0.58203125,
       "learning_rate": 0.00015488429582847192,
-      "loss": 1.1125,
+      "loss": 1.0743,
       "step": 280
     },
     {
       "epoch": 3.904109589041096,
-      "grad_norm": 0.51953125,
+      "grad_norm": 0.37109375,
       "learning_rate": 0.0001528702179831338,
-      "loss": 1.0925,
+      "loss": 1.0544,
       "step": 285
     },
     {
       "epoch": 3.9726027397260273,
-      "grad_norm": 0.3671875,
+      "grad_norm": 0.39453125,
       "learning_rate": 0.00015082591979208976,
-      "loss": 1.0995,
+      "loss": 1.0605,
       "step": 290
     },
     {
       "epoch": 4.0,
-      "eval_loss": 2.506415843963623,
-      "eval_runtime": 0.5555,
-      "eval_samples_per_second": 18.001,
-      "eval_steps_per_second": 1.8,
+      "eval_loss": 2.5120725631713867,
+      "eval_runtime": 0.5541,
+      "eval_samples_per_second": 18.046,
+      "eval_steps_per_second": 1.805,
       "step": 292
     },
     {
       "epoch": 4.041095890410959,
-      "grad_norm": 0.515625,
+      "grad_norm": 0.41796875,
       "learning_rate": 0.00014875256976571135,
-      "loss": 1.0993,
+      "loss": 1.0555,
       "step": 295
     },
     {
       "epoch": 4.109589041095891,
-      "grad_norm": 0.375,
+      "grad_norm": 0.3671875,
       "learning_rate": 0.00014665135302025035,
-      "loss": 1.0789,
+      "loss": 1.0315,
       "step": 300
     },
     {
       "epoch": 4.178082191780822,
       "grad_norm": 0.3984375,
       "learning_rate": 0.00014452347060043237,
-      "loss": 1.0724,
+      "loss": 1.0247,
       "step": 305
     },
     {
       "epoch": 4.2465753424657535,
-      "grad_norm": 0.447265625,
+      "grad_norm": 0.45703125,
       "learning_rate": 0.0001423701387929459,
-      "loss": 1.0874,
+      "loss": 1.0381,
       "step": 310
     },
     {
       "epoch": 4.315068493150685,
       "grad_norm": 0.357421875,
       "learning_rate": 0.00014019258843121893,
-      "loss": 1.0913,
+      "loss": 1.0427,
       "step": 315
     },
     {
       "epoch": 4.383561643835616,
-      "grad_norm": 0.35546875,
+      "grad_norm": 0.36328125,
       "learning_rate": 0.00013799206419188103,
-      "loss": 1.085,
+      "loss": 1.038,
       "step": 320
     },
     {
       "epoch": 4.4520547945205475,
-      "grad_norm": 0.46875,
+      "grad_norm": 0.388671875,
       "learning_rate": 0.0001357698238833126,
-      "loss": 1.0733,
+      "loss": 1.0269,
       "step": 325
     },
     {
       "epoch": 4.52054794520548,
-      "grad_norm": 0.44140625,
+      "grad_norm": 0.404296875,
       "learning_rate": 0.00013352713772668765,
-      "loss": 1.0799,
+      "loss": 1.0337,
       "step": 330
     },
     {
       "epoch": 4.589041095890411,
-      "grad_norm": 0.412109375,
+      "grad_norm": 0.421875,
       "learning_rate": 0.00013126528762992247,
-      "loss": 1.0774,
+      "loss": 1.031,
       "step": 335
     },
     {
       "epoch": 4.657534246575342,
-      "grad_norm": 0.34375,
+      "grad_norm": 0.53125,
       "learning_rate": 0.00012898556645494325,
-      "loss": 1.0742,
+      "loss": 1.0291,
       "step": 340
     },
     {
       "epoch": 4.726027397260274,
-      "grad_norm": 0.408203125,
+      "grad_norm": 0.404296875,
       "learning_rate": 0.0001266892772786929,
-      "loss": 1.0806,
+      "loss": 1.0362,
       "step": 345
     },
     {
       "epoch": 4.794520547945205,
-      "grad_norm": 0.38671875,
+      "grad_norm": 0.34765625,
       "learning_rate": 0.00012437773264829897,
-      "loss": 1.0892,
+      "loss": 1.0442,
       "step": 350
     },
     {
       "epoch": 4.863013698630137,
-      "grad_norm": 0.482421875,
+      "grad_norm": 0.328125,
       "learning_rate": 0.00012205225383082843,
-      "loss": 1.0805,
+      "loss": 1.0357,
       "step": 355
     },
     {
       "epoch": 4.931506849315069,
-      "grad_norm": 0.3828125,
+      "grad_norm": 0.5703125,
       "learning_rate": 0.00011971417005805818,
-      "loss": 1.0786,
+      "loss": 1.0336,
       "step": 360
     },
     {
       "epoch": 5.0,
-      "grad_norm": 0.37890625,
+      "grad_norm": 0.359375,
       "learning_rate": 0.00011736481776669306,
-      "loss": 1.0809,
+      "loss": 1.0362,
       "step": 365
     },
     {
       "epoch": 5.0,
-      "eval_loss": 2.5177359580993652,
-      "eval_runtime": 0.545,
-      "eval_samples_per_second": 18.35,
-      "eval_steps_per_second": 1.835,
+      "eval_loss": 2.540134906768799,
+      "eval_runtime": 0.5557,
+      "eval_samples_per_second": 17.997,
+      "eval_steps_per_second": 1.8,
       "step": 365
     },
     {
       "epoch": 5.068493150684931,
-      "grad_norm": 0.388671875,
+      "grad_norm": 0.408203125,
       "learning_rate": 0.00011500553983446527,
-      "loss": 1.0503,
+      "loss": 0.9955,
       "step": 370
     },
     {
       "epoch": 5.136986301369863,
-      "grad_norm": 0.38671875,
+      "grad_norm": 0.375,
       "learning_rate": 0.00011263768481255264,
-      "loss": 1.0578,
+      "loss": 1.0022,
       "step": 375
     },
     {
       "epoch": 5.205479452054795,
-      "grad_norm": 0.41015625,
+      "grad_norm": 0.35546875,
       "learning_rate": 0.00011026260615475333,
-      "loss": 1.0601,
+      "loss": 1.0047,
       "step": 380
     },
     {
       "epoch": 5.273972602739726,
-      "grad_norm": 0.40625,
+      "grad_norm": 0.388671875,
       "learning_rate": 0.00010788166144385888,
-      "loss": 1.0519,
+      "loss": 0.9985,
       "step": 385
     },
     {
       "epoch": 5.342465753424658,
-      "grad_norm": 0.515625,
+      "grad_norm": 0.455078125,
       "learning_rate": 0.0001054962116156667,
-      "loss": 1.0555,
+      "loss": 1.002,
       "step": 390
     },
     {
       "epoch": 5.410958904109589,
-      "grad_norm": 0.56640625,
+      "grad_norm": 0.57421875,
       "learning_rate": 0.0001031076201810762,
-      "loss": 1.0673,
+      "loss": 1.0136,
       "step": 395
     },
     {
       "epoch": 5.47945205479452,
-      "grad_norm": 0.48046875,
+      "grad_norm": 0.39453125,
       "learning_rate": 0.00010071725244671282,
-      "loss": 1.0495,
+      "loss": 0.9965,
       "step": 400
     },
     {
       "epoch": 5.5479452054794525,
-      "grad_norm": 0.478515625,
+      "grad_norm": 0.431640625,
       "learning_rate": 9.83264747345259e-05,
-      "loss": 1.054,
+      "loss": 1.0025,
       "step": 405
     },
     {
       "epoch": 5.616438356164384,
-      "grad_norm": 0.408203125,
+      "grad_norm": 0.474609375,
       "learning_rate": 9.593665360080599e-05,
-      "loss": 1.0631,
+      "loss": 1.0099,
       "step": 410
     },
     {
       "epoch": 5.684931506849315,
-      "grad_norm": 0.376953125,
+      "grad_norm": 0.42578125,
       "learning_rate": 9.354915505506839e-05,
-      "loss": 1.0599,
+      "loss": 1.0071,
       "step": 415
     },
     {
       "epoch": 5.7534246575342465,
-      "grad_norm": 0.373046875,
+      "grad_norm": 0.44921875,
       "learning_rate": 9.116534377924883e-05,
-      "loss": 1.0554,
+      "loss": 1.0035,
       "step": 420
     },
     {
       "epoch": 5.821917808219178,
-      "grad_norm": 0.40625,
+      "grad_norm": 0.359375,
       "learning_rate": 8.878658234765858e-05,
-      "loss": 1.0503,
+      "loss": 0.9981,
       "step": 425
     },
     {
       "epoch": 5.890410958904109,
-      "grad_norm": 0.39453125,
+      "grad_norm": 0.384765625,
       "learning_rate": 8.641423044814374e-05,
-      "loss": 1.0632,
+      "loss": 1.0106,
       "step": 430
     },
     {
       "epoch": 5.958904109589041,
-      "grad_norm": 0.431640625,
+      "grad_norm": 0.396484375,
       "learning_rate": 8.404964410489485e-05,
-      "loss": 1.0583,
+      "loss": 1.0052,
       "step": 435
     },
     {
       "epoch": 6.0,
-      "eval_loss": 2.541933536529541,
-      "eval_runtime": 0.5507,
-      "eval_samples_per_second": 18.16,
-      "eval_steps_per_second": 1.816,
+      "eval_loss": 2.5710809230804443,
+      "eval_runtime": 0.5523,
+      "eval_samples_per_second": 18.105,
+      "eval_steps_per_second": 1.81,
       "step": 438
     },
     {
       "epoch": 6.027397260273973,
-      "grad_norm": 0.3828125,
+      "grad_norm": 0.373046875,
       "learning_rate": 8.169417490335007e-05,
-      "loss": 1.0504,
+      "loss": 0.9954,
       "step": 440
     },
     {
       "epoch": 6.095890410958904,
-      "grad_norm": 0.38671875,
+      "grad_norm": 0.37109375,
       "learning_rate": 7.934916921763628e-05,
-      "loss": 1.0301,
+      "loss": 0.9689,
       "step": 445
     },
     {
       "epoch": 6.164383561643835,
-      "grad_norm": 0.3828125,
+      "grad_norm": 0.37109375,
       "learning_rate": 7.701596744098818e-05,
-      "loss": 1.0358,
+      "loss": 0.9747,
       "step": 450
     },
     {
       "epoch": 6.232876712328767,
-      "grad_norm": 0.40625,
+      "grad_norm": 0.36328125,
       "learning_rate": 7.469590321958662e-05,
-      "loss": 1.0316,
+      "loss": 0.9715,
       "step": 455
     },
     {
       "epoch": 6.301369863013699,
-      "grad_norm": 0.59765625,
+      "grad_norm": 0.5078125,
       "learning_rate": 7.239030269025311e-05,
-      "loss": 1.0415,
+      "loss": 0.9809,
       "step": 460
     },
     {
       "epoch": 6.36986301369863,
-      "grad_norm": 0.482421875,
+      "grad_norm": 0.484375,
       "learning_rate": 7.010048372243698e-05,
-      "loss": 1.0336,
+      "loss": 0.9735,
       "step": 465
     },
     {
       "epoch": 6.438356164383562,
-      "grad_norm": 0.40625,
+      "grad_norm": 0.39453125,
       "learning_rate": 6.782775516492771e-05,
-      "loss": 1.0496,
+      "loss": 0.9895,
       "step": 470
     },
     {
       "epoch": 6.506849315068493,
-      "grad_norm": 0.38671875,
+      "grad_norm": 0.44140625,
       "learning_rate": 6.5573416097724e-05,
-      "loss": 1.04,
+      "loss": 0.9805,
       "step": 475
     },
     {
       "epoch": 6.575342465753424,
-      "grad_norm": 0.46875,
+      "grad_norm": 0.41015625,
       "learning_rate": 6.333875508948593e-05,
-      "loss": 1.0461,
+      "loss": 0.9865,
       "step": 480
     },
     {
       "epoch": 6.6438356164383565,
-      "grad_norm": 0.4296875,
+      "grad_norm": 0.52734375,
       "learning_rate": 6.112504946099604e-05,
-      "loss": 1.0439,
+      "loss": 0.9844,
       "step": 485
     },
     {
       "epoch": 6.712328767123288,
-      "grad_norm": 0.42578125,
+      "grad_norm": 0.373046875,
       "learning_rate": 5.8933564555049105e-05,
-      "loss": 1.0358,
+      "loss": 0.9775,
       "step": 490
     },
     {
       "epoch": 6.780821917808219,
-      "grad_norm": 0.5234375,
+      "grad_norm": 0.52734375,
       "learning_rate": 5.6765553013188766e-05,
-      "loss": 1.045,
+      "loss": 0.9852,
       "step": 495
     },
     {
       "epoch": 6.8493150684931505,
-      "grad_norm": 0.365234375,
+      "grad_norm": 0.376953125,
       "learning_rate": 5.462225405970401e-05,
-      "loss": 1.0329,
+      "loss": 0.9741,
       "step": 500
     },
     {
       "epoch": 6.917808219178082,
-      "grad_norm": 0.388671875,
+      "grad_norm": 0.416015625,
       "learning_rate": 5.2504892793295e-05,
-      "loss": 1.0442,
+      "loss": 0.9851,
       "step": 505
     },
     {
       "epoch": 6.986301369863014,
-      "grad_norm": 0.37109375,
+      "grad_norm": 0.3828125,
       "learning_rate": 5.041467948681269e-05,
-      "loss": 1.04,
+      "loss": 0.9813,
       "step": 510
     },
     {
       "epoch": 7.0,
-      "eval_loss": 2.5487561225891113,
-      "eval_runtime": 0.5536,
-      "eval_samples_per_second": 18.064,
-      "eval_steps_per_second": 1.806,
+      "eval_loss": 2.591195583343506,
+      "eval_runtime": 0.5557,
+      "eval_samples_per_second": 17.996,
+      "eval_steps_per_second": 1.8,
       "step": 511
     },
     {
       "epoch": 7.054794520547945,
-      "grad_norm": 0.384765625,
+      "grad_norm": 0.3984375,
       "learning_rate": 4.835280889547351e-05,
-      "loss": 1.0298,
+      "loss": 0.9656,
       "step": 515
     },
     {
       "epoch": 7.123287671232877,
-      "grad_norm": 0.376953125,
+      "grad_norm": 0.37109375,
       "learning_rate": 4.6320459573942856e-05,
-      "loss": 1.0213,
+      "loss": 0.9546,
       "step": 520
     },
     {
       "epoch": 7.191780821917808,
-      "grad_norm": 0.390625,
+      "grad_norm": 0.40625,
       "learning_rate": 4.431879320267972e-05,
-      "loss": 1.0334,
+      "loss": 0.9675,
       "step": 525
     },
     {
       "epoch": 7.260273972602739,
-      "grad_norm": 0.390625,
+      "grad_norm": 0.41015625,
       "learning_rate": 4.2348953923925916e-05,
-      "loss": 1.0208,
+      "loss": 0.9552,
       "step": 530
     },
     {
       "epoch": 7.328767123287671,
       "grad_norm": 0.365234375,
       "learning_rate": 4.041206768772022e-05,
-      "loss": 1.0323,
+      "loss": 0.9671,
       "step": 535
     },
     {
       "epoch": 7.397260273972603,
-      "grad_norm": 0.41015625,
+      "grad_norm": 0.392578125,
       "learning_rate": 3.850924160831115e-05,
-      "loss": 1.0295,
+      "loss": 0.9625,
       "step": 540
     },
     {
       "epoch": 7.465753424657534,
-      "grad_norm": 0.373046875,
+      "grad_norm": 0.38671875,
       "learning_rate": 3.6641563331336125e-05,
-      "loss": 1.0266,
+      "loss": 0.9607,
       "step": 545
     },
     {
       "epoch": 7.534246575342466,
-      "grad_norm": 0.3828125,
+      "grad_norm": 0.37109375,
       "learning_rate": 3.4810100412128747e-05,
-      "loss": 1.0218,
+      "loss": 0.9566,
       "step": 550
     },
     {
       "epoch": 7.602739726027397,
-      "grad_norm": 0.375,
+      "grad_norm": 0.376953125,
       "learning_rate": 3.3015899705509734e-05,
-      "loss": 1.0369,
+      "loss": 0.9709,
       "step": 555
     },
     {
       "epoch": 7.671232876712329,
-      "grad_norm": 0.40234375,
+      "grad_norm": 0.404296875,
       "learning_rate": 3.125998676740987e-05,
-      "loss": 1.0308,
+      "loss": 0.9659,
       "step": 560
     },
     {
       "epoch": 7.739726027397261,
-      "grad_norm": 0.3671875,
+      "grad_norm": 0.37890625,
       "learning_rate": 2.9543365268667867e-05,
-      "loss": 1.0225,
+      "loss": 0.9582,
       "step": 565
     },
     {
       "epoch": 7.808219178082192,
-      "grad_norm": 0.36328125,
+      "grad_norm": 0.365234375,
       "learning_rate": 2.7867016421336776e-05,
-      "loss": 1.0236,
+      "loss": 0.9593,
       "step": 570
     },
     {
       "epoch": 7.876712328767123,
-      "grad_norm": 0.3828125,
+      "grad_norm": 0.375,
       "learning_rate": 2.6231898417828603e-05,
-      "loss": 1.0272,
+      "loss": 0.9619,
       "step": 575
     },
     {
       "epoch": 7.945205479452055,
-      "grad_norm": 0.37890625,
+      "grad_norm": 0.369140625,
       "learning_rate": 2.4638945883216235e-05,
-      "loss": 1.0248,
+      "loss": 0.9593,
       "step": 580
     },
     {
       "epoch": 8.0,
-      "eval_loss": 2.557366132736206,
-      "eval_runtime": 0.554,
-      "eval_samples_per_second": 18.051,
-      "eval_steps_per_second": 1.805,
+      "eval_loss": 2.613189220428467,
+      "eval_runtime": 0.5557,
+      "eval_samples_per_second": 17.996,
+      "eval_steps_per_second": 1.8,
       "step": 584
     },
     {
       "epoch": 8.013698630136986,
-      "grad_norm": 0.365234375,
+      "grad_norm": 0.357421875,
       "learning_rate": 2.3089069341006565e-05,
-      "loss": 1.0197,
+      "loss": 0.954,
       "step": 585
     },
     {
       "epoch": 8.082191780821917,
       "grad_norm": 0.359375,
       "learning_rate": 2.1583154692689976e-05,
-      "loss": 1.0234,
+      "loss": 0.9542,
       "step": 590
     },
     {
       "epoch": 8.150684931506849,
-      "grad_norm": 0.435546875,
+      "grad_norm": 0.396484375,
       "learning_rate": 2.0122062711363532e-05,
-      "loss": 1.0212,
+      "loss": 0.9519,
       "step": 595
     },
     {
       "epoch": 8.219178082191782,
-      "grad_norm": 0.357421875,
+      "grad_norm": 0.361328125,
       "learning_rate": 1.8706628549717452e-05,
-      "loss": 1.0168,
+      "loss": 0.9478,
       "step": 600
     },
     {
       "epoch": 8.287671232876713,
       "grad_norm": 0.37109375,
       "learning_rate": 1.7337661262666294e-05,
-      "loss": 1.0172,
+      "loss": 0.9475,
       "step": 605
     },
     {
       "epoch": 8.356164383561644,
       "grad_norm": 0.373046875,
       "learning_rate": 1.601594334489702e-05,
-      "loss": 1.0167,
+      "loss": 0.9484,
       "step": 610
     },
     {
       "epoch": 8.424657534246576,
-      "grad_norm": 0.3828125,
+      "grad_norm": 0.37109375,
       "learning_rate": 1.474223028359939e-05,
-      "loss": 1.0271,
+      "loss": 0.9574,
       "step": 615
     },
     {
       "epoch": 8.493150684931507,
-      "grad_norm": 0.369140625,
+      "grad_norm": 0.37109375,
       "learning_rate": 1.3517250126632986e-05,
-      "loss": 1.0233,
+      "loss": 0.9537,
       "step": 620
     },
     {
       "epoch": 8.561643835616438,
-      "grad_norm": 0.3671875,
+      "grad_norm": 0.36328125,
       "learning_rate": 1.2341703066379074e-05,
-      "loss": 1.0209,
+      "loss": 0.9515,
       "step": 625
     },
     {
       "epoch": 8.63013698630137,
       "grad_norm": 0.376953125,
       "learning_rate": 1.1216261039514087e-05,
-      "loss": 1.0143,
+      "loss": 0.9459,
       "step": 630
     },
     {
       "epoch": 8.698630136986301,
-      "grad_norm": 0.380859375,
+      "grad_norm": 0.37109375,
       "learning_rate": 1.0141567342934132e-05,
-      "loss": 1.0231,
+      "loss": 0.9538,
       "step": 635
     },
     {
       "epoch": 8.767123287671232,
-      "grad_norm": 0.35546875,
+      "grad_norm": 0.359375,
       "learning_rate": 9.118236266049707e-06,
-      "loss": 1.014,
+      "loss": 0.9437,
       "step": 640
     },
     {
       "epoch": 8.835616438356164,
-      "grad_norm": 0.375,
+      "grad_norm": 0.37109375,
       "learning_rate": 8.146852739661105e-06,
-      "loss": 1.016,
+      "loss": 0.948,
       "step": 645
     },
     {
       "epoch": 8.904109589041095,
-      "grad_norm": 0.373046875,
+      "grad_norm": 0.369140625,
       "learning_rate": 7.2279720016148244e-06,
-      "loss": 1.0274,
+      "loss": 0.9595,
       "step": 650
     },
     {
       "epoch": 8.972602739726028,
-      "grad_norm": 0.392578125,
+      "grad_norm": 0.38671875,
       "learning_rate": 6.36211927943271e-06,
-      "loss": 1.021,
+      "loss": 0.953,
       "step": 655
     },
     {
       "epoch": 9.0,
-      "eval_loss": 2.5613903999328613,
-      "eval_runtime": 0.5421,
-      "eval_samples_per_second": 18.447,
-      "eval_steps_per_second": 1.845,
+      "eval_loss": 2.61802339553833,
+      "eval_runtime": 0.5528,
+      "eval_samples_per_second": 18.091,
+      "eval_steps_per_second": 1.809,
       "step": 657
     },
     {
       "epoch": 9.04109589041096,
-      "grad_norm": 0.361328125,
+      "grad_norm": 0.359375,
       "learning_rate": 5.549789490094304e-06,
-      "loss": 1.0166,
+      "loss": 0.9475,
       "step": 660
     },
     {
       "epoch": 9.10958904109589,
-      "grad_norm": 0.365234375,
+      "grad_norm": 0.369140625,
       "learning_rate": 4.79144695714504e-06,
-      "loss": 1.0251,
+      "loss": 0.9546,
       "step": 665
     },
     {
       "epoch": 9.178082191780822,
       "grad_norm": 0.390625,
       "learning_rate": 4.087525145291204e-06,
-      "loss": 1.0213,
+      "loss": 0.9509,
       "step": 670
     },
     {
       "epoch": 9.246575342465754,
-      "grad_norm": 0.404296875,
+      "grad_norm": 0.40234375,
       "learning_rate": 3.4384264126337328e-06,
-      "loss": 1.0181,
+      "loss": 0.9491,
       "step": 675
     },
     {
       "epoch": 9.315068493150685,
-      "grad_norm": 0.357421875,
+      "grad_norm": 0.353515625,
       "learning_rate": 2.8445217806824077e-06,
-      "loss": 1.0202,
+      "loss": 0.95,
       "step": 680
     },
     {
       "epoch": 9.383561643835616,
-      "grad_norm": 0.361328125,
+      "grad_norm": 0.359375,
       "learning_rate": 2.30615072228183e-06,
-      "loss": 1.0139,
+      "loss": 0.945,
       "step": 685
     },
     {
       "epoch": 9.452054794520548,
       "grad_norm": 0.359375,
       "learning_rate": 1.8236209675705274e-06,
-      "loss": 1.0292,
+      "loss": 0.9579,
       "step": 690
     },
     {
       "epoch": 9.520547945205479,
-      "grad_norm": 0.36328125,
+      "grad_norm": 0.369140625,
       "learning_rate": 1.397208328083921e-06,
-      "loss": 1.016,
+      "loss": 0.9463,
       "step": 695
     },
     {
       "epoch": 9.58904109589041,
-      "grad_norm": 0.35546875,
+      "grad_norm": 0.357421875,
       "learning_rate": 1.0271565391018922e-06,
-      "loss": 1.0206,
+      "loss": 0.9501,
       "step": 700
     },
     {
       "epoch": 9.657534246575342,
-      "grad_norm": 0.369140625,
+      "grad_norm": 0.365234375,
       "learning_rate": 7.136771203310245e-07,
-      "loss": 1.0181,
+      "loss": 0.9471,
       "step": 705
     },
     {
       "epoch": 9.726027397260275,
-      "grad_norm": 0.359375,
+      "grad_norm": 0.357421875,
       "learning_rate": 4.569492550008603e-07,
-      "loss": 1.0123,
+      "loss": 0.9428,
       "step": 710
     },
     {
       "epoch": 9.794520547945206,
       "grad_norm": 0.3515625,
       "learning_rate": 2.5711968744382974e-07,
-      "loss": 1.0101,
+      "loss": 0.941,
       "step": 715
     },
     {
       "epoch": 9.863013698630137,
       "grad_norm": 0.3671875,
       "learning_rate": 1.143026392168789e-07,
-      "loss": 1.0103,
+      "loss": 0.9407,
       "step": 720
     },
     {
       "epoch": 9.931506849315069,
-      "grad_norm": 0.365234375,
+      "grad_norm": 0.3671875,
       "learning_rate": 2.8579743813006432e-08,
-      "loss": 1.0282,
+      "loss": 0.9568,
       "step": 725
     },
     {
       "epoch": 10.0,
-      "grad_norm": 0.36328125,
+      "grad_norm": 0.365234375,
       "learning_rate": 0.0,
-      "loss": 1.0179,
+      "loss": 0.9482,
       "step": 730
     },
     {
       "epoch": 10.0,
-      "eval_loss": 2.5617620944976807,
-      "eval_runtime": 0.5453,
-      "eval_samples_per_second": 18.339,
-      "eval_steps_per_second": 1.834,
+      "eval_loss": 2.6190853118896484,
+      "eval_runtime": 0.5494,
+      "eval_samples_per_second": 18.201,
+      "eval_steps_per_second": 1.82,
       "step": 730
     },
     {
       "epoch": 10.0,
       "step": 730,
-      "total_flos": 4.287825372721971e+17,
-      "train_loss": 1.1969801510850044,
-      "train_runtime": 3891.9077,
-      "train_samples_per_second": 9.003,
-      "train_steps_per_second": 0.188
+      "total_flos": 4.3089359154446336e+17,
+      "train_loss": 1.133717959547696,
+      "train_runtime": 3901.5754,
+      "train_samples_per_second": 8.981,
+      "train_steps_per_second": 0.187
     }
   ],
   "logging_steps": 5,
@@ -1144,7 +1144,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.287825372721971e+17,
+  "total_flos": 4.3089359154446336e+17,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null