Model save

README.md

@@ -1,13 +1,11 @@
 ---
 library_name: peft
 tags:
-- alignment-handbook
 - trl
 - dpo
+- alignment-handbook
 - generated_from_trainer
 base_model: NbAiLab/nb-gpt-j-6B-v2
-datasets:
-- hugodk-sch/aftonposten_title_prefs
 model-index:
 - name: aftonposten-6b-align-scan
   results: []
@@ -18,17 +16,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # aftonposten-6b-align-scan
 
-This model is a fine-tuned version of [data/ap-gpt-j-6b-sft-qlora-04-08](https://huggingface.co/data/ap-gpt-j-6b-sft-qlora-04-08) on the hugodk-sch/aftonposten_title_prefs dataset.
+This model is a fine-tuned version of [NbAiLab/nb-gpt-j-6B-v2](https://huggingface.co/NbAiLab/nb-gpt-j-6B-v2) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4422
-- Rewards/chosen: 0.0026
-- Rewards/rejected: 0.0097
-- Rewards/accuracies: 0.4568
-- Rewards/margins: -0.0071
-- Logps/rejected: -37.5045
-- Logps/chosen: -34.0313
-- Logits/rejected: -2.2388
-- Logits/chosen: -2.2437
+- Logits/chosen: -2.2435
+- Logits/rejected: -2.2386
+- Logps/chosen: -34.0177
+- Logps/rejected: -37.5002
+- Loss: 0.9998
+- Rewards/accuracies: 0.4904
+- Rewards/chosen: 0.0135
+- Rewards/margins: 0.0004
+- Rewards/rejected: 0.0132
 
 ## Model description
 
@@ -63,9 +61,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Logits/chosen | Logits/rejected | Logps/chosen | Logps/rejected | Validation Loss | Rewards/accuracies | Rewards/chosen | Rewards/margins | Rewards/rejected |
 |:-------------:|:-----:|:----:|:-------------:|:---------------:|:------------:|:--------------:|:---------------:|:------------------:|:--------------:|:---------------:|:----------------:|
-| 0.3805        | 0.26  | 100  | -2.2441       | -2.2393         | -34.0171     | -37.5145       | 0.4061          | 0.5395             | 0.0139         | 0.0122          | 0.0017           |
-| 0.371         | 0.52  | 200  | -2.2439       | -2.2390         | -34.0231     | -37.5019       | 0.4323          | 0.5253             | 0.0091         | -0.0027         | 0.0118           |
-| 0.3839        | 0.78  | 300  | -2.2436       | -2.2387         | -34.0238     | -37.5004       | 0.4345          | 0.4809             | 0.0086         | -0.0044         | 0.0130           |
+| 1.0026        | 0.26  | 100  | -2.2433       | -2.2384         | -34.0232     | -37.4972       | 1.0066          | 0.4983             | 0.0091         | -0.0064         | 0.0155           |
+| 0.9923        | 0.52  | 200  | -2.2438       | -2.2389         | -34.0187     | -37.5175       | 0.9864          | 0.5328             | 0.0127         | 0.0134          | -0.0007          |
+| 0.9775        | 0.78  | 300  | -2.2435       | -2.2386         | -34.0177     | -37.5002       | 0.9998          | 0.4904             | 0.0135         | 0.0004          | 0.0132           |
 
 
 ### Framework versions

adapter_config.json

@@ -19,13 +19,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
+    "k_proj",
     "down_proj",
-    "up_proj",
     "o_proj",
     "gate_proj",
-    "q_proj",
-    "k_proj"
+    "up_proj",
+    "v_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_rslora": false

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd4d4dffa9b30971c702f1f39924c707a8b1925a5d67f8de841d1128fe03486b
+oid sha256:726700f20fcb28f7858cd1aa59e0cb2bd53a0fe4b7d2b662333531d42f0321bf
 size 176183216

all_results.json

@@ -1,20 +1,20 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.24367356300354,
-    "eval_logits/rejected": -2.2387921810150146,
-    "eval_logps/chosen": -34.03134536743164,
-    "eval_logps/rejected": -37.504520416259766,
-    "eval_loss": 0.4422106444835663,
-    "eval_rewards/accuracies": 0.4568106234073639,
-    "eval_rewards/chosen": 0.002567538060247898,
-    "eval_rewards/margins": -0.0071130492724478245,
-    "eval_rewards/rejected": 0.009680588729679585,
-    "eval_runtime": 145.9043,
+    "eval_logits/chosen": -2.2434139251708984,
+    "eval_logits/rejected": -2.2385354042053223,
+    "eval_logps/chosen": -34.0224494934082,
+    "eval_logps/rejected": -37.51002883911133,
+    "eval_loss": 0.9960440397262573,
+    "eval_rewards/accuracies": 0.5141196250915527,
+    "eval_rewards/chosen": 0.0096820630133152,
+    "eval_rewards/margins": 0.004409347660839558,
+    "eval_rewards/rejected": 0.005272718146443367,
+    "eval_runtime": 145.6562,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.351,
+    "eval_samples_per_second": 2.355,
     "eval_steps_per_second": 0.295,
-    "train_loss": 0.08930327675559303,
-    "train_runtime": 627.8357,
+    "train_loss": 0.21957750196580764,
+    "train_runtime": 627.8531,
     "train_samples": 3079,
     "train_samples_per_second": 4.904,
     "train_steps_per_second": 0.613

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.24367356300354,
-    "eval_logits/rejected": -2.2387921810150146,
-    "eval_logps/chosen": -34.03134536743164,
-    "eval_logps/rejected": -37.504520416259766,
-    "eval_loss": 0.4422106444835663,
-    "eval_rewards/accuracies": 0.4568106234073639,
-    "eval_rewards/chosen": 0.002567538060247898,
-    "eval_rewards/margins": -0.0071130492724478245,
-    "eval_rewards/rejected": 0.009680588729679585,
-    "eval_runtime": 145.9043,
+    "eval_logits/chosen": -2.2434139251708984,
+    "eval_logits/rejected": -2.2385354042053223,
+    "eval_logps/chosen": -34.0224494934082,
+    "eval_logps/rejected": -37.51002883911133,
+    "eval_loss": 0.9960440397262573,
+    "eval_rewards/accuracies": 0.5141196250915527,
+    "eval_rewards/chosen": 0.0096820630133152,
+    "eval_rewards/margins": 0.004409347660839558,
+    "eval_rewards/rejected": 0.005272718146443367,
+    "eval_runtime": 145.6562,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.351,
+    "eval_samples_per_second": 2.355,
     "eval_steps_per_second": 0.295
 }

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 1.0,
-    "train_loss": 0.08930327675559303,
-    "train_runtime": 627.8357,
+    "train_loss": 0.21957750196580764,
+    "train_runtime": 627.8531,
     "train_samples": 3079,
     "train_samples_per_second": 4.904,
     "train_steps_per_second": 0.613

trainer_state.json

@@ -15,7 +15,7 @@
       "logits/rejected": -1.7377450466156006,
       "logps/chosen": -29.553977966308594,
       "logps/rejected": -42.813133239746094,
-      "loss": 0.3906,
+      "loss": 1.0,
       "rewards/accuracies": 0.0,
       "rewards/chosen": 0.0,
       "rewards/margins": 0.0,
@@ -25,311 +25,311 @@
     {
       "epoch": 0.03,
       "learning_rate": 1.2820512820512818e-07,
-      "logits/chosen": -1.8662992715835571,
-      "logits/rejected": -1.8706117868423462,
-      "logps/chosen": -36.97681427001953,
-      "logps/rejected": -33.66523361206055,
-      "loss": 0.3447,
-      "rewards/accuracies": 0.5416666865348816,
-      "rewards/chosen": 0.023829741403460503,
-      "rewards/margins": 0.05195777863264084,
-      "rewards/rejected": -0.028128040954470634,
+      "logits/chosen": -1.866065502166748,
+      "logits/rejected": -1.8703795671463013,
+      "logps/chosen": -36.988380432128906,
+      "logps/rejected": -33.66728210449219,
+      "loss": 0.9557,
+      "rewards/accuracies": 0.5694444179534912,
+      "rewards/chosen": 0.014575847424566746,
+      "rewards/margins": 0.04434243589639664,
+      "rewards/rejected": -0.02976658195257187,
       "step": 10
     },
     {
       "epoch": 0.05,
       "learning_rate": 2.5641025641025636e-07,
-      "logits/chosen": -1.997193694114685,
-      "logits/rejected": -1.9998573064804077,
-      "logps/chosen": -29.65359878540039,
-      "logps/rejected": -29.054311752319336,
-      "loss": 0.4534,
+      "logits/chosen": -1.997332215309143,
+      "logits/rejected": -1.999983549118042,
+      "logps/chosen": -29.625896453857422,
+      "logps/rejected": -29.035802841186523,
+      "loss": 1.0091,
       "rewards/accuracies": 0.4375,
-      "rewards/chosen": -0.009124360978603363,
-      "rewards/margins": -0.016431041061878204,
-      "rewards/rejected": 0.007306680083274841,
+      "rewards/chosen": 0.013037634082138538,
+      "rewards/margins": -0.009074507281184196,
+      "rewards/rejected": 0.02211214043200016,
       "step": 20
     },
     {
       "epoch": 0.08,
       "learning_rate": 3.8461538461538463e-07,
-      "logits/chosen": -1.9197280406951904,
-      "logits/rejected": -1.9170429706573486,
-      "logps/chosen": -31.414775848388672,
-      "logps/rejected": -33.24064254760742,
-      "loss": 0.4025,
-      "rewards/accuracies": 0.5625,
-      "rewards/chosen": 0.0010542668169364333,
-      "rewards/margins": 0.018799806013703346,
-      "rewards/rejected": -0.01774553768336773,
+      "logits/chosen": -1.9199495315551758,
+      "logits/rejected": -1.917249321937561,
+      "logps/chosen": -31.421478271484375,
+      "logps/rejected": -33.2115364074707,
+      "loss": 1.0098,
+      "rewards/accuracies": 0.4625000059604645,
+      "rewards/chosen": -0.004308355040848255,
+      "rewards/margins": -0.009847955778241158,
+      "rewards/rejected": 0.005539602134376764,
       "step": 30
     },
     {
       "epoch": 0.1,
       "learning_rate": 4.999896948438433e-07,
-      "logits/chosen": -2.0171010494232178,
-      "logits/rejected": -2.008350372314453,
-      "logps/chosen": -32.59648513793945,
-      "logps/rejected": -32.50862121582031,
-      "loss": 0.4532,
+      "logits/chosen": -2.0169284343719482,
+      "logits/rejected": -2.008178949356079,
+      "logps/chosen": -32.59435272216797,
+      "logps/rejected": -32.49193572998047,
+      "loss": 1.0306,
       "rewards/accuracies": 0.44999998807907104,
-      "rewards/chosen": -0.015740757808089256,
-      "rewards/margins": -0.018979396671056747,
-      "rewards/rejected": 0.0032386414241045713,
+      "rewards/chosen": -0.014034454710781574,
+      "rewards/margins": -0.030623534694314003,
+      "rewards/rejected": 0.016589080914855003,
       "step": 40
     },
     {
       "epoch": 0.13,
       "learning_rate": 4.987541037542186e-07,
-      "logits/chosen": -1.8645904064178467,
-      "logits/rejected": -1.8538070917129517,
-      "logps/chosen": -33.542823791503906,
-      "logps/rejected": -35.43744659423828,
-      "loss": 0.4304,
-      "rewards/accuracies": 0.5,
-      "rewards/chosen": 0.011724205687642097,
-      "rewards/margins": -0.00019102543592453003,
-      "rewards/rejected": 0.011915231123566628,
+      "logits/chosen": -1.86457097530365,
+      "logits/rejected": -1.8537908792495728,
+      "logps/chosen": -33.56566619873047,
+      "logps/rejected": -35.423240661621094,
+      "loss": 1.0298,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.006550998892635107,
+      "rewards/margins": -0.029834842309355736,
+      "rewards/rejected": 0.02328384295105934,
       "step": 50
     },
     {
       "epoch": 0.16,
       "learning_rate": 4.954691471941118e-07,
-      "logits/chosen": -1.9461469650268555,
-      "logits/rejected": -1.948094129562378,
-      "logps/chosen": -32.587215423583984,
-      "logps/rejected": -33.21445846557617,
-      "loss": 0.3539,
-      "rewards/accuracies": 0.637499988079071,
-      "rewards/chosen": 0.010374858975410461,
-      "rewards/margins": 0.04483799636363983,
-      "rewards/rejected": -0.03446313738822937,
+      "logits/chosen": -1.9449050426483154,
+      "logits/rejected": -1.9468472003936768,
+      "logps/chosen": -32.59955596923828,
+      "logps/rejected": -33.1828498840332,
+      "loss": 0.9903,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": 0.0005022765835747123,
+      "rewards/margins": 0.009675229899585247,
+      "rewards/rejected": -0.00917295552790165,
       "step": 60
     },
     {
       "epoch": 0.18,
       "learning_rate": 4.901618883413548e-07,
-      "logits/chosen": -2.079591989517212,
-      "logits/rejected": -2.0845742225646973,
-      "logps/chosen": -34.00908279418945,
-      "logps/rejected": -36.58150863647461,
-      "loss": 0.4506,
+      "logits/chosen": -2.079878330230713,
+      "logits/rejected": -2.084862232208252,
+      "logps/chosen": -33.98878860473633,
+      "logps/rejected": -36.574462890625,
+      "loss": 0.9958,
       "rewards/accuracies": 0.512499988079071,
-      "rewards/chosen": -0.01648041605949402,
-      "rewards/margins": -0.006370754446834326,
-      "rewards/rejected": -0.010109663009643555,
+      "rewards/chosen": -0.00024233898147940636,
+      "rewards/margins": 0.004231657367199659,
+      "rewards/rejected": -0.004473996348679066,
       "step": 70
     },
     {
       "epoch": 0.21,
       "learning_rate": 4.828760511501322e-07,
-      "logits/chosen": -1.9423307180404663,
-      "logits/rejected": -1.9454963207244873,
-      "logps/chosen": -34.39698028564453,
-      "logps/rejected": -34.59761428833008,
-      "loss": 0.3889,
+      "logits/chosen": -1.9425569772720337,
+      "logits/rejected": -1.9457191228866577,
+      "logps/chosen": -34.40068054199219,
+      "logps/rejected": -34.5762939453125,
+      "loss": 0.9912,
       "rewards/accuracies": 0.5249999761581421,
-      "rewards/chosen": 0.024663064628839493,
-      "rewards/margins": 0.028829574584960938,
-      "rewards/rejected": -0.00416650902479887,
+      "rewards/chosen": 0.021701809018850327,
+      "rewards/margins": 0.00881027989089489,
+      "rewards/rejected": 0.012891530990600586,
       "step": 80
     },
     {
       "epoch": 0.23,
       "learning_rate": 4.736716601303429e-07,
-      "logits/chosen": -1.9508873224258423,
-      "logits/rejected": -1.9553953409194946,
-      "logps/chosen": -32.48583221435547,
-      "logps/rejected": -32.35867691040039,
-      "loss": 0.4445,
-      "rewards/accuracies": 0.4625000059604645,
-      "rewards/chosen": -0.014726865105330944,
-      "rewards/margins": -0.014909917488694191,
-      "rewards/rejected": 0.00018304325931239873,
+      "logits/chosen": -1.9507396221160889,
+      "logits/rejected": -1.9552500247955322,
+      "logps/chosen": -32.460357666015625,
+      "logps/rejected": -32.354434967041016,
+      "loss": 0.9979,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": 0.005654479376971722,
+      "rewards/margins": 0.002074276329949498,
+      "rewards/rejected": 0.0035802037455141544,
       "step": 90
     },
     {
       "epoch": 0.26,
       "learning_rate": 4.62624545834521e-07,
-      "logits/chosen": -2.0491719245910645,
-      "logits/rejected": -2.047179937362671,
-      "logps/chosen": -32.22635269165039,
-      "logps/rejected": -31.287487030029297,
-      "loss": 0.3805,
-      "rewards/accuracies": 0.550000011920929,
-      "rewards/chosen": 0.011584864929318428,
-      "rewards/margins": 0.02860759198665619,
-      "rewards/rejected": -0.01702272891998291,
+      "logits/chosen": -2.0492873191833496,
+      "logits/rejected": -2.0472888946533203,
+      "logps/chosen": -32.23810958862305,
+      "logps/rejected": -31.260278701782227,
+      "loss": 1.0026,
+      "rewards/accuracies": 0.4625000059604645,
+      "rewards/chosen": 0.0021828364115208387,
+      "rewards/margins": -0.002563956892117858,
+      "rewards/rejected": 0.004746791906654835,
       "step": 100
     },
     {
       "epoch": 0.26,
-      "eval_logits/chosen": -2.244148015975952,
-      "eval_logits/rejected": -2.2392663955688477,
-      "eval_logps/chosen": -34.01713943481445,
-      "eval_logps/rejected": -37.514495849609375,
-      "eval_loss": 0.4061162769794464,
-      "eval_rewards/accuracies": 0.5394518375396729,
-      "eval_rewards/chosen": 0.013928660191595554,
-      "eval_rewards/margins": 0.012230273336172104,
-      "eval_rewards/rejected": 0.0016983875539153814,
-      "eval_runtime": 146.1091,
-      "eval_samples_per_second": 2.348,
+      "eval_logits/chosen": -2.243312358856201,
+      "eval_logits/rejected": -2.238436222076416,
+      "eval_logps/chosen": -34.023216247558594,
+      "eval_logps/rejected": -37.49723434448242,
+      "eval_loss": 1.0065712928771973,
+      "eval_rewards/accuracies": 0.49833887815475464,
+      "eval_rewards/chosen": 0.009070919826626778,
+      "eval_rewards/margins": -0.006438401062041521,
+      "eval_rewards/rejected": 0.015509321354329586,
+      "eval_runtime": 146.1739,
+      "eval_samples_per_second": 2.347,
       "eval_steps_per_second": 0.294,
       "step": 100
     },
     {
       "epoch": 0.29,
       "learning_rate": 4.4982572012636904e-07,
-      "logits/chosen": -2.005580186843872,
-      "logits/rejected": -2.0031564235687256,
-      "logps/chosen": -33.24415588378906,
-      "logps/rejected": -33.99993133544922,
-      "loss": 0.4792,
-      "rewards/accuracies": 0.4375,
-      "rewards/chosen": -0.00026987865567207336,
-      "rewards/margins": -0.02382112666964531,
-      "rewards/rejected": 0.023551244288682938,
+      "logits/chosen": -2.005169153213501,
+      "logits/rejected": -2.0027499198913574,
+      "logps/chosen": -33.2365837097168,
+      "logps/rejected": -34.01953125,
+      "loss": 1.0021,
+      "rewards/accuracies": 0.4625000059604645,
+      "rewards/chosen": 0.005784005858004093,
+      "rewards/margins": -0.002089007291942835,
+      "rewards/rejected": 0.007873011752963066,
       "step": 110
     },
     {
       "epoch": 0.31,
       "learning_rate": 4.353806263777677e-07,
-      "logits/chosen": -2.01680326461792,
-      "logits/rejected": -2.0084121227264404,
-      "logps/chosen": -32.45465850830078,
-      "logps/rejected": -32.1729850769043,
-      "loss": 0.4286,
-      "rewards/accuracies": 0.4375,
-      "rewards/chosen": -0.007803081069141626,
-      "rewards/margins": 0.0013800703454762697,
-      "rewards/rejected": -0.009183152578771114,
+      "logits/chosen": -2.0166728496551514,
+      "logits/rejected": -2.0083022117614746,
+      "logps/chosen": -32.457847595214844,
+      "logps/rejected": -32.18357467651367,
+      "loss": 0.9927,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": -0.010353903286159039,
+      "rewards/margins": 0.0072973729111254215,
+      "rewards/rejected": -0.017651278525590897,
       "step": 120
     },
     {
       "epoch": 0.34,
       "learning_rate": 4.194082707715275e-07,
-      "logits/chosen": -2.0466856956481934,
-      "logits/rejected": -2.0386409759521484,
-      "logps/chosen": -30.494335174560547,
-      "logps/rejected": -32.051963806152344,
-      "loss": 0.4497,
-      "rewards/accuracies": 0.4749999940395355,
-      "rewards/chosen": -0.014895597472786903,
-      "rewards/margins": -0.015729816630482674,
-      "rewards/rejected": 0.0008342192741110921,
+      "logits/chosen": -2.0462448596954346,
+      "logits/rejected": -2.038203001022339,
+      "logps/chosen": -30.475027084350586,
+      "logps/rejected": -32.046302795410156,
+      "loss": 1.0048,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": 0.0005506269517354667,
+      "rewards/margins": -0.004812211729586124,
+      "rewards/rejected": 0.005362839438021183,
       "step": 130
     },
     {
       "epoch": 0.36,
       "learning_rate": 4.020402418666621e-07,
-      "logits/chosen": -1.977065086364746,
-      "logits/rejected": -1.987343430519104,
-      "logps/chosen": -31.37868309020996,
-      "logps/rejected": -32.54730224609375,
-      "loss": 0.3589,
-      "rewards/accuracies": 0.574999988079071,
-      "rewards/chosen": 0.03278205543756485,
-      "rewards/margins": 0.04271895810961723,
-      "rewards/rejected": -0.009936909191310406,
+      "logits/chosen": -1.976986289024353,
+      "logits/rejected": -1.9872560501098633,
+      "logps/chosen": -31.407278060913086,
+      "logps/rejected": -32.543296813964844,
+      "loss": 0.9834,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": 0.009907384403049946,
+      "rewards/margins": 0.016641682013869286,
+      "rewards/rejected": -0.006734299007803202,
       "step": 140
     },
     {
       "epoch": 0.39,
       "learning_rate": 3.8341962650351185e-07,
-      "logits/chosen": -1.8916466236114502,
-      "logits/rejected": -1.8927490711212158,
-      "logps/chosen": -34.209651947021484,
-      "logps/rejected": -34.766143798828125,
-      "loss": 0.4325,
-      "rewards/accuracies": 0.42500001192092896,
-      "rewards/chosen": -0.013569498434662819,
-      "rewards/margins": -0.00645996630191803,
-      "rewards/rejected": -0.007109532598406076,
+      "logits/chosen": -1.8905508518218994,
+      "logits/rejected": -1.891632080078125,
+      "logps/chosen": -34.20501708984375,
+      "logps/rejected": -34.77235412597656,
+      "loss": 0.9978,
+      "rewards/accuracies": 0.4625000059604645,
+      "rewards/chosen": -0.009861050173640251,
+      "rewards/margins": 0.0022162157110869884,
+      "rewards/rejected": -0.012077265419065952,
       "step": 150
     },
     {
       "epoch": 0.42,
       "learning_rate": 3.636998309800572e-07,
-      "logits/chosen": -1.9427303075790405,
-      "logits/rejected": -1.939252495765686,
-      "logps/chosen": -36.14452362060547,
-      "logps/rejected": -32.73284149169922,
-      "loss": 0.3796,
-      "rewards/accuracies": 0.550000011920929,
-      "rewards/chosen": 0.021723434329032898,
-      "rewards/margins": 0.02979486621916294,
-      "rewards/rejected": -0.008071433752775192,
+      "logits/chosen": -1.94281005859375,
+      "logits/rejected": -1.939327597618103,
+      "logps/chosen": -36.144107818603516,
+      "logps/rejected": -32.72822570800781,
+      "loss": 0.9736,
+      "rewards/accuracies": 0.5375000238418579,
+      "rewards/chosen": 0.02205902896821499,
+      "rewards/margins": 0.026437148451805115,
+      "rewards/rejected": -0.004378115758299828,
       "step": 160
     },
     {
       "epoch": 0.44,
       "learning_rate": 3.430433172111807e-07,
-      "logits/chosen": -2.0424513816833496,
-      "logits/rejected": -2.035060405731201,
-      "logps/chosen": -33.786170959472656,
-      "logps/rejected": -31.34820556640625,
-      "loss": 0.4172,
-      "rewards/accuracies": 0.512499988079071,
-      "rewards/chosen": 0.009350698441267014,
-      "rewards/margins": 0.006434415467083454,
-      "rewards/rejected": 0.0029162843711674213,
+      "logits/chosen": -2.0420708656311035,
+      "logits/rejected": -2.0346803665161133,
+      "logps/chosen": -33.771934509277344,
+      "logps/rejected": -31.371145248413086,
+      "loss": 0.9638,
+      "rewards/accuracies": 0.5375000238418579,
+      "rewards/chosen": 0.020745161920785904,
+      "rewards/margins": 0.036180444061756134,
+      "rewards/rejected": -0.015435280278325081,
       "step": 170
     },
     {
       "epoch": 0.47,
       "learning_rate": 3.216202642830543e-07,
-      "logits/chosen": -2.0475661754608154,
-      "logits/rejected": -2.052834987640381,
-      "logps/chosen": -32.528114318847656,
-      "logps/rejected": -32.50902557373047,
-      "loss": 0.3766,
-      "rewards/accuracies": 0.4749999940395355,
-      "rewards/chosen": 0.013338183984160423,
-      "rewards/margins": 0.03244578838348389,
-      "rewards/rejected": -0.019107606261968613,
+      "logits/chosen": -2.0475738048553467,
+      "logits/rejected": -2.0528526306152344,
+      "logps/chosen": -32.524593353271484,
+      "logps/rejected": -32.510643005371094,
+      "loss": 0.9634,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": 0.016154326498508453,
+      "rewards/margins": 0.036555372178554535,
+      "rewards/rejected": -0.020401041954755783,
       "step": 180
     },
     {
       "epoch": 0.49,
       "learning_rate": 2.9960716642946403e-07,
-      "logits/chosen": -2.048661231994629,
-      "logits/rejected": -2.045872926712036,
-      "logps/chosen": -31.496755599975586,
-      "logps/rejected": -31.314464569091797,
-      "loss": 0.4475,
+      "logits/chosen": -2.048490524291992,
+      "logits/rejected": -2.0457024574279785,
+      "logps/chosen": -31.492746353149414,
+      "logps/rejected": -31.319293975830078,
+      "loss": 1.0022,
       "rewards/accuracies": 0.5,
-      "rewards/chosen": -0.00624021515250206,
-      "rewards/margins": -0.009242123924195766,
-      "rewards/rejected": 0.003001909703016281,
+      "rewards/chosen": -0.003033895045518875,
+      "rewards/margins": -0.002170243766158819,
+      "rewards/rejected": -0.0008636537822894752,
       "step": 190
     },
     {
       "epoch": 0.52,
       "learning_rate": 2.771853789806683e-07,
-      "logits/chosen": -1.9193611145019531,
-      "logits/rejected": -1.9240529537200928,
-      "logps/chosen": -31.579212188720703,
-      "logps/rejected": -32.80603790283203,
-      "loss": 0.371,
-      "rewards/accuracies": 0.6000000238418579,
-      "rewards/chosen": 0.02165621519088745,
-      "rewards/margins": 0.033739686012268066,
-      "rewards/rejected": -0.012083468958735466,
+      "logits/chosen": -1.9185683727264404,
+      "logits/rejected": -1.9232347011566162,
+      "logps/chosen": -31.5926513671875,
+      "logps/rejected": -32.78697204589844,
+      "loss": 0.9923,
+      "rewards/accuracies": 0.5874999761581421,
+      "rewards/chosen": 0.01090637780725956,
+      "rewards/margins": 0.007740379776805639,
+      "rewards/rejected": 0.0031659987289458513,
       "step": 200
     },
     {
       "epoch": 0.52,
-      "eval_logits/chosen": -2.243856191635132,
-      "eval_logits/rejected": -2.2389791011810303,
-      "eval_logps/chosen": -34.023136138916016,
-      "eval_logps/rejected": -37.501853942871094,
-      "eval_loss": 0.4322855770587921,
-      "eval_rewards/accuracies": 0.5253322720527649,
-      "eval_rewards/chosen": 0.009133166633546352,
-      "eval_rewards/margins": -0.002679171971976757,
-      "eval_rewards/rejected": 0.01181233860552311,
-      "eval_runtime": 145.9618,
+      "eval_logits/chosen": -2.243807315826416,
+      "eval_logits/rejected": -2.2389235496520996,
+      "eval_logps/chosen": -34.018714904785156,
+      "eval_logps/rejected": -37.517478942871094,
+      "eval_loss": 0.9864435195922852,
+      "eval_rewards/accuracies": 0.5328072905540466,
+      "eval_rewards/chosen": 0.012671924196183681,
+      "eval_rewards/margins": 0.013361074030399323,
+      "eval_rewards/rejected": -0.0006891504744999111,
+      "eval_runtime": 145.9847,
       "eval_samples_per_second": 2.35,
       "eval_steps_per_second": 0.295,
       "step": 200
@@ -337,277 +337,277 @@
     {
       "epoch": 0.55,
       "learning_rate": 2.5453962426402e-07,
-      "logits/chosen": -2.0320448875427246,
-      "logits/rejected": -2.0427393913269043,
-      "logps/chosen": -31.947372436523438,
-      "logps/rejected": -33.899864196777344,
-      "loss": 0.3622,
-      "rewards/accuracies": 0.6000000238418579,
-      "rewards/chosen": 0.019143681973218918,
-      "rewards/margins": 0.04887578636407852,
-      "rewards/rejected": -0.029732098802924156,
+      "logits/chosen": -2.0318503379821777,
+      "logits/rejected": -2.042539596557617,
+      "logps/chosen": -31.948400497436523,
+      "logps/rejected": -33.86983871459961,
+      "loss": 0.976,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.018323851749300957,
+      "rewards/margins": 0.024035325273871422,
+      "rewards/rejected": -0.005711475852876902,
       "step": 210
     },
     {
       "epoch": 0.57,
       "learning_rate": 2.318564697655179e-07,
-      "logits/chosen": -1.9257709980010986,
-      "logits/rejected": -1.940639853477478,
-      "logps/chosen": -30.079341888427734,
-      "logps/rejected": -31.5616397857666,
-      "loss": 0.3884,
-      "rewards/accuracies": 0.612500011920929,
-      "rewards/chosen": 0.025758206844329834,
-      "rewards/margins": 0.028527002781629562,
-      "rewards/rejected": -0.0027687971014529467,
+      "logits/chosen": -1.9251388311386108,
+      "logits/rejected": -1.9399843215942383,
+      "logps/chosen": -30.099853515625,
+      "logps/rejected": -31.55409812927246,
+      "loss": 0.9939,
+      "rewards/accuracies": 0.5249999761581421,
+      "rewards/chosen": 0.009348717518150806,
+      "rewards/margins": 0.006087350193411112,
+      "rewards/rejected": 0.003261365694925189,
       "step": 220
     },
     {
       "epoch": 0.6,
       "learning_rate": 2.093227910899832e-07,
-      "logits/chosen": -1.9829959869384766,
-      "logits/rejected": -1.9869670867919922,
-      "logps/chosen": -33.4053955078125,
-      "logps/rejected": -31.562353134155273,
-      "loss": 0.3918,
-      "rewards/accuracies": 0.5874999761581421,
-      "rewards/chosen": 0.009657363407313824,
-      "rewards/margins": 0.02010050043463707,
-      "rewards/rejected": -0.010443134233355522,
+      "logits/chosen": -1.9835479259490967,
+      "logits/rejected": -1.9875112771987915,
+      "logps/chosen": -33.387638092041016,
+      "logps/rejected": -31.554845809936523,
+      "loss": 0.9717,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": 0.023866940289735794,
+      "rewards/margins": 0.028299665078520775,
+      "rewards/rejected": -0.00443272665143013,
       "step": 230
     },
     {
       "epoch": 0.62,
       "learning_rate": 1.8712423238279356e-07,
-      "logits/chosen": -1.9836658239364624,
-      "logits/rejected": -1.9616800546646118,
-      "logps/chosen": -34.175201416015625,
-      "logps/rejected": -34.966102600097656,
-      "loss": 0.4346,
-      "rewards/accuracies": 0.5375000238418579,
-      "rewards/chosen": -0.017159918323159218,
-      "rewards/margins": -0.004913450218737125,
-      "rewards/rejected": -0.012246469967067242,
+      "logits/chosen": -1.9826898574829102,
+      "logits/rejected": -1.9607274532318115,
+      "logps/chosen": -34.158443450927734,
+      "logps/rejected": -34.963783264160156,
+      "loss": 0.9934,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": -0.0037571662105619907,
+      "rewards/margins": 0.0066327848471701145,
+      "rewards/rejected": -0.01038995012640953,
       "step": 240
     },
     {
       "epoch": 0.65,
       "learning_rate": 1.654436768970182e-07,
-      "logits/chosen": -2.0244596004486084,
-      "logits/rejected": -2.021144390106201,
-      "logps/chosen": -32.932373046875,
-      "logps/rejected": -36.2264289855957,
-      "loss": 0.4397,
-      "rewards/accuracies": 0.512499988079071,
-      "rewards/chosen": -0.005845203995704651,
-      "rewards/margins": -0.004640025552362204,
-      "rewards/rejected": -0.001205177279189229,
+      "logits/chosen": -2.024381160736084,
+      "logits/rejected": -2.0210862159729004,
+      "logps/chosen": -32.9254035949707,
+      "logps/rejected": -36.251712799072266,
+      "loss": 0.9788,
+      "rewards/accuracies": 0.5375000238418579,
+      "rewards/chosen": -0.0002662956831045449,
+      "rewards/margins": 0.02116318605840206,
+      "rewards/rejected": -0.021429482847452164,
       "step": 250
     },
     {
       "epoch": 0.68,
       "learning_rate": 1.444597403062196e-07,
-      "logits/chosen": -1.891283392906189,
-      "logits/rejected": -1.8888483047485352,
-      "logps/chosen": -34.20048522949219,
-      "logps/rejected": -35.507022857666016,
-      "loss": 0.4441,
-      "rewards/accuracies": 0.5625,
-      "rewards/chosen": -0.009669994935393333,
-      "rewards/margins": -0.0077504729852080345,
-      "rewards/rejected": -0.0019195213681086898,
+      "logits/chosen": -1.8911311626434326,
+      "logits/rejected": -1.8886839151382446,
+      "logps/chosen": -34.194557189941406,
+      "logps/rejected": -35.51445770263672,
+      "loss": 0.9971,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": -0.004926486872136593,
+      "rewards/margins": 0.002939151134341955,
+      "rewards/rejected": -0.007865638472139835,
       "step": 260
     },
     {
       "epoch": 0.7,
       "learning_rate": 1.2434529917578887e-07,
-      "logits/chosen": -1.8760111331939697,
-      "logits/rejected": -1.8734772205352783,
-      "logps/chosen": -34.39020919799805,
-      "logps/rejected": -31.75579261779785,
-      "loss": 0.4295,
-      "rewards/accuracies": 0.42500001192092896,
-      "rewards/chosen": 0.0059113698080182076,
-      "rewards/margins": -0.006519269198179245,
-      "rewards/rejected": 0.012430639937520027,
+      "logits/chosen": -1.8759450912475586,
+      "logits/rejected": -1.8734045028686523,
+      "logps/chosen": -34.40558624267578,
+      "logps/rejected": -31.752349853515625,
+      "loss": 1.0216,
+      "rewards/accuracies": 0.4625000059604645,
+      "rewards/chosen": -0.0063878437504172325,
+      "rewards/margins": -0.02157408930361271,
+      "rewards/rejected": 0.015186244621872902,
       "step": 270
     },
     {
       "epoch": 0.73,
       "learning_rate": 1.0526606671603521e-07,
-      "logits/chosen": -1.9797089099884033,
-      "logits/rejected": -1.9690707921981812,
-      "logps/chosen": -35.31258773803711,
-      "logps/rejected": -31.837697982788086,
-      "loss": 0.3868,
-      "rewards/accuracies": 0.5249999761581421,
-      "rewards/chosen": 0.03524213656783104,
-      "rewards/margins": 0.0326063297688961,
-      "rewards/rejected": 0.0026358035393059254,
+      "logits/chosen": -1.980015754699707,
+      "logits/rejected": -1.9693737030029297,
+      "logps/chosen": -35.33230209350586,
+      "logps/rejected": -31.845691680908203,
+      "loss": 0.9768,
+      "rewards/accuracies": 0.512499988079071,
+      "rewards/chosen": 0.019471382722258568,
+      "rewards/margins": 0.023230455815792084,
+      "rewards/rejected": -0.0037590693682432175,
       "step": 280
     },
     {
       "epoch": 0.75,
       "learning_rate": 8.737922755071453e-08,
-      "logits/chosen": -2.0755209922790527,
-      "logits/rejected": -2.060497760772705,
-      "logps/chosen": -30.90865135192871,
-      "logps/rejected": -32.64521026611328,
-      "loss": 0.41,
-      "rewards/accuracies": 0.44999998807907104,
-      "rewards/chosen": 0.012379061430692673,
-      "rewards/margins": 0.007080032490193844,
-      "rewards/rejected": 0.005299028940498829,
+      "logits/chosen": -2.0756278038024902,
+      "logits/rejected": -2.060606002807617,
+      "logps/chosen": -30.907390594482422,
+      "logps/rejected": -32.64055252075195,
+      "loss": 0.9956,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": 0.01338677667081356,
+      "rewards/margins": 0.0043600378558039665,
+      "rewards/rejected": 0.00902673788368702,
       "step": 290
     },
     {
       "epoch": 0.78,
       "learning_rate": 7.08321427484816e-08,
-      "logits/chosen": -1.9463651180267334,
-      "logits/rejected": -1.9438308477401733,
-      "logps/chosen": -32.89823532104492,
-      "logps/rejected": -30.81850814819336,
-      "loss": 0.3839,
-      "rewards/accuracies": 0.5874999761581421,
-      "rewards/chosen": 0.02309919334948063,
-      "rewards/margins": 0.024458223953843117,
-      "rewards/rejected": -0.0013590289745479822,
+      "logits/chosen": -1.946616768836975,
+      "logits/rejected": -1.9440828561782837,
+      "logps/chosen": -32.894561767578125,
+      "logps/rejected": -30.812387466430664,
+      "loss": 0.9775,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.026041675359010696,
+      "rewards/margins": 0.022502180188894272,
+      "rewards/rejected": 0.003539492143318057,
       "step": 300
     },
     {
       "epoch": 0.78,
-      "eval_logits/chosen": -2.2435548305511475,
-      "eval_logits/rejected": -2.2386720180511475,
-      "eval_logps/chosen": -34.023826599121094,
-      "eval_logps/rejected": -37.50039291381836,
-      "eval_loss": 0.4344586431980133,
-      "eval_rewards/accuracies": 0.4808970093727112,
-      "eval_rewards/chosen": 0.00858243927359581,
-      "eval_rewards/margins": -0.004399063065648079,
-      "eval_rewards/rejected": 0.012981505133211613,
-      "eval_runtime": 145.9473,
-      "eval_samples_per_second": 2.35,
+      "eval_logits/chosen": -2.2434821128845215,
+      "eval_logits/rejected": -2.2386035919189453,
+      "eval_logps/chosen": -34.017669677734375,
+      "eval_logps/rejected": -37.50018310546875,
+      "eval_loss": 0.9997690916061401,
+      "eval_rewards/accuracies": 0.490448534488678,
+      "eval_rewards/chosen": 0.013507848605513573,
+      "eval_rewards/margins": 0.000357679498847574,
+      "eval_rewards/rejected": 0.0131501704454422,
+      "eval_runtime": 145.7344,
+      "eval_samples_per_second": 2.354,
       "eval_steps_per_second": 0.295,
       "step": 300
     },
     {
       "epoch": 0.81,
       "learning_rate": 5.576113578589034e-08,
-      "logits/chosen": -1.9286285638809204,
-      "logits/rejected": -1.9253568649291992,
-      "logps/chosen": -31.583232879638672,
-      "logps/rejected": -33.75123596191406,
-      "loss": 0.3881,
-      "rewards/accuracies": 0.5375000238418579,
-      "rewards/chosen": 0.016091840341687202,
-      "rewards/margins": 0.023359118029475212,
-      "rewards/rejected": -0.007267280016094446,
+      "logits/chosen": -1.9287067651748657,
+      "logits/rejected": -1.925451636314392,
+      "logps/chosen": -31.603496551513672,
+      "logps/rejected": -33.734046936035156,
+      "loss": 1.0066,
+      "rewards/accuracies": 0.4749999940395355,
+      "rewards/chosen": -0.00011934786743950099,
+      "rewards/margins": -0.006604696158319712,
+      "rewards/rejected": 0.006485348101705313,
       "step": 310
     },
     {
       "epoch": 0.83,
       "learning_rate": 4.229036944380912e-08,
-      "logits/chosen": -1.9803975820541382,
-      "logits/rejected": -1.9680900573730469,
-      "logps/chosen": -34.58079147338867,
-      "logps/rejected": -33.575767517089844,
-      "loss": 0.3778,
+      "logits/chosen": -1.9802377223968506,
+      "logits/rejected": -1.9679291248321533,
+      "logps/chosen": -34.585323333740234,
+      "logps/rejected": -33.57084274291992,
+      "loss": 0.9658,
       "rewards/accuracies": 0.5375000238418579,
-      "rewards/chosen": 0.012225466780364513,
-      "rewards/margins": 0.041770923882722855,
-      "rewards/rejected": -0.029545456171035767,
+      "rewards/chosen": 0.00860034953802824,
+      "rewards/margins": 0.034203141927719116,
+      "rewards/rejected": -0.02560279332101345,
       "step": 320
     },
     {
       "epoch": 0.86,
       "learning_rate": 3.053082288996112e-08,
-      "logits/chosen": -2.015634298324585,
-      "logits/rejected": -2.0141713619232178,
-      "logps/chosen": -33.45996856689453,
-      "logps/rejected": -32.48029327392578,
-      "loss": 0.4032,
-      "rewards/accuracies": 0.5625,
-      "rewards/chosen": 0.020660031586885452,
-      "rewards/margins": 0.021225089207291603,
-      "rewards/rejected": -0.0005650619277730584,
+      "logits/chosen": -2.015906810760498,
+      "logits/rejected": -2.014427661895752,
+      "logps/chosen": -33.49116516113281,
+      "logps/rejected": -32.47978973388672,
+      "loss": 1.0041,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0042965407483279705,
+      "rewards/margins": -0.004129170440137386,
+      "rewards/rejected": -0.00016737030819058418,
       "step": 330
     },
     {
       "epoch": 0.88,
       "learning_rate": 2.05793773749158e-08,
-      "logits/chosen": -2.1035873889923096,
-      "logits/rejected": -2.0877768993377686,
-      "logps/chosen": -34.162208557128906,
-      "logps/rejected": -33.095733642578125,
-      "loss": 0.439,
-      "rewards/accuracies": 0.4625000059604645,
-      "rewards/chosen": 0.013746557757258415,
-      "rewards/margins": -0.004091509617865086,
-      "rewards/rejected": 0.017838066443800926,
+      "logits/chosen": -2.1030631065368652,
+      "logits/rejected": -2.0872654914855957,
+      "logps/chosen": -34.18492889404297,
+      "logps/rejected": -33.08319854736328,
+      "loss": 1.0323,
+      "rewards/accuracies": 0.42500001192092896,
+      "rewards/chosen": -0.00442854966968298,
+      "rewards/margins": -0.03229869529604912,
+      "rewards/rejected": 0.027870142832398415,
       "step": 340
     },
     {
       "epoch": 0.91,
       "learning_rate": 1.251801807404168e-08,
-      "logits/chosen": -1.9746344089508057,
-      "logits/rejected": -1.9736888408660889,
-      "logps/chosen": -33.253448486328125,
-      "logps/rejected": -32.458499908447266,
-      "loss": 0.3954,
-      "rewards/accuracies": 0.5249999761581421,
-      "rewards/chosen": 0.02291743829846382,
-      "rewards/margins": 0.017600122839212418,
-      "rewards/rejected": 0.005317316390573978,
+      "logits/chosen": -1.9745471477508545,
+      "logits/rejected": -1.9736032485961914,
+      "logps/chosen": -33.23271942138672,
+      "logps/rejected": -32.4765510559082,
+      "loss": 0.9514,
+      "rewards/accuracies": 0.612500011920929,
+      "rewards/chosen": 0.0395023450255394,
+      "rewards/margins": 0.04862459748983383,
+      "rewards/rejected": -0.009122253395617008,
       "step": 350
     },
     {
       "epoch": 0.94,
       "learning_rate": 6.41315865106129e-09,
-      "logits/chosen": -1.9309253692626953,
-      "logits/rejected": -1.941303014755249,
-      "logps/chosen": -32.21305465698242,
-      "logps/rejected": -35.309688568115234,
-      "loss": 0.4095,
-      "rewards/accuracies": 0.42500001192092896,
-      "rewards/chosen": -0.006690672133117914,
-      "rewards/margins": 0.010053041391074657,
-      "rewards/rejected": -0.01674371212720871,
+      "logits/chosen": -1.9305438995361328,
+      "logits/rejected": -1.940913438796997,
+      "logps/chosen": -32.22040939331055,
+      "logps/rejected": -35.28728103637695,
+      "loss": 1.0138,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": -0.012576865963637829,
+      "rewards/margins": -0.013758530840277672,
+      "rewards/rejected": 0.0011816651094704866,
       "step": 360
     },
     {
       "epoch": 0.96,
       "learning_rate": 2.3150941078050324e-09,
-      "logits/chosen": -2.069108009338379,
-      "logits/rejected": -2.0625574588775635,
-      "logps/chosen": -33.637176513671875,
-      "logps/rejected": -29.221187591552734,
-      "loss": 0.4059,
-      "rewards/accuracies": 0.48750001192092896,
-      "rewards/chosen": 0.011833530850708485,
-      "rewards/margins": 0.00764369685202837,
-      "rewards/rejected": 0.004189834464341402,
+      "logits/chosen": -2.069648265838623,
+      "logits/rejected": -2.0630898475646973,
+      "logps/chosen": -33.63695526123047,
+      "logps/rejected": -29.226470947265625,
+      "loss": 0.988,
+      "rewards/accuracies": 0.5625,
+      "rewards/chosen": 0.012009462341666222,
+      "rewards/margins": 0.012046033516526222,
+      "rewards/rejected": -3.6573597753886133e-05,
       "step": 370
     },
     {
       "epoch": 0.99,
       "learning_rate": 2.575864278703266e-10,
-      "logits/chosen": -1.9292182922363281,
-      "logits/rejected": -1.9313886165618896,
-      "logps/chosen": -34.24732971191406,
-      "logps/rejected": -30.893402099609375,
-      "loss": 0.4086,
-      "rewards/accuracies": 0.4749999940395355,
-      "rewards/chosen": -0.006622877903282642,
-      "rewards/margins": 0.009996414184570312,
-      "rewards/rejected": -0.01661929115653038,
+      "logits/chosen": -1.928865671157837,
+      "logits/rejected": -1.9310123920440674,
+      "logps/chosen": -34.243560791015625,
+      "logps/rejected": -30.892742156982422,
+      "loss": 0.9875,
+      "rewards/accuracies": 0.5249999761581421,
+      "rewards/chosen": -0.003606534795835614,
+      "rewards/margins": 0.012486499734222889,
+      "rewards/rejected": -0.01609303429722786,
       "step": 380
     },
     {
       "epoch": 1.0,
       "step": 385,
       "total_flos": 0.0,
-      "train_loss": 0.08930327675559303,
-      "train_runtime": 627.8357,
+      "train_loss": 0.21957750196580764,
+      "train_runtime": 627.8531,
       "train_samples_per_second": 4.904,
       "train_steps_per_second": 0.613
     }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5dcdc999fe444f3765c1759b69b7b761c87318fff78403404a07fca4e2fd2e2f
+oid sha256:4cff3f16728b8d9571d3e16284fa5d4a5a3f3e9139edda68dfc9a3846a2ca29a
 size 4984