Model save

all_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 0.992,
     "total_flos": 0.0,
-    "train_loss": 0.7046444348750576,
-    "train_runtime": 776.543,
+    "train_loss": 0.704919635288177,
+    "train_runtime": 933.9268,
     "train_samples": 1995,
-    "train_samples_per_second": 2.569,
-    "train_steps_per_second": 0.08
+    "train_samples_per_second": 2.136,
+    "train_steps_per_second": 0.066
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e6acd4f2edc1f7730dafd2f0b9084c16a3a0adcc2faa46a35335dc6252f085a4
+oid sha256:f49dbd2f2d490ba057516e1f9306a55e7c0637326b6e6be86d5f53b14b3637ef
 size 4877660776

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6ee8e2dbcd414ebb8630e4c1ac5e7209033b2fb3b34df06198dbc64ed9183779
+oid sha256:29c45de9a29dc8ec25c25e96284e78b3e97d2a2f19fe77e883d231a533d5e778
 size 4932751008

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2baca308706598f8809344bb46fd2112c3cb080db9f195f72ff72f75a008a344
+oid sha256:9cb80f49591f876daed95ba28de7d10b23c200dc9f039b6e0b5322191f23bcd8
 size 4330865200

model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee80d907122ea50cc20eeb4dbc840beb40dde7400013b1739ebd60fcb7036307
+oid sha256:74edcb8c5e932ce550f5f8b73c7604a2aff2e5083b38fdfdd724c151c8b6e109
 size 1089994880

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 0.992,
     "total_flos": 0.0,
-    "train_loss": 0.7046444348750576,
-    "train_runtime": 776.543,
+    "train_loss": 0.704919635288177,
+    "train_runtime": 933.9268,
     "train_samples": 1995,
-    "train_samples_per_second": 2.569,
-    "train_steps_per_second": 0.08
+    "train_samples_per_second": 2.136,
+    "train_steps_per_second": 0.066
 }

trainer_state.json

@@ -10,7 +10,7 @@
   "log_history": [
     {
       "epoch": 0.016,
-      "grad_norm": 50.82963308997749,
+      "grad_norm": 50.835563259148245,
       "learning_rate": 7.142857142857142e-08,
       "logits/generated": -1.7165532112121582,
       "logits/real": -1.5090866088867188,
@@ -25,102 +25,102 @@
     },
     {
       "epoch": 0.16,
-      "grad_norm": 41.3948314326989,
+      "grad_norm": 41.66364421784157,
       "learning_rate": 4.727272727272727e-07,
-      "logits/generated": -1.8417009115219116,
-      "logits/real": -1.5857137441635132,
-      "logps/generated": -281.29766845703125,
-      "logps/real": -260.3644104003906,
-      "loss": 0.9043,
-      "rewards/accuracies": 0.5416666865348816,
-      "rewards/generated": 0.23805660009384155,
-      "rewards/margins": 0.10852599143981934,
-      "rewards/real": 0.3465825915336609,
+      "logits/generated": -1.8415862321853638,
+      "logits/real": -1.5861574411392212,
+      "logps/generated": -281.33892822265625,
+      "logps/real": -260.4600830078125,
+      "loss": 0.9056,
+      "rewards/accuracies": 0.5138888955116272,
+      "rewards/generated": 0.23393188416957855,
+      "rewards/margins": 0.10307849943637848,
+      "rewards/real": 0.33701038360595703,
       "step": 10
     },
     {
       "epoch": 0.32,
-      "grad_norm": 35.46442576215001,
+      "grad_norm": 35.24058113049003,
       "learning_rate": 3.818181818181818e-07,
-      "logits/generated": -1.7797447443008423,
-      "logits/real": -1.4822289943695068,
-      "logps/generated": -265.62164306640625,
-      "logps/real": -228.3395233154297,
-      "loss": 0.7337,
-      "rewards/accuracies": 0.824999988079071,
-      "rewards/generated": 1.8866024017333984,
-      "rewards/margins": 0.8881384134292603,
-      "rewards/real": 2.7747409343719482,
+      "logits/generated": -1.7791579961776733,
+      "logits/real": -1.480543613433838,
+      "logps/generated": -265.5545959472656,
+      "logps/real": -228.2611083984375,
+      "loss": 0.7334,
+      "rewards/accuracies": 0.8374999761581421,
+      "rewards/generated": 1.893307089805603,
+      "rewards/margins": 0.8892760276794434,
+      "rewards/real": 2.782582998275757,
       "step": 20
     },
     {
       "epoch": 0.48,
-      "grad_norm": 34.49996220301337,
+      "grad_norm": 34.94460500011804,
       "learning_rate": 2.909090909090909e-07,
-      "logits/generated": -1.678046464920044,
-      "logits/real": -1.3650352954864502,
-      "logps/generated": -255.2372589111328,
-      "logps/real": -221.446044921875,
-      "loss": 0.6873,
+      "logits/generated": -1.688865303993225,
+      "logits/real": -1.3790943622589111,
+      "logps/generated": -254.9552764892578,
+      "logps/real": -221.18008422851562,
+      "loss": 0.6877,
       "rewards/accuracies": 0.7250000238418579,
-      "rewards/generated": 3.282224178314209,
-      "rewards/margins": 0.8962146639823914,
-      "rewards/real": 4.178439140319824,
+      "rewards/generated": 3.3104233741760254,
+      "rewards/margins": 0.8946127891540527,
+      "rewards/real": 4.205036163330078,
       "step": 30
     },
     {
       "epoch": 0.64,
-      "grad_norm": 31.3654281108449,
+      "grad_norm": 31.70597635761484,
       "learning_rate": 2e-07,
-      "logits/generated": -1.5494762659072876,
-      "logits/real": -1.2621484994888306,
-      "logps/generated": -233.6564483642578,
-      "logps/real": -204.22552490234375,
-      "loss": 0.6527,
-      "rewards/accuracies": 0.7875000238418579,
-      "rewards/generated": 4.193800926208496,
-      "rewards/margins": 1.0007737874984741,
-      "rewards/real": 5.19457483291626,
+      "logits/generated": -1.5597318410873413,
+      "logits/real": -1.2736680507659912,
+      "logps/generated": -233.81716918945312,
+      "logps/real": -204.30459594726562,
+      "loss": 0.6532,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/generated": 4.177728176116943,
+      "rewards/margins": 1.0089408159255981,
+      "rewards/real": 5.18666934967041,
       "step": 40
     },
     {
       "epoch": 0.8,
-      "grad_norm": 32.94242474259375,
+      "grad_norm": 33.862934848228804,
       "learning_rate": 1.0909090909090908e-07,
-      "logits/generated": -1.5461094379425049,
-      "logits/real": -1.345146894454956,
-      "logps/generated": -235.3571319580078,
-      "logps/real": -204.53231811523438,
-      "loss": 0.6424,
+      "logits/generated": -1.547353982925415,
+      "logits/real": -1.3472106456756592,
+      "logps/generated": -235.5671844482422,
+      "logps/real": -204.73304748535156,
+      "loss": 0.6417,
       "rewards/accuracies": 0.75,
-      "rewards/generated": 4.165228843688965,
-      "rewards/margins": 1.0194575786590576,
-      "rewards/real": 5.184686183929443,
+      "rewards/generated": 4.144225597381592,
+      "rewards/margins": 1.0203888416290283,
+      "rewards/real": 5.164614677429199,
       "step": 50
     },
     {
       "epoch": 0.96,
-      "grad_norm": 35.41079003724441,
+      "grad_norm": 35.459665070843094,
       "learning_rate": 1.818181818181818e-08,
-      "logits/generated": -1.635496735572815,
-      "logits/real": -1.3326714038848877,
-      "logps/generated": -236.8266143798828,
-      "logps/real": -199.64089965820312,
-      "loss": 0.6352,
+      "logits/generated": -1.6352506875991821,
+      "logits/real": -1.3348934650421143,
+      "logps/generated": -237.15017700195312,
+      "logps/real": -200.07449340820312,
+      "loss": 0.6375,
       "rewards/accuracies": 0.800000011920929,
-      "rewards/generated": 4.119868278503418,
-      "rewards/margins": 1.221605658531189,
-      "rewards/real": 5.341473579406738,
+      "rewards/generated": 4.087512969970703,
+      "rewards/margins": 1.2106025218963623,
+      "rewards/real": 5.2981157302856445,
       "step": 60
     },
     {
       "epoch": 0.992,
       "step": 62,
       "total_flos": 0.0,
-      "train_loss": 0.7046444348750576,
-      "train_runtime": 776.543,
-      "train_samples_per_second": 2.569,
-      "train_steps_per_second": 0.08
+      "train_loss": 0.704919635288177,
+      "train_runtime": 933.9268,
+      "train_samples_per_second": 2.136,
+      "train_steps_per_second": 0.066
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:63a6939c71cf543e1a6e6c96c8332c9bfcb5a06ec0c8ed9721adf0d5a899c491
+oid sha256:381d74d8c9ecba27e99d67d9e76341be694008be74d7d18d61299f7e09f15680
 size 6392