Model save

all_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 0.992,
     "total_flos": 0.0,
-    "train_loss": 0.8402291834354401,
-    "train_runtime": 793.2299,
+    "train_loss": 0.8468392750909252,
+    "train_runtime": 825.6155,
     "train_samples": 1992,
-    "train_samples_per_second": 2.511,
-    "train_steps_per_second": 0.078
+    "train_samples_per_second": 2.413,
+    "train_steps_per_second": 0.075
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d13c8c8dfa84c4ddf8a2033ea7399c58dcbf2fc3dce12d14477378aba235903b
+oid sha256:10c280e00bda02b932bc6be5bb3c427dab166a4643b7a1bde8b638145a9721fd
 size 4877660776

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8aaa1cad6d1561059558324827b51800c231bc8c39574177c402478fbcb6bc02
+oid sha256:f27b0f1ebe33ef5845955dd6e349804a2dd7d925d1e4710e0d9f001cbcdae83c
 size 4932751008

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7862e63b4356ca02482b85242d7dd6e801fc5ba829ebfade13af57f8d35b1494
+oid sha256:41905b0d955f33ca2a2eb39e4eb699030e81cae98f616470a5a94e02194e691d
 size 4330865200

model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a386a7f8e8217c746f02b845c32cd0ebebf8a64efe47d2a24d2053a2bc892b2e
+oid sha256:e8926b37bd395e1785e34cecc3da5e8af4ee70aef48a98c983cb36388eab2636
 size 1089994880

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 0.992,
     "total_flos": 0.0,
-    "train_loss": 0.8402291834354401,
-    "train_runtime": 793.2299,
+    "train_loss": 0.8468392750909252,
+    "train_runtime": 825.6155,
     "train_samples": 1992,
-    "train_samples_per_second": 2.511,
-    "train_steps_per_second": 0.078
+    "train_samples_per_second": 2.413,
+    "train_steps_per_second": 0.075
 }

trainer_state.json

@@ -10,7 +10,7 @@
   "log_history": [
     {
       "epoch": 0.016,
-      "grad_norm": 54.68133828489169,
+      "grad_norm": 54.80237307470059,
       "learning_rate": 7.142857142857142e-08,
       "logits/generated": -1.112015962600708,
       "logits/real": -0.9725905656814575,
@@ -25,102 +25,102 @@
     },
     {
       "epoch": 0.16,
-      "grad_norm": 78.08601698235012,
+      "grad_norm": 627.6270610217415,
       "learning_rate": 4.727272727272727e-07,
-      "logits/generated": -0.9562382102012634,
-      "logits/real": -0.9224010109901428,
-      "logps/generated": -202.88710021972656,
-      "logps/real": -216.11236572265625,
-      "loss": 0.8993,
-      "rewards/accuracies": 0.4722222089767456,
-      "rewards/generated": 0.16792699694633484,
-      "rewards/margins": 0.015514086931943893,
-      "rewards/real": 0.18344110250473022,
+      "logits/generated": -0.9520471096038818,
+      "logits/real": -0.920363187789917,
+      "logps/generated": -202.67523193359375,
+      "logps/real": -215.9678497314453,
+      "loss": 0.8982,
+      "rewards/accuracies": 0.4861111044883728,
+      "rewards/generated": 0.18911296129226685,
+      "rewards/margins": 0.008780588395893574,
+      "rewards/real": 0.19789355993270874,
       "step": 10
     },
     {
       "epoch": 0.32,
-      "grad_norm": 52.37055337168455,
+      "grad_norm": 50.96666235352248,
       "learning_rate": 3.818181818181818e-07,
-      "logits/generated": -0.9773967862129211,
-      "logits/real": -0.9421946406364441,
-      "logps/generated": -205.1346435546875,
-      "logps/real": -216.21044921875,
-      "loss": 0.8759,
-      "rewards/accuracies": 0.612500011920929,
-      "rewards/generated": 1.164185881614685,
-      "rewards/margins": 0.03111693263053894,
-      "rewards/real": 1.1953028440475464,
+      "logits/generated": -1.0125733613967896,
+      "logits/real": -0.9775249361991882,
+      "logps/generated": -208.7979278564453,
+      "logps/real": -219.83773803710938,
+      "loss": 0.8919,
+      "rewards/accuracies": 0.625,
+      "rewards/generated": 0.7978582382202148,
+      "rewards/margins": 0.03471279889345169,
+      "rewards/real": 0.8325710296630859,
       "step": 20
     },
     {
       "epoch": 0.48,
-      "grad_norm": 45.56925653007757,
+      "grad_norm": 46.130447914310665,
       "learning_rate": 2.909090909090909e-07,
-      "logits/generated": -1.0400158166885376,
-      "logits/real": -0.8696712255477905,
-      "logps/generated": -174.08009338378906,
-      "logps/real": -180.80616760253906,
-      "loss": 0.8296,
-      "rewards/accuracies": 0.6000000238418579,
-      "rewards/generated": 1.7348871231079102,
-      "rewards/margins": 0.13560794293880463,
-      "rewards/real": 1.870495080947876,
+      "logits/generated": -1.005338191986084,
+      "logits/real": -0.8279012441635132,
+      "logps/generated": -176.2639617919922,
+      "logps/real": -183.01602172851562,
+      "loss": 0.8324,
+      "rewards/accuracies": 0.675000011920929,
+      "rewards/generated": 1.5164999961853027,
+      "rewards/margins": 0.1330079436302185,
+      "rewards/real": 1.6495081186294556,
       "step": 30
     },
     {
       "epoch": 0.64,
-      "grad_norm": 47.856709569705075,
+      "grad_norm": 51.04951332078808,
       "learning_rate": 2e-07,
-      "logits/generated": -1.0196596384048462,
-      "logits/real": -0.8638933897018433,
-      "logps/generated": -176.03512573242188,
-      "logps/real": -207.9772491455078,
-      "loss": 0.824,
+      "logits/generated": -1.0122597217559814,
+      "logits/real": -0.854594349861145,
+      "logps/generated": -179.2203826904297,
+      "logps/real": -211.2075958251953,
+      "loss": 0.8282,
       "rewards/accuracies": 0.675000011920929,
-      "rewards/generated": 2.319814920425415,
-      "rewards/margins": 0.2503294050693512,
-      "rewards/real": 2.5701441764831543,
+      "rewards/generated": 2.0012855529785156,
+      "rewards/margins": 0.24582402408123016,
+      "rewards/real": 2.2471091747283936,
       "step": 40
     },
     {
       "epoch": 0.8,
-      "grad_norm": 45.32128168371637,
+      "grad_norm": 46.26306651530627,
       "learning_rate": 1.0909090909090908e-07,
-      "logits/generated": -1.049173355102539,
-      "logits/real": -0.8328621983528137,
-      "logps/generated": -177.70408630371094,
-      "logps/real": -205.44827270507812,
-      "loss": 0.808,
-      "rewards/accuracies": 0.6499999761581421,
-      "rewards/generated": 2.6795811653137207,
-      "rewards/margins": 0.25618499517440796,
-      "rewards/real": 2.9357664585113525,
+      "logits/generated": -1.0062984228134155,
+      "logits/real": -0.7997950315475464,
+      "logps/generated": -180.25086975097656,
+      "logps/real": -208.59616088867188,
+      "loss": 0.823,
+      "rewards/accuracies": 0.625,
+      "rewards/generated": 2.424903154373169,
+      "rewards/margins": 0.1960751712322235,
+      "rewards/real": 2.620978355407715,
       "step": 50
     },
     {
       "epoch": 0.96,
-      "grad_norm": 55.117145590325535,
+      "grad_norm": 44.84748297178086,
       "learning_rate": 1.818181818181818e-08,
-      "logits/generated": -1.1112569570541382,
-      "logits/real": -1.001379132270813,
-      "logps/generated": -170.5053253173828,
-      "logps/real": -174.80856323242188,
-      "loss": 0.8158,
-      "rewards/accuracies": 0.6625000238418579,
-      "rewards/generated": 2.7520089149475098,
-      "rewards/margins": 0.20990166068077087,
-      "rewards/real": 2.9619107246398926,
+      "logits/generated": -1.0795314311981201,
+      "logits/real": -0.9678686857223511,
+      "logps/generated": -172.9720001220703,
+      "logps/real": -177.3105926513672,
+      "loss": 0.8174,
+      "rewards/accuracies": 0.637499988079071,
+      "rewards/generated": 2.505340099334717,
+      "rewards/margins": 0.2063644826412201,
+      "rewards/real": 2.711704730987549,
       "step": 60
     },
     {
       "epoch": 0.992,
       "step": 62,
       "total_flos": 0.0,
-      "train_loss": 0.8402291834354401,
-      "train_runtime": 793.2299,
-      "train_samples_per_second": 2.511,
-      "train_steps_per_second": 0.078
+      "train_loss": 0.8468392750909252,
+      "train_runtime": 825.6155,
+      "train_samples_per_second": 2.413,
+      "train_steps_per_second": 0.075
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a71e8e8b17237369da6edc61c634343a2a0512db288396a56fbf85dbe4242447
+oid sha256:b5547eb26040d6f01b03222eaf4deebffef336be6b12b16cce383bf86ef90ad1
 size 6392