Model save

README.md

@@ -15,15 +15,15 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1366
-- Rewards/real: 7.8246
-- Rewards/generated: -2.3247
+- Loss: 0.1369
+- Rewards/real: 7.7762
+- Rewards/generated: -2.5293
 - Rewards/accuracies: 1.0
-- Rewards/margins: 10.1492
-- Logps/generated: -263.3269
-- Logps/real: -127.7119
-- Logits/generated: -0.9835
-- Logits/real: -0.6921
+- Rewards/margins: 10.3055
+- Logps/generated: -265.3735
+- Logps/real: -128.1957
+- Logits/generated: -0.9822
+- Logits/real: -0.7243
 
 ## Model description
 
@@ -60,7 +60,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss | Rewards/real | Rewards/generated | Rewards/accuracies | Rewards/margins | Logps/generated | Logps/real | Logits/generated | Logits/real |
 |:-------------:|:------:|:----:|:---------------:|:------------:|:-----------------:|:------------------:|:---------------:|:---------------:|:----------:|:----------------:|:-----------:|
-| 0.1463        | 0.6410 | 50   | 0.1366          | 7.8246       | -2.3247           | 1.0                | 10.1492         | -263.3269       | -127.7119  | -0.9835          | -0.6921     |
+| 0.1463        | 0.6410 | 50   | 0.1369          | 7.7762       | -2.5293           | 1.0                | 10.3055         | -265.3735       | -128.1957  | -0.9822          | -0.7243     |
 
 
 ### Framework versions

all_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 0.0,
-    "train_loss": 0.2174399999471811,
-    "train_runtime": 1268.5918,
+    "train_loss": 0.21996817527673182,
+    "train_runtime": 1475.9075,
     "train_samples": 2484,
-    "train_samples_per_second": 1.958,
-    "train_steps_per_second": 0.061
+    "train_samples_per_second": 1.683,
+    "train_steps_per_second": 0.053
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c1f8fe80a90ed968b51cf2b2acdf8e1374d7a5083a3b2d825cbd4b7df31ce9bd
+oid sha256:ba4621422fcaa408b43c7e596d25660518be4fe007afc8a9f25eca1b67638b7a
 size 4877660776

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3f2e1cfcef2525458b91722ea4094b7b9cb732b90fb8f764139d7857db8725ee
+oid sha256:73baa311f2049027e0467a484b08689e8f299c29325caf8b332cff245e9a3795
 size 4932751008

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7c12526a5353c6624fcbc0f507bedafa96f01e46949884d705cf4466838fbf3c
+oid sha256:1c3ef277ef8ac9dfe10c2ed2e7f4f7311f5f0b0d4fef8f8322a168309c89c6c6
 size 4330865200

model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7e54b3e79abf32e06bf2cb5b5a5e8445e07215071711fa58308d71c1c0ccee30
+oid sha256:11bffdf067690c282a41c2df69394f808572d595cca1b5cb5c1015eaec2f9139
 size 1089994880

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 0.0,
-    "train_loss": 0.2174399999471811,
-    "train_runtime": 1268.5918,
+    "train_loss": 0.21996817527673182,
+    "train_runtime": 1475.9075,
     "train_samples": 2484,
-    "train_samples_per_second": 1.958,
-    "train_steps_per_second": 0.061
+    "train_samples_per_second": 1.683,
+    "train_steps_per_second": 0.053
 }

trainer_state.json

@@ -10,7 +10,7 @@
   "log_history": [
     {
       "epoch": 0.01282051282051282,
-      "grad_norm": 169.85406133026603,
+      "grad_norm": 169.86816716315738,
       "learning_rate": 6.25e-08,
       "logits/generated": -1.1294811964035034,
       "logits/real": -0.7686214447021484,
@@ -25,133 +25,133 @@
     },
     {
       "epoch": 0.1282051282051282,
-      "grad_norm": 63.23414694225829,
+      "grad_norm": 81.55279578883227,
       "learning_rate": 4.857142857142857e-07,
-      "logits/generated": -0.9032017588615417,
-      "logits/real": -0.9559243321418762,
-      "logps/generated": -279.2696838378906,
-      "logps/real": -224.7206268310547,
-      "loss": 0.6755,
-      "rewards/accuracies": 0.8333333134651184,
-      "rewards/generated": -0.39545938372612,
-      "rewards/margins": 0.8560593724250793,
-      "rewards/real": 0.4606000483036041,
+      "logits/generated": -0.9103206992149353,
+      "logits/real": -0.9607930779457092,
+      "logps/generated": -278.6439208984375,
+      "logps/real": -225.61936950683594,
+      "loss": 0.6932,
+      "rewards/accuracies": 0.8472222089767456,
+      "rewards/generated": -0.3328787088394165,
+      "rewards/margins": 0.7036054730415344,
+      "rewards/real": 0.3707267642021179,
       "step": 10
     },
     {
       "epoch": 0.2564102564102564,
-      "grad_norm": 4.693240657102186,
+      "grad_norm": 7.080598136148484,
       "learning_rate": 4.142857142857143e-07,
-      "logits/generated": -0.951930820941925,
-      "logits/real": -0.8856027722358704,
-      "logps/generated": -284.35430908203125,
-      "logps/real": -156.9384002685547,
-      "loss": 0.1771,
+      "logits/generated": -0.9712547063827515,
+      "logits/real": -0.9375473260879517,
+      "logps/generated": -284.4731140136719,
+      "logps/real": -156.74594116210938,
+      "loss": 0.1789,
       "rewards/accuracies": 1.0,
-      "rewards/generated": -1.183829426765442,
-      "rewards/margins": 5.8197126388549805,
-      "rewards/real": 4.635883808135986,
+      "rewards/generated": -1.1957100629806519,
+      "rewards/margins": 5.850839138031006,
+      "rewards/real": 4.655129432678223,
       "step": 20
     },
     {
       "epoch": 0.38461538461538464,
-      "grad_norm": 1.1912756226499355,
+      "grad_norm": 1.0725410774049144,
       "learning_rate": 3.4285714285714286e-07,
-      "logits/generated": -0.5055528283119202,
-      "logits/real": -0.7188040018081665,
-      "logps/generated": -301.6423645019531,
-      "logps/real": -140.4794921875,
-      "loss": 0.1408,
+      "logits/generated": -0.5535954833030701,
+      "logits/real": -0.7542158365249634,
+      "logps/generated": -300.3318786621094,
+      "logps/real": -141.39231872558594,
+      "loss": 0.1418,
       "rewards/accuracies": 1.0,
-      "rewards/generated": -2.185486316680908,
-      "rewards/margins": 9.369997024536133,
-      "rewards/real": 7.184512138366699,
+      "rewards/generated": -2.0544393062591553,
+      "rewards/margins": 9.147669792175293,
+      "rewards/real": 7.093230247497559,
       "step": 30
     },
     {
       "epoch": 0.5128205128205128,
-      "grad_norm": 1.2464411282623236,
+      "grad_norm": 1.1609685873740068,
       "learning_rate": 2.714285714285714e-07,
-      "logits/generated": -0.5057858228683472,
-      "logits/real": -0.5780868530273438,
-      "logps/generated": -299.8144226074219,
-      "logps/real": -154.14199829101562,
-      "loss": 0.1457,
+      "logits/generated": -0.5164953470230103,
+      "logits/real": -0.5888045430183411,
+      "logps/generated": -299.4565734863281,
+      "logps/real": -154.54141235351562,
+      "loss": 0.1462,
       "rewards/accuracies": 1.0,
-      "rewards/generated": -2.8365063667297363,
-      "rewards/margins": 10.34560489654541,
-      "rewards/real": 7.509098052978516,
+      "rewards/generated": -2.800717353820801,
+      "rewards/margins": 10.269874572753906,
+      "rewards/real": 7.469156742095947,
       "step": 40
     },
     {
       "epoch": 0.6410256410256411,
-      "grad_norm": 1.0548152399288286,
+      "grad_norm": 1.1798404000846567,
       "learning_rate": 2e-07,
-      "logits/generated": -0.38232049345970154,
-      "logits/real": -0.6496118307113647,
-      "logps/generated": -303.3151550292969,
-      "logps/real": -143.59616088867188,
+      "logits/generated": -0.40408602356910706,
+      "logits/real": -0.677122950553894,
+      "logps/generated": -303.20843505859375,
+      "logps/real": -143.86439514160156,
       "loss": 0.1463,
       "rewards/accuracies": 1.0,
-      "rewards/generated": -2.8232903480529785,
-      "rewards/margins": 10.742189407348633,
-      "rewards/real": 7.9188995361328125,
+      "rewards/generated": -2.8126168251037598,
+      "rewards/margins": 10.704693794250488,
+      "rewards/real": 7.8920769691467285,
       "step": 50
     },
     {
       "epoch": 0.6410256410256411,
-      "eval_logits/generated": -0.9835317730903625,
-      "eval_logits/real": -0.6921402812004089,
-      "eval_logps/generated": -263.3268737792969,
-      "eval_logps/real": -127.7119369506836,
-      "eval_loss": 0.13658176362514496,
+      "eval_logits/generated": -0.9821963906288147,
+      "eval_logits/real": -0.7243232131004333,
+      "eval_logps/generated": -265.37347412109375,
+      "eval_logps/real": -128.19570922851562,
+      "eval_loss": 0.13692454993724823,
       "eval_rewards/accuracies": 1.0,
-      "eval_rewards/generated": -2.3246724605560303,
-      "eval_rewards/margins": 10.149224281311035,
-      "eval_rewards/real": 7.824552059173584,
-      "eval_runtime": 10.8908,
-      "eval_samples_per_second": 4.591,
-      "eval_steps_per_second": 0.367,
+      "eval_rewards/generated": -2.5293374061584473,
+      "eval_rewards/margins": 10.305512428283691,
+      "eval_rewards/real": 7.776175022125244,
+      "eval_runtime": 15.6793,
+      "eval_samples_per_second": 3.189,
+      "eval_steps_per_second": 0.255,
       "step": 50
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 1.1141881809347227,
+      "grad_norm": 1.1860788753919929,
       "learning_rate": 1.2857142857142855e-07,
-      "logits/generated": -0.503508448600769,
-      "logits/real": -0.6755964159965515,
-      "logps/generated": -294.5934143066406,
-      "logps/real": -133.47311401367188,
-      "loss": 0.1358,
+      "logits/generated": -0.5171874165534973,
+      "logits/real": -0.6936507225036621,
+      "logps/generated": -294.03045654296875,
+      "logps/real": -133.57363891601562,
+      "loss": 0.136,
       "rewards/accuracies": 1.0,
-      "rewards/generated": -3.0683863162994385,
-      "rewards/margins": 10.883561134338379,
-      "rewards/real": 7.815173149108887,
+      "rewards/generated": -3.012089252471924,
+      "rewards/margins": 10.81721019744873,
+      "rewards/real": 7.805120944976807,
       "step": 60
     },
     {
       "epoch": 0.8974358974358975,
-      "grad_norm": 1.0596728323269835,
+      "grad_norm": 1.1142769804706112,
       "learning_rate": 5.714285714285714e-08,
-      "logits/generated": -0.6190778017044067,
-      "logits/real": -0.6327217221260071,
-      "logps/generated": -289.4383850097656,
-      "logps/real": -148.02481079101562,
-      "loss": 0.145,
+      "logits/generated": -0.6302377581596375,
+      "logits/real": -0.6565964818000793,
+      "logps/generated": -289.1152038574219,
+      "logps/real": -148.23373413085938,
+      "loss": 0.1452,
       "rewards/accuracies": 1.0,
-      "rewards/generated": -2.6634089946746826,
-      "rewards/margins": 10.635811805725098,
-      "rewards/real": 7.972402095794678,
+      "rewards/generated": -2.631093740463257,
+      "rewards/margins": 10.58260440826416,
+      "rewards/real": 7.951510429382324,
       "step": 70
     },
     {
       "epoch": 1.0,
       "step": 78,
       "total_flos": 0.0,
-      "train_loss": 0.2174399999471811,
-      "train_runtime": 1268.5918,
-      "train_samples_per_second": 1.958,
-      "train_steps_per_second": 0.061
+      "train_loss": 0.21996817527673182,
+      "train_runtime": 1475.9075,
+      "train_samples_per_second": 1.683,
+      "train_steps_per_second": 0.053
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c2c5f648b014f6c1acc157d81fa5f5fb4cf76525e7160b32b8fc67294a1a9a31
+oid sha256:d54fba7b1ccc66f5d916428efb0a621e64059933f1084988291ad80aa0201e7d
 size 6392