Model save

README.md

@@ -1,11 +1,9 @@
 ---
 base_model: Qwen/Qwen2.5-0.5B
-datasets: open-r1/OpenR1-Math-220k
 library_name: transformers
 model_name: Qwen2.5-0.5B-Open-R1-Code-GRPO
 tags:
 - generated_from_trainer
-- open-r1
 - trl
 - grpo
 licence: license
@@ -13,7 +11,7 @@ licence: license
 
 # Model Card for Qwen2.5-0.5B-Open-R1-Code-GRPO
 
-This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) on the [open-r1/OpenR1-Math-220k](https://huggingface.co/datasets/open-r1/OpenR1-Math-220k) dataset.
+This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -29,7 +27,7 @@ print(output["generated_text"])
 
 ## Training procedure
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/js2024-teleai/huggingface/runs/v4b6t9k2) 
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/js2024-teleai/huggingface/runs/nas9us5b) 
 
 
 This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0,
-    "train_runtime": 1.4257,
+    "train_loss": 0.00930938980280871,
+    "train_runtime": 450.9542,
     "train_samples": 16,
-    "train_samples_per_second": 11.223,
-    "train_steps_per_second": 2.806
+    "train_samples_per_second": 0.035,
+    "train_steps_per_second": 0.009
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0,
-    "train_runtime": 1.4257,
+    "train_loss": 0.00930938980280871,
+    "train_runtime": 450.9542,
     "train_samples": 16,
-    "train_samples_per_second": 11.223,
-    "train_steps_per_second": 2.806
+    "train_samples_per_second": 0.035,
+    "train_steps_per_second": 0.009
 }

trainer_state.json

@@ -12,7 +12,7 @@
       "clip_ratio": 0.0,
       "completion_length": 912.75,
       "epoch": 0.25,
-      "grad_norm": 1.1149588823318481,
+      "grad_norm": 1.1153451204299927,
       "kl": 0.0,
       "learning_rate": 1e-06,
       "loss": 0.0475,
@@ -27,7 +27,7 @@
       "clip_ratio": 0.0,
       "completion_length": 1019.5,
       "epoch": 0.5,
-      "grad_norm": 0.8432132005691528,
+      "grad_norm": 0.8429189324378967,
       "kl": 0.0,
       "learning_rate": 7.75e-07,
       "loss": -0.0102,
@@ -40,10 +40,10 @@
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 578.25,
+      "completion_length": 1197.5,
       "epoch": 0.75,
-      "grad_norm": 0.027600621804594994,
-      "kl": 0.0020586252212524414,
+      "grad_norm": 0.01639331877231598,
+      "kl": 0.0010622739791870117,
       "learning_rate": 3.250000000000001e-07,
       "loss": 0.0,
       "reward": 0.0,
@@ -55,10 +55,10 @@
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 1318.5,
+      "completion_length": 1408.5,
       "epoch": 1.0,
-      "grad_norm": 0.005318487994372845,
-      "kl": 8.308887481689453e-05,
+      "grad_norm": 0.005612128414213657,
+      "kl": 5.0902366638183594e-05,
       "learning_rate": 1e-07,
       "loss": 0.0,
       "reward": 0.0,
@@ -72,10 +72,10 @@
       "epoch": 1.0,
       "step": 4,
       "total_flos": 0.0,
-      "train_loss": 0.0,
-      "train_runtime": 1.4257,
-      "train_samples_per_second": 11.223,
-      "train_steps_per_second": 2.806
+      "train_loss": 0.00930938980280871,
+      "train_runtime": 450.9542,
+      "train_samples_per_second": 0.035,
+      "train_steps_per_second": 0.009
     }
   ],
   "logging_steps": 1,