Model save

README.md

@@ -1,12 +1,15 @@
 ---
 base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
 library_name: transformers
+model_name: Qwen2.5-1.5B-Open-R1-GRPO
 tags:
 - generated_from_trainer
+- trl
+- grpo
 licence: license
 ---
 
-# Model Card for None
+# Model Card for Qwen2.5-1.5B-Open-R1-GRPO
 
 This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B).
 It has been trained using [TRL](https://github.com/huggingface/trl).
@@ -24,7 +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/ruikangzhao-danmarks-tekniske-universitet-dtu/huggingface/runs/u8581ung) 
+[<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/ruikangzhao-danmarks-tekniske-universitet-dtu/huggingface/runs/bare070u) 
 
 
 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": 1.862645149230957e-09,
-    "train_runtime": 387.7333,
-    "train_samples": 3,
-    "train_samples_per_second": 0.008,
-    "train_steps_per_second": 0.003
+    "train_loss": 4.6566128730773926e-09,
+    "train_runtime": 1048.9444,
+    "train_samples": 10,
+    "train_samples_per_second": 0.01,
+    "train_steps_per_second": 0.001
 }

config.json

@@ -21,7 +21,7 @@
   "rope_theta": 10000,
   "sliding_window": null,
   "tie_word_embeddings": false,
-  "torch_dtype": "float32",
+  "torch_dtype": "bfloat16",
   "transformers_version": "4.48.2",
   "use_cache": false,
   "use_mrope": false,

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16bea0a35dba47a078ac1750e001512a02837da37c70c3c806244efbe9efefc8
+size 3554214752

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.862645149230957e-09,
-    "train_runtime": 387.7333,
-    "train_samples": 3,
-    "train_samples_per_second": 0.008,
-    "train_steps_per_second": 0.003
+    "train_loss": 4.6566128730773926e-09,
+    "train_runtime": 1048.9444,
+    "train_samples": 10,
+    "train_samples_per_second": 0.01,
+    "train_steps_per_second": 0.001
 }

trainer_state.json

@@ -9,19 +9,19 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 955.5833333333334,
+      "completion_length": 877.275,
       "epoch": 1.0,
       "kl": 0.0,
-      "reward": 0.25,
-      "reward_std": 0.15430335203806558,
-      "rewards/accuracy_reward": 0.25,
+      "reward": 0.15,
+      "reward_std": 0.1498008370399475,
+      "rewards/accuracy_reward": 0.15,
       "rewards/format_reward": 0.0,
       "step": 1,
       "total_flos": 0.0,
-      "train_loss": 1.862645149230957e-09,
-      "train_runtime": 387.7333,
-      "train_samples_per_second": 0.008,
-      "train_steps_per_second": 0.003
+      "train_loss": 4.6566128730773926e-09,
+      "train_runtime": 1048.9444,
+      "train_samples_per_second": 0.01,
+      "train_steps_per_second": 0.001
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1f0ba137453abddbfd819c5f1476a21083d44a541e4ecd3b4e7af905fe6f8160
-size 5752
+oid sha256:1ced2970672832e566c0331145ccd0d2531ff8c5440a13c61f0f1c2a369ca256
+size 5880