Create README.md

README.md

@@ -0,0 +1,17 @@
+---
+base_model:
+- Qwen/Qwen2.5-7B-Instruct
+---
+
+unsloth GRPO LoRA training with ```final_reward_data.json```
+
+Parameters: 
+
+  * ```context```: 4096
+  * ```maxgen```: 512
+  * ```batch```: 4
+  * ```rollout```: 4
+  * ```LoRA 16 bit rank```: 64
+  * ```gradient```: 4
+
+Further details at https://wandb.ai/furkansahinuc-personal/expert-reward-models/runs/m08y4h1i