Update README.md
Browse files
README.md
CHANGED
|
@@ -39,5 +39,7 @@ We use our [code review dataset](https://huggingface.co/datasets/AGENTDARS/gener
|
|
| 39 |
| Compute Environment | DeepSpeed for memory-efficient distributed training |
|
| 40 |
| Compute Infrastructure | 8x H100 |
|
| 41 |
|
|
|
|
|
|
|
| 42 |
## Results
|
| 43 |
Using this model as a reviewer with DARS trajectories generated using Claude 3.5 Sonnet V2 achieves 38.7% on SWE-Bench Lite.
|
|
|
|
| 39 |
| Compute Environment | DeepSpeed for memory-efficient distributed training |
|
| 40 |
| Compute Infrastructure | 8x H100 |
|
| 41 |
|
| 42 |
+
We use training script provided in [Qwen-2.5 codebase](https://github.com/QwenLM/Qwen2.5-Coder).
|
| 43 |
+
|
| 44 |
## Results
|
| 45 |
Using this model as a reviewer with DARS trajectories generated using Claude 3.5 Sonnet V2 achieves 38.7% on SWE-Bench Lite.
|