---
base_model:
- Qwen/Qwen3-4B-Instruct-2507
---

```
int4_block_size=32
int4_accuracy_level=4
```