---
license: apache-2.0
base_model: AliesTaha/fable-traces
library_name: mlx
tags:
- mlx
- qwen3
- text-generation
- 4-bit
- mxfp4
---

# fable-traces-mlx-mxfp4

MLX MXFP4 conversion of `AliesTaha/fable-traces`.

Base: `Qwen/Qwen3-4B-Instruct-2507`  
Context: inherits base model context  
License: Apache-2.0

```python
from mlx_lm import load, generate

model, tokenizer = load("gtoxlili/fable-traces-mlx-mxfp4")
print(generate(model, tokenizer, prompt="Hello", max_tokens=64))
```