Qwen3-Med-MLX
This model is a quantized version of Qwen2.5-7B-Medical-Instruct, optimized for Apple Silicon using MLX.
Usage
This model is optimized for MLX (Apple Silicon). When using this model, please set the following parameters to avoid repetition loops:
- Repetition Penalty: 1.2 or higher
- Temperature: 0.7
Quantization details
- Format: 4-bit (MLX)
- Creator: a-m4-user
- Downloads last month
- 13
Model size
8B params
Tensor type
U32
·
F16 ·
Hardware compatibility
Log In to add your hardware
4-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support