How to use RepublicOfKorokke/granite-4.1-3b-oQ4-fp16 with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download RepublicOfKorokke/granite-4.1-3b-oQ4-fp16 --local-dir granite-4.1-3b-oQ4-fp16
https://huggingface.co/RepublicOfKorokke/granite-4.1-3b-oQ4-fp16/resolve/main/README.md
hf download hf://RepublicOfKorokke/granite-4.1-3b-oQ4-fp16/README.md
curl -L -o README.md https://huggingface.co/RepublicOfKorokke/granite-4.1-3b-oQ4-fp16/resolve/main/README.md
library_name: mlx tags: - mlx - oq - quantized base_model: - ibm-granite/granite-4.1-3b
This model was quantized using oQ (oMLX v0.3.8rc1korokke) mixed-precision quantization.
