Text-to-Image
MLX
Safetensors
English
Chinese
moss_image
apple-silicon
moss-image
text-rendering
layout
quantized
8-bit precision
Instructions to use VisualInference/MOSS-Image-SFT-MLX-compact with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use VisualInference/MOSS-Image-SFT-MLX-compact with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download VisualInference/MOSS-Image-SFT-MLX-compact --local-dir MOSS-Image-SFT-MLX-compact
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
File size: 350 Bytes
a53047a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 | {
"model_type": "moss_image",
"library_name": "moss-image-mlx",
"format": 1,
"tier": "compact",
"source_model": "OpenMOSS-Team/MOSS-Image-SFT",
"transformer_quantization": "affine8g64",
"text_encoder_quantization": "affine6g64",
"components": [
"transformer",
"text_encoder",
"tokenizer",
"vae",
"scheduler"
]
}
|