Instructions to use atkinschang/TableFormerV2-MLX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use atkinschang/TableFormerV2-MLX with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir TableFormerV2-MLX atkinschang/TableFormerV2-MLX
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Publish MLX-converted weights of docling-project/TableFormerV2 at 51559fad3946873e26a6f9b8e912f948e8745bef
79a9ab1 Download preprocessor_config.json from atkinschang/TableFormerV2-MLX: direct link, hf CLI and curl.
- Browser
- Download file 407 Bytes
-
https://huggingface.co/atkinschang/TableFormerV2-MLX/resolve/main/preprocessor_config.json
- Command line
-
hf download hf://atkinschang/TableFormerV2-MLX/preprocessor_config.json
-
curl -L -o preprocessor_config.json https://huggingface.co/atkinschang/TableFormerV2-MLX/resolve/main/preprocessor_config.json
407 Bytes
| { | |
| "do_convert_rgb": true, | |
| "do_resize": true, | |
| "size": { | |
| "height": 448, | |
| "width": 448 | |
| }, | |
| "interpolation": "bilinear", | |
| "antialias": true, | |
| "do_rescale": true, | |
| "rescale_factor": 0.00392156862745098, | |
| "do_normalize": true, | |
| "image_mean": [ | |
| 0.485, | |
| 0.456, | |
| 0.406 | |
| ], | |
| "image_std": [ | |
| 0.229, | |
| 0.224, | |
| 0.225 | |
| ], | |
| "output_dtype": "float32", | |
| "output_layout": "NHWC" | |
| } | |