Image-Text-to-Text
MLX
Safetensors
Chinese
ocr
rare-characters
ids
ideographic-description-sequence
glyphwiki
palaeography
Instructions to use OpenGlyph/OpenGlyph-OCR-IDS with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use OpenGlyph/OpenGlyph-OCR-IDS with MLX:
# Make sure mlx-vlm is installed # pip install --upgrade mlx-vlm from mlx_vlm import load, generate from mlx_vlm.prompt_utils import apply_chat_template from mlx_vlm.utils import load_config # Load the model model, processor = load("OpenGlyph/OpenGlyph-OCR-IDS") config = load_config("OpenGlyph/OpenGlyph-OCR-IDS") # Prepare input image = ["http://images.cocodataset.org/val2017/000000039769.jpg"] prompt = "Describe this image." # Apply chat template formatted_prompt = apply_chat_template( processor, config, prompt, num_images=1 ) # Generate output output = generate(model, processor, formatted_prompt, image) print(output) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
Download reading.json from OpenGlyph/OpenGlyph-OCR-IDS: direct link, hf CLI and curl.
- Browser
- Download file 1.39 kB
-
https://huggingface.co/OpenGlyph/OpenGlyph-OCR-IDS/resolve/main/reading.json
- Command line
-
hf download hf://OpenGlyph/OpenGlyph-OCR-IDS/reading.json
-
curl -L -o reading.json https://huggingface.co/OpenGlyph/OpenGlyph-OCR-IDS/resolve/main/reading.json
1.39 kB
| { | |
| "top": 3, | |
| "rule": { | |
| "a": 1, | |
| "b": 1, | |
| "how": "mean", | |
| "t": -1, | |
| "m": 0.5, | |
| "flat": true | |
| }, | |
| "offset": 8, | |
| "gate": null, | |
| "gate_b": -0.5, | |
| "gates_by": { | |
| "a coded character | a glyph token": -0.25, | |
| "a coded character | a rare character": -2.0, | |
| "a glyph with a token | a glyph token": null, | |
| "a glyph with a token | a rare character": -2.0, | |
| "a new glyph | a glyph token": -0.5, | |
| "a new glyph | a rare character": -2.0, | |
| "forced: a coded character | a glyph token": -0.25, | |
| "forced: a coded character | a rare character": -0.25, | |
| "forced: a glyph with a token | a glyph token": null, | |
| "forced: a new glyph | a glyph token": null, | |
| "forced: a new glyph | a rare character": -0.1 | |
| }, | |
| "open_hard": "speller", | |
| "note": "the settings OpenGlyph-OCR-IDS was measured with, chosen on tuning lines (never on a test): offset = what is added to the opener's evidence that an IDS begins; gate = how surely the unfolding must read for its IDS to stand alone (null: only where the speller writes the same glyph); gate_b and gates_by = how surely the speller must spell for its IDS to stand, by the kind of place (\"forced: ...\" = a place the opener did not ask for, read because the reader writes a rare character or a glyph token there; null = never); open_hard = such places are read by the speller; top = the structures tried; rule = how the unfolding chooses among its cuts" | |
| } | |