Support standard mlx-vlm model loading

Embed the layout and OCR configurations and add a standard safetensors index for the existing stage files. Map tensor keys to their parser submodules and document direct load_model usage without a preparation step.

Files changed (3) hide show

README.md +7 -3
config.json +0 -0
model.safetensors.index.json +0 -0

README.md CHANGED Viewed

@@ -56,7 +56,7 @@ This repository contains two models:
 ## Usage
 Use an [mlx-vlm](https://github.com/Blaizzy/mlx-vlm) checkout that includes
-`indic_ocr` and `pp_doclayout_v3` support. From that checkout:
 ```sh
 python -m pip install -e .
@@ -65,15 +65,19 @@ python -m pip install -e .
 Replace `page.png` with a document image:
 ```python
-from mlx_vlm.models.indic_ocr import IndicOCRParser
-with IndicOCRParser.from_pretrained("HashNuke/indic-ocr-mlx") as parser:
     page = parser.parse("page.png")
 print(page.markdown)
 page.save("page.json")
 ```
 The pipeline detects and cleans the layout, resolves nested equations, and
 transcribes eligible blocks using greedy decoding. Skipped regions remain in
 the JSON with empty text. Closing the parser releases its model references.

 ## Usage
 Use an [mlx-vlm](https://github.com/Blaizzy/mlx-vlm) checkout that includes
+`indic_ocr` support for the standard `load_model` interface. From that checkout:
 ```sh
 python -m pip install -e .
 Replace `page.png` with a document image:
 ```python
+from mlx_vlm.utils import get_model_path, load_model
+with load_model(get_model_path("HashNuke/indic-ocr-mlx")) as parser:
     page = parser.parse("page.png")
 print(page.markdown)
 page.save("page.json")
 ```
+The root config embeds both stage configurations, and a standard safetensors
+index points to their existing weight files. No conversion or preparation step
+is needed. The weight files are unchanged.
 The pipeline detects and cleans the layout, resolves nested equations, and
 transcribes eligible blocks using greedy decoding. Skipped regions remain in
 the JSON with empty text. Closing the parser releases its model references.

config.json CHANGED Viewed

The diff for this file is too large to render. See raw diff

model.safetensors.index.json ADDED Viewed

The diff for this file is too large to render. See raw diff