Is there a simple and unified way to load all the required models from a custom location?

[huanggefan]

I hope to configure something similar to the following, enabling the loading of all models from the D drive. This way, I can centrally manage various large model files:

analyzer_config = {
    "model_name": "mfd",
    "model_type": "yolov7_tiny",
    "root": AIGC_ENTRIES_BASE_DIR / "huggingface.co" / "breezedeus" / "cnstd-cnocr-models" / "models" / "cnstd",
    "model_backend": "pytorch",
}

clf_config = {
    "base_model_name": "mobilenet_v2",
    "categories": IMAGE_TYPES,
    "transform_configs": {
        "crop_size": [150, 450],
        "resize_size": 160,
        "resize_max_size": 1000,
    },
    "model_dir": AIGC_ENTRIES_BASE_DIR / "huggingface.co" / "breezedeus" / "cnstd-cnocr-models" / "models" / "pix2text" / "0.2" / "clf",
    "model_fp": None
}

general_config = {
    "rec_model_name": "ch_PP-OCRv3",
    "det_model_name": "ch_PP-OCRv3_det",
    "rec_root": AIGC_ENTRIES_BASE_DIR / "huggingface.co" / "breezedeus" / "cnstd-cnocr-models" / "models" / "cnocr",
    "det_root": AIGC_ENTRIES_BASE_DIR / "huggingface.co" / "breezedeus" / "cnstd-cnocr-models" / "models" / "cnstd",
    "rec_model_fp": None,
    "det_model_fp": None,
    "rec_model_backend": "onnx",
    "det_model_backend": "onnx",
}

english_config = {
    "rec_model_name": "en_PP-OCRv3",
    "det_model_name": "en_PP-OCRv3_det",
    "rec_root": AIGC_ENTRIES_BASE_DIR / "huggingface.co" / "breezedeus" / "cnstd-cnocr-models" / "models" / "cnocr",
    "det_root": AIGC_ENTRIES_BASE_DIR / "huggingface.co" / "breezedeus" / "cnstd-cnocr-models" / "models" / "cnstd",
    "rec_model_fp": None,
    "det_model_fp": None,
    "rec_model_backend": "onnx",
    "det_model_backend": "onnx",
}

formula_config = {
    "config": LATEX_CONFIG_FP,
    "mfr_checkpoint": AIGC_ENTRIES_BASE_DIR / "github.com" / "lukas-blecher" / "LaTeX-OCR" / "weights.pth",
    "resizer_checkpoint": AIGC_ENTRIES_BASE_DIR / "github.com" / "lukas-blecher" / "LaTeX-OCR" / "image_resizer.pth",
    "no_resize": False,
}

Now, after executing git clone https://huggingface.co/breezedeus/cnstd-cnocr-models, it is necessary to unzip the contents of models\cnocr\2.2, models\cnstd\1.2, and models\pix2text\0.2 into the current folder.

[FHYQ-Dong]

import os
from pix2text import Pix2Text

text_formula_config = dict(
    languages=('en', 'ch_sim'),  # 设置识别的语言
    mfd=dict(  # 声明 LayoutAnalyzer 的初始化参数
        model_type='yolov7',  # 表示使用的是 YoloV7 模型，而不是 YoloV7_Tiny 模型
        model_fp=os.path.expanduser(
            '~/.cnstd/1.2/analysis/mfd-yolov7-epoch224-20230613.pt'
        ),  # 注：修改成你的模型文件所存储的路径
    ),
    formula=dict(
        model_name='mfr-pro',
        model_backend='onnx',
        model_dir=os.path.expanduser(
            '~/.pix2text/1.0/mfr-pro-onnx'
        ),  # 注：修改成你的模型文件所存储的路径
    ),
    text=dict(
        rec_model_name='doc-densenet_lite_666-gru_large',
        rec_model_backend='onnx',
        rec_model_fp=os.path.expanduser(
            '~/.cnocr/2.3/doc-densenet_lite_666-gru_large/cnocr-v2.3-doc-densenet_lite_666-gru_large-epoch=005-ft-model.onnx'
            # noqa
        ),  # 注：修改成你的模型文件所存储的路径
    ),
)
total_config = {
    'layout': {'scores_thresh': 0.45},
    'text_formula': text_formula_config,
}
p2t = Pix2Text.from_config(total_configs=total_config)

This may be what you're looking for. You can change the value of model_fp or model_dir to custom your model location. See more @ Usage