Image-to-Text
MLX
Safetensors
mlx-weights
paddlepaddle-ocr
ppocrv5
ppocrv6
ppdoclayoutv3
pp-structure
apple-silicon
Instructions to use plaincompute/ppocr-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use plaincompute/ppocr-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir ppocr-mlx plaincompute/ppocr-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
File size: 2,134 Bytes
1e1b9bd | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 | {
"model_type": "pp_ocrv6_medium_det",
"mode": "large",
"upsample_mode": "nearest",
"upsample_align_mode": 1,
"backbone_config": {
"model_type": "pp_lcnet_v4",
"stem_channels": [3, 64, 128],
"stem_type": "large",
"out_features": ["stage1", "stage2", "stage3", "stage4"],
"out_indices": [1, 2, 3, 4],
"block_configs": [
[[3, 128, 128, 1, true], [3, 128, 128, 1, false]],
[
[3, 128, 256, 2, false],
[3, 256, 256, 1, true],
[3, 256, 256, 1, false]
],
[
[3, 256, 512, 2, false],
[3, 512, 512, 1, true],
[3, 512, 512, 1, false],
[3, 512, 512, 1, true],
[3, 512, 512, 1, false]
],
[
[3, 512, 896, 2, false],
[3, 896, 896, 1, true],
[3, 896, 896, 1, false]
]
]
},
"use_lab": false,
"use_last_conv": true,
"class_expand": 2048,
"class_num": 1000,
"out_indices": [0, 1, 2, 3],
"neck_out_channels": 256,
"reduce_factor": 2,
"intraclass_block_number": 4,
"intraclass_block_config": {
"reduce_channel": [1, 1, 0],
"return_channel": [1, 1, 0],
"vertical_long_to_small_conv_longratio": [[7, 1], [1, 1], [3, 0]],
"vertical_long_to_small_conv_midratio": [[5, 1], [1, 1], [2, 0]],
"vertical_long_to_small_conv_shortratio": [[3, 1], [1, 1], [1, 0]],
"horizontal_small_to_long_conv_longratio": [[1, 7], [1, 1], [0, 3]],
"horizontal_small_to_long_conv_midratio": [[1, 5], [1, 1], [0, 2]],
"horizontal_small_to_long_conv_shortratio": [[1, 3], [1, 1], [0, 1]],
"symmetric_conv_long_longratio": [[7, 7], [1, 1], [3, 3]],
"symmetric_conv_long_midratio": [[5, 5], [1, 1], [2, 2]],
"symmetric_conv_long_shortratio": [[3, 3], [1, 1], [1, 1]]
},
"head_in_channels": 1024,
"scale_factor": 2,
"scale_factor_list": [1, 2, 4, 8],
"hidden_act": "relu",
"kernel_list": [3, 2, 2]
} |