English | 中文
MindOCR can support the inference of third-party models, and this document displays a list of adapted models.
After downloading the model file, it needs to be converted to a model file supported by ACL/MindSpore Lite inference (OM
or MindIR). Please refer to the model conversion tutorial.
The original model files involved are as follows:
type | format | description |
---|---|---|
pp-train | .pdparams、.pdopt、.states | PaddlePaddle trained model, which can store information such as model structure, weights, optimizer status, etc |
pp-infer | inference.pdmodel、inference.pdiparams | PaddlePaddle inference model, which can be derived from its trained model, saving the network structure and weights. |
pth | .pth | Pytorch model, which can store information such as model structure, weights, optimizer status, etc |
name | model | backbone | config | download | reference | source |
---|---|---|---|---|---|---|
ch_pp_server_det_v2.0 | DB | ResNet18_vd | yaml | pp-infer | ch_ppocr_server_v2.0_det | PaddleOCR |
ch_pp_det_OCRv3 | DB | MobileNetV3 | yaml | pp-infer | ch_PP-OCRv3_det | PaddleOCR |
ch_pp_det_OCRv2 | DB | MobileNetV3 | yaml | pp-infer | ch_PP-OCRv2_det | PaddleOCR |
ch_pp_mobile_det_v2.0_slim | DB | MobileNetV3 | yaml | pp-infer | ch_ppocr_mobile_slim_v2.0_det | PaddleOCR |
ch_pp_mobile_det_v2.0 | DB | MobileNetV3 | yaml | pp-infer | ch_ppocr_mobile_v2.0_det | PaddleOCR |
en_pp_det_OCRv3 | DB | MobileNetV3 | yaml | pp-infer | en_PP-OCRv3_det | PaddleOCR |
ml_pp_det_OCRv3 | DB | MobileNetV3 | yaml | pp-infer | ml_PP-OCRv3_det | PaddleOCR |
en_pp_det_dbnet_resnet50vd | DB | ResNet50_vd | yaml | pp-train | DBNet | PaddleOCR |
en_pp_det_psenet_resnet50vd | PSE | ResNet50_vd | yaml | pp-train | PSE | PaddleOCR |
en_pp_det_east_resnet50vd | EAST | ResNet50_vd | yaml | pp-train | EAST | PaddleOCR |
en_pp_det_sast_resnet50vd | SAST | ResNet50_vd | yaml | pp-train | SAST | PaddleOCR |
en_mm_det_denetpp_resnet50 | DB++ | ResNet50 | yaml | pth | DBNetpp | MMOCR |
en_mm_det_fcenet_resnet50 | FCENet | ResNet50 | yaml | pth | FCENet | MMOCR |
name | model | backbone | dict file | config | download | reference | source |
---|---|---|---|---|---|---|---|
ch_pp_server_rec_v2.0 | CRNN | ResNet34 | ppocr_keys_v1.txt | yaml | pp-infer | ch_ppocr_server_v2.0_rec | PaddleOCR |
ch_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | ppocr_keys_v1.txt | yaml | pp-infer | ch_PP-OCRv3_rec | PaddleOCR |
ch_pp_rec_OCRv2 | CRNN | MobileNetV1Enhance | ppocr_keys_v1.txt | yaml | pp-infer | ch_PP-OCRv2_rec | PaddleOCR |
ch_pp_mobile_rec_v2.0 | CRNN | MobileNetV3 | ppocr_keys_v1.txt | yaml | pp-infer | ch_ppocr_mobile_v2.0_rec | PaddleOCR |
en_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | en_dict.txt | yaml | pp-infer | en_PP-OCRv3_rec | PaddleOCR |
en_pp_mobile_rec_number_v2.0_slim | CRNN | MobileNetV3 | en_dict.txt | yaml | pp-infer | en_number_mobile_slim_v2.0_rec | PaddleOCR |
en_pp_mobile_rec_number_v2.0 | CRNN | MobileNetV3 | en_dict.txt | yaml | pp-infer | en_number_mobile_v2.0_rec | PaddleOCR |
korean_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | korean_dict.txt | yaml | pp-infer | korean_PP-OCRv3_rec | PaddleOCR |
japan_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | japan_dict.txt | yaml | pp-infer | japan_PP-OCRv3_rec | PaddleOCR |
chinese_cht_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | chinese_cht_dict.txt | yaml | pp-infer | chinese_cht_PP-OCRv3_rec | PaddleOCR |
te_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | te_dict.txt | yaml | pp-infer | te_PP-OCRv3_rec | PaddleOCR |
ka_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | ka_dict.txt | yaml | pp-infer | ka_PP-OCRv3_rec | PaddleOCR |
ta_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | ta_dict.txt | yaml | pp-infer | ta_PP-OCRv3_rec | PaddleOCR |
latin_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | latin_dict.txt | yaml | pp-infer | latin_PP-OCRv3_rec | PaddleOCR |
arabic_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | arabic_dict.txt | yaml | pp-infer | arabic_PP-OCRv3_rec | PaddleOCR |
cyrillic_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | cyrillic_dict.txt | yaml | pp-infer | cyrillic_PP-OCRv3_rec | PaddleOCR |
devanagari_pp_rec_OCRv3 | SVTR | MobileNetV1Enhance | devanagari_dict.txt | yaml | pp-infer | devanagari_PP-OCRv3_rec | PaddleOCR |
en_pp_rec_crnn_resnet34vd | CRNN | ResNet34_vd | ic15_dict.txt | yaml | pp-train | CRNN | PaddleOCR |
en_pp_rec_rosetta_resnet34vd | Rosetta | Resnet34_vd | ic15_dict.txt | yaml | pp-train | Rosetta | PaddleOCR |
en_pp_rec_vitstr_vitstr | ViTSTR | ViTSTR | EN_symbol_dict.txt | yaml | pp-train | ViTSTR | PaddleOCR |
en_mm_rec_nrtr_resnet31 | NRTR | ResNet31 | english_digits_symbols.txt | nrtr_resnet31-1by8-1by4_6e_st_mj.yaml | pth | NRTR | MMOCR |
en_mm_rec_satrn_shallowcnn | SATRN | ShallowCNN | english_digits_symbols.txt | satrn_shallow_5e_st_mj.yaml | pth | SATRN | MMOCR |
name | model | config | download | reference | source |
---|---|---|---|---|---|
ch_pp_mobile_cls_v2.0 | MobileNetV3 | yaml | pp-infer | ch_ppocr_mobile_v2.0_cls | PaddleOCR |
Please refer to the third-party model inference test performance table.
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》