The prediction result of OCR model is wrong by using mkldnn.

[baoachun]

Download ch_PP-OCRv2_rec, ch_ppocr_mobile_v2.0_rec and ch_ppocr_server_v2.0_rec from https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.3/doc/doc_en/models_list_en.md.

Then
git clone https://github.com/PaddlePaddle/PaddleOCR.git
cd PaddleOCR
git checkout develop
python3 tools/infer/predict_rec.py --image_dir=doc/imgs_words/ch/word_1.jpg --rec_model_dir=./rec_models_ch/ch_PP-OCRv2_rec_infer --rec_image_shape="3, 32, 100" --rec_char_type="ch" --rec_char_dict_path="./ppocr/utils/ppocr_keys_v1.txt" --rec_algorithm=CRNN --use_gpu=False --use_tensorrt=False --enable_mkldnn=True --rec_batch_num=1

[paddle-bot-old]

您好，我们已经收到了您的问题，会安排技术人员尽快解答您的问题，请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时，您也可以通过查看官网API文档、常见问题、历史Issue、AI社区来寻求解答。祝您生活愉快～

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the API，FAQ，Github Issue and AI community to get the answer.Have a nice day!

[jakpiase]

Hi @baoachun, I have downloaded ch_PP-OCRv2_rec_infer.tar, I have also changed filenames from "inference..pdiparams" to "params" and from "inference.pdmodel" to "model" and I am getting "list index out of range" error when I run
python tools/infer/predict_rec.py --image_dir=doc/imgs_words/ch/word_1.jpg --rec_model_dir=./rec_models_ch/ch_PP-OCRv2_rec_infer --rec_image_shape="3, 32, 100" --rec_char_type="ch" --rec_char_dict_path="./ppocr/utils/ppocr_keys_v1.txt" --rec_algorithm=CRNN --use_gpu=False --use_tensorrt=False --enable_mkldnn=True --rec_batch_num=1
Could you please give me more detailed steps how to reproduce this issue?

[jakpiase]

Hi @baoachun, @JiaXiao243 sent a screenshot in #35781 that disabling reshape and squeeze ops fixes both ch_PP-OCRv2_rec_infer and shufflenetv2, can I close these two issues?

[baoachun]

@jakpiase
Yep，thank you.