Zhaokai Wang

> 抱歉...‘captcha文件夹和exceptions替换至zhongxinghong的PKUAutoElective，并对照修改loop.py，这样可以正常运行。’ 没太懂是什么意思 是把这些文件夹和文件全都替换成原来的zhongxinghong的相应同名文件和文件夹吗？ 看楼上这个同学的库就行，实测可以正常使用 https://github.com/Somoku/PKUElective2022Spring

我记得写了一个AI类，不过可能没写完（

Hi, for feature extraction we follow [M4C](https://mmf.sh/docs/projects/m4c/#:~:text=Note%20that%20the%20object%20Faster%20R%2DCNN%20features%20are%20extracted%20with%20extract_features_vmb.py%20and%20the%20OCR%20Faster%20R%2DCNN%20features%20are%20extracted%20with%20extract_ocr_frcn_feature.py.). It is performed on the original dataset, noting that both TextVQA and TextCaps use images from OpenImages, so features extracted on TextVQA can...

For OCR systems you can refer to their official repos ([CRAFT](https://github.com/clovaai/CRAFT-pytorch), [ABCNet](https://github.com/Yuliang-Liu/bezier_curve_text_spotting) and [four-stage STR](https://github.com/Yuliang-Liu/bezier_curve_text_spotting)).

Yes, [extract_features_vmb.py](https://github.com/facebookresearch/mmf/blob/main/tools/scripts/features/extract_features_vmb.py) and [extract_ocr_frcn_feature.py](https://github.com/facebookresearch/mmf/blob/main/projects/m4c/scripts/extract_ocr_frcn_feature.py) to be precise

I think you should probably open issues in MMF repo for this

> > Yes, [extract_features_vmb.py](https://github.com/facebookresearch/mmf/blob/main/tools/scripts/features/extract_features_vmb.py) and [extract_ocr_frcn_feature.py](https://github.com/facebookresearch/mmf/blob/main/projects/m4c/scripts/extract_ocr_frcn_feature.py) to be precise > > Hello, first of all, thank you for the code you provided. I still want to ask you some questions....