How can we extract image features instead of downloading large files?

[lhwcv]

Can pythia provide pretrained image feature model thus we extract the features from the original image, no need to download image features' file , that's pretty large. Thank you!

[endernewton]

How long does it take to download the feature files now? If it can solve your problem, we can upload the features to somewhere that are easy for you to download.

[lhwcv]

@endernewton Upload to BaiduYun maybe? Thank you very much!

[lhwcv]

@endernewton For file tools/extract_detectron_feat.py , which config file should we use ? It seems like there is no such config file we can get in the repo. I guess we can extact features from raw image by this file.

[YuJiang01]

Add help information under data_prep/data_preprocess.md,

[lhwcv]

@YuJiang01 Thanks!

[wangpengnorman]

Hi, I want to extract image features, however, "data_prep/data_preprocess.md" is not available any more.

[apsdehal]

@wangpengnorman check https://github.com/facebookresearch/pythia/blob/0.1/data_prep/data_preprocess.md in branch 0.1.