Closed leobxpan closed 5 years ago
@flauted is this something that should have been added in #1409 ?
I actually think the import might be unused but: https://github.com/Cadene/pretrained-models.pytorch You can pip install pretrainedmodels
thanks. it's actually used as pretrainedmodels.resnet152
is used as feature extractor in your script.
Hi @flauted, thanks for contributing to the captioning! I also raised a new issue regarding this, but I'm wondering what kind of input file (also file format) should I provide to translate.py
if I wanna use it for beam search decoding for my customized video captioning model (the captioning module of which is a transformer)? Thanks!
I believe the full question is on #1565.
You can leave this open because pretrainedmodels should be listed in the requirements opt file. I'm short on time right now, so if you have free time it would be awesome if you could contribute a PR adding that line to the requirements file.
Sure. added and PR submitted. I'm closing this issue. Thanks.
Hi, for the
tools/vid_feature_extractor.py
for video feature extraction a modulepretrainedmodels
it's importing cannot be found. Thanks.