Could you provide a pip requirements

[WujiangXu]

Hello, thank you for your work. I want try to use this code to extract the i3d feature and vggish feature in the BMT branch. But the environment is hard for me to build. Cause some package can not be found in the anaconda environment. So, I want to know whether you can provide a pip requirements list.

[v-iashin]

Hello. Thanks for trying out the project.

You provide too little information to understand your problem.

The reason why your conda cannot find some packages is, most likely, because you are on Windows. You might need to look for correct conda binaries yourself since the project is built with Linux in mind. Another option is to build a Docker. This repo provides all versions along with the package names which will guide you.

[ohjho]

@WujiangXu I used pipreqs to make an requirements.txt for this repo. You might want to add in opencv-python==4.4.0.44 manually.

if you are going to use PWCNet you will need cupy as well. Installing that might be tricky, since I have multiple CUDA_versions and the doc explicitly mentioned that CUDA 11 is not supported. I installed mine like this: CUDA_PATH=/usr/local/cuda-10.2 pip install cupy==5.1.0. Probably best to refer to their docs though for your specific setup. Note that this cupy version is pretty old... latest version as of May 2022 is 10.4.0.

here's the content of my requirements.txt excluding cupy:

av==9.2.0
numpy==1.19.2
omegaconf==2.1.2
Pillow==9.1.0
resampy==0.2.2
scipy==1.6.1
soundfile==0.10.3.post1
torch==1.11.0
torchvision==0.12.0
tqdm==4.46.0
opencv-python==4.4.0.44

[v-iashin]

I think this one should be fixed after we will add a docker supporting feature extraction.

[v-iashin]

I wanted to get back to this. I think it can be related to #25. Thanks for bringing this up.

Also, if you or others have this same issue, it, perhaps, can be solved with a docker environment. Recently I made a docker image that has both environments in #53 (see the documentation as well). Another, perhaps more invasive, solution is to switch to miniconda.

[v-iashin]

I am going to close this issue for now as one could find a way to install these https://v-iashin.github.io/video_features/meta/install_conda/ with pip