video2numpy
video2numpy copied to clipboard
Optimized library for large-scale extraction of frames and audio from video.
[pyAV](https://github.com/PyAV-Org/PyAV) - untested [VPF](https://github.com/NVIDIA/VideoProcessingFramework) - untested [decord](https://github.com/dmlc/decord): initial testing showed no clear performance advantage over video2numpy
in particular https://github.com/iejMac/clip-video-encode
Hello, is there any easy way to use the frame extraction method within the dataset/dataloader for pytorch? My understanding that it is not trivial given the use of multiprocessing here:...
this is an important feature of this lib. N frames/s on hardware X is more useful than just "optimized"
A place holder pull request
currently we only do intra-video batching but for short videos this leads to 1 padded batch. Should be a way to batch a few videos into one batch and split...
like here - https://github.com/rom1504/img2dataset/blob/main/img2dataset/resizer.py#L81