tvl
tvl copied to clipboard
anibali
Torch video loading library with support for GPU decoding
I did this to change torch tensor storage of decoded frames to cupy. This is because the memory footprint of cupy is much lower than pytorch. Since I plan on...
The opencv `VideoCapture` class has `grab` and `retrieve` methods that allow it to ingest codec data but not actually decode a frame, which saves compute power (not 100% though ofc)....
Attempting to create an FFFR CPU backend after importing PyAV (`import av`) will result in a freeze. PyAV interacts with the ffmpeg libraries which FFFR also uses, so this is...
It seems that the NVDec backend has much higher GPU utilisation than the NVVL backend. I ran this benchmarking script that @anibali gave me as reference to begin investigating why...
