Small fix in inference cli: Do final multiplication in Pytorch instead#578
Open
sognetic wants to merge 1 commit into
Open
Small fix in inference cli: Do final multiplication in Pytorch instead#578sognetic wants to merge 1 commit into
sognetic wants to merge 1 commit into
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hi, first of all: Thanks for the nice software and especially the great docs, running the model is really a breeze with this tooling.
I've run this in standalone mode on a RTX 5090 server node to process a >1h video files in ~10 minute chunks and noticed noticeable slowdown (more than 10 minutes) after each chunk, with only a single core out of 192 actually doing anything. I think the culprit is doing this final transformation in numpy instead of e.g. pytorch, multiplying first in pytorch and then casting to numpy removed the bottleneck. The
clampis just to make the transformation a bit more obvious wrt. value range.I've only tested this with a video file in a single-GPU setup on that specific machine and there might be ways to do this better (e.g. multiplying entirely on the GPU) but I think this already solves the issue sufficiently.
Let me know what you think and thanks again for the great project!