Skip to content

Conversation

@scotts
Copy link
Contributor

@scotts scotts commented Dec 4, 2025

Working on #1094 made me realize that if _make_transform_specs() continues to live in _video_decoder.py, it will linearly grow in size as we add transforms, and the additions will be boiler-plate code. We should try to keep the code in _video_decoder.py specific to the actual VideoDecoder public API. l think making _make_transform_specs() live in torchcodec.transforms is cleaner, particularly since we have a clear API (transforms go in, string specs come out).

I also realized that over time, we might want the transforms files to start mirroring how they're structured in TorchVision, but we can put that off we have enough transforms it starts being an organizational problem.

@meta-cla meta-cla bot added the CLA Signed This label is managed by the Meta Open Source bot. label Dec 4, 2025
@scotts scotts marked this pull request as ready for review December 4, 2025 16:45
@scotts scotts merged commit 4e412b7 into meta-pytorch:main Dec 4, 2025
74 checks passed
@scotts scotts deleted the refactor_make_spec branch December 4, 2025 20:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Meta Open Source bot.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants