Turn BytesContext into FromTensorContext #721
Merged
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
facebook-github-bot
added
the
CLA Signed
scotts
commented
Jun 11, 2025
scotts
commented
Jun 11, 2025
NicolasHug
approved these changes
Jun 11, 2025
test/test_decoders.py
Outdated
Comment on lines
129
to
130
| def test_create_bytes_ownership(self): | ||
| # Note that the bytes object we use to instantiate the decoder does not |
There was a problem hiding this comment.
Suggested change
| def test_create_bytes_ownership(self): | |
| # Note that the bytes object we use to instantiate the decoder does not | |
| def test_create_bytes_ownership(self): | |
| # Non-regression test for https://github.com/pytorch/torchcodec/issues/720 | |
| # Note that the bytes object we use to instantiate the decoder does not |
|
|
||
| SingleStreamDecoder::SeekMode realSeek = SingleStreamDecoder::SeekMode::exact; | ||
| if (seek_mode.has_value()) { | ||
| realSeek = seekModeFromString(seek_mode.value()); | ||
| } | ||
|
|
||
| auto contextHolder = std::make_unique<AVIOBytesContext>(data, length); | ||
| auto contextHolder = std::make_unique<AVIOFromTensorContext>(video_tensor); |
There was a problem hiding this comment.
Can you confirm my understanding that the crux of the fix is this line, where we now pass-in the tensor which is ref-counted, instead of passing the raw underlying data which wasn't ref-counted and thus freed?
There was a problem hiding this comment.
Yes, that is correct!
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Fixes #720.
Reading from a raw
void*is problematic because we have no way of claiming ownership of it. This PR makes sure reading from bytes only happens through a tensor. The tensor is effectively a smart pointer, so it ensures that we actually keep the data around. We now have both a tensor-reading context and a tensor-writing context. This PR also refactors both to share code - which we thought we may eventually need to do.The test added by this PR fails without these changes in the C++ level.