Fix: Prevent Invalid next_state References When optimize_memory=True #918
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
- Restrict sampling indices to valid transitions when buffer is not full, avoiding uninitialized memory access in next_state - Maintain full sampling range when buffer is full (circular behavior)
s1lent4gnt
force-pushed
the
lilkm/fix-next-state-uninitialized-memory
branch
from
12f7db5 to
610ac46
Compare
Collaborator
|
lgtm thank you 😄 |
AdilZouitine
merged commit 66c3672
into
huggingface:user/adil-zouitine/2025-1-7-port-hil-serl-new
1 check failed
michel-aractingi
pushed a commit
that referenced
this pull request
Apr 17, 2025
michel-aractingi
pushed a commit
that referenced
this pull request
Apr 18, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
What this does
When
optimize_memory=True, theReplayBufferavoids redundantnext_statestorage by referencingstate[i+1]. However, if the buffer is not full, sampling indices nearself.sizecould reference uninitialized memory, leading to invalidnext_statevalues (e.g., black images).To address this, the sampling logic has been adjusted to:
Restrict indices to valid transitions when the buffer is not full.
Allow full sampling when the buffer is full (circular buffer behavior).
How it was tested
A training experiment was run, confirming no errors related to black images.