DirectML Out of Memory Retry #707
Merged
Conversation
This file contains 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
|
Found it helps quite a bit to clear the traceback before retrying. Now it'll more often recover and will usually finish 25 steps of SDXL without requiring offloading. The decoding stage still doesn't complete on my GPU without offloading or tiling though. That should be it for this PR. |
carson-katri
approved these changes
Sep 22, 2023
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Outdated suggestions cannot be applied.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.
Allows DML execution in submodules to retry running when it runs out of memory. Sometimes this can allow generation to finish, but often it'll occur again on a later step and not recover from it.
Found and fixed some DML issues caused from merging the last branch.
It can run SDXL, but may require offloading depending on what GPU is used. Using half VAE tiling is also effective and maintains good decoded image accuracy.