Flux2klein pipeline part2#440
Open
amepas wants to merge 2 commits into
Open
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This is part 2 of the Flux2 Klein onboarding - it contains the full end-to-end pipeline for inference w/ smoke tests and parity tests against CPU PyTorch (in both float32 and bfloat16 to isolate rounding errors vs. algorithmic issues).
This relies on pull request 439.
On a v7x (1x1x1), with batch-size 8, we achieve the following latency w/ splash attention. This timing is for the 9B model (both 4B and 9B are supported)
The E2E latency after compilation is broken down into the following:
Future PRs will investigate faster model loading, cold-start and bringing down latency of inference