feat(flux): diffusion training primitives (forward step, schedulers, loss)#816
Draft
jasainio wants to merge 5 commits into
Draft
feat(flux): diffusion training primitives (forward step, schedulers, loss)#816jasainio wants to merge 5 commits into
jasainio wants to merge 5 commits into
Conversation
… on draft PRs Bump PRIMUS_TURBO_COMMIT to Primus-Turbo main (0.3.x, validated build) which exposes the mxfp4 gemm_fp4_impl(..., preshuffled=...) fast path, and PRIMUS_TURBO_AITER_COMMIT to AITER v0.1.14.post1. Mirror the turbo commit into benchmark.yaml; the docker build receives both via ci.yaml build-args. Add a draft-skip guard to the expensive jobs (build-docker, run-unittest-torch, run-unittest-jax) so the stacked Flux PRs can sit in Draft without firing the full pipeline; code-lint stays ungated. Add ready_for_review to the pull_request trigger types so flipping a draft to Ready triggers the first run. MEGATRON_PATH export intentionally omitted: the Flux diffusion unit tests run in-process (no megatron-importing subprocess remains after the te_vs_local in-process conversion), and tests/conftest.py auto-adds third_party/Megatron-LM to sys.path when MEGATRON_PATH is unset. CI already checks out submodules: recursive, so megatron imports without any extra env.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Part of an 18-PR series splitting the Flux diffusion-training feature (training Flux, a DiT text-to-image diffusion model, on Primus/Megatron) out of one large branch for reviewability. Targets
feat/flux/flux— review after it. One of the parents of the trainers PR.What this changes
The diffusion training primitives: the forward step, loss computation, noise utilities, timestep sampling, and the flow-matching schedulers.
Why it's stacked here
The forward step imports
flux.utils, so it bases onfeat/flux/fluxrather thanfeat/flux/model-common.Dependencies
Sequenced after the CI-pins PR (
feat/flux/ci-env); builds onfeat/flux/flux.Test plan
pytest tests/unit_tests/backends/megatron/diffusion/training -k "forward_step or loss or scheduler". Validated locally on an AMD GPU container: 20 passed.Files
13 (forward step, loss, noise/timestep sampling, flow-matching schedulers + tests).