Fix VAE spatial sharding dynamic calculation bug in Wan pipeline.#407
Open
ninatu wants to merge 1 commit into
Open
Fix VAE spatial sharding dynamic calculation bug in Wan pipeline.#407ninatu wants to merge 1 commit into
ninatu wants to merge 1 commit into
Conversation
eltsai
previously approved these changes
May 15, 2026
Perseus14
reviewed
May 16, 2026
Previously, setting `vae_spatial: -1` in the config (intended to trigger dynamic calculation of the VAE spatial sharding axis size) was ineffective because `pyconfig.py` prematurely overrode any `-1` or missing `vae_spatial` value to `1`. Furthermore, the dynamic calculation formula in `wan_pipeline.py` (`vae_spatial = (2 * total_devices) // dp_size`) was not robust. On single-device runs (where `total_devices=1` and `dp_size=1`) or configurations with odd data parallel (DP) sizes, it would calculate a `vae_spatial` value (e.g., 2) that does not divide `total_devices`, failing the mesh validation assertion.
ninatu
force-pushed
the
ninatu/wan_vae_sharding_bug
branch
from
69cbe91 to
f9c1253
Compare
Perseus14
approved these changes
May 19, 2026
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
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Previously, setting
vae_spatial: -1in the config (intended to trigger dynamic calculation of the VAE spatial sharding axis size) was ineffective becausepyconfig.pyprematurely overrode any-1or missingvae_spatialvalue to1.Furthermore, the dynamic calculation formula in
wan_pipeline.py(vae_spatial = (2 * total_devices) // dp_size) was not robust. On single-device runs (wheretotal_devices=1anddp_size=1) or configurations with odd data parallel (DP) sizes, it would calculate avae_spatialvalue (e.g., 2) that does not dividetotal_devices, failing the mesh validation assertion.