Ethanfel f99d2666e8 fix: interpolate sync_cond to match audio sequence length in transformer ...

Sync_MLP interpolates sync features based on video duration, but audio
latent length depends on the user-set audio duration. When video != audio
duration, the sequences diverge. Resample sync_cond to x's length before
the gated addition so any video/audio duration combo works.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

2026-03-27 21:21:39 +01:00

..

mmmodules

feat: extract prismaudio_core model modules (DiT, conditioners, VAE, diffusion)

2026-03-27 17:31:22 +01:00

__init__.py

feat: extract prismaudio_core model modules (DiT, conditioners, VAE, diffusion)

2026-03-27 17:31:22 +01:00

adp.py

feat: extract prismaudio_core model modules (DiT, conditioners, VAE, diffusion)

2026-03-27 17:31:22 +01:00

autoencoders.py

fix: resolve critical bugs and quality issues in prismaudio_core/models

2026-03-27 17:56:02 +01:00

blocks.py

fix: resolve critical bugs and quality issues in prismaudio_core/models

2026-03-27 17:56:02 +01:00

bottleneck.py

feat: extract prismaudio_core model modules (DiT, conditioners, VAE, diffusion)

2026-03-27 17:31:22 +01:00

conditioners.py

fix: clean up dead code paths and debug artifacts in prismaudio_core/models

2026-03-27 17:49:57 +01:00

diffusion.py

fix: resolve critical bugs and quality issues in prismaudio_core/models

2026-03-27 17:56:02 +01:00

dit.py

fix: clean up dead code paths and debug artifacts in prismaudio_core/models

2026-03-27 17:49:57 +01:00

local_attention.py

fix: resolve critical bugs and quality issues in prismaudio_core/models

2026-03-27 17:56:02 +01:00

pqmf.py

feat: extract prismaudio_core model modules (DiT, conditioners, VAE, diffusion)

2026-03-27 17:31:22 +01:00

pretransforms.py

fix: clean up dead code paths and debug artifacts in prismaudio_core/models

2026-03-27 17:49:57 +01:00

transformer.py

fix: interpolate sync_cond to match audio sequence length in transformer

2026-03-27 21:21:39 +01:00

utils.py

fix: resolve critical bugs and quality issues in prismaudio_core/models

2026-03-27 17:56:02 +01:00