Ethanfel c0b7ccbcee fix: substitute empty_clip_feat for video features when no video present ...

Zero features through bias-free Cond_MLP produce near-zero activations,
not the learned null signal the model was trained with. Use empty_clip_feat
(the learned null video embedding) just like empty_sync_feat for sync.
Also improve text_prompt tooltip to encourage detailed CoT descriptions.

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

2026-03-27 22:13:22 +01:00

..

__init__.py

feat: project scaffolding with shared utils and node registration

2026-03-27 16:59:21 +01:00

feature_extractor.py

perf: replace PIL+PNG frame files with direct ffmpeg stdin pipe

2026-03-27 21:20:00 +01:00

feature_loader.py

feat: PrismAudioFeatureLoader node for pre-computed .npz files

2026-03-27 18:04:32 +01:00

model_loader.py

debug: fix VAE load_state_dict diagnostic — load into .model directly

2026-03-27 21:56:06 +01:00

sampler.py

fix: substitute empty_clip_feat for video features when no video present

2026-03-27 22:13:22 +01:00

text_only.py

fix: substitute empty_clip_feat for video features when no video present

2026-03-27 22:13:22 +01:00

utils.py

feat: project scaffolding with shared utils and node registration

2026-03-27 16:59:21 +01:00