facok ec9941b5fc Use model's latent_format for space conversion instead of hardcoded FLUX constants ...

Replaces hardcoded SCALE_FACTOR/SHIFT_FACTOR with model.latent_format.process_in/out
so LCS works with any model (FLUX, LTXV, SD, etc). Adds LTXAV packed tensor
unpack/repack support for audio+video models. All hooks (color, tone, sharpness,
observer) now use the model-aware conversion from args["model"].

2026-03-21 21:30:22 +08:00

..

__pycache__

Add progress output to calibration

2026-03-20 03:02:04 +08:00

__init__.py

Add sharpness intervention via PCA subspace in FLUX VAE patch space

2026-03-20 13:01:23 +08:00

calibrate.py

Merge LCSCalibrate into LCSLoadData with per-VAE auto-caching

2026-03-17 06:27:55 +08:00

intervene.py

Use model's latent_format for space conversion instead of hardcoded FLUX constants

2026-03-21 21:30:22 +08:00

observe.py

Use model's latent_format for space conversion instead of hardcoded FLUX constants

2026-03-21 21:30:22 +08:00

sharpen.py

Use model's latent_format for space conversion instead of hardcoded FLUX constants

2026-03-21 21:30:22 +08:00