AEmotionStudio cb391fa5d9 fix: Consistent shader registration and tensor-aware interpolation ...

- shader_noise_ksampler.py: register_shader_generator now registers to both
  the legacy SHADER_GENERATORS dict and the new centralized registry,
  matching the behavior in __init__.py for backward compatibility

- core/transforms.py: resize_noise_spatial now considers INPUT tensor
  dimensionality when selecting interpolation mode, not just target_size.
  For 5D video tensors with 2D target_size (H, W), uses frame-by-frame
  bilinear interpolation via new _resize_5d_tensor_spatial helper.
  Previously would incorrectly select bilinear mode for 5D tensors.

2026-01-27 00:32:53 -08:00

..

__init__.py

refactor: Major codebase restructuring for separation of concerns

2026-01-26 19:48:23 -08:00

blending.py

fix: Handle spatial dimensions before channels in shape correction

2026-01-27 00:08:31 -08:00

constants.py

refactor: Major codebase restructuring for separation of concerns

2026-01-26 19:48:23 -08:00

debug.py

refactor: Major codebase restructuring for separation of concerns

2026-01-26 19:48:23 -08:00

logging_config.py

refactor: Major codebase restructuring for separation of concerns

2026-01-26 19:48:23 -08:00

model_compat.py

fix: Resolve shader registration, unused variable, and parameter mismatch

2026-01-26 21:29:12 -08:00

params.py

fix: Address video channel detection and type consistency issues

2026-01-26 21:12:35 -08:00

sampler.py

fix: Handle spatial dimensions before channels in shape correction

2026-01-27 00:08:31 -08:00

transforms.py

fix: Consistent shader registration and tensor-aware interpolation

2026-01-27 00:32:53 -08:00