AI/diffusers
1
0
Fork 0
mirror of https://github.com/huggingface/diffusers.git synced 2026-01-27 17:22:53 +03:00
Code Activity

[fix] refer use_framewise_encoding on AutoencoderKLHunyuanVideo._encode (#10600)

Browse Source 
* fix: refer to use_framewise_encoding on AutoencoderKLHunyuanVideo._encode

* fix: comment about tile_sample_min_num_frames

---------

Co-authored-by: Aryan <aryan@huggingface.co>
This commit is contained in:
Hanch Han
2025-01-28 10:21:27 +09:00
committed by GitHub
parent 658e24e86c
commit f295e2eefc
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py
View File

@@ -786,7 +786,7 @@ class AutoencoderKLHunyuanVideo(ModelMixin, ConfigMixin):
self.use_tiling = False
# When decoding temporally long video latents, the memory requirement is very high. By decoding latent frames
# at a fixed frame batch size (based on `self.num_latent_frames_batch_sizes`), the memory requirement can be lowered.
# at a fixed frame batch size (based on `self.tile_sample_min_num_frames`), the memory requirement can be lowered.
self.use_framewise_encoding = True
self.use_framewise_decoding = True
@@ -868,7 +868,7 @@ class AutoencoderKLHunyuanVideo(ModelMixin, ConfigMixin):
def _encode(self, x: torch.Tensor) -> torch.Tensor:
batch_size, num_channels, num_frames, height, width = x.shape
if self.use_framewise_decoding and num_frames > self.tile_sample_min_num_frames:
if self.use_framewise_encoding and num_frames > self.tile_sample_min_num_frames:
return self._temporal_tiled_encode(x)
if self.use_tiling and (width > self.tile_sample_min_width or height > self.tile_sample_min_height):