From ef61448f00df33753e89c2bbd999d7498f45105e Mon Sep 17 00:00:00 2001 From: Hanch Han <51526347+hanchchch@users.noreply.github.com> Date: Fri, 17 Jan 2025 11:35:22 +0900 Subject: [PATCH 1/2] fix: refer to use_framewise_encoding on AutoencoderKLHunyuanVideo._encode --- .../models/autoencoders/autoencoder_kl_hunyuan_video.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py b/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py index e2236a7f20ad..4e2cfb39903c 100644 --- a/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py +++ b/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py @@ -868,7 +868,7 @@ def disable_slicing(self) -> None: 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): From 1adc2a126cb2a0132359e308afa2ef56530e991a Mon Sep 17 00:00:00 2001 From: Hanch Han <51526347+hanchchch@users.noreply.github.com> Date: Fri, 17 Jan 2025 11:44:39 +0900 Subject: [PATCH 2/2] fix: comment about tile_sample_min_num_frames --- .../models/autoencoders/autoencoder_kl_hunyuan_video.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py b/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py index 4e2cfb39903c..9be79cfe7dc9 100644 --- a/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py +++ b/src/diffusers/models/autoencoders/autoencoder_kl_hunyuan_video.py @@ -786,7 +786,7 @@ def __init__( 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