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: enable unet_3d_condition to support time_cond_proj_dim (#7364)

Browse Source 
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
This commit is contained in:
Yuanhao Zhai
2024-03-20 22:16:32 -04:00
committed by GitHub
parent 3028089e5e
commit 405a1facd2
1 changed files with 4 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/diffusers/models/unets/unet_3d_condition.py
View File

@@ -91,6 +91,8 @@ class UNet3DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin)
cross_attention_dim (`int`, *optional*, defaults to 1024): The dimension of the cross attention features.
attention_head_dim (`int`, *optional*, defaults to 64): The dimension of the attention heads.
num_attention_heads (`int`, *optional*): The number of attention heads.
time_cond_proj_dim (`int`, *optional*, defaults to `None`):
The dimension of `cond_proj` layer in the timestep embedding.
"""
_supports_gradient_checkpointing = False
@@ -123,6 +125,7 @@ class UNet3DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin)
cross_attention_dim: int = 1024,
attention_head_dim: Union[int, Tuple[int]] = 64,
num_attention_heads: Optional[Union[int, Tuple[int]]] = None,
time_cond_proj_dim: Optional[int] = None,
):
super().__init__()
@@ -174,6 +177,7 @@ class UNet3DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin)
timestep_input_dim,
time_embed_dim,
act_fn=act_fn,
cond_proj_dim=time_cond_proj_dim,
)
self.transformer_in = TransformerTemporalModel(