Skip to content

[attention] Fix attention #2656

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Mar 13, 2023
Merged

[attention] Fix attention #2656

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
79b4075
[attention] Fix attention
patrickvonplaten Mar 13, 2023
ad106be
fix
patrickvonplaten Mar 13, 2023
665322e
correct
patrickvonplaten Mar 13, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 5 additions & 2 deletions src/diffusers/models/attention.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -271,9 +271,10 @@ def __init__(
def forward(
self,
hidden_states,
attention_mask=None,
encoder_hidden_states=None,
encoder_attention_mask=None,
timestep=None,
attention_mask=None,
cross_attention_kwargs=None,
class_labels=None,
):
Expand Down Expand Up @@ -302,12 +303,14 @@ def forward(
norm_hidden_states = (
self.norm2(hidden_states, timestep) if self.use_ada_layer_norm else self.norm2(hidden_states)
)
# TODO (Birch-San): Here we should prepare the encoder_attention mask correctly
# prepare attention mask here

# 2. Cross-Attention
attn_output = self.attn2(
norm_hidden_states,
encoder_hidden_states=encoder_hidden_states,
attention_mask=attention_mask,
attention_mask=encoder_attention_mask,
**cross_attention_kwargs,
)
hidden_states = attn_output + hidden_states
Expand Down
2 changes: 1 addition & 1 deletion tests/pipelines/stable_diffusion/test_stable_diffusion.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -737,7 +737,7 @@ def test_stable_diffusion_vae_tiling(self):

# make sure that more than 4 GB is allocated
mem_bytes = torch.cuda.max_memory_allocated()
assert mem_bytes > 4e9
assert mem_bytes > 5e9
assert np.abs(image_chunked.flatten() - image.flatten()).max() < 1e-2

def test_stable_diffusion_fp16_vs_autocast(self):
Expand Down