Skip to content

Fix: Align auxiliary output features with input view order #183

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

Open
diankun-wu wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix: Align auxiliary output features with input view order #183

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
6 changes: 5 additions & 1 deletion src/depth_anything_3/model/dinov2/vision_transformer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -345,7 +345,11 @@ def _get_intermediate_layers_not_chunked(self, x, n=1, export_feat_layers=[], **
out_x = restore_original_order(out_x, b_idx)
output.append((out_x[:, :, 0], out_x))
if i in export_feat_layers:
aux_output.append(x)
# Apply the same restoration logic to auxiliary features to align with input view order
aux_x = x.clone()
if x.shape[1] >= THRESH_FOR_REF_SELECTION and self.alt_start != -1 and 'b_idx' in locals():
aux_x = restore_original_order(aux_x, b_idx)
aux_output.append(aux_x)
return output, aux_output

def process_attention(self, x, block, attn_type="global", pos=None, attn_mask=None):
Expand Down