Skip to content
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

Fix med/coco caption #252

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Fix med/coco caption #252

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
93e2cf4
fix on med.py: ensure the prompt is also being repeated as the visual…
Luodian Apr 12, 2023
537a1de
fix on med.py: add determine statement before repeat tokenized_prompt.
Luodian Apr 12, 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
5 changes: 5 additions & 0 deletions lavis/models/med.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1332,6 +1332,11 @@ def generate_from_encoder(
num_beams = num_beams
visual_embeds = visual_embeds.repeat_interleave(num_beams, dim=0)

if visual_embeds.size(0) != tokenized_prompt.input_ids.size(0):
tokenized_prompt.input_ids = tokenized_prompt.input_ids.repeat_interleave(num_beams, dim=0)
# Make sure that the prompt is repeated same number of times as the visual_embeds
assert visual_embeds.size(0) == tokenized_prompt.input_ids.size(0)

image_atts = torch.ones(visual_embeds.size()[:-1], dtype=torch.long).to(
self.device
)
Expand Down