Skip to content

[VLM][Bugfix] enable internvl running with num_scheduler_steps > 1 #8614

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
youkaichao merged 6 commits into from
Sep 25, 2024
Merged

[VLM][Bugfix] enable internvl running with num_scheduler_steps > 1 #8614

Changes from 4 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
890d1c6
fix error while running internvl with num_scheduler_steps > 1
DefTruth Sep 19, 2024
ca53757
Merge branch 'vllm-project:main' into internvl-num-scheduler-steps
DefTruth Sep 24, 2024
3ce3a69
enable internvl run with num_scheduler_steps > 1
DefTruth Sep 24, 2024
b3140a9
Merge branch 'vllm-project:main' into internvl-num-scheduler-steps
DefTruth Sep 25, 2024
a2b19b2
fix ci
DefTruth Sep 25, 2024
1e31a44
fix ci
DefTruth Sep 25, 2024
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
2 changes: 2 additions & 0 deletions vllm/model_executor/models/internvl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -376,6 +376,8 @@ def __init__(self,
self.make_empty_intermediate_tensors = (
self.language_model.make_empty_intermediate_tensors)

self.sampler = self.language_model.sampler

def pixel_shuffle(self, x, scale_factor=0.5):
n, w, h, c = x.size()
# N, W, H, C --> N, W, H * scale, C // scale
Expand Down