enforce original size to be a list#33564
Conversation
zucchini-nlp
left a comment
There was a problem hiding this comment.
@dom-dziela Thanks for fixing this! Can you run make fix-copies to make our CI happy?
|
If I run |
|
@dom-dziela Ah, I forgot this is LLaVa-OV. The modified function has a |
9a55164 to
6b4ff66
Compare
|
@dom-dziela Thanks a lot! Can you pls also propagate changes to llava-next-video model? My bad that it didn't propagare automatically with Also, requested review from core maintainer, after their approval we can merge |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
* enforce original size to be a list * formatting * apply datatype change to unpad_image in llava_next
What does this PR do?
Fixes #33531
Changes the datatype of
original_sizefromtorch.tensor(original_size, dtype = torch.int64)to a list. This prevents possible calculation inaccuracies induced by hardware- and torch-versions, as in issue above, by conducting calculation in base python.Before submitting
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
@amyeroberts
@zucchini-nlp