mirror of
https://github.com/openai/CLIP.git
synced 2026-11-05 01:37:26 +00:00
[GH-ISSUE #118] what text you use in (image, text) pairs #108
Labels
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
openai/CLIP#108
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @CastleCC on GitHub (Jun 19, 2021).
Original GitHub issue: https://github.com/openai/CLIP/issues/118
Thanks for this amazing work.
This is a question related to the paper, In paper section 8, "CLIP only uses full text sequences co-occuring with images as supervision rather than just the queries". I am curious what “full text sequences” refer to, the image google title or the webpage text along with the image, or other text.
@jongwook commented on GitHub (Jun 30, 2021):
The dataset is a mixture of image-text pairs from various sources. The "full text sequences" are usually title + description concatenated using whatever is available about the image, usually being a sentence or two and not the whole webpage. Similar work from google (ALIGN) uses alt texts.