[GH-ISSUE #32] Image feature clustering #29

New issue

Closed

opened 2026-10-31 21:10:03 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-31 21:10:03 +00:00

Owner

Originally created by @haofanwang on GitHub (Feb 1, 2021).
Original GitHub issue: https://github.com/openai/CLIP/issues/32

Now, the feature dimension is 512 for both image and text. I'd like to know whether it is possible to do dimension reduction to speed up the inference time. After I do PCA on the image feature, is it possible to use clip features to do clustering?

Originally created by @haofanwang on GitHub (Feb 1, 2021). Original GitHub issue: https://github.com/openai/CLIP/issues/32 Now, the feature dimension is 512 for both image and text. I'd like to know whether it is possible to do dimension reduction to speed up the inference time. After I do PCA on the image feature, is it possible to use clip features to do clustering?

bsakaguchi closed this issue

2026-10-31 21:10:03 +00:00

bsakaguchi commented

2026-10-31 21:10:04 +00:00

Author

Owner

@jongwook commented on GitHub (Feb 12, 2021):

We don't have data for this. In general, you'll need to find a trade-off point between the dimensionality/speed and feature quality. You might also want to look into approximate nearest neighbor methods.

@jongwook commented on GitHub (Feb 12, 2021): We don't have data for this. In general, you'll need to find a trade-off point between the dimensionality/speed and feature quality. You might also want to look into [approximate nearest neighbor](http://ann-benchmarks.com/) methods.

No labels

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

openai/CLIP#29

No description provided.

Rows
Columns