[GH-ISSUE #10913] image to base64 conversion too slow on Mac Mini M4 Pro? #32938

Open
opened 2026-04-22 14:55:05 -05:00 by GiteaMirror · 2 comments
Owner

Originally created by @eUGENE-BNX on GitHub (May 30, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10913

A webp image with 720x416 resolution and a file size of 26 KB. When I want to analyze this image with Gemma3:12b model, it seems to wait 10-12 seconds just for this conversion.

Actually model is quite fast 23.5tokens/s. If I continue the analysis with the same image, the total time is shortened by 10-12 seconds. But if I want it to analyze with a new image, I wait 10-15 seconds even though it is a small image. Do you think it is normal for this process to be this slow?

Originally created by @eUGENE-BNX on GitHub (May 30, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10913 A webp image with 720x416 resolution and a file size of 26 KB. When I want to analyze this image with Gemma3:12b model, it seems to wait 10-12 seconds just for this conversion. Actually model is quite fast 23.5tokens/s. If I continue the analysis with the same image, the total time is shortened by 10-12 seconds. But if I want it to analyze with a new image, I wait 10-15 seconds even though it is a small image. Do you think it is normal for this process to be this slow?
Author
Owner

@SpiderMan95 commented on GitHub (Jun 9, 2025):

I also encountered the same problem.

<!-- gh-comment-id:2954696805 --> @SpiderMan95 commented on GitHub (Jun 9, 2025): I also encountered the same problem.
Author
Owner

@eUGENE-BNX commented on GitHub (Jun 10, 2025):

I'm sure this cycle can be done much faster. Is time being wasted elsewhere?

<!-- gh-comment-id:2958378140 --> @eUGENE-BNX commented on GitHub (Jun 10, 2025): I'm sure this cycle can be done much faster. Is time being wasted elsewhere?
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#32938