[GH-ISSUE #2058] Prompt Eval Count is 1 when image is included in multimodal request #1190

Closed
opened 2026-04-12 10:58:19 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @Dillon-Yun on GitHub (Jan 18, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2058

When I hit the /api/generate endpoint for a multimodal model, such as bakllava, I am seeing that prompt_eval_count == 1 when an image is included. Is this correct behavior? How can I retrieve the total amount of tokens in the request for a multimodal request?

Originally created by @Dillon-Yun on GitHub (Jan 18, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2058 When I hit the `/api/generate` endpoint for a multimodal model, such as bakllava, I am seeing that `prompt_eval_count == 1` when an image is included. Is this correct behavior? How can I retrieve the total amount of tokens in the request for a multimodal request?
Author
Owner

@dhiltgen commented on GitHub (Oct 23, 2024):

Confirmed this is fixed in the Go server in the 0.4.0 RC.

<!-- gh-comment-id:2432992688 --> @dhiltgen commented on GitHub (Oct 23, 2024): Confirmed this is fixed in the Go server in the 0.4.0 RC.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#1190