Response Cut-off Issue in GPT-4 Vision Preview #204

Closed
opened 2025-11-11 14:11:14 -06:00 by GiteaMirror · 2 comments
Owner

Originally created by @CreatorGhost on GitHub (Jan 18, 2024).

Bug Report

Description

Bug Summary:
Responses from the GPT-4 Vision Preview feature are being cut off after the first line, preventing users from receiving complete answers to their prompts.

Steps to Reproduce:

  1. Go to the GPT-4 Vision Preview feature within the Ollama project.
  2. Input a prompt that is expected to generate a multi-line response.
  3. Observe that the generated response is incomplete, with only the first line being displayed before it gets cut off.

Expected Behavior:
The GPT-4 Vision Preview should generate and display the full response to the prompt, providing a complete and coherent answer.

Actual Behavior:
The response is truncated after the first line, leaving the user with an incomplete answer and hindering the functionality of the GPT-4 Vision Preview feature.

Environment

  • Operating System: macOS Big Sur
  • Browser (if applicable): Chrome
Originally created by @CreatorGhost on GitHub (Jan 18, 2024). # Bug Report ## Description **Bug Summary:** Responses from the GPT-4 Vision Preview feature are being cut off after the first line, preventing users from receiving complete answers to their prompts. **Steps to Reproduce:** 1. Go to the GPT-4 Vision Preview feature within the Ollama project. 2. Input a prompt that is expected to generate a multi-line response. 3. Observe that the generated response is incomplete, with only the first line being displayed before it gets cut off. **Expected Behavior:** The GPT-4 Vision Preview should generate and display the full response to the prompt, providing a complete and coherent answer. **Actual Behavior:** The response is truncated after the first line, leaving the user with an incomplete answer and hindering the functionality of the GPT-4 Vision Preview feature. ## Environment - **Operating System:** macOS Big Sur - **Browser (if applicable):** Chrome
Author
Owner

@tjbck commented on GitHub (Jan 19, 2024):

Hi, Thanks for creating this issue! I've just checked the GPT-4 Vision Preview api myself, and it seems like the issue is from openAI prematurely sending a data: [DONE] response, not the webui.

image

I'll be closing this issue as it seems like there isn't anything we can do from the webui, but if you manage to find a fix, feel free to make a PR!

@tjbck commented on GitHub (Jan 19, 2024): Hi, Thanks for creating this issue! I've just checked the `GPT-4 Vision Preview` api myself, and it seems like the issue is from openAI prematurely sending a `data: [DONE]` response, not the webui. <img width="555" alt="image" src="https://github.com/ollama-webui/ollama-webui/assets/25473318/4e89e309-401e-4f0f-9c88-d936c5effde0"> I'll be closing this issue as it seems like there isn't anything we can do from the webui, but if you manage to find a fix, feel free to make a PR!
Author
Owner

@CreatorGhost commented on GitHub (Jan 19, 2024):

I have fixed the issue and raised the PR kindly check and merge.
https://github.com/ollama-webui/ollama-webui/pull/537
image

@CreatorGhost commented on GitHub (Jan 19, 2024): I have fixed the issue and raised the PR kindly check and merge. https://github.com/ollama-webui/ollama-webui/pull/537 <img width="1431" alt="image" src="https://github.com/ollama-webui/ollama-webui/assets/46842909/52db73ef-3693-4447-9c42-b60bf523e0c7">
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#204