github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-17 13:47:08 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #3135] LLava returns non-unique responses #79606

New Issue
Open
opened 2026-05-09 05:58:35 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-09 05:58:35 -05:00
Owner
Copy Link

Originally created by @jeffreylm on GitHub (Mar 14, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3135

Dear Ollama,

I am using LLava with WebUI, and have already spoken to WebUi github about this, and as here is no errors in the docker error logs, they advised me to contact the Ollama team's Github.

When I type : Describe Image
LLava give me a cloned answer no matter the images I give it. The first image/response (after a system reboot) will be unique, but all future request with new images will give the same exact answer, word for word/a clone response.

It would go off for a few minutes to think, and comeback with the prior text. Whether I use the LLava:7b or the LLava:34b
My graphics card is an AMD RX580 8GB (I understand it is on the low-end side), but nonetheless, I don't understand why it is giving a cloned answer.

Originally created by @jeffreylm on GitHub (Mar 14, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3135 Dear Ollama, I am using LLava with WebUI, and have already spoken to WebUi github about this, and as here is no errors in the docker error logs, they advised me to contact the Ollama team's Github. When I type : Describe Image LLava give me a cloned answer no matter the images I give it. The first image/response (after a system reboot) will be unique, but all future request with new images will give the same exact answer, word for word/a clone response. It would go off for a few minutes to think, and comeback with the prior text. Whether I use the LLava:7b or the LLava:34b My graphics card is an AMD RX580 8GB (I understand it is on the low-end side), but nonetheless, I don't understand why it is giving a cloned answer.
GiteaMirror added the bug label 2026-05-09 05:58:35 -05:00
GiteaMirror commented 2026-05-09 06:02:44 -05:00
Author
Owner
Copy Link

@RandomGitUser321 commented on GitHub (Mar 15, 2024):

It sounds like a bad template or something. Which exact model version are you using? If you're on windows, open cmd and type ollama list and paste the exact name here like llava:7b-v1.6-mistral-q4_K_M in my case.

When you installed the model, how did you do it? Did you do it the correct way through Ollama with a command like ollama run llava:7b-v1.6-mistral-q4_K_M or ollama pull llava:7b-v1.6-mistral-q4_K_M

You can find the models in the https://ollama.com/library search for llava, click the drop down, pick a version/quant you want and click the copy command icon toward the top right of the results. Paste that into the comand prompt and it will get everything you need, including the templates.

<!-- gh-comment-id:1998884375 --> @RandomGitUser321 commented on GitHub (Mar 15, 2024): It sounds like a bad template or something. Which exact model version are you using? If you're on windows, open `cmd` and type `ollama list` and paste the exact name here like `llava:7b-v1.6-mistral-q4_K_M` in my case. When you installed the model, how did you do it? Did you do it the correct way through Ollama with a command like `ollama run llava:7b-v1.6-mistral-q4_K_M` or `ollama pull llava:7b-v1.6-mistral-q4_K_M` You can find the models in the `https://ollama.com/library` search for llava, click the drop down, pick a version/quant you want and click the copy command icon toward the top right of the results. Paste that into the comand prompt and it will get everything you need, including the templates.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#79606
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?