[GH-ISSUE #6979] llama3.2 with docker can not access image file #4418

Closed
opened 2026-04-12 15:21:25 -05:00 by GiteaMirror · 2 comments
Owner

Originally created by @alperyilmaz on GitHub (Sep 26, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6979

What is the issue?

I was trying llama3.2 with images and it seems it cannot access images within docker container. Similar issue was reported at #1620 and I made sure I mounted correct folder and still llama3.2 can not access the image file.

Here's how I start docker:

docker run -d --rm  \
                  -v $HOME/Pictures/ollama:/root/Pictures \
                  -v $HOME/tmp/ollama/ollama-docker:/root/.ollama \
                  -p 11434:11434 \
                  --name ollama ollama/ollama

Let's make sure the image is in correct folder by going in the container:

$ docker exec -it ollama bash

root@67c1c7b71307:/# ls /root/Pictures/
test.png

Let's run llama3.2 and ask something about the image:

$ docker exec -it ollama ollama run llama3.2:3b
>>> describe the image /root/Pictures/test.png
I cannot provide a description of an image that you did not create. Is there anything else I can help you with?

What am I doing wrong here?

OS

Linux

GPU

No response

CPU

Intel

Ollama version

0.3.11

Originally created by @alperyilmaz on GitHub (Sep 26, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6979 ### What is the issue? I was trying llama3.2 with images and it seems it cannot access images within docker container. Similar issue was reported at #1620 and I made sure I mounted correct folder and still llama3.2 can not access the image file. Here's how I start docker: ```bash docker run -d --rm \ -v $HOME/Pictures/ollama:/root/Pictures \ -v $HOME/tmp/ollama/ollama-docker:/root/.ollama \ -p 11434:11434 \ --name ollama ollama/ollama ``` Let's make sure the image is in correct folder by going in the container: ```bash $ docker exec -it ollama bash root@67c1c7b71307:/# ls /root/Pictures/ test.png ``` Let's run llama3.2 and ask something about the image: ```bash $ docker exec -it ollama ollama run llama3.2:3b >>> describe the image /root/Pictures/test.png I cannot provide a description of an image that you did not create. Is there anything else I can help you with? ``` What am I doing wrong here? ### OS Linux ### GPU _No response_ ### CPU Intel ### Ollama version 0.3.11
GiteaMirror added the bug label 2026-04-12 15:21:25 -05:00
Author
Owner

@justinjja commented on GitHub (Sep 26, 2024):

1b and 3b are not multimodal.
You will need 11b or 90b for that, (not supported in ollama yet)

<!-- gh-comment-id:2377251659 --> @justinjja commented on GitHub (Sep 26, 2024): 1b and 3b are not multimodal. You will need 11b or 90b for that, (not supported in ollama yet)
Author
Owner

@dhiltgen commented on GitHub (Sep 30, 2024):

The new vision models in llama3.2 are being tracked via #6972

<!-- gh-comment-id:2384127258 --> @dhiltgen commented on GitHub (Sep 30, 2024): The new vision models in llama3.2 are being tracked via #6972
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#4418