github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7682] OpenCoder's template doesn't make sense for an instruct model #30663

New Issue
Closed
opened 2026-04-22 10:33:06 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 10:33:06 -05:00
Owner
Copy Link

Originally created by @ProjectMoon on GitHub (Nov 15, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7682

What is the issue?

I am testing out OpenCoder 8b. The information on the page on ollama.com implies that it's a chat model, but the template does not reference the system prompt at all. Setting the system prompt in OpenWebUI seems to do funny things with the chat (the model has a conversation with itself, because it's not generating the right stop tokens). The template as it is at the moment seems to be for a code generation model? Since it has the suffix stuff in it. Is OpenCoder on the website a chat model or a base model?

OS

Linux

GPU

AMD

CPU

AMD

Ollama version

0.4.1

Originally created by @ProjectMoon on GitHub (Nov 15, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7682 ### What is the issue? I am testing out OpenCoder 8b. The information on the page on ollama.com implies that it's a chat model, but the template does not reference the system prompt at all. Setting the system prompt in OpenWebUI seems to do funny things with the chat (the model has a conversation with itself, because it's not generating the right stop tokens). The template as it is at the moment seems to be for a code generation model? Since it has the suffix stuff in it. Is OpenCoder on the website a chat model or a base model? ### OS Linux ### GPU AMD ### CPU AMD ### Ollama version 0.4.1
GiteaMirror added the bug label 2026-04-22 10:33:06 -05:00
GiteaMirror commented 2026-04-22 10:33:08 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Nov 15, 2024):

The embedded chat_template uses system and doesn't mention FIM, so it looks like the ollama template is wrong.

{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system
You are OpenCoder, created by OpenCoder Team.<|im_end|>
' }}{% endif %}{{'<|im_start|>' + message['role'] + '
' + message['content'] + '<|im_end|>' + '
'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant
' }}{% endif %}
<!-- gh-comment-id:2478267789 --> @rick-github commented on GitHub (Nov 15, 2024): The embedded `chat_template` uses system and doesn't mention FIM, so it looks like the ollama template is wrong. ``` {% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system You are OpenCoder, created by OpenCoder Team.<|im_end|> ' }}{% endif %}{{'<|im_start|>' + message['role'] + ' ' + message['content'] + '<|im_end|>' + ' '}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant ' }}{% endif %} ```
GiteaMirror commented 2026-04-22 10:33:08 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Nov 18, 2024):

It should be a chat model – it doesn't always listen to the system prompt but it does work sometimes:

% ollama run opencoder
>>> /set system only answer in french  
Set system message.
>>> hi
Bonjour ! Comment puis-je vous aider aujourd'hui ?
 
(The assistant greets the user and waits for instructions.)

I've since removed the FIM template – it seems the tokens are present, and it works, but was not trained for this 😊

<!-- gh-comment-id:2482152537 --> @jmorganca commented on GitHub (Nov 18, 2024): It should be a chat model – it doesn't always listen to the system prompt but it does work sometimes: ``` % ollama run opencoder >>> /set system only answer in french Set system message. >>> hi Bonjour ! Comment puis-je vous aider aujourd'hui ? (The assistant greets the user and waits for instructions.) ``` I've since removed the FIM template – it seems the tokens are present, and it works, but was not trained for this 😊
GiteaMirror commented 2026-04-22 10:33:09 -05:00
Author
Owner
Copy Link

@ProjectMoon commented on GitHub (Nov 18, 2024):

So, this seems to work much better now. That is, until a system prompt is set. If you set a system prompt, in OpenWebUI at least, the model starts having conversations with itself.

<!-- gh-comment-id:2482863544 --> @ProjectMoon commented on GitHub (Nov 18, 2024): So, this seems to work much better now. That is, until a system prompt is set. If you set a system prompt, in OpenWebUI at least, the model starts having conversations with itself.
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#30663
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?