github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7356] Console show formation backslash chars #66732

New Issue
Closed
opened 2026-05-04 07:59:30 -05:00 by GiteaMirror · 10 comments
GiteaMirror commented 2026-05-04 07:59:30 -05:00
Owner
Copy Link

Originally created by @lsalamon on GitHub (Oct 25, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7356

What is the issue?

Why any question show text not formatted at console like this:
A number \( a \) has a multiplicative inverse modulo \( n \) if there exists an integer \( b \) such that:
\[ (a \cdot b) \equiv 1 \pmod{n} \]

Ollama Win64 0.3.14.0.
Old version 0.3.10.0 has same issue.

I noticed that sometimes, in daily use, the problem doesn't always appear.

OS

Windows

GPU

Other

CPU

AMD

Ollama version

0.3.14.0

Originally created by @lsalamon on GitHub (Oct 25, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7356 ### What is the issue? Why any question show text not formatted at console like this: <code>A number \\( a \\) has a multiplicative inverse modulo \\( n \\) if there exists an integer \\( b \\) such that: \\[ (a \\cdot b) \\equiv 1 \\pmod{n} \\]</code> Ollama Win64 0.3.14.0. Old version 0.3.10.0 has same issue. I noticed that sometimes, in daily use, the problem doesn't always appear. ### OS Windows ### GPU Other ### CPU AMD ### Ollama version 0.3.14.0
GiteaMirror added the bug label 2026-05-04 07:59:30 -05:00
GiteaMirror commented 2026-05-04 07:59:37 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 25, 2024):

This is LaTeX markup. If you put this in a latex renderer, you get a nicely formatted equation.

<!-- gh-comment-id:2437658439 --> @rick-github commented on GitHub (Oct 25, 2024): This is [LaTeX](https://en.wikipedia.org/wiki/LaTeX) markup. If you put this in a [latex renderer](https://arachnoid.com/latex/?equ=%5B%20(a%20%5Ccdot%20b)%20%5Cequiv%201%20%5Cpmod%7Bn%7D%20%5D), you get a nicely formatted equation.
GiteaMirror commented 2026-05-04 07:59:40 -05:00
Author
Owner
Copy Link

@lsalamon commented on GitHub (Oct 25, 2024):

This is Latex markup. If you put this in a latex renderer, you get a nicely formatted equation.

Your tip is useful, but I need the text output in the console to be without formatting characters.

<!-- gh-comment-id:2437673103 --> @lsalamon commented on GitHub (Oct 25, 2024): > This is [Latex](https://en.wikipedia.org/wiki/LaTeX) markup. If you put this in a [latex renderer](https://arachnoid.com/latex/?equ=%5B%20(a%20%5Ccdot%20b)%20%5Cequiv%201%20%5Cpmod%7Bn%7D%20%5D), you get a nicely formatted equation. Your tip is useful, but I need the text output in the console to be without formatting characters.
GiteaMirror commented 2026-05-04 07:59:43 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 25, 2024):

This is how the model was trained. You can try telling it to not use latex in the prompt that you give it, or in the system message. Which model are you using?

<!-- gh-comment-id:2437679526 --> @rick-github commented on GitHub (Oct 25, 2024): This is how the model was trained. You can try telling it to not use latex in the prompt that you give it, or in the system message. Which model are you using?
GiteaMirror commented 2026-05-04 07:59:45 -05:00
Author
Owner
Copy Link

@lsalamon commented on GitHub (Oct 25, 2024):

This is how the model was trained. You can try telling it to not use latex in the prompt that you give it, or in the system message. Which model are you using?

Today is qwen2-math:72b.
Just yesterday I got the answers correctly!

<!-- gh-comment-id:2437683483 --> @lsalamon commented on GitHub (Oct 25, 2024): > This is how the model was trained. You can try telling it to not use latex in the prompt that you give it, or in the system message. Which model are you using? Today is qwen2-math:72b. Just yesterday I got the answers correctly!
GiteaMirror commented 2026-05-04 07:59:47 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 25, 2024):

Models answer on a probabilistic basis, unless you set the seed and temperature you will always get variations in the answers.

<!-- gh-comment-id:2437699438 --> @rick-github commented on GitHub (Oct 25, 2024): Models answer on a probabilistic basis, unless you set the `seed` and `temperature` you will always get variations in the answers.
GiteaMirror commented 2026-05-04 07:59:51 -05:00
Author
Owner
Copy Link

@lsalamon commented on GitHub (Oct 25, 2024):

Models answer on a probabilistic basis, unless you set the seed and temperature you will always get variations in the answers.

Ok. How I set this on my case?

<!-- gh-comment-id:2437724128 --> @lsalamon commented on GitHub (Oct 25, 2024): > Models answer on a probabilistic basis, unless you set the `seed` and `temperature` you will always get variations in the answers. Ok. How I set this on my case?
GiteaMirror commented 2026-05-04 07:59:53 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 25, 2024):

/set parameter seed 1
/set parameter temperature 0

This will reduce the variability in the responses but it won't make the model stop using latex. If the model was trained with data in latex format, that's how it's going to respond.

<!-- gh-comment-id:2437751675 --> @rick-github commented on GitHub (Oct 25, 2024): ``` /set parameter seed 1 /set parameter temperature 0 ``` This will reduce the variability in the responses but it won't make the model stop using latex. If the model was trained with data in latex format, that's how it's going to respond.
GiteaMirror commented 2026-05-04 07:59:56 -05:00
Author
Owner
Copy Link

@lsalamon commented on GitHub (Oct 25, 2024):

/set parameter seed 1
/set parameter temperature 0

This will reduce the variability in the responses but it won't make the model stop using latex. If the model was trained with data in latex format, that's how it's going to respond.

What could cause responses to not always contain Latex formatting characters?

<!-- gh-comment-id:2438234539 --> @lsalamon commented on GitHub (Oct 25, 2024): > ``` > /set parameter seed 1 > /set parameter temperature 0 > ``` > > This will reduce the variability in the responses but it won't make the model stop using latex. If the model was trained with data in latex format, that's how it's going to respond. What could cause responses to not always contain Latex formatting characters?
GiteaMirror commented 2026-05-04 07:59:58 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 25, 2024):

Models answer on a probabilistic basis. Some of the training data was in latex format, some wasn't. Sometimes the answer will be based on the latex data, sometimes not.

<!-- gh-comment-id:2438272594 --> @rick-github commented on GitHub (Oct 25, 2024): Models answer on a probabilistic basis. Some of the training data was in latex format, some wasn't. Sometimes the answer will be based on the latex data, sometimes not.
GiteaMirror commented 2026-05-04 08:00:00 -05:00
Author
Owner
Copy Link

@lsalamon commented on GitHub (Oct 26, 2024):

/set parameter seed 1
/set parameter temperature 0

This will reduce the variability in the responses but it won't make the model stop using latex. If the model was trained with data in latex format, that's how it's going to respond.

What could cause responses to not always contain Latex formatting characters?

Setting the parameters stabilized the responses without markers.

<!-- gh-comment-id:2439766115 --> @lsalamon commented on GitHub (Oct 26, 2024): > > ``` > > /set parameter seed 1 > > /set parameter temperature 0 > > ``` > > > > > > > > > > > > > > > > > > > > > > > > This will reduce the variability in the responses but it won't make the model stop using latex. If the model was trained with data in latex format, that's how it's going to respond. > > What could cause responses to not always contain Latex formatting characters? Setting the parameters stabilized the responses without markers.
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#66732
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?