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 #10194] Add Human Feedback #68745

New Issue
Closed
opened 2026-05-04 15:03:35 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-04 15:03:35 -05:00
Owner
Copy Link

Originally created by @KevinKrueger on GitHub (Apr 9, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10194

Hello,

the idea is that you say, okay, I liked the answer or not.
possibly with a reason (optional) and you have the possibility to improve your model further.

This should then be possible on the part of the API.

I know of course that human feedback can be much more than just “yes, it was good or not” and a reason why (if that makes any sense at all) but I am a friend of approaching the topic with small steps.

What do you think about this?

Originally created by @KevinKrueger on GitHub (Apr 9, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10194 Hello, the idea is that you say, okay, I liked the answer or not. possibly with a reason (optional) and you have the possibility to improve your model further. This should then be possible on the part of the API. I know of course that human feedback can be much more than just “yes, it was good or not” and a reason why (if that makes any sense at all) but I am a friend of approaching the topic with small steps. What do you think about this?
GiteaMirror added the feature request label 2026-05-04 15:03:35 -05:00
GiteaMirror commented 2026-05-04 15:03:40 -05:00
Author
Owner
Copy Link

@KevinKrueger commented on GitHub (Apr 9, 2025):

The Technical idea of the request:
{
"Output": "The model output text",
"Evaluation": 0 // 0 = bad, 1 = good etc...
"Reason": "The Reason why is this bad or good etc",
"Solution" "The Solution to make the model better"
}

<!-- gh-comment-id:2788691273 --> @KevinKrueger commented on GitHub (Apr 9, 2025): The Technical idea of the request: { "Output": "The model output text", "Evaluation": 0 // 0 = bad, 1 = good etc... "Reason": "The Reason why is this bad or good etc", "Solution" "The Solution to make the model better" }
GiteaMirror commented 2026-05-04 15:03:41 -05:00
Author
Owner
Copy Link

@apunkt commented on GitHub (Apr 9, 2025):

Ollama is a backend system.
What you want to add is a feature of a frontend!
It is implemented in OpenWebUI for example, but this is just collecting the HumanFeedback (HF).
In order to reflect this in the models, you need to finetune the model yourself with the HF and ReinforcementLearning (RL), with a technique called Reinforcement Learning from Human Feedback (RLHF).

For this you need a) massive amounts of HF (>thousands) and b) massive amount of hardware as this is much more difficult to do than just the simple inference that Ollama does.

<!-- gh-comment-id:2790022534 --> @apunkt commented on GitHub (Apr 9, 2025): Ollama is a backend system. What you want to add is a feature of a frontend! It is implemented in OpenWebUI for example, but this is just collecting the HumanFeedback (HF). In order to reflect this in the models, you need to finetune the model yourself with the HF and ReinforcementLearning (RL), with a technique called Reinforcement Learning from Human Feedback (RLHF). For this you need a) massive amounts of HF (>thousands) and b) massive amount of hardware as this is much more difficult to do than just the simple inference that Ollama does.
GiteaMirror commented 2026-05-04 15:03:42 -05:00
Author
Owner
Copy Link

@KevinKrueger commented on GitHub (Apr 9, 2025):

Okay, yes, I understand. Let me put it this way, if it were easy I would do it myself.

But that would be something if Ollama could do both ;)

<!-- gh-comment-id:2790816588 --> @KevinKrueger commented on GitHub (Apr 9, 2025): Okay, yes, I understand. Let me put it this way, if it were easy I would do it myself. But that would be something if Ollama could do both ;)
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 feature request
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#68745
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?