github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 13:19:40 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #2678] Understanding Modelfile template with respect to conversational history #79272

New Issue
Closed
opened 2026-05-09 04:39:38 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-05-09 04:39:38 -05:00
Owner
Copy Link

Originally created by @nikhil0360 on GitHub (Feb 22, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2678

[proposed Label] question

Hello, I want to understand how does the conversational history is feed back into the template from the model file.

for example, for llama2:chat

TEMPLATE """[INST] <<SYS>>{{ .System }}<</SYS>>

{{ .Prompt }} [/INST]
"""

I am able to do conversational question answering on the terminal, but I am not sure how does the template take care of the history.

On the contrary, how do I disable such behaviour? I want to run a fine-tune model which just answer to user based on the current question. I don't want it to get influenced by history/previous question. I tried using /set nohistory but it doesn't seem to work (for llama2)

Please help.

I tried looking in docs, online forms, github issues. Was unable to find any solutions.

Originally created by @nikhil0360 on GitHub (Feb 22, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2678 [proposed Label] question Hello, I want to understand how does the conversational history is feed back into the template from the model file. for example, for llama2:chat ``` TEMPLATE """[INST] <<SYS>>{{ .System }}<</SYS>> {{ .Prompt }} [/INST] """ ``` I am able to do conversational question answering on the terminal, but I am not sure how does the template take care of the history. On the contrary, how do I disable such behaviour? I want to run a fine-tune model which just answer to user based on the current question. I don't want it to get influenced by history/previous question. I tried using `/set nohistory` but it doesn't seem to work (for llama2) Please help. I tried looking in docs, online forms, github issues. Was unable to find any solutions.
GiteaMirror added the question label 2026-05-09 04:39:38 -05:00
GiteaMirror commented 2026-05-09 04:39:43 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Mar 12, 2024):

Hey @nikhil0360 There's no way to disable this in the repl, although if you want to clear out the current context you can just use /load <model>. If you load the same model again it will remain in memory (i.e. it won't be "re-loaded") and it will make the model forget anything from before.

<!-- gh-comment-id:1992713720 --> @pdevine commented on GitHub (Mar 12, 2024): Hey @nikhil0360 There's no way to disable this in the repl, although if you want to clear out the current context you can just use `/load <model>`. If you load the same model again it will remain in memory (i.e. it won't be "re-loaded") and it will make the model forget anything from before.
GiteaMirror commented 2026-05-09 04:39:44 -05:00
Author
Owner
Copy Link

@nikhil0360 commented on GitHub (Mar 17, 2024):

@pdevine Thanks for the answer.

Just for more details for anyone who comes across here, I want to share that this behaviour (influenced by history/previous question.) is not followed in the API call, which is what I finally used for my work.

<!-- gh-comment-id:2002555196 --> @nikhil0360 commented on GitHub (Mar 17, 2024): @pdevine Thanks for the answer. Just for more details for anyone who comes across here, I want to share that this behaviour (influenced by history/previous question.) is not followed in the API call, which is what I finally used for my work.
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 question
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#79272
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?