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 #10941] Models don't have memory. #7204

New Issue
Closed
opened 2026-04-12 19:12:05 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 19:12:05 -05:00
Owner
Copy Link

Originally created by @o1243535241 on GitHub (Jun 1, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10941

I want trein models then close ollama, then open ollama, and models must remember all. Now they don't remember what about I wrote before.

Can you fix this, please?

Originally created by @o1243535241 on GitHub (Jun 1, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10941 I want trein models then close ollama, then open ollama, and models must remember all. Now they don't remember what about I wrote before. Can you fix this, please?
GiteaMirror added the bug label 2026-04-12 19:12:05 -05:00
GiteaMirror commented 2026-04-12 19:12:06 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jun 1, 2025):

Memory is not a function of the server. The client implements it, normally by recording the messages[] sent to the server. Using the ollama CLI you can save state with the /save command and re-load it later with the /load command.

<!-- gh-comment-id:2927952041 --> @rick-github commented on GitHub (Jun 1, 2025): Memory is not a function of the server. The client implements it, normally by recording the `messages[]` sent to the server. Using the ollama CLI you can save state with the [`/save`](https://github.com/ollama/ollama/blob/5c42800fca4da07d1c362c0f190429993e53c3b5/cmd/interactive.go#L38) command and re-load it later with the `/load` command.
GiteaMirror commented 2026-04-12 19:12:06 -05:00
Author
Owner
Copy Link

@GreazySpoon commented on GitHub (Jun 10, 2025):

  1. To close ollama, open it again to start a new chat session is the normal and expected behaviour.
  2. And ollama is an inference server, it is not supposed to manage the model memory.
  3. You must handle the memory on the client side, you can either concat messages and send them with each request, or give the agent a tool to navigate history or you might have an other approach, depends on the client. Remember, managing context and memory, is the responsability of the framework, or the developper in general.
  4. As mentionned above, in a chat session you can use /save /load commands.
<!-- gh-comment-id:2958688664 --> @GreazySpoon commented on GitHub (Jun 10, 2025): 1. To close ollama, open it again to start a new chat session is the normal and expected behaviour. 2. And ollama is an inference server, it is not supposed to manage the model memory. 3. You must handle the memory on the client side, you can either concat messages and send them with each request, or give the agent a tool to navigate history or you might have an other approach, depends on the client. Remember, managing context and memory, is the responsability of the framework, or the developper in general. 4. As mentionned above, in a chat session you can use /save /load commands.
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#7204
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?