github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-15 12:48:05 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #3115] Switching between models is very slow? #1916

New Issue
Closed
opened 2026-04-12 12:02:43 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 12:02:43 -05:00
Owner
Copy Link

Originally created by @eliranwong on GitHub (Mar 13, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3115

I noted that switching between models is very slow. Is it possible to load two different models in memory at the same time?

Originally created by @eliranwong on GitHub (Mar 13, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3115 I noted that switching between models is very slow. Is it possible to load two different models in memory at the same time?
GiteaMirror added the question label 2026-04-12 12:02:43 -05:00
GiteaMirror commented 2026-04-12 12:02:44 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Mar 13, 2024):

It's not possible yet to load two models simultaneously, but it's something we've been contemplating for a while (and there are a couple issues on this already). There are a couple reasons for the slow switching time, but loading between system memory and the GPU can be pretty slow, particularly if you're loading/unloading large models and/or if you have older hardware.

Going to close out the issue, but feel free to keep commenting.

<!-- gh-comment-id:1995168598 --> @pdevine commented on GitHub (Mar 13, 2024): It's not possible *yet* to load two models simultaneously, but it's something we've been contemplating for a while (and there are a couple issues on this already). There are a couple reasons for the slow switching time, but loading between system memory and the GPU can be pretty slow, particularly if you're loading/unloading large models and/or if you have older hardware. Going to close out the issue, but feel free to keep commenting.
GiteaMirror commented 2026-04-12 12:02:45 -05:00
Author
Owner
Copy Link

@ivanenev commented on GitHub (Apr 24, 2024):

I am running two models simultaneously but they do not run in parallel, they respond consecutively. I've found a way of running two models,mistral and llama (or two instances of llama) in parallel by changing the OLLAMA_HOST variable for the second instance. First copy the model:
ollama cp llama3 llama3-2
For the second instance,change the ip and port like so:
OLLAMA_HOST=0.0.0.0:11435 ollama serve
OLLAMA_HOST=0.0.0.0:11435 ollama run ollama

<!-- gh-comment-id:2074938437 --> @ivanenev commented on GitHub (Apr 24, 2024): I am running two models simultaneously but they do not run in parallel, they respond consecutively. I've found a way of running two models,mistral and llama (or two instances of llama) in parallel by changing the OLLAMA_HOST variable for the second instance. First copy the model: ollama cp llama3 llama3-2 For the second instance,change the ip and port like so: OLLAMA_HOST=0.0.0.0:11435 ollama serve OLLAMA_HOST=0.0.0.0:11435 ollama run ollama
GiteaMirror commented 2026-04-12 12:02:45 -05:00
Author
Owner
Copy Link

@FarooqAlaulddin commented on GitHub (Jun 12, 2025):

Running LangFlow pipeline with two models (14b + 7b) in a batch run and here is what I saw. My system is Macbook M4 Pro Chip + 24GB.

Image

<!-- gh-comment-id:2965637129 --> @FarooqAlaulddin commented on GitHub (Jun 12, 2025): Running LangFlow pipeline with two models (14b + 7b) in a batch run and here is what I saw. My system is Macbook M4 Pro Chip + 24GB. ![Image](https://github.com/user-attachments/assets/b16d3841-bb66-4aec-991f-9dad0980e20e)
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#1916
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?