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 #8778] Feature: Parallel processing of embeddings? #5700

New Issue
Closed
opened 2026-04-12 16:59:38 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-12 16:59:38 -05:00
Owner
Copy Link

Originally created by @AncientMystic on GitHub (Feb 2, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8778

When utilizing embedding models for documents in UIs for ollama, they are pretty slow and i am seeing, 1.2GB vram usage from the model and 20% cpu and 15% gpu total usage on the whole system.

Would it be possible to somehow allow parallel processing or requests for documents?

Maybe at least 2-4+ at once instead of one, even if we cannot increase the speed of processing, maybe we could at least process a few at once? (Even if that means loading the model more than once to handle each instance)

(I also have ollama_num_parallel set and max queue, max loaded models, etc but it doesn't seem to work the same as it did in older versions, even multiple small models wont load together it seems but thats another issue)

Originally created by @AncientMystic on GitHub (Feb 2, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8778 When utilizing embedding models for documents in UIs for ollama, they are pretty slow and i am seeing, 1.2GB vram usage from the model and 20% cpu and 15% gpu total usage on the whole system. Would it be possible to somehow allow parallel processing or requests for documents? Maybe at least 2-4+ at once instead of one, even if we cannot increase the speed of processing, maybe we could at least process a few at once? (Even if that means loading the model more than once to handle each instance) (I also have ollama_num_parallel set and max queue, max loaded models, etc but it doesn't seem to work the same as it did in older versions, even multiple small models wont load together it seems but thats another issue)
GiteaMirror added the feature request label 2026-04-12 16:59:38 -05:00
GiteaMirror commented 2026-04-12 16:59:39 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 3, 2025):

ollama currently doesn't allow parallel completions for embedding models, which is why OLLAMA_NUM_PARALLEL has no effect. Loading a model multiple times is not supported but there is a ticket open for this issue. The only way to accomplish what you want at the moment is to run multiple servers and offer a uniform interface with a reverse proxy (eg litellm, ollama_proxy, nginx).

<!-- gh-comment-id:2630949027 --> @rick-github commented on GitHub (Feb 3, 2025): ollama currently [doesn't allow](https://github.com/ollama/ollama/blob/ad22ace439eb3fab7230134e56bb6276a78347e4/server/sched.go#L196) parallel completions for embedding models, which is why `OLLAMA_NUM_PARALLEL` has no effect. Loading a model multiple times is not supported but there is a [ticket](https://github.com/ollama/ollama/issues/3902) open for this issue. The only way to accomplish what you want at the moment is to run multiple servers and offer a uniform interface with a reverse proxy (eg [litellm](https://github.com/BerriAI/litellm), [ollama_proxy](https://github.com/ParisNeo/ollama_proxy_server), [nginx](https://github.com/ollama/ollama/issues/8186#issuecomment-2560443545)).
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#5700
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?