Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 14:13:35 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #5712] [POSSIBLE BUG]: embedding brutally slow/unusable?? #5248

New Issue
Closed
opened 2026-06-05 14:52:54 -04:00 by yindo · 1 comment
yindo commented 2026-06-05 14:52:54 -04:00
Owner
Copy Link
Copy Source

Originally created by @ohromeo on GitHub (May 27, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5712

How are you running AnythingLLM?

Docker (local)

What happened?

has something changed in 1.13 with embedding. its taking a massive amount of time to embed a document via the api - at times forever to never.

the docs are small 1-2 pages test pdf's, this was all fine in 1.121, albeit i don't have a real server its a lenovo desktop running vmware 32gb ram, vm's are tumbleweed linux, i7, but in 1.12.1 this particular function to review a document at most took 45 sectonds to 1 minute if that, as noted now the function is unsable.

just curious if this is correct behaviour now

Are there known steps to reproduce?

No response

LLM Provider & Model (if applicable)

deekseek, deekseek-chat

Embedder Provider & Model (if applicable)

aLLM default
ollama - momic-embed-text:lastest

Originally created by @ohromeo on GitHub (May 27, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5712 ### How are you running AnythingLLM? Docker (local) ### What happened? has something changed in 1.13 with embedding. its taking a massive amount of time to embed a document via the api - at times forever to never. the docs are small 1-2 pages test pdf's, this was all fine in 1.121, albeit i don't have a real server its a lenovo desktop running vmware 32gb ram, vm's are tumbleweed linux, i7, but in 1.12.1 this particular function to review a document at most took 45 sectonds to 1 minute if that, as noted now the function is unsable. just curious if this is correct behaviour now ### Are there known steps to reproduce? _No response_ ### LLM Provider & Model (if applicable) deekseek, deekseek-chat ### Embedder Provider & Model (if applicable) aLLM default ollama - momic-embed-text:lastest
yindo added the possible bugneeds info / can't replicate labels 2026-06-05 14:52:55 -04:00
yindo closed this issue 2026-06-05 14:52:55 -04:00
yindo commented 2026-06-05 14:52:55 -04:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (May 27, 2026):

This sounds like the virtualization layer is killing the performance here? However since it is happening with only 1.13.0 that is worth looking into. Since the documents are small this should still be very quick. The only thing changed was there is now a queue for embeddings to actually prevent conflicts when embedding a ton of documents.

if you tried to embed like 20 documents it would actually do 20 parallel calls and could easily overwhelm some lower-spec servers. All we did was add a FIFO queue that streams back progress now so we can handle larger workloads - this keeps the embedder model hot loaded as well so that its not constantly loading/reloading.

Does reverting to 1.12.1 go back to faster speeds?

<!-- gh-comment-id:4556092249 --> @timothycarambat commented on GitHub (May 27, 2026): This sounds like the virtualization layer is killing the performance here? However since it is happening with only 1.13.0 that is worth looking into. Since the documents are small this should still be very quick. The only thing changed was there is now a queue for embeddings to actually prevent conflicts when embedding a ton of documents. if you tried to embed like 20 documents it would actually do 20 parallel calls and could easily overwhelm some lower-spec servers. All we did was add a FIFO queue that streams back progress now so we can handle larger workloads - this keeps the embedder model hot loaded as well so that its not constantly loading/reloading. Does reverting to 1.12.1 go back to faster speeds?
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Desktop
Docker
Integration Request
Integration Request
OS: Linux
OS: Mobile
OS: Windows
UI/UX
blocked
bug
bug
core-team-only
documentation
duplicate
embed-widget
enhancement
feature request
github_actions
good first issue
investigating
needs info / can't replicate
possible bug
pull-request
Mirrored from GitHub Pull Request
 question
stage: specifications
wontfix
No Label needs info / can't replicate possible bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#5248
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?