Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-21 01:25:28 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #858] [BUG]: too slow of the collector to transfrom the docx file to data #507

New Issue
Closed
opened 2026-02-22 18:19:49 -05:00 by yindo · 5 comments
yindo commented 2026-02-22 18:19:49 -05:00
Owner
Copy Link
Copy Source

Originally created by @sddzcuigc on GitHub (Mar 4, 2024).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/858

How are you running AnythingLLM?

Docker (local)

What happened?

I use the local anythingLLM and ollama , I use very small chinese docx into the database,but it takes too long
and when it is done. I don't know how to ask due to the data I upload. When I ask any question It shows ollama failed to embed undefined: undefined.

Are there known steps to reproduce?

No response

Originally created by @sddzcuigc on GitHub (Mar 4, 2024). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/858 ### How are you running AnythingLLM? Docker (local) ### What happened? I use the local anythingLLM and ollama , I use very small chinese docx into the database,but it takes too long and when it is done. I don't know how to ask due to the data I upload. When I ask any question It shows ollama failed to embed undefined: undefined. ### Are there known steps to reproduce? _No response_
yindo added the possible bug label 2026-02-22 18:19:49 -05:00
yindo closed this issue 2026-02-22 18:19:49 -05:00
yindo commented 2026-02-22 18:19:50 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 6, 2024):

Taking too long would indicate a resource constraint on Docker. Since local embedding run on CPU we should first check that the docker container has enough resources to work with (including RAM- which is likely limiting you).

Also, given you are in docker you should be able to view the storage/documents folder - which will show a JSON object with a pageContent key that will show any text that was parsed. It is possible the document was not parseable.

Can you check those two details and report your findings?

@timothycarambat commented on GitHub (Mar 6, 2024): Taking too long would indicate a resource constraint on Docker. Since local embedding run on CPU we should first check that the docker container has enough resources to work with (including RAM- which is likely limiting you). Also, given you are in docker you should be able to view the `storage/documents` folder - which will show a JSON object with a `pageContent` key that will show any text that was parsed. It is possible the document was not parseable. Can you check those two details and report your findings?
yindo commented 2026-02-22 18:19:50 -05:00
Author
Owner
Copy Link
Copy Source

@wallartup commented on GitHub (Mar 6, 2024):

@timothycarambat are you working on moving the local embeddings to GPU? That would make things much much faster.

One thing that would make sense is to be able to embed files in terminal via a command like: ingest (then anything in hotdir2 gets ingested and embedded)

I am doing the above in another project of mine which is python based.

@wallartup commented on GitHub (Mar 6, 2024): @timothycarambat are you working on moving the local embeddings to GPU? That would make things much much faster. One thing that would make sense is to be able to embed files in terminal via a command like: ingest (then anything in hotdir2 gets ingested and embedded) I am doing the above in another project of mine which is python based.
yindo commented 2026-02-22 18:19:50 -05:00
Author
Owner
Copy Link
Copy Source

@sddzcuigc commented on GitHub (Mar 7, 2024):

yes，using gpu is faster,and makes less error. I use the ollama as the local engine,
when I ask questions, if the questions I force to stop the AI answer streaming,cpu/gpu will not be free for a very long time, that means when I stop the question, the backgroud thread will not force to stop, and the question will not record.

发件人: wallartup @.>
发送时间: 2024年3月7日 3:14
收件人: Mintplex-Labs/anything-llm @.>
抄送: Yang xiaoyu @.>; Author @.>
主题: Re: [Mintplex-Labs/anything-llm] [BUG]: too slow of the collector to transfrom the docx file to data (Issue #858)

@timothycarambathttps://github.com/timothycarambat are you working in moving the local embeddings to GPU? That would make things much much faster.

―
Reply to this email directly, view it on GitHubhttps://github.com/Mintplex-Labs/anything-llm/issues/858#issuecomment-1981600385, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AUP6NMOXKULT2TOTBMLGJEDYW5TINAVCNFSM6AAAAABEGHWZ5WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOBRGYYDAMZYGU.
You are receiving this because you authored the thread.Message ID: @.***>

@sddzcuigc commented on GitHub (Mar 7, 2024): yes，using gpu is faster,and makes less error. I use the ollama as the local engine, when I ask questions, if the questions I force to stop the AI answer streaming,cpu/gpu will not be free for a very long time, that means when I stop the question, the backgroud thread will not force to stop, and the question will not record. ________________________________ 发件人: wallartup ***@***.***> 发送时间: 2024年3月7日 3:14 收件人: Mintplex-Labs/anything-llm ***@***.***> 抄送: Yang xiaoyu ***@***.***>; Author ***@***.***> 主题: Re: [Mintplex-Labs/anything-llm] [BUG]: too slow of the collector to transfrom the docx file to data (Issue #858) @timothycarambat<https://github.com/timothycarambat> are you working in moving the local embeddings to GPU? That would make things much much faster. ― Reply to this email directly, view it on GitHub<https://github.com/Mintplex-Labs/anything-llm/issues/858#issuecomment-1981600385>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AUP6NMOXKULT2TOTBMLGJEDYW5TINAVCNFSM6AAAAABEGHWZ5WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOBRGYYDAMZYGU>. You are receiving this because you authored the thread.Message ID: ***@***.***>
yindo commented 2026-02-22 18:19:50 -05:00
Author
Owner
Copy Link
Copy Source

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

I got the same problem, and solve it by reducing chunk size setting to 2048 ( from 8196) in anythingLLM's admin pannel.
Maybe you can have I try. BTW, I think it may helps also to splitting large text/ pdf file to smaller pieces.

@airylinus commented on GitHub (Mar 17, 2024): I got the same problem, and solve it by reducing chunk size setting to 2048 ( from 8196) in anythingLLM's admin pannel. Maybe you can have I try. BTW, I think it may helps also to splitting large text/ pdf file to smaller pieces.
yindo commented 2026-02-22 18:19:50 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 18, 2024):

This seems more like a setup/config issue on the Ollama side where putting the placeholder value (which is arbitrary) too high for the model provided, which is also memory dependent in addition to model-constrained.

Glad there was a path around. AnythingLLM will only work with the settings given since there is no way to "ask" a model how big its context window is

@timothycarambat commented on GitHub (Mar 18, 2024): This seems more like a setup/config issue on the Ollama side where putting the placeholder value (which is arbitrary) too high for the model provided, which is also memory dependent in addition to model-constrained. Glad there was a path around. AnythingLLM will only work with the settings given since there is no way to "ask" a model how big its context window is
yindo changed title from [BUG]: too slow of the collector to transfrom the docx file to data to [GH-ISSUE #858] [BUG]: too slow of the collector to transfrom the docx file to data 2026-06-05 14:35:40 -04:00
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 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#507
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?