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 #3081] [UI/UX]: Continue generation button for limited context output #1973

New Issue
Closed
opened 2026-02-22 18:27:29 -05:00 by yindo · 2 comments
yindo commented 2026-02-22 18:27:29 -05:00
Owner
Copy Link
Copy Source

Originally created by @clover1980 on GitHub (Jan 31, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3081

How are you running AnythingLLM?

Local development

What happened?

We all know that Deepseek R1 makes a reasoning portion of chat, which in the AnythingLLM abruptly ending unfinished text generation and there's no method in application to continue generation.
Using spacebar in request STARTS NEW GENERATION and previous text context is lost completely.
Oobabooga had Continue button for years, but they added a new recently on visible place specifically for Deepseek. This is clearly a bug of functionality.

Are there known steps to reproduce?

starting Deepseek R1 in any quant in 4096 or 8192 token context sizes.

Originally created by @clover1980 on GitHub (Jan 31, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3081 ### How are you running AnythingLLM? Local development ### What happened? We all know that Deepseek R1 makes a reasoning portion of chat, which in the AnythingLLM abruptly ending unfinished text generation and there's no method in application to continue generation. Using spacebar in request STARTS NEW GENERATION and previous text context is lost completely. Oobabooga had Continue button for years, but they added a new recently on visible place specifically for Deepseek. This is clearly a bug of functionality. ### Are there known steps to reproduce? starting Deepseek R1 in any quant in 4096 or 8192 token context sizes.
yindo added the enhancementUI/UX labels 2026-02-22 18:27:29 -05:00
yindo closed this issue 2026-02-22 18:27:29 -05:00
yindo commented 2026-02-22 18:27:30 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Feb 1, 2025):

This would mean that your input is either so long that it cuts the output or that the output is so long that the model has to stop generating tokens.

You are not specific about what provider you are using either - which would matter here. A Continue button would simply re-prompt the context from that singular message until it runs out of generation token space.

@timothycarambat commented on GitHub (Feb 1, 2025): This would mean that your input is either so long that it cuts the output or that the output is so long that the model has to stop generating tokens. You are not specific about what provider you are using either - which would matter here. A `Continue` button would simply re-prompt the context from that singular message until it runs out of generation token space.
yindo commented 2026-02-22 18:27:30 -05:00
Author
Owner
Copy Link
Copy Source

@clover1980 commented on GitHub (Feb 2, 2025):

This would mean that your input is either so long that it cuts the output or that the output is so long that the model has to stop generating tokens.

You are not specific about what provider you are using either - which would matter here. A Continue button would simply re-prompt the context from that singular message until it runs out of generation token space.

Hello.
I'm using oobabooga API because Ollama still can't work with model files in parts.
Usually when i click "continue" in other llm launchers it's literally continued, the continuation is coherent and correct (esp in code writing), i've never seen fully new generation by that (maybe it's possible on very low quants with high hallucination).
I need continue function to use full Deepseek R1 Q6 from other computer by local API in ComfyUI on current PC (you can't fit large diffusion models & big deepseek on one PC, no resources for that). For now the stage of connecting to local API works.

@clover1980 commented on GitHub (Feb 2, 2025): > This would mean that your input is either so long that it cuts the output or that the output is so long that the model has to stop generating tokens. > > You are not specific about what provider you are using either - which would matter here. A `Continue` button would simply re-prompt the context from that singular message until it runs out of generation token space. Hello. I'm using oobabooga API because Ollama still can't work with model files in parts. Usually when i click "continue" in other llm launchers it's literally continued, the continuation is coherent and correct (esp in code writing), i've never seen fully new generation by that (maybe it's possible on very low quants with high hallucination). I need continue function to use full Deepseek R1 Q6 from other computer by local API in ComfyUI on current PC (you can't fit large diffusion models & big deepseek on one PC, no resources for that). For now the stage of connecting to local API works.
yindo changed title from [UI/UX]: Continue generation button for limited context output to [GH-ISSUE #3081] [UI/UX]: Continue generation button for limited context output 2026-06-05 14:43:46 -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 UI/UX enhancement
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#1973
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?