Skip to content

[UPDATE][llamacppqwen3827bggufv3][1.0.1] llm-init v1.3.10, reasoning env, llama.cpp b10548 - #3452

Closed
harveyff wants to merge 1 commit into
beclab:mainfrom
harveyff:update-llamacppqwen3827bggufv3-1.0.5
Closed

[UPDATE][llamacppqwen3827bggufv3][1.0.1] llm-init v1.3.10, reasoning env, llama.cpp b10548#3452
harveyff wants to merge 1 commit into
beclab:mainfrom
harveyff:update-llamacppqwen3827bggufv3-1.0.5

Conversation

@harveyff

@harveyff harveyff commented Aug 21, 2026

Copy link
Copy Markdown
Member

Summary

  • Bump chart from beclab main 1.0.0 to 1.0.1 for test market (market.test)
  • Upgrade llm-init to v1.3.10
  • Add non-editable env vars MODEL_REASONING_EFFORT=low,medium,xhigh and MODEL_REASONING_EFFORT_DEFAULT=xhigh
  • Upgrade llama.cpp engine image to beclab/ggml-org-llama.cpp:server-cuda12-b10548
  • Lower NVIDIA host-memory reservation (engine request 21Gi to 6Gi, requiredMemory 23Gi to 7Gi)

Test plan

  • GitBot checks pass
  • Install from market.test on olarestest001
  • Model download completes and Model Console reaches ready
  • Router lists reasoning_effort options on /v1/models

@Terminus-Gitbot

Copy link
Copy Markdown
Collaborator

After the modification is completed, please click ready for review to submit the PR

…env, llama.cpp b10548

Bump from beclab main 1.0.0: llm-init 1.3.10, non-editable reasoning effort env vars, llama.cpp server-cuda12-b10548, and lower NVIDIA host-memory reservation for 1.12.6 nodes.

Co-authored-by: Cursor <cursoragent@cursor.com>
@harveyff harveyff changed the title [UPDATE][llamacppqwen3827bggufv3][1.0.5] llm-init v1.3.10, reasoning env, llama.cpp b10548 [UPDATE][llamacppqwen3827bggufv3][1.0.1] llm-init v1.3.10, reasoning env, llama.cpp b10548 Aug 21, 2026
@harveyff
harveyff force-pushed the update-llamacppqwen3827bggufv3-1.0.5 branch from c530545 to 0bb810f Compare August 21, 2026 13:46
@Terminus-Gitbot

Copy link
Copy Markdown
Collaborator

After the modification is completed, please click ready for review to submit the PR

1 similar comment
@Terminus-Gitbot

Copy link
Copy Markdown
Collaborator

After the modification is completed, please click ready for review to submit the PR

@harveyff

Copy link
Copy Markdown
Member Author

Closing: wrong base version (1.0.1 vs market.test 1.0.4). Resubmitting as 1.0.5 for test market.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants