mudler / mudler/vllm.cpp

MODEL-MM-GLM53-FLASH: establish a runnable pinned GLM5Next oracle

Open
#3,047 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C++
Stars
423
Forks
53
Avg merge
20h 26m
Merged PRs (30d)
310

Description

Row: MODEL-MM-GLM53-FLASH

Own the pending real-model oracle build and identical-workload run under .agents/specs/glm5-next-upstream-reconciliation.md and the Flash campaign spec. Merged vLLM source 98ed0856f31fa3aaf5e27464e2b4ef5a8ee6b2f5 registers GLM5Next, but registration is not model gateability. Preserve global parity pin e126687a9a828d513c01a07cd69f025f27d63280 until its separate reconciliation obligations are satisfied.

The inherited #1998 reference currently returns GraphQL and REST 404 while #2410 and #3019 are readable. The campaign already recorded the same uncertainty under O32. This issue provides a verified current owner; it does not assert deletion or rewrite historical citations.

Acceptance: record exact source and dependency pins, authorized checkpoint identity, leased environment, successful oracle build and real checkpoint execution, then identical-workload correctness evidence. Until those measurements exist, gateability stays PENDING. Any checkpoint or resource authorization not already recorded remains a gate, not an assumption.

Spec: .agents/specs/glm5-next-upstream-reconciliation.md
Owning row: MODEL-MM-GLM53-FLASH. Documentation binding repair is tracked by #3045; this issue stays open for the actual oracle evidence.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Read .agents/specs/glm5-next-upstream-reconciliation.md and the Flash campaign spec first, then review the recorded vLLM source and parity pins and the O32 uncertainty. Confirm the authorized checkpoint and leased environment before attempting the oracle build and real checkpoint run. Done means the exact pins, authorization, environment, successful execution, and identical-workload correctness evidence are recorded so gateability can move beyond PENDING.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
ai, testing-qa
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.