Skip to content

fix(stt): fall back on CUBLAS_STATUS_NOT_SUPPORTED - #17559

Closed
liuhao1024 wants to merge 3 commits into
NousResearch:mainfrom
liuhao1024:fix/issue-17526-cublas-not-supported-stt
Closed

fix(stt): fall back on CUBLAS_STATUS_NOT_SUPPORTED#17559
liuhao1024 wants to merge 3 commits into
NousResearch:mainfrom
liuhao1024:fix/issue-17526-cublas-not-supported-stt

Conversation

@liuhao1024

@liuhao1024 liuhao1024 commented Apr 29, 2026

Copy link
Copy Markdown
Contributor

What does this PR do?

  • Treat CUBLAS_STATUS_NOT_SUPPORTED as a CUDA library/runtime error in local STT.
  • Add regression coverage for Blackwell/faster-whisper failures falling back to CPU.

Related Issue

N/A

Type of Change

  • 🐛 Bug fix (non-breaking change that fixes an issue)

Changes Made

  • See commit messages for detailed changes

How to Test

  1. Run pytest tests/ -q — all tests should pass
  2. Verify the specific scenario described above is resolved

Checklist

Code

  • I've read the Contributing Guide
  • My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
  • I searched for existing PRs to make sure this isn't a duplicate
  • My PR contains only changes related to this fix/feature (no unrelated commits)
  • I've run pytest tests/ -q and all tests pass
  • I've added tests for my changes (required for bug fixes, strongly encouraged for features)
  • I've tested on my platform: macOS 26.4.1

Documentation & Housekeeping

  • I've updated relevant documentation (README, docs/, docstrings) — or N/A
  • I've updated cli-config.yaml.example if I added/changed config keys — or N/A
  • I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture and workflows — or N/A
  • I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A

- Add Blackwell-specific cuBLAS error marker to _CUDA_LIB_ERROR_MARKERS
- Allows CPU fallback on RTX 5090 (sm_120) when faster-whisper
  reports CUBLAS_STATUS_NOT_SUPPORTED instead of loading successfully
- Add regression test for CUBLAS_STATUS_NOT_SUPPORTED path

Closes NousResearch#17526
@alt-glitch alt-glitch added type/bug Something isn't working P3 Low — cosmetic, nice to have tool/tts Text-to-speech and transcription labels Apr 29, 2026
@liuhao1024 liuhao1024 closed this May 2, 2026
@liuhao1024 liuhao1024 reopened this May 2, 2026
@teknium1

Copy link
Copy Markdown
Contributor

Thanks for the focused regression fix. Current main still leaves CUBLAS_STATUS_NOT_SUPPORTED out of _CUDA_LIB_ERROR_MARKERS (tools/transcription_tools.py:1065-1086), so the existing runtime fallback re-raises this exact Blackwell error at tools/transcription_tools.py:1141-1163 instead of retrying on CPU.

The diff adds the exact missing classifier value and tests the existing cached-model eviction and CPU-int8 retry path. The implementation is narrowly scoped, matches the existing fallback design introduced in 4350668ae, and does not add configuration or tool surface.

Automated hermes-sweeper review.

@teknium1 teknium1 added sweeper:risk-compatibility Sweeper risk: may break existing users, config, migrations, defaults, or upgrades sweeper:blast-moderate Sweeper blast radius: moderate — a subsystem or single platform labels Jul 12, 2026
@teknium1

Copy link
Copy Markdown
Contributor

Merged into main via consolidated salvage PR #73510 (merge c0c5dac531). Your CUBLAS_STATUS_NOT_SUPPORTED marker addition (Blackwell/RTX-5090 CPU fallback) was cherry-picked with your authorship and sabotage-verified by its regression test.

Your contribution is credited to you in git history. Thank you! Closing this PR as merged-via-salvage.

@teknium1 teknium1 closed this Jul 29, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

P3 Low — cosmetic, nice to have sweeper:blast-moderate Sweeper blast radius: moderate — a subsystem or single platform sweeper:risk-compatibility Sweeper risk: may break existing users, config, migrations, defaults, or upgrades tool/tts Text-to-speech and transcription type/bug Something isn't working

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants