build(whisper): pin model source and revision, pin Python dependencies - #21
Merged
Merged
Conversation
Download large-v3-turbo from the repository that actually serves it (dropbox-dash/faster-whisper-large-v3-turbo, the target of the faster-whisper alias redirect) at a fixed commit and verify model.bin against its SHA-256. Pin all Python dependencies in requirements.txt, the base image by digest, and keep Renovate away from av. Closes #18
eksrha
force-pushed
the
fix/whisper-pin-model-source
branch
from
October 4, 2026 20:11
0165ce1 to
a6d30c3
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #18
What
snapshot_download(repo_id=dropbox-dash/faster-whisper-large-v3-turbo, revision=0a363e91...)instead of faster-whisper'sdownload_model('large-v3-turbo'), which resolves tomobiuslabsgmbh/faster-whisper-large-v3-turboand is redirected (HTTP 307) to the dropbox-dash repository.Systran/faster-whisper-large-v3-turboanswers HTTP 401.model.binis verified against its SHA-256, so a changed source fails the build. Source, revision, hash and the update procedure are documented in the Containerfile and README. Model license: MIT.deploy/whisper/requirements.txtpins faster-whisper, av (16.x required, with reason), ctranslate2, huggingface_hub, numpy, onnxruntime, tokenizers, fastapi, uvicorn, python-multipart. Base image pinned by digest. Renovate ignoresav.ARG WHISPER_MODEL=large-v3-turbostays untouched (the release workflow reads it).Verified locally
Built with docker, started read-only with all capabilities dropped,
/healthok, German test clip transcribed correctly, pinned versions confirmed in the image (av 16.1.0, faster-whisper 1.2.1). Image size 2.24 GB.