Skip to content

Add Cohere CoreML ASR provider integration#244

Merged
altic-dev merged 22 commits intomainfrom
B/coreml-asr-provider
Apr 1, 2026
Merged

Add Cohere CoreML ASR provider integration#244
altic-dev merged 22 commits intomainfrom
B/coreml-asr-provider

Conversation

@altic-dev
Copy link
Copy Markdown
Owner

Summary

  • add the Cohere CoreML ASR provider integration and external artifact flow
  • wire Cohere through onboarding, settings, model presentation, and live preview
  • include the follow-up runtime and packaging fixes needed for validation and archive stability

Notes

@altic-dev
Copy link
Copy Markdown
Owner Author

@codex can you fix it and run lint and formatting

@chatgpt-codex-connector
Copy link
Copy Markdown

To use Codex here, create an environment for this repo.

Copy link
Copy Markdown
Owner Author

Updated summary note for reviewers:

  • This branch includes split compute configuration for Cohere CoreML loading, using Neural Engine and GPU across different stages.
  • It also includes async overlap/prefetch in the Cohere chunked decode path, which can improve throughput for multi-chunk transcriptions. Actual speedup may vary by device and audio length.

The PR body should treat this as a qualified performance improvement, not a blanket speed claim for every machine or every short utterance.

@altic-dev altic-dev marked this pull request as ready for review April 1, 2026 06:42
@altic-dev altic-dev merged commit a6b6c5f into main Apr 1, 2026
2 checks passed
Copy link
Copy Markdown

@chatgpt-codex-connector chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e104859624

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

revision: spec.repositoryRevision,
requiredItems: spec.requiredEntries.map { .init(path: $0, isDirectory: $0.hasSuffix(".mlpackage")) }
)
try await downloader.ensureModelsPresent(at: directory) { progress, item in
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Purge invalid artifacts before retrying external model download

If artifact validation fails, this path retries with ensureModelsPresent, but that downloader only fills missing paths and does not overwrite existing files. In the common failure case where all required files exist but are incompatible/corrupted (for example a stale coreml_manifest.json or interrupted prior download), the subsequent validateArtifactsOrThrow will fail on every retry, so users cannot recover from the app and must manually delete cache files. Clearing/replacing managed artifacts before retrying would make download recovery reliable.

Useful? React with 👍 / 👎.

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant