Parakeet v3 often tries to transcribe in English instead of Portuguese
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 231
- Forks
- 27
- Avg merge
- 4h 24m
- Merged PRs (30d)
- 24
Description
First of all, thanks for making this awesome open source project, it's much appreciated.
I want to use it mostly to transcribe Portuguese, so I'm using the "Portuguese - Parakeet v3" model. However, it often tries to transcribe my speech to English, which fails badly. I'm not sure if it's completely a problem with the model or if something could be improved in how Hex prompts the model.
I understand how difficult it is to test languages other than your native one, but I'm more than happy to help test and troubleshoot this in any way I can.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the issue with the "Portuguese - Parakeet v3" model and Portuguese speech, then investigate whether the behavior comes from the model or how Hex prompts it. Done means Portuguese speech is reliably transcribed as Portuguese, with the cause and any needed change supported by repeatable testing.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100