238 25 66

Yoach Lacombe

ylacombe

AI & ML interests

None yet

Recent Activity

upvoted a collection about 12 hours ago

Speakers

upvoted a collection 6 days ago

Hibiki fr-en

updated a collection 18 days ago

Audio Work Highlights

View all activity

Organizations

ylacombe's activity

upvoted a collection about 12 hours ago

Speakers

Collection

A set of models dedicated to voice descriptions, with special focus on speaker traits (timbral and non-timbral characteristics, etc) • 3 items • Updated 1 day ago • 4

upvoted a collection 6 days ago

Hibiki fr-en

Collection

Hibiki is a model for streaming speech translation , which can run on device! See https://github.com/kyutai-labs/hibiki. • 5 items • Updated 5 days ago • 45

updated a collection 18 days ago

Audio Work Highlights

Collection

An informative compilation of my work at 🤗 Hugging Face as a ML Engineer • 13 items • Updated 18 days ago • 3

commented on Fine-Tune W2V2-Bert for low-resource ASR with 🤗 Transformers 18 days ago

Hey @nicccobb ,
I see that you feel the post and code are lacking.
Could you specify which parts seem incorrectly coded? I’m willing to revise and give more depth where needed. Also, I’ve already discovered one clear mistake: my calculations for the time span each CTC token covers (ms per token) were off.
I’d love to hear any other points you think should be fixed or expanded upon. Constructive feedback is always appreciated —I’m open to correcting and learning.