Cohere transcribe Sota open source transcription model running in the browser :) Weights on Hugging Face link below

@cohere @huggingface Marketing for open models just hits different

@cohere @huggingface why not use the actual latency on this demo? the transcriptions appear before the audio at some points, makes no sense

@cohere @huggingface its not a steaming real time API. though its wickedly fast.

@cohere @huggingface thats fine i wouldnt expect it to be, but why not actually demo it as it is? "it can run in the browser, as it is now" sounds like you are actually demoing it

@MoritzLaurer @cohere @huggingface good demo sir

@cohere @huggingface lol, great demo! Seriously, @Apple @tim_cook get on this right now.

@cohere @huggingface love to see Cohere releasing models again!

@1vnzh @cohere @huggingface Incredible work guys, eager to try it out

@cohere @huggingface That’s the way to do demo

@cohere @huggingface Very cool congrats

@ClementDelangue @cohere @huggingface Great work Nick ! You do Canada proud!

@_akhaliq @cohere @huggingface Bro broke out the bodhran

@huggingface @cohere me when you started playing the drum thingy

@cohere @huggingface This needs to be a series where the background noise gets increasingly more unhinged

@huggingface @cohere Can it do singing?

@cohere @huggingface bodhrán deas!

@aidangomez @cohere @huggingface amazing

@cohere @huggingface Hey mate, I’ve built this free tool relying on it to dictate anywhere on your mac

@cohere @huggingface Any plans for streaming real time? How is it doing with Arabic?

@_akhaliq @cohere @huggingface canadian company btw

@badlogicgames @cohere @huggingface ooo how do we run this in the browser?

@huggingface @cohere Going to tie it into Pocket TTS subtitles and drop Whisper API. Thanks.

@_akhaliq @cohere @huggingface a fellow bodhrán player on my feed? say no more, I will try your model

@cohere @huggingface @CloudflareDev if you guys could just... thank you

@cohere @huggingface @CloudflareDev if you want this on a cloud offering its on Coheres API and on our secure single tenant model vault as well

@dasfacc @cohere @huggingface @CloudflareDev Yup here’s the reference to get started on our API: You should get insane speeds. Should you need anything, feel free to email us on [email protected] always here to help! :)

@dasfacc @cohere @huggingface @CloudflareDev For secure model vault deployments visit > Model Vault > Select Audio > Cohere Transcribe model during deployment.

@nickfrosst @cohere @huggingface @CloudflareDev idk what `secure model vault deployments` are, but much appreciated guys! I've a bit of a 'google meet for code' that runs on workers, so I use nova-3 through workers AI but it breaks down easily with messy sound inputs

@ClementDelangue @cohere @huggingface @geerlingguy

@ClementDelangue @cohere @huggingface 🎉🔥

@cohere @huggingface Hey @nickrfrosst can you post some benchmark

@cohere @huggingface For DGX Spark owners

@cohere @huggingface whats the difference between this and whisper

@ClementDelangue @cohere @huggingface Let’s fucking go

@GeoffreyHuntley @cohere @huggingface Curious if its encore or encore ESP?

@cohere @huggingface Nick thanks for sharing this 🙌🏿 I was able to make a /voice-memos skill using your model that can provide summaries on conversations recorded on one’s iPhone

@cohere @huggingface Problem case: transcription that works while making coffee

@cohere @huggingface best demo I've seen in a while!

@cohere @huggingface Awesome demo post!

@cohere @huggingface super video