This is pretty cool! Just yesterday I finished a similar demo, also using transformers.js. I am currently in the process of adding real-time transcription, do you plan on adding that?
Thanks! I actually initially had it doing transcription in 5- and 10-second chunks for close-to-realtime results, but the CPU usage on my laptop (which admittedly doesn't have the best specs) was a bit higher than i wanted. 30-second blocks gave me the best balance of semi-real-time and good performance, especially since the whisper model is built for 30-second chunks. If you get real-time working smoothly though, i'd love to take a look!
Comments
This is pretty cool! Just yesterday I finished a similar demo, also using transformers.js. I am currently in the process of adding real-time transcription, do you plan on adding that?
Thanks! I actually initially had it doing transcription in 5- and 10-second chunks for close-to-realtime results, but the CPU usage on my laptop (which admittedly doesn't have the best specs) was a bit higher than i wanted. 30-second blocks gave me the best balance of semi-real-time and good performance, especially since the whisper model is built for 30-second chunks. If you get real-time working smoothly though, i'd love to take a look!