Were there any other techniques or dataset used outside commonvoice?
#1
by
munasco
- opened
Curious how you achieved 25 WER
Yes there were. Refer to my write up in the link below. Also all the code is on GitHub. The GitHub link is also in the write up.
Hello, thanks for your research and open source. I am working on finetuning multiple languages dataset too. I wanna to know how you define the language code?
All details can be found here:
Advancing Multilingual Speech Recognition: Fine-Tuning Whisper for Enhanced Low-Resource…