CT2 format?
Can you please convert this model to CT2 format (model.bin) ?
I would like to use it with https://github.com/rhasspy/wyoming-faster-whisper HomeAssistant add-on...
Hey! I am preparing a better dataset for fine tuning and then there will be a second generation model. Then I’ll try to convert it to different formats.
Great! Looking forward.
Hey! I am preparing a better dataset for fine tuning and then there will be a second generation model. Then I’ll try to convert it to different formats.
Hi, do you have any shedule, when this update can be released? Really looking forward for an updated version. Thanks!
It's not finished yet because I'm still working on the dataset. However, the current version is uploaded with tag 'v2.0alpha1'
Please notice that I don't recommend using this v2.0alpha1 release version in production.
For anyone interested in CT2 format of this model:
- float16 ver: https://huggingface.co/bzikst/faster-whisper-large-v3-russian
- int8(bfloat16) ver: https://huggingface.co/bzikst/faster-whisper-large-v3-russian-int8
Int8 version is 1,6Gb and provides for me the best quality/speed ratio for faster whisper speech recognition of Russian language!