Asking support for distil-large-v3.5

[MaBoCoMark]

I accidentally find out that there is a distilled large v3.5 model on Hugging Face and it is showing that it is 1.5 times faster than large V3 turbo.

I would like to try out some faster models, Because sometimes if I set too much lot of things in a row, it will likely look as stuck when using V3 turbo (~10s waiting or more).

The tiny model do significantly fast, but the accuracy is even worse than Apple dictation.

I've been seeing a comparison graph between different models about speed versus accuracy. So I'm highly wanted to try out the distilled large V3.5 model.

[SeoFood]

Thanks for raising this. We added the available WhisperKit CoreML turbo distil variant in #539: distil-whisper_distil-large-v3_turbo now shows up as “Distil Large v3 Turbo” in the WhisperKit model list.

distil-large-v3.5 itself is not available in argmaxinc/whisperkit-coreml yet, so TypeWhisper cannot load it through the current CoreML download path until those artifacts exist.