Compatibility check
Can I run Lite Whisper Large v3 Acc on Apple M2 Ultra (76-core GPU)?
Yes — it runs.best quant:
Decode speed
660 tok/sest
Usable context
262K
Memory
192 GB
Fits in memory660 tok/sest · 262K ctx
Fits in memory562 tok/sest · 262K ctx
Fits in memory377 tok/sest · 262K ctx