Use openai-whisper on CPU vs MPS

gistfile1.txt

Python 3.10.8 (main, Nov 24 2022, 08:08:27) [Clang 14.0.6 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> import whisper
>>> torch.__version__
'2.0.0a0+git01de5dd'
>>> model = whisper.load_model("base")
>>> audio = whisper.load_audio("c1.mp3") # downloaded from https://www.mobydickbigread.com/chapter-1-loomings/
>>> audio = whisper.pad_or_trim(audio)
>>> model.transcribe(audio)["text"]
/Users/malfet/miniconda3/lib/python3.10/site-packages/whisper/transcribe.py:78: UserWarning: FP16 is not supported on CPU; using FP32 instead
  warnings.warn("FP16 is not supported on CPU; using FP32 instead")
' Mobey Dick or The Whale by Herman Melville. Chapter 1. Loomings. Call me Ishmael. Some years ago, never mind how long precisely, having little or no money in my purse and nothing particular to interest me on shore, I thought I would sail about a little and see the watery part of the world. So I have of driving off the spleen and regulating the circulation.'
>>> model.to("mps").transcribe(audio)["text"]
/Users/malfet/miniconda3/lib/python3.10/site-packages/whisper/decoding.py:633: UserWarning: 0MPS: no support for int64 repeats mask, casting it to int32 (Triggered internally at /Users/malfet/git/pytorch/pytorch/aten/src/ATen/native/mps/operations/Repeat.mm:236.)
  audio_features = audio_features.repeat_interleave(self.n_group, dim=0)
/Users/malfet/miniconda3/lib/python3.10/site-packages/whisper/decoding.py:439: UserWarning: 1MPS: no support for int64 min/max ops, casting it to int32 (Triggered internally at /Users/malfet/git/pytorch/pytorch/aten/src/ATen/native/mps/operations/ReduceOps.mm:1260.)
  max_text_token_logprob = logprobs[k, : self.tokenizer.timestamp_begin].max()
'<|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|la|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|><|en|>'

Works now, using
python -c "import torch;import whisper;model=whisper.load_model('base').to('mps');audio=whisper.load_audio('c1.mp3');print(model.transcribe(whisper.pad_or_trim(audio))['text'])"