Testing for optimal **options in sequence

[dgoryeo]

I am trying to determine an optimal set of options for transcribing few old Japanese anime. For a test flow of various options, would the model behave differently if it is loaded and called in below methods A and B?

• Test 3 sets of options on same audio file:

========= moethod A ==================
model = whisper.load_model("large")
result1 = model.transcribe("file.wav", **options1)
result2 = model.transcribe("file.wav", **options2)
result3 = model.transcribe("file.wav", **options3)

========= method B ===================
model = whisper.load_model("large")
result1 = model.transcribe("file.wav", **options1)

model = whisper.load_model("large")
result2 = model.transcribe("file.wav", **options2)

model = whisper.load_model("large")
result3 = model.transcribe("file.wav", **options3)

=====================================

Thanks for any pointers.