Looks like torch has to be newer then 2.0:
import gigaam
model_name = "v2_rnnt" # Options: "v2_ctc" or "ctc", "v2_rnnt" or "rnnt", "v1_ctc", "v1_rnnt"
model = gigaam.load_model(model_name)
transcription = model.transcribe(my_mp3_path)
~/.local/lib/python3.8/site-packages/gigaam/preprocess.py in load_audio(audio_path, sample_rate, return_format)
38
39 if return_format == "float":
---> 40 return torch.frombuffer(audio, dtype=torch.int16).float() / 32768.0
41
42 return torch.frombuffer(audio, dtype=torch.int16)
**AttributeError: module 'torch' has no attribute 'frombuffer'**
Looks like torch has to be newer then 2.0: