Allow mps device but set PYTORCH_ENABLE_MPS_FALLBACK

This is needed since the operator 'aten::unfold_backward' is not currently implemented for MPS in pytorch - see pytorch/pytorch#77764

noisereduce/spectralgate/streamed_torch_gate.py

@@ -1,3 +1,4 @@
+import os
 import torch
 from noisereduce.spectralgate.base import SpectralGate
 from noisereduce.torchgate import TorchGate as TG
@@ -50,7 +51,12 @@ def __init__(
             n_jobs=n_jobs,
         )
 
-        self.device = torch.device(device if torch.cuda.is_available() else 'cpu')
+        if "cuda" in device and not torch.cuda.is_available():
+            device = "cpu"
+        elif device == "mps":
+            os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
+
+        self.device = torch.device(device)
 
         # noise convert to torch if needed
         if y_noise is not None: