torchaudio.save("t.wav", tensor.unsqueeze(0), sampling_rate, bits_per_sample=16)

.gitignore

@@ -0,0 +1 @@
+t.wav

train.py

@@ -1,4 +1,4 @@
-# pip install numpy==1.26.4 transformers==4.45.2 datasets==3.0.1 librosa==0.10.2.post1 jiwer==3.0.4 bitsandbytes==0.38.0.post1 accelerate==0.26.0
+# pip install numpy==1.26.4 torchaudio==0.13.1 transformers==4.45.2 datasets==3.0.1 librosa==0.10.2.post1 jiwer==3.0.4 bitsandbytes==0.38.0.post1 accelerate==0.26.0
 
 dataset_name = "dlxjj/common_voice_17_0"
 language = "Japanese"
@@ -14,4 +14,11 @@ print(common_voice["train"][0])
 
 array = common_voice["train"][0]["audio"]["array"]
 
+import torch, numpy as np, torchaudio
+ 
+tensor = torch.tensor(np.array( array, dtype='float32'))
+
+sampling_rate = 48000
+torchaudio.save("t.wav", tensor.unsqueeze(0), sampling_rate, bits_per_sample=16)
+
 pass