refactor wav2mel module

Author: yistLin

data/wav2mel.py

@@ -1,16 +1,17 @@
 """Wav2Mel for processing audio data."""
 
 import torch
+import torch.nn as nn
 from torchaudio.sox_effects import apply_effects_tensor
 from torchaudio.transforms import MelSpectrogram
 
 
-class Wav2Mel(torch.nn.Module):
+class Wav2Mel(nn.Module):
     """Transform audio file into mel spectrogram tensors."""
 
     def __init__(
         self,
-        sample_rate: float = 16000,
+        sample_rate: int = 16000,
         norm_db: float = -3.0,
         sil_threshold: float = 1.0,
         sil_duration: float = 0.1,
@@ -41,7 +42,7 @@ def forward(self, wav_tensor: torch.Tensor, sample_rate: int) -> torch.Tensor:
         return mel_tensor
 
 
-class SoxEffects(torch.nn.Module):
+class SoxEffects(nn.Module):
     """Transform waveform tensors."""
 
     def __init__(
@@ -72,12 +73,12 @@ def forward(self, wav_tensor: torch.Tensor, sample_rate: int) -> torch.Tensor:
         return wav_tensor
 
 
-class LogMelspectrogram(torch.nn.Module):
+class LogMelspectrogram(nn.Module):
     """Transform waveform tensors into log mel spectrogram tensors."""
 
     def __init__(
         self,
-        sample_rate: float,
+        sample_rate: int,
         fft_window_ms: float,
         fft_hop_ms: float,
         f_min: float,
@@ -94,4 +95,4 @@ def __init__(
 
     def forward(self, wav_tensor: torch.Tensor) -> torch.Tensor:
         mel_tensor = self.melspectrogram(wav_tensor).squeeze(0).T  # (time, n_mels)
-        return torch.log(mel_tensor.squeeze(0) + 1e-9)
+        return torch.log(torch.clamp(mel_tensor, min=1e-9))