Update data_utils.py

data_utils.py

@@ -23,7 +23,7 @@ class TextAudioSpeakerLoader(torch.utils.data.Dataset):
         3) computes spectrograms from audio files.
     """
 
-    def __init__(self, audiopaths, hparams, all_in_mem: bool = False):
+    def __init__(self, audiopaths, hparams):
         self.audiopaths = load_filepaths_and_text(audiopaths)
         self.max_wav_value = hparams.data.max_wav_value
         self.sampling_rate = hparams.data.sampling_rate
@@ -37,10 +37,6 @@ class TextAudioSpeakerLoader(torch.utils.data.Dataset):
 
         random.seed(1234)
         random.shuffle(self.audiopaths)
-        
-        self.all_in_mem = all_in_mem
-        if self.all_in_mem:
-            self.cache = [self.get_audio(p[0]) for p in self.audiopaths]
 
     def get_audio(self, filename):
         filename = filename.replace("\\", "/")
@@ -51,8 +47,6 @@ class TextAudioSpeakerLoader(torch.utils.data.Dataset):
         audio_norm = audio / self.max_wav_value
         audio_norm = audio_norm.unsqueeze(0)
         spec_filename = filename.replace(".wav", ".spec.pt")
-
-        # Ideally, all data generated after Mar 25 should have .spec.pt
         if os.path.exists(spec_filename):
             spec = torch.load(spec_filename)
         else:
@@ -79,10 +73,6 @@ class TextAudioSpeakerLoader(torch.utils.data.Dataset):
         assert abs(audio_norm.shape[1]-lmin * self.hop_length) < 3 * self.hop_length
         spec, c, f0, uv = spec[:, :lmin], c[:, :lmin], f0[:lmin], uv[:lmin]
         audio_norm = audio_norm[:, :lmin * self.hop_length]
-
-        return c, f0, spec, audio_norm, spk, uv
-
-    def random_slice(self, c, f0, spec, audio_norm, spk, uv):
         # if spec.shape[1] < 30:
         #     print("skip too short audio:", filename)
         #     return None
@@ -95,10 +85,7 @@ class TextAudioSpeakerLoader(torch.utils.data.Dataset):
         return c, f0, spec, audio_norm, spk, uv
 
     def __getitem__(self, index):
-        if self.all_in_mem:
-            return self.random_slice(*self.cache[index])
-        else:
-            return self.random_slice(*self.get_audio(self.audiopaths[index][0]))
+        return self.get_audio(self.audiopaths[index][0])
 
     def __len__(self):
         return len(self.audiopaths)