Add unsupervised_audio_dataset

codes/data/__init__.py

@@ -75,6 +75,8 @@ def create_dataset(dataset_opt, return_collate=False):
         collate = C(dataset_opt)
     elif mode == 'wavfile_clips':
         from data.audio.wavfile_dataset import WavfileDataset as D
+    elif mode == 'unsupervised_audio':
+        from data.audio.unsupervised_audio_dataset import UnsupervisedAudioDataset as D
     elif mode == 'stop_prediction':
         from models.tacotron2.hparams import create_hparams
         default_params = create_hparams()

codes/data/audio/unsupervised_audio_dataset.py

@@ -0,0 +1,154 @@
+import os
+import pathlib
+import random
+
+import torch
+import torch.utils.data
+import torch.nn.functional as F
+import torchaudio
+from audio2numpy import open_audio
+from tqdm import tqdm
+
+from data.audio.wav_aug import WavAugmentor
+from data.util import find_files_of_type, is_wav_file, is_audio_file
+from models.tacotron2.taco_utils import load_wav_to_torch
+from utils.util import opt_get
+
+
+def load_audio(audiopath, sampling_rate):
+    if audiopath[:-4] == '.wav':
+        audio, lsr = load_wav_to_torch(audiopath)
+    else:
+        audio, lsr = open_audio(audiopath)
+        audio = torch.FloatTensor(audio)
+
+    # Remove any channel data.
+    if len(audio.shape) > 1:
+        if audio.shape[0] < 5:
+            audio = audio[0]
+        else:
+            assert audio.shape[1] < 5
+            audio = audio[:, 0]
+
+    if lsr != sampling_rate:
+        if lsr < sampling_rate:
+            print(f'{audiopath} has a sample rate of {sampling_rate} which is lower than the requested sample rate of {sampling_rate}. This is not a good idea.')
+        audio = torch.nn.functional.interpolate(audio.unsqueeze(0).unsqueeze(1), scale_factor=sampling_rate/lsr, mode='nearest', recompute_scale_factor=False).squeeze()
+
+    # Check some assumptions about audio range. This should be automatically fixed in load_wav_to_torch, but might not be in some edge cases, where we should squawk.
+    # '2' is arbitrarily chosen since it seems like audio will often "overdrive" the [-1,1] bounds.
+    if torch.any(audio > 2) or not torch.any(audio < 0):
+        print(f"Error with {audiopath}. Max={audio.max()} min={audio.min()}")
+    audio.clip_(-1, 1)
+
+    return audio.unsqueeze(0)
+
+
+class UnsupervisedAudioDataset(torch.utils.data.Dataset):
+
+    def __init__(self, opt):
+        path = opt['path']
+        cache_path = opt['cache_path']  # Will fail when multiple paths specified, must be specified in this case.
+        if not isinstance(path, list):
+            path = [path]
+        if os.path.exists(cache_path):
+            self.audiopaths = torch.load(cache_path)
+        else:
+            print("Building cache..")
+            self.audiopaths = []
+            for p in path:
+                self.audiopaths.extend(find_files_of_type('img', p, qualifier=is_audio_file)[0])
+            torch.save(self.audiopaths, cache_path)
+
+        # Parse options
+        self.sampling_rate = opt_get(opt, ['sampling_rate'], 22050)
+        self.pad_to = opt_get(opt, ['pad_to_seconds'], None)
+        if self.pad_to is not None:
+            self.pad_to *= self.sampling_rate
+        self.pad_to = opt_get(opt, ['pad_to_samples'], self.pad_to)
+
+        self.extra_samples = opt_get(opt, ['extra_samples'], 0)
+        self.extra_sample_len = opt_get(opt, ['extra_sample_length'], 2)
+        self.extra_sample_len *= self.sampling_rate
+
+    def get_audio_for_index(self, index):
+        audiopath = self.audiopaths[index]
+        audio = load_audio(audiopath, self.sampling_rate)
+        return audio, audiopath
+
+    def get_related_audio_for_index(self, index):
+        if self.extra_samples <= 0:
+            return None
+        audiopath = self.audiopaths[index]
+        related_files = find_files_of_type('img', os.path.dirname(audiopath), qualifier=is_audio_file)[0]
+        assert audiopath in related_files
+        assert len(related_files) < 50000  # Sanity check to ensure we aren't loading "related files" that aren't actually related.
+        related_files.remove(audiopath)
+        related_clips = []
+        random.shuffle(related_clips)
+        for j, related_file in enumerate(related_files):
+            rel_clip = load_audio(related_file, self.sampling_rate)
+            gap = rel_clip.shape[-1] - self.extra_sample_len
+            if gap < 0:
+                rel_clip = F.pad(rel_clip, pad=(0, abs(gap)))
+            elif gap > 0:
+                rand_start = random.randint(0, gap)
+                rel_clip = rel_clip[:, rand_start:rand_start+self.extra_sample_len]
+            related_clips.append(rel_clip)
+            if j >= self.extra_samples:
+                break
+        actual_extra_samples = j
+        while j < self.extra_samples:
+            related_clips.append(torch.zeros(1, self.extra_sample_len))
+            j += 1
+        return torch.stack(related_clips, dim=0), actual_extra_samples
+
+    def __getitem__(self, index):
+        # Split audio_norm into two tensors of equal size.
+        audio_norm, filename = self.get_audio_for_index(index)
+        alt_files, actual_samples = self.get_related_audio_for_index(index)
+
+        # This is required when training to make sure all clips align.
+        if self.pad_to is not None:
+            if audio_norm.shape[-1] <= self.pad_to:
+                audio_norm = torch.nn.functional.pad(audio_norm, (0, self.pad_to - audio_norm.shape[-1]))
+            else:
+                gap = audio_norm.shape[-1] - self.pad_to
+                start = random.randint(0, gap-1)
+                audio_norm = audio_norm[:, start:start+self.pad_to]
+
+        output = {
+            'clip': audio_norm,
+            'alt_clips': alt_files,
+            'num_alt_clips': actual_samples,  # We need to pad so that the dataloader can combine these.
+            'path': filename,
+        }
+        return output
+
+    def __len__(self):
+        return len(self.audiopaths)
+
+
+if __name__ == '__main__':
+    params = {
+        'mode': 'unsupervised_audio',
+        'path': ['Z:\\split\\cleaned\\books0'],
+        'cache_path': 'E:\\audio\\remote-cache.pth',
+        'sampling_rate': 22050,
+        'pad_to_seconds': 5,
+        'phase': 'train',
+        'n_workers': 0,
+        'batch_size': 16,
+        'extra_samples': 4,
+    }
+    from data import create_dataset, create_dataloader, util
+
+    ds = create_dataset(params)
+    dl = create_dataloader(ds, params)
+    i = 0
+    for b in tqdm(dl):
+        for b_ in range(16):
+            pass
+            #torchaudio.save(f'{i}_clip1_{b_}.wav', b['clip1'][b_], ds.sampling_rate)
+            #torchaudio.save(f'{i}_clip2_{b_}.wav', b['clip2'][b_], ds.sampling_rate)
+            #i += 1