feat: add librispeech dataset, add text info option to commonvoice

Author: flavioschneider

README.md

@@ -44,8 +44,38 @@ dataset[0] # (1, 158621)
 dataset[1] # (1, 153757)
 ```
 
+#### Full API:
+```py
+LJSpeechDataset(
+    root: str = "./data", # The root where the dataset will be downloaded
+    transforms: Optional[Callable] = None, # Transforms to apply to audio files
+)
+```
+
+### LibriSpeech Dataset
+Wrapper for the [LibriSpeech](https://www.openslr.org/12) dataset (EN only). Requires `pip install datasets`. Note that this dataset requires several GBs of storage.
+
+```py
+from audio_data_pytorch import LibriSpeechDataset
+
+dataset = LibriSpeechDataset(
+    root="./data",
+)
+
+dataset[0] # (1, 222336)
+```
+
+#### Full API:
+```py
+LibriSpeechDataset(
+    root: str = "./data", # The root where the dataset will be downloaded
+    with_info: bool = False, # Whether to return info (i.e. text, sampling rate, speaker_id)
+    transforms: Optional[Callable] = None, # Transforms to apply to audio files
+)
+```
+
 ### Common Voice Dataset
-Multilanguage wrapper for the [Common Voice](https://commonvoice.mozilla.org/) dataset with voice-only data. Requires `pip install datasets`. Note that each language requires several GBs of storage, and that you have to confirm access for each distinct version you use e.g. [here](https://huggingface.co/datasets/mozilla-foundation/common_voice_10_0), to validate your Huggingface access token. You can provide a list of `languages` and to avoid an unbalanced dataset the values will be interleaved by downsampling the majority language to have the same number of samples as the minority language.
+Multilanguage wrapper for the [Common Voice](https://commonvoice.mozilla.org/). Requires `pip install datasets`. Note that each language requires several GBs of storage, and that you have to confirm access for each distinct version you use e.g. [here](https://huggingface.co/datasets/mozilla-foundation/common_voice_10_0), to validate your Huggingface access token. You can provide a list of `languages` and to avoid an unbalanced dataset the values will be interleaved by downsampling the majority language to have the same number of samples as the minority language.
 
 ```py
 from audio_data_pytorch import CommonVoiceDataset
@@ -66,7 +96,7 @@ CommonVoiceDataset(
     sub_version: int = 0, # Subversion: common_voice_{version}_{sub_version}
     root: str = "./data", # The root where the dataset will be downloaded
     languages: Sequence[str] = ['en'], # List of languages to include in the dataset
-    with_sample_rate: bool = False,  # Returns sample rate as second argument
+    with_info: bool = False,  #  Whether to return info (i.e. text, sampling rate, age, gender, accent, locale)
     transforms: Optional[Callable] = None, # Transforms to apply to audio files
 )
 ```

audio_data_pytorch/datasets/__init__.py

@@ -1,4 +1,5 @@
 from .common_voice_dataset import CommonVoiceDataset
-from .ljspeech_dataset import LJSpeechDataset
+from .libri_speech_dataset import LibriSpeechDataset
+from .lj_speech_dataset import LJSpeechDataset
 from .wav_dataset import WAVDataset
 from .youtube_dataset import YoutubeDataset

audio_data_pytorch/datasets/common_voice_dataset.py

@@ -1,5 +1,5 @@
 import os
-from typing import Callable, Optional, Sequence, Tuple, Union
+from typing import Callable, Dict, Optional, Sequence, Tuple, Union
 
 import torch
 from torch import Tensor
@@ -14,10 +14,10 @@ def __init__(
         sub_version: int = 0,
         root: str = "./data",
         languages: Sequence[str] = ["en"],
-        with_sample_rate: bool = False,
+        with_info: bool = False,
         transforms: Optional[Callable] = None,
     ):
-        self.with_sample_rate = with_sample_rate
+        self.with_info = with_info
         self.transforms = transforms
 
         from datasets import interleave_datasets, load_dataset
@@ -37,15 +37,24 @@ def __init__(
 
     def __getitem__(
         self, idx: Union[Tensor, int]
-    ) -> Union[Tensor, Tuple[Tensor, Tensor]]:
+    ) -> Union[Tensor, Tuple[Tensor, Dict]]:
         idx = idx.tolist() if torch.is_tensor(idx) else idx  # type: ignore
         data = self.dataset[idx]
+
         waveform = torch.tensor(data["audio"]["array"]).view(1, -1)
-        sample_rate = data["audio"]["sampling_rate"]
+
+        info = dict(
+            sample_rate=data["audio"]["sampling_rate"],
+            text=data["sentence"],
+            age=data["age"],
+            accent=data["accent"],
+            gender=data["gender"],
+            locale=data["locale"],
+        )
 
         if self.transforms:
             waveform = self.transforms(waveform)
-        return (waveform, sample_rate) if self.with_sample_rate else waveform
+        return (waveform, info) if self.with_info else waveform
 
     def __len__(self) -> int:
         return len(self.dataset)

audio_data_pytorch/datasets/libri_speech_dataset.py

@@ -0,0 +1,44 @@
+import os
+from typing import Callable, Dict, Optional, Tuple, Union
+
+import torch
+from torch import Tensor
+from torch.utils.data import Dataset
+
+
+class LibriSpeechDataset(Dataset):
+    def __init__(
+        self,
+        root: str = "./data",
+        with_info: bool = False,
+        transforms: Optional[Callable] = None,
+    ):
+        self.with_info = with_info
+        self.transforms = transforms
+
+        from datasets import load_dataset
+
+        self.dataset = load_dataset(
+            "librispeech_asr",
+            "clean",
+            split="train.100",
+            cache_dir=os.path.join(root, "librispeech_dataset"),
+        )
+
+    def __getitem__(
+        self, idx: Union[Tensor, int]
+    ) -> Union[Tensor, Tuple[Tensor, Dict]]:
+        idx = idx.tolist() if torch.is_tensor(idx) else idx  # type: ignore
+        data = self.dataset[idx]
+        waveform = torch.tensor(data["audio"]["array"]).view(1, -1)
+        info = dict(
+            sample_rate=data["audio"]["sampling_rate"],
+            text=data["text"],
+            speaker_id=data["speaker_id"],
+        )
+        if self.transforms:
+            waveform = self.transforms(waveform)
+        return (waveform, info) if self.with_info else waveform
+
+    def __len__(self) -> int:
+        return len(self.dataset)

audio_data_pytorch/datasets/ljspeech_dataset.py renamed to audio_data_pytorch/datasets/lj_speech_dataset.py

@@ -3,7 +3,7 @@
 setup(
     name="audio-data-pytorch",
     packages=find_packages(exclude=[]),
-    version="0.0.10",
+    version="0.0.11",
     license="MIT",
     description="Audio Data - PyTorch",
     long_description_content_type="text/markdown",