Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Refactor LibriSpeech Lightning datamodule to accommodate different dataset implementations #2437

Closed
wants to merge 1 commit into from
Closed

Refactor LibriSpeech Lightning datamodule to accommodate different dataset implementations #2437

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 9 additions & 6 deletions examples/asr/librispeech_conformer_rnnt/data_module.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,6 +103,8 @@ def __len__(self):


class LibriSpeechDataModule(LightningDataModule):
librispeech_cls = torchaudio.datasets.LIBRISPEECH

def __init__(
self,
*,
Expand All @@ -116,6 +118,7 @@ def __init__(
train_shuffle=True,
num_workers=10,
):
super().__init__()
self.librispeech_path = librispeech_path
self.train_dataset_lengths = None
self.val_dataset_lengths = None
Expand All @@ -130,9 +133,9 @@ def __init__(

def train_dataloader(self):
datasets = [
torchaudio.datasets.LIBRISPEECH(self.librispeech_path, url="train-clean-360"),
torchaudio.datasets.LIBRISPEECH(self.librispeech_path, url="train-clean-100"),
torchaudio.datasets.LIBRISPEECH(self.librispeech_path, url="train-other-500"),
self.librispeech_cls(self.librispeech_path, url="train-clean-360"),
self.librispeech_cls(self.librispeech_path, url="train-clean-100"),
self.librispeech_cls(self.librispeech_path, url="train-other-500"),
]

if not self.train_dataset_lengths:
Expand Down Expand Up @@ -161,8 +164,8 @@ def train_dataloader(self):

def val_dataloader(self):
datasets = [
torchaudio.datasets.LIBRISPEECH(self.librispeech_path, url="dev-clean"),
torchaudio.datasets.LIBRISPEECH(self.librispeech_path, url="dev-other"),
self.librispeech_cls(self.librispeech_path, url="dev-clean"),
self.librispeech_cls(self.librispeech_path, url="dev-other"),
]

if not self.val_dataset_lengths:
Expand All @@ -185,7 +188,7 @@ def val_dataloader(self):
return dataloader

def test_dataloader(self):
dataset = torchaudio.datasets.LIBRISPEECH(self.librispeech_path, url="test-clean")
dataset = self.librispeech_cls(self.librispeech_path, url="test-clean")
dataset = TransformDataset(dataset, self.test_transform)
dataloader = torch.utils.data.DataLoader(dataset, batch_size=None)
return dataloader