def from_urban8k(
batch_size: int = 4,
**data_module_kwargs,
) -> AudioClassificationData:
"""Downloads and loads the Urban 8k sounds images data set."""
download_data("https://pl-flash-data.s3.amazonaws.com/urban8k_images.zip",
"./data")
return AudioClassificationData.from_folders(
train_folder="data/urban8k_images/train",
val_folder="data/urban8k_images/val",
batch_size=batch_size,
**data_module_kwargs,
)
def test_from_folders_only_train(tmpdir):
seed_everything(42)
train_dir = Path(tmpdir / "train")
train_dir.mkdir()
(train_dir / "a").mkdir()
_rand_image().save(train_dir / "a" / "1.png")
_rand_image().save(train_dir / "a" / "2.png")
(train_dir / "b").mkdir()
_rand_image().save(train_dir / "b" / "1.png")
_rand_image().save(train_dir / "b" / "2.png")
spectrograms_data = AudioClassificationData.from_folders(train_dir,
batch_size=1)
data = next(iter(spectrograms_data.train_dataloader()))
imgs, labels = data["input"], data["target"]
assert imgs.shape == (1, 3, 128, 128)
assert labels.shape == (1, )
def test_from_folders_train_val(tmpdir):
seed_everything(42)
train_dir = Path(tmpdir / "train")
train_dir.mkdir()
(train_dir / "a").mkdir()
_rand_image().save(train_dir / "a" / "1.png")
_rand_image().save(train_dir / "a" / "2.png")
(train_dir / "b").mkdir()
_rand_image().save(train_dir / "b" / "1.png")
_rand_image().save(train_dir / "b" / "2.png")
spectrograms_data = AudioClassificationData.from_folders(
train_dir,
val_folder=train_dir,
test_folder=train_dir,
batch_size=2,
num_workers=0,
)
data = next(iter(spectrograms_data.train_dataloader()))
imgs, labels = data["input"], data["target"]
assert imgs.shape == (2, 3, 128, 128)
assert labels.shape == (2, )
assert list(labels.numpy()) == [0, 1]
data = next(iter(spectrograms_data.val_dataloader()))
imgs, labels = data["input"], data["target"]
assert imgs.shape == (2, 3, 128, 128)
assert labels.shape == (2, )
assert list(labels.numpy()) == [0, 0]
data = next(iter(spectrograms_data.test_dataloader()))
imgs, labels = data["input"], data["target"]
assert imgs.shape == (2, 3, 128, 128)
assert labels.shape == (2, )
assert list(labels.numpy()) == [0, 0]
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
import flash
from flash.audio import AudioClassificationData
from flash.core.data.utils import download_data
from flash.core.finetuning import FreezeUnfreeze
from flash.image import ImageClassifier
# 1. Create the DataModule
download_data("https://pl-flash-data.s3.amazonaws.com/urban8k_images.zip", "./data")
datamodule = AudioClassificationData.from_folders(
train_folder="data/urban8k_images/train",
val_folder="data/urban8k_images/val",
spectrogram_size=(64, 64),
)
# 2. Build the model.
model = ImageClassifier(backbone="resnet18", num_classes=datamodule.num_classes)
# 3. Create the trainer and finetune the model
trainer = flash.Trainer(max_epochs=3, gpus=torch.cuda.device_count())
trainer.finetune(model, datamodule=datamodule, strategy=FreezeUnfreeze(unfreeze_epoch=1))
# 4. Predict what's on few images! air_conditioner, children_playing, siren e.t.c
predictions = model.predict(
[
"data/urban8k_images/test/air_conditioner/13230-0-0-5.wav.jpg",
"data/urban8k_images/test/children_playing/9223-2-0-15.wav.jpg",
# See the License for the specific language governing permissions and
# limitations under the License.
import torch
import flash
from flash.audio import AudioClassificationData
from flash.core.data.utils import download_data
from flash.image import ImageClassifier
# 1. Create the DataModule
download_data("https://pl-flash-data.s3.amazonaws.com/urban8k_images.zip",
"./data")
datamodule = AudioClassificationData.from_folders(
train_folder="data/urban8k_images/train",
val_folder="data/urban8k_images/val",
transform_kwargs=dict(spectrogram_size=(64, 64)),
batch_size=4,
)
# 2. Build the model.
model = ImageClassifier(backbone="resnet18", labels=datamodule.labels)
# 3. Create the trainer and finetune the model
trainer = flash.Trainer(max_epochs=3, gpus=torch.cuda.device_count())
trainer.finetune(model, datamodule=datamodule, strategy=("freeze_unfreeze", 1))
# 4. Predict what's on few images! air_conditioner, children_playing, siren etc.
datamodule = AudioClassificationData.from_files(
predict_files=[
"data/urban8k_images/test/air_conditioner/13230-0-0-5.wav.jpg",
"data/urban8k_images/test/children_playing/9223-2-0-15.wav.jpg",