Skip to content

Data API

Complete API reference for data operations.

DataLoader Class

class DataLoader:
    def __init__(self, data_path, batch_size=32, shuffle=True)

Methods

split(train_ratio, val_ratio, test_ratio)

Split data into sets.

train_loader, val_loader, test_loader = loader.split(0.7, 0.15, 0.15)

add_transform(transform)

Add preprocessing transform.

loader.add_transform(my_transform)

set_batch_size(size)

Change batch size.

loader.set_batch_size(64)

__iter__()

Iterate over batches.

for batch in loader:
    # Process batch

Dataset Class

class Dataset:
    def __init__(self, data, targets=None)

Methods

__len__()

Get dataset size.

size = len(dataset)

__getitem__(index)

Get single item.

item = dataset[0]

Transforms

from iovalence.data import transforms

# Resize
transforms.Resize(size=(224, 224))

# Normalize
transforms.Normalize(mean=[0.5], std=[0.5])

# Random Crop
transforms.RandomCrop(size=200)

# Random Flip
transforms.RandomFlip()

Data Formats Supported

  • Images (PNG, JPEG, TIFF)
  • Text (TXT, CSV, JSON)
  • Tabular (CSV, Parquet)
  • Audio (WAV, MP3)
  • Time Series (NPY, HDF5)

See Also