latest (dev)
Copy
Latest development documentation · Updated 2026-10-15
tensorplay.vision
tensorplay.vision is the computer-vision toolkit of TensorPlay. The listing
below is a static overview; the
package requires optional dependencies (Pillow) for some entry points, so this
page intentionally does not use autodoc.
Top-level functions
to_tensor,from_image,from_file— tensor/image conversionmake_grid,save_image— visualization helpersset_backend,get_backend— image backend selection (PIL)
Submodules
tensorplay.vision.datasets— dataset classes and utilities (MNIST,CIFAR10/CIFAR100,ImageFolder,Folder,DatasetFolder,UCF101, …)tensorplay.vision.transforms— composable transforms (Compose,ToTensor,Normalize,Resize,CenterCrop,RandomCrop,RandomHorizontalFlip, …)tensorplay.vision.models— reference model architectures (AlexNet,VGG,ResNetvariants, …)tensorplay.vision.io— image reading/writing backendstensorplay.vision.ops— vision-specific operatorstensorplay.vision.utils— helpers shared across the package
Help improve this page
Found an error, an unclear step, or a missing example?
tensorplay.utils.data
At the heart of TensorPlay data loading utility is the tensorplay.utils.data.DataLoader class. It represents a Python iterable over a dataset, with support for
Type information
tensorplay.finfo and tensorplay.iinfo describe the limits of a floating-point or integer data type. You can query a dtype, or construct one from a tensor.