torchvision

torchvision is an extension for torch providing image loading, transformations, common architectures for computer vision, pre-trained weights and access to commonly used datasets.

Installation

The CRAN release can be installed with:

install.packages("torchvision")

You can install the development version from GitHub with:

remotes::install_github("mlverse/torchvision@main")

RF100 Dataset Catalog

torchvision includes 34 datasets from the RoboFlow 100 benchmark, organized into 6 collections. Use the catalog to easily discover and search for datasets:

library(torchvision)

# Search for datasets by keyword
search_rf100("solar")       # Find solar/photovoltaic datasets
search_rf100("cell")        # Find cell-related datasets
search_rf100("medical")     # Find medical imaging datasets

# Browse by collection
search_rf100(collection = "biology")    # All biology datasets
search_rf100(collection = "medical")    # All medical datasets

# View complete catalog
catalog <- get_rf100_catalog()
View(catalog)

Available Collections

• Biology (9 datasets): Microscopy, cells, bacteria, parasites, plant diseases
• Medical (8 datasets): X-rays, MRI, pathology, tumor detection
• Infrared (4 datasets): Thermal imaging, FLIR cameras, solar panels
• Damage (3 datasets): Infrastructure damage, defect detection
• Underwater (4 datasets): Marine life, coral reefs, underwater objects
• Document (6 datasets): OCR, document parsing, diagrams

See vignette("rf100-datasets") for the complete catalog and detailed information.

Example Usage

# Search for a dataset
search_rf100("blood")

# Load the dataset
ds <- rf100_biology_collection(
  dataset = "blood_cell",
  split = "train",
  download = TRUE
)

# Visualize a sample
item <- ds[1]
boxed <- draw_bounding_boxes(item)
tensor_image_browse(boxed)