How to find functionality

For some kinds of functionality, FastAI.jl provides feature registries that allow you to search for and use features. The following registries currently exist:

datasets to download and unpack datasets,
datarecipes to load datasets into data containers that are compatible with a learning task; and
learningtasks to find learning tasks that are compatible with a dataset

Domain packages

Functionality is registered by domain packages such as FastVision and FastTabular . You need to import the respective packages to be able to find their functionality in their registry.

To load functionality:

Get an entry using its ID


			
			
			
			using
			
			 

	
			FastAI
			,
			
			 
			FastVision
			

			
			entry
			 
			=
			 
			
			

	
			datasets
			(
			)
			[
			
			"
			mnist_var_size_tiny
			"
			]


			RegistryEntry(                                                                                                  
             id  =  "mnist_var_size_tiny"                         (String)                       
    description  =  missing                                       (String)                       
           size  =  missing                                       (String)                       
           tags  =  String[]                                      (Vector{String})               
                                                                
        package  =  FastAI                                        (Module)                       
     downloaded  =  ⨯                                             (Bool)                         
         loader  =  type_format (generic function with 1 method)  (FastAI.Datasets.DatasetLoader)
                                                                
)

And load it


			
			
			
			load
			(
			entry
			)


			
7-Zip (a) [64] 17.04 : Copyright (c) 1999-2021 Igor Pavlov : 2017-08-28
p7zip Version 17.04 (locale=C.UTF-8,Utf16=on,HugeFiles=on,64 bits,4 CPUs AMD EPYC 7763 64-Core Processor                 (A00F11),ASM,AES-NI)


Extracting archive: 
--
Path = 
Type = tar
Code Page = UTF-8

Everything is Ok

Folders: 9
Files: 1431
Size:       878654
Compressed: 747008


			/home/runner/.julia/datadeps/fastai-mnist_var_size_tiny

Datasets


			
			
			
			using
			
			 

	
			FastAI
			

			

	
			datasets
			(
			)

Datasets
ID	Description	Size	Tags	Package	Is downloaded	Dataset loader
:id	:description	:size	:tags	:package	:downloaded	:loader
"CUB_200_2011"	missing	1GiB	String[]	`FastAI`	false	type_format
"bedroom"	missing	4.25GiB	String[]	`FastAI`	false	type_format
"caltech_101"	missing	126MiB	String[]	`FastAI`	false	type_format
"cifar10"	missing	missing	String[]	`FastAI`	false	type_format
"cifar100"	missing	missing	String[]	`FastAI`	false	type_format
"food-101"	101 food categories, with 101,000 images; 250 test images and 750 training images per class. The training images were not cleaned. All images were rescaled to have a maximum side length of 512 pixels.	5.3GB	String[]	`FastAI`	false	type_format
"imagenette-160"	A subset of 10 easily classified classes from Imagenet: tench, English springer, cassette player, chain saw, church, French horn, garbage truck, gas pump, golf ball, parachute	1.45GiB	String[]	`FastAI`	false	type_format
"imagenette-320"	A subset of 10 easily classified classes from Imagenet: tench, English springer, cassette player, chain saw, church, French horn, garbage truck, gas pump, golf ball, parachute	missing	String[]	`FastAI`	false	type_format
"imagenette"	A subset of 10 easily classified classes from Imagenet: tench, English springer, cassette player, chain saw, church, French horn, garbage truck, gas pump, golf ball, parachute	missing	String[]	`FastAI`	false	type_format
"imagenette2-160"	A subset of 10 easily classified classes from Imagenet: tench, English springer, cassette player, chain saw, church, French horn, garbage truck, gas pump, golf ball, parachute	missing	String[]	`FastAI`	false	type_format
"imagenette2-320"	A subset of 10 easily classified classes from Imagenet: tench, English springer, cassette player, chain saw, church, French horn, garbage truck, gas pump, golf ball, parachute	missing	String[]	`FastAI`	false	type_format
"imagenette2"	A subset of 10 easily classified classes from Imagenet: tench, English springer, cassette player, chain saw, church, French horn, garbage truck, gas pump, golf ball, parachute	missing	String[]	`FastAI`	false	type_format
"imagewang-160"	missing	182MiB	String[]	`FastAI`	false	type_format
"imagewang-320"	missing	639MiB	String[]	`FastAI`	false	type_format
"imagewang"	missing	missing	String[]	`FastAI`	false	type_format
"imagewoof-160"	A subset of 10 harder to classify classes from Imagenet (all dog breeds): Australian terrier, Border terrier, Samoyed, beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, dingo, golden retriever, Old English sheepdog	missing	String[]	`FastAI`	false	type_format
"imagewoof-320"	A subset of 10 harder to classify classes from Imagenet (all dog breeds): Australian terrier, Border terrier, Samoyed, beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, dingo, golden retriever, Old English sheepdog	missing	String[]	`FastAI`	false	type_format
"imagewoof"	A subset of 10 harder to classify classes from Imagenet (all dog breeds): Australian terrier, Border terrier, Samoyed, beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, dingo, golden retriever, Old English sheepdog	missing	String[]	`FastAI`	false	type_format
"imagewoof2-160"	A subset of 10 harder to classify classes from Imagenet (all dog breeds): Australian terrier, Border terrier, Samoyed, beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, dingo, golden retriever, Old English sheepdog	missing	String[]	`FastAI`	false	type_format
"imagewoof2-320"	A subset of 10 harder to classify classes from Imagenet (all dog breeds): Australian terrier, Border terrier, Samoyed, beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, dingo, golden retriever, Old English sheepdog	313MB	String[]	`FastAI`	false	type_format
"imagewoof2"	A subset of 10 harder to classify classes from Imagenet (all dog breeds): Australian terrier, Border terrier, Samoyed, beagle, Shih-Tzu, English foxhound, Rhodesian ridgeback, dingo, golden retriever, Old English sheepdog	1.25GiB	String[]	`FastAI`	false	type_format
"mnist_png"	missing	missing	String[]	`FastAI`	false	type_format
"mnist_var_size_tiny"	missing	missing	String[]	`FastAI`	false	type_format
"oxford-102-flowers"	missing	missing	String[]	`FastAI`	false	type_format
"oxford-iiit-pet"	missing	missing	String[]	`FastAI`	false	type_format
"stanford-cars"	missing	missing	String[]	`FastAI`	false	type_format
"ag_news_csv"	missing	11MB	String[]	`FastAI`	false	type_format
"amazon_review_full_csv"	missing	600MB	String[]	`FastAI`	false	type_format
"amazon_review_polarity_csv"	missing	600MB	String[]	`FastAI`	false	type_format
"dbpedia_csv"	missing	65MB	String[]	`FastAI`	false	type_format
"giga-fren"	missing	2.4GB	String[]	`FastAI`	false	type_format
"imdb"	missing	140MB	String[]	`FastAI`	false	type_format
"sogou_news_csv"	missing	360MB	String[]	`FastAI`	false	type_format
"wikitext-103"	missing	181MB	String[]	`FastAI`	false	type_format
"wikitext-2"	missing	4MB	String[]	`FastAI`	false	type_format
"yahoo_answers_csv"	missing	305MB	String[]	`FastAI`	false	type_format
"yelp_review_full_csv"	missing	187MB	String[]	`FastAI`	false	type_format
"yelp_review_polarity_csv"	missing	158MB	String[]	`FastAI`	false	type_format
"biwi_head_pose"	missing	430MiB	String[]	`FastAI`	false	type_format
"camvid"	missing	571MB	String[]	`FastAI`	false	type_format
"pascal-voc"	missing	4.3GB	String[]	`FastAI`	false	type_format
"pascal_2007"	missing	missing	String[]	`FastAI`	false	type_format
"pascal_2012"	missing	missing	String[]	`FastAI`	false	type_format
"siim_small"	missing	missing	String[]	`FastAI`	false	type_format
"skin-lesion"	missing	missing	String[]	`FastAI`	false	type_format
"tcga-small"	missing	missing	String[]	`FastAI`	false	type_format
"adult_sample"	missing	3.8MB	String[]	`FastAI`	false	type_format
"biwi_sample"	missing	missing	String[]	`FastAI`	false	type_format
"camvid_tiny"	missing	missing	String[]	`FastAI`	false	type_format
"dogscats"	missing	800MiB	String[]	`FastAI`	false	type_format
"human_numbers"	missing	missing	String[]	`FastAI`	false	type_format
"imdb_sample"	missing	4KB	String[]	`FastAI`	false	type_format
"mnist_sample"	missing	3MB	String[]	`FastAI`	false	type_format
"mnist_tiny"	missing	300KB	String[]	`FastAI`	false	type_format
"movie_lens_sample"	missing	missing	String[]	`FastAI`	false	type_format
"planet_sample"	missing	14.8MB	String[]	`FastAI`	false	type_format
"planet_tiny"	missing	1MB	String[]	`FastAI`	false	type_format
"coco_sample"	missing	3GB	String[]	`FastAI`	false	type_format
"coco-train2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-val2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-test2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-unlabeled2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-image_info_test2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-image_info_unlabeled2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-annotations_trainval2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-stuff_annotations_trainval2017"	missing	missing	String[]	`FastAI`	false	type_format
"coco-panoptic_annotations_trainval2017"	missing	missing	String[]	`FastAI`	false	type_format
"ecg5000"	The original dataset for "ECG5000" is a 20-hour long ECG downloaded from Physionet. The name is BIDMC Congestive Heart Failure Database(chfdb) and it is record "chf07".	10MB	String[]	`FastAI`	false	type_format
"atrial"	This is a physionet dataset of two-channel ECG recordings has been created from data used in the Computers in Cardiology Challenge 2004, an open competition with the goal of developing automated methods for predicting spontaneous termination of atrial fibrillation (AF).	226KB	String[]	`FastAI`	false	type_format
"natops"	The data is generated by sensors on the hands, elbows, wrists and thumbs. The data are the x,y,z coordinates for each of the eight locations.	5.1MB	String[]	`FastAI`	false	type_format
"appliances_energy"	The goal of this dataset is to predict total energy usage in kWh of a house.	15MB	String[]	`FastAI`	false	type_format

Data recipes


			
			
			
			using
			
			 

	
			FastAI
			

			

	
			datarecipes
			(
			)

Dataset recipes
ID	Block types	Description	Is downloaded	Dataset ID	Package	Recipe
:id	:blocks	:description	:downloaded	:datasetid	:package	:recipe
"CUB_200_2011"	`(Image{2}, Label)`	missing	false	CUB_200_2011	FastVision	ImageFolders
"imagenette"	`(Image{2}, Label)`	missing	false	imagenette	FastVision	ImageFolders
"imagenette2"	`(Image{2}, Label)`	missing	false	imagenette2	FastVision	ImageFolders
"imagewang-320"	`(Image{2}, Label)`	missing	false	imagewang-320	FastVision	ImageFolders
"mnist_sample"	`(Image{2}, Label)`	missing	false	mnist_sample	FastVision	ImageFolders
"pascal_2007"	`(Image{2}, LabelMulti)`	missing	false	pascal_2007	FastVision	ImageTableMultiLabel
"camvid"	`(Image{2}, Mask{2})`	missing	false	camvid	FastVision	ImageSegmentationFolders
"camvid_tiny"	`(Image{2}, Mask{2})`	missing	false	camvid_tiny	FastVision	ImageSegmentationFolders
"imagenette-160"	`(Image{2}, Label)`	missing	false	imagenette-160	FastVision	ImageFolders
"imagenette-320"	`(Image{2}, Label)`	missing	false	imagenette-320	FastVision	ImageFolders
"imagewoof2-160"	`(Image{2}, Label)`	missing	false	imagewoof2-160	FastVision	ImageFolders
"cifar100"	`(Image{2}, Label)`	missing	false	cifar100	FastVision	ImageFolders
"imagewang-160"	`(Image{2}, Label)`	missing	false	imagewang-160	FastVision	ImageFolders
"mnist_var_size_tiny"	`(Image{2}, Label)`	missing	false	mnist_var_size_tiny	FastVision	ImageFolders
"cifar10"	`(Image{2}, Label)`	missing	false	cifar10	FastVision	ImageFolders
"mnist_png"	`(Image{2}, Label)`	missing	false	mnist_png	FastVision	ImageFolders
"caltech_101"	`(Image{2}, Label)`	missing	false	caltech_101	FastVision	ImageFolders
"food-101"	`(Image{2}, Label)`	missing	false	food-101	FastVision	ImageFolders
"imagenette2-320"	`(Image{2}, Label)`	missing	false	imagenette2-320	FastVision	ImageFolders
"imagewoof2"	`(Image{2}, Label)`	missing	false	imagewoof2	FastVision	ImageFolders
"imagewoof2-320"	`(Image{2}, Label)`	missing	false	imagewoof2-320	FastVision	ImageFolders
"imagewoof-160"	`(Image{2}, Label)`	missing	false	imagewoof-160	FastVision	ImageFolders
"imagewoof-320"	`(Image{2}, Label)`	missing	false	imagewoof-320	FastVision	ImageFolders
"imagewang"	`(Image{2}, Label)`	missing	false	imagewang	FastVision	ImageFolders
"mnist_tiny"	`(Image{2}, Label)`	missing	false	mnist_tiny	FastVision	ImageFolders
"imagewoof"	`(Image{2}, Label)`	missing	false	imagewoof	FastVision	ImageFolders
"imagenette2-160"	`(Image{2}, Label)`	missing	false	imagenette2-160	FastVision	ImageFolders
"adult_sample"	`TableRow`	missing	false	adult_sample	FastTabular	TableDatasetRecipe
"adult_sample/clf_salary"	`(TableRow, Label)`	missing	false	adult_sample	FastTabular	TableClassificationRecipe
"adult_sample/reg_age"	`(TableRow, Continuous)`	missing	false	adult_sample	FastTabular	TableRegressionRecipe
"imdb_sample"	`TableRow`	missing	false	imdb_sample	FastTabular	TableDatasetRecipe
"imdb_sample/clf"	`(TableRow, Label)`	missing	false	imdb_sample	FastTabular	TableClassificationRecipe

Learning tasks


			
			
			
			using
			
			 

	
			FastAI
			

			

	
			learningtasks
			(
			)

Learning tasks
ID	Name	Block types	Category	Description	Learning task	Package
:id	:name	:blocks	:category	:description	:constructor	:package
"vision/imageclfmulti"	Image classification (multi-label)	`(Image, LabelMulti)`	supervised	Multi-label image classification task where every image can have multiple class labels associated with it.	`ImageClassificationMulti`	`FastVision`
"vision/imagekeypoint"	Image keypoint regression	`(Image, Keypoints)`	supervised	Keypoint regression task with a fixed number of keypoints to be detected.	`ImageKeypointRegression`	`FastVision`
"vision/imagesegmentation"	Image segmentation	`(Image, Mask)`	supervised	Semantic segmentation task in which every pixel in an image is classified.	`ImageSegmentation`	`FastVision`
"vision/imageclfsingle"	Image classification (single-label)	`(Image, Label)`	supervised	Single-label image classification task where every image has a single class label associated with it.	`ImageClassificationSingle`	`FastVision`
"tabular/clfsingle"	Tabular classification (single-label)	`(TableRow, Label)`	supervised	Task where a table row with categorical and continuous variables is classified as one of a number of classes.	`TabularClassificationSingle`	`FastTabular`
"tabular/regression"	Tabular regression	`(TableRow, Continuous)`	supervised	Task where a number of continuous variables are regressed from a table row with categorical and continuous variables.	`TabularClassificationSingle`	`FastTabular`