Machine Learning Datasets Machine Learning Datasets
  • GitHub
  • Slack
  • Documentation
Get Started
Machine Learning Datasets Machine Learning Datasets
Get Started
Machine Learning Datasets
  • GitHub
  • Slack
  • Documentation

Docy

Machine Learning Datasets

  • Folder icon closed Folder open iconDatasets
    • MNIST
    • ImageNet Dataset
    • COCO Dataset
    • CIFAR 10 Dataset
    • CIFAR 100 Dataset
    • FFHQ Dataset
    • Places205 Dataset
    • GTZAN Genre Dataset
    • GTZAN Music Speech Dataset
    • The Street View House Numbers (SVHN) Dataset
    • Caltech 101 Dataset
    • LibriSpeech Dataset
    • dSprites Dataset
    • PUCPR Dataset
    • RAVDESS Dataset
    • GTSRB Dataset
    • CSSD Dataset
    • ATIS Dataset
    • Free Spoken Digit Dataset (FSDD)
    • not-MNIST Dataset
    • ECSSD Dataset
    • COCO-Text Dataset
    • CoQA Dataset
    • FGNET Dataset
    • ESC-50 Dataset
    • GlaS Dataset
    • UTZappos50k Dataset
    • Pascal VOC 2012 Dataset
    • Pascal VOC 2007 Dataset
    • Omniglot Dataset
    • HMDB51 Dataset
    • Chest X-Ray Image Dataset
    • NIH Chest X-ray Dataset
    • Fashionpedia Dataset
    • DRIVE Dataset
    • Kaggle Cats & Dogs Dataset
    • Lincolnbeet Dataset
    • Sentiment-140 Dataset
    • MURA Dataset
    • LIAR Dataset
    • Stanford Cars Dataset
    • SWAG Dataset
    • HASYv2 Dataset
    • WFLW Dataset
    • Visdrone Dataset
    • 11k Hands Dataset
    • QuAC Dataset
    • LFW Deep Funneled Dataset
    • LFW Funneled Dataset
    • Office-Home Dataset
    • LFW Dataset
    • PlantVillage Dataset
    • Optical Handwritten Digits Dataset
    • UCI Seeds Dataset
    • STN-PLAD Dataset
    • FER2013 Dataset
    • Adience Dataset
    • PPM-100 Dataset
    • CelebA Dataset
    • Fashion MNIST Dataset
    • Google Objectron Dataset
    • CARPK Dataset
    • CACD Dataset
    • Flickr30k Dataset
    • Kuzushiji-Kanji (KKanji) dataset
    • KMNIST
    • EMNIST Dataset
    • USPS Dataset
    • MARS Dataset
    • HICO Classification Dataset
    • NSynth Dataset
    • RESIDE dataset
    • Electricity Dataset
    • DRD Dataset
    • Caltech 256 Dataset
    • AFW Dataset
    • PACS Dataset
    • TIMIT Dataset
    • KTH Actions Dataset
    • WIDER Face Dataset
    • WISDOM Dataset
    • DAISEE Dataset
    • WIDER Dataset
    • LSP Dataset
    • UCF Sports Action Dataset
    • Wiki Art Dataset
    • FIGRIM Dataset
    • ANIMAL (ANIMAL10N) Dataset
    • OPA Dataset
    • DomainNet Dataset
    • HAM10000 Dataset
    • Tiny ImageNet Dataset
    • Speech Commands Dataset
    • 300w Dataset
    • Food 101 Dataset
    • VCTK Dataset
    • LOL Dataset
    • AQUA Dataset
    • LFPW Dataset
    • ARID Video Action dataset
    • NABirds Dataset
    • SQuAD Dataset
    • ICDAR 2013 Dataset
    • Animal Pose Dataset
  • Folder icon closed Folder open iconDeep Lake Docs Home
  • Folder icon closed Folder open iconDataset Visualization
  • API Basics
  • Storage & Credentials
  • Getting Started
  • Tutorials (w Colab)
  • Playbooks
  • Data Layout
  • Folder icon closed Folder open iconShuffling in ds.pytorch()
  • Folder icon closed Folder open iconStorage Synchronization
  • Folder icon closed Folder open iconTensor Relationships
  • Folder icon closed Folder open iconQuickstart
  • Folder icon closed Folder open iconHow to Contribute

FFHQ Dataset

Estimated reading: 4 minutes

Visualization of the FFHQ Dataset in the Deep Lake UI

FFHQ (Flickr-Faces-HQ) Dataset

What is FFHQ Dataset?

The FFHQ Dataset (Flickr-Faces-HQ) dataset is a high-quality image set containing 70,000 PNG images at 1024×1024 resolution, it has been put considerable effort to include as many attributes as possible and variations these, so expect a wide range of different ages and ethnicity.

Importantly, this dataset is not intended for, and should not be used for, the development or improvement of facial recognition technologies.

Download FFHQ Dataset in Python

Instead of downloading the FFHQ dataset in Python, you can effortlessly load it in Python via our Deep Lake open source with just one line of code

Load FFHQ Dataset in Python

				
					import deeplake
ds = deeplake.load("hub://activeloop/ffhq")
				
			

FFHQ Dataset Structure

FFHQ Data Fields
  • images_1024/face_landmarks: an int32 tensor containing 204 coco points
  • images_1024/image: a uint8 tensor containing a 1024 x 1024 png image
  • images_128/image: a uint8 tensor containing a 128 x 128 png image
  • images_metadata: an str tensor containing the 70,000 image metadata
  • images_wild/face_landmarks: a int32 tensor containing 204 coco points
  • images_wild/face_quad: a float32 tensor containing 4 generic objects
  • images_wild/face_rect: a int32 tensor containing 1 box object
  • images_wild/image: a uint8 tensor containing a 1062 x 1072 compressed png image
FFHQ Data Splits

While the FFHQ dataset is not explicitly split between a training and validation set, it is customary to use the first 60,000 images as a training set and the remaining 10,000 for validation. You can achieve that in the following way:

				
					import deeplake
ds_train = deeplake.load("hub://activeloop/ffhq")[60000]
ds_val = deeplake.load("hub://activeloop/ffhq")[10000]
				
			

How to use FFHQ Dataset with PyTorch and TensorFlow in Python

Train a model on FFHQ dataset with PyTorch in Python

Let’s use Deep Lake built-in PyTorch one-line dataloader to connect the data to the compute:

				
					dataloader = ds.pytorch(num_workers=0, batch_size=4, shuffle=False)
				
			
Train a model on FFHQ dataset with TensorFlow in Python
				
					dataloader = ds.tensorflow()
				
			

FFHQ Dataset Creation

Data collection and Normalization of images

These images were collected from Flickr under a permissive license, they were cropped and aligned using dlib and various filters used to prune the set. Finally, Amazon Mechanical Turk was used to remove the occasional statues, paintings, or photos of photos.

Additional information about FFHQ Dataset

FFHQ Dataset Description

  • Homepage: N/A
  • Repository: https://github.com/NVlabs/ffhq-dataset
  • Paper: T. Karras, S. Laine and T. Aila, “A Style-Based Generator Architecture for Generative Adversarial Networks,” 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 4396-4405, doi: 10.1109/CVPR.2019.00453.
  • Point of Contact: [email protected]
FFHQ Dataset Curators

Terros Karras, Samuli Laine, Timo Aila

FFHQ Dataset Licensing information

Creative Commons Attribution-Share Alike 4.0 license

FFHQ Dataset Citation Information
				
					@article{,
  title={A Style-Based Generator Architecture for Generative Adversarial Networks},
  author={Tero Karras, Samuli Laine, Timo Aila},
  journal={IEEE[Online]. Avaliable: https://ieeexplore.ieee.org/document/8953766},
  volume={3},
  year={2019}
}
				
			

FFHQ Datasets FAQs

What is the FFHQ dataset for Python?

The FFHQ dataset(Flickr-Faces-HQ) is a popular high-quality image dataset scraped from Twitter. FFHQ dataset provides 70,000 PNG images of a wide range of people of different features, these images are provided with a resolution of 1024 x 1024.

What is the FFHQ dataset used for?

FFHQ is used as a means for the benchmarking of generative adversarial network models, but the wide range of age and other features allows for multiple uses on this dataset such as face aging, face generation, and metric learning.

How to download the FFHQ dataset in Python?

You can load the FFHQ dataset fast with one line of code using the open-source package Activeloop Deep Lake in Python. See detailed instructions on how to load the FFHQ dataset.

How can i use FFHQ dataset in Pytorch or TensorFlow?

You can stream the FFHQ dataset while training a model in PyTorch or TensorFlow with one line of code using the open-source package Activeloop Deep Lake in Python. See detailed instructions on how to train a model on the FFHQ dataset with Pytorch in Python or train a model on the FFHQ dataset with TensorFlow in Python.

Deep Lake community member Leon Oswaldo has contributed to this documentation. You rock, Leon! 🙂

Datasets - Previous Places205 Dataset Next - Datasets CARPK Dataset
Leaf Illustration

© 2022 All Rights Reserved by Snark AI, inc dba Activeloop