Tools to generate and use multi-object datasets. The datasets consist of images and a dictionary of labels, where each image is labeled with 1) the number of objects in it and 2) each object's attributes.
Using datasets only requires numpy as datasets are .npz.
Generating sprites requires scikit-image. Tools for using the
datasets in PyTorch are provided, with usage examples.
- Either download one of the datasets in
generated/, or generate a new one. - Place the
.npzdataset in/path/to/data/. pip install multiobject- Usage in PyTorch:
from multiobject.pytorch import MultiObjectDataLoader, MultiObjectDataset dataset_path = '/path/to/data/some_dataset.npz' train_set = MultiObjectDataset(dataset_path, train=True) test_set = MultiObjectDataset(dataset_path, train=False) train_loader = MultiObjectDataLoader(train_set, batch_size=batch_size, shuffle=True) test_loader = MultiObjectDataLoader(test_set, batch_size=test_batch_size)
conda create --name multiobject python=3.7
conda activate multiobject
pip install -r requirements.txt
CUDA_VISIBLE_DEVICES=0 python demo_vae.py
CUDA_VISIBLE_DEVICES=0 python demo_count.py
Datasets are available as .npz files in ./generated/.
dSprites1
Binary RGB images with monochromatic dSprites on a black canvas. Sprites can overlap (sum and clip).
| n images | size | max object size |
objects per image |
sprite colors | file size |
|---|---|---|---|---|---|
| 100k | 64x64 | 18x18 | 1 | 7 | 10.6 MB |
| 100k | 64x64 | 28x28 | 1 | 7 | 12.4 MB |
| 100k | 64x64 | 18x18 | 0–2 (uniformly) | 7 | 11.0 MB |
Binary 64x64 single-channel images with MNIST digits on a black canvas. Digits are rescaled to 18x18 and binarized, and they can overlap (sum and clip). Only digits from the MNIST training set are used (60k).
| n images | size | max object size |
objects per image |
file size |
|---|---|---|---|---|
| 100k | 64x64 | 18x18 | 1 | 4.5 MB |
| 100k | 64x64 | 18x18 | 0–2 (uniformly) | 4.8 MB |
-
Clone this repo.
-
See requirements, or set up a virtual environment as follows:
conda create --name multiobject python=3.7 conda activate multiobject pip install -r requirements.txt -
Optional: generate a new type of sprites:
- create a file
sprites/xyz.pycontaining a functiongenerate_xyz(), where "xyz" denotes the new sprite type - in
generate_dataset.py, add a call togenerate_xyz()to generate the correct sprites, and add'xyz'to the list of supported sprites
- create a file
-
Call
generate_dataset.pywith the desired sprite type as--typeargument. Example:python generate_dataset.py --type dsprites
The sprite attributes are managed automatically when generating a dataset from a set of sprites that have per-sprite labels. However, since they are dataset-specific, they have to be defined when creating the sprites.
Note. For now, the following parameters have to be customized in generate_dataset.py directly:
- probability distribution over number of objects
- image size
- sprite size
- dataset size
- whether sprites can overlap
To generate datasets:
numpy==1.18.1
matplotlib==3.1.2
scikit_image==0.16.2
tqdm==4.41.1
pillow==7.0.0
To run the examples or use the pytorch tools:
torch==1.4.0
torchvision==0.5.0
1 This is actually an extension of the original dSprites dataset to many objects and to color images. ↩
