📔 Table of Contents
🌟 About the Project
thingsvision is a Python package that let's you easily extract image representations from many state-of-the-art neural networks for computer vision. In a nutshell, you feed thingsvision with a directory of images and tell it which neural network you are interested in. thingsvision will then give you the representation of the indicated neural network for each image so that you will end up with one feature vector per image. You can use these feature vectors for further analyses. We use the word features for short when we mean "image representation".
README and the documentation.
🦾 Functionality
With thingsvision, you can:
- extract features for any imageset from many popular networks.
- extract features for any imageset from your custom networks.
- extract features for >26,000 images from the THINGS image database.
- optionally turn off the standard center cropping performed by many networks before extracting features.
- extract features from HDF5 datasets directly (e.g. NSD stimuli)
- conduct basic Representational Similarity Analysis (RSA) after feature extraction.
- perform Centered Kernel Alignment (CKA) to compare image features across model-module combinations.
🗄️ Model collection
Neural networks come from different sources. With thingsvision, you can extract image representations of all models from:
- torchvision
- Keras
- timm
ssl(Self-Supervised Learning Models)- OpenCLIP
- both original CLIP variants (
ViT-B/32andRN50) - a few custom models (Alexnet, VGG-16, Resnet50, and Inception_v3) trained on Ecoset rather than ImageNet and one Alexnet pretrained on ImageNet and fine-tuned on SalObjSub
- each of the many CORnet versions
- Harmonization models from the official repo. The default variant is
ViT_B16. However, the following encoders are additionally available:ResNet50,VGG16,EfficientNetB0,tiny_ConvNeXT,tiny_MaxViT,LeViT_small
🏃 Getting Started
💻 Setting up your environment
Working locally.
First, create a new conda environment with Python version 3.8, 3.9, or 3.10 e.g. by using conda:
$ conda create -n thingsvision python=3.9
$ conda activate thingsvisionThen, activate the environment and simply install thingsvision via running the following pip command in your terminal.
$ pip install --upgrade thingsvision
$ pip install git+https://github.com/openai/CLIP.gitIf you want to extract features for harmonized models from the Harmonization repo, you have to additionally run the following pip command in your thingsvision environment (FYI: as of now, this seems to be working smoothly on Ubuntu only but not on macOS),
$ pip install git+https://github.com/serre-lab/Harmonization.git
$ pip install keras-cv-attention-models>=1.3.5Google Colab.
Alternatively, you can use Google Colab to play around with thingsvision by uploading your image data to Google Drive (via directory mounting).
You can find the jupyter notebook using PyTorch here and the TensorFlow example here.
🔍 Basic usage
Command Line Interface (CLI)
thingsvision was designed to simplify feature extraction. If you have some folder of images (e.g., ./images) and want to extract features for each of these images without opening a Jupyter Notebook instance or writing a Python script, it's probably easiest to use our CLI. The interface includes two options,
thingsvision show-modelthingsvision extract-features
Example calls might look as follows:
thingsvision show-model --model-name "alexnet" --source "torchvision"
thingsvision extract_features --image-root "./data" --model-name "alexnet" --module-name "features.10" --batch-size 32 --device "cuda" --source "torchvision" --file-format "npy" --out-path "./features"See thingsvision show-model -h and thingsvision extract-features -h for a list of all possible arguments. Note that the CLI provides just the basic extraction functionalities but is probably enough for most users that don't want to dive too deep into various models and modules. If you need more fine-grained control over the extraction itself, we recommend to use the python package directly and write your own Python script.
Python commands
To do this start by importing all the necessary components and instantiating a thingsvision extractor. Here we're using AlexNet from the torchvision library as the model to extract features from and also load the model to GPU for faster inference,
import torch
from thingsvision import get_extractor
from thingsvision.utils.storing import save_features
from thingsvision.utils.data import ImageDataset, DataLoader
model_name = 'alexnet'
source = 'torchvision'
device = 'cuda' if torch.cuda.is_available() else 'cpu'
extractor = get_extractor(
model_name=model_name,
source=source,
device=device,
pretrained=True
)As a next step, create both dataset and dataloader for your images. We assume that all of your images are in a single root directory which can contain subfolders (e.g., for individual classes). Therefore, we leverage the ImageDataset class.
root='path/to/root/img/directory' # (e.g., './images/)
batch_size = 32
dataset = ImageDataset(
root=root,
out_path='path/to/features',
backend=extractor.get_backend(), # backend framework of model
transforms=extractor.get_transformations(resize_dim=256, crop_dim=224) # set input dimensionality to whatever is required for your pretrained model
)
batches = DataLoader(
dataset=dataset,
batch_size=batch_size,
backend=extractor.get_backend() # backend framework of model
)Now all that is left is to extract the image features and store them to disk! Here we're extracting features from the last convolutional layer of AlexNet (features.10), but if you don't know which modules are available for a given model, just call extractor.show_model() to print all modules.
module_name = 'features.10'
features = extractor.extract_features(
batches=batches,
module_name=module_name,
flatten_acts=True, # flatten 2D feature maps from convolutional layer
output_type="ndarray", # or "tensor" (only applicable to PyTorch models)
)
save_features(features, out_path='path/to/features', file_format='npy') # file_format can be set to "npy", "txt", "mat", "pt", or "hdf5"For more examples on the many models available in thingsvision and explanations of additional functionality like how to optionally turn off center cropping, how to use HDF5 datasets (e.g. NSD stimuli), how to perform RSA or CKA, or how to easily extract features for the THINGS image database, please refer to the Documentation.
👋 How to contribute
If you come across problems or have suggestions please submit an issue!
⚠️ License
This GitHub repository is licensed under the MIT License - see the LICENSE.md file for details.
📃 Citation
If you use this GitHub repository (or any modules associated with it), please cite our paper for the initial version of thingsvision as follows:
@article{Muttenthaler_2021,
author = {Muttenthaler, Lukas and Hebart, Martin N.},
title = {THINGSvision: A Python Toolbox for Streamlining the Extraction of Activations From Deep Neural Networks},
journal ={Frontiers in Neuroinformatics},
volume = {15},
pages = {45},
year = {2021},
url = {https://www.frontiersin.org/article/10.3389/fninf.2021.679838},
doi = {10.3389/fninf.2021.679838},
issn = {1662-5196},
}💎 Contributions
This library is based on the groundwork laid by Lukas Muttenthaler and Martin N. Hebart, who are both still actively involved, but has been extended and refined into its current form with the help of our many contributors,
- Alex Murphy (software dev.)
- Florian Mahner (software dev.)
- Hannes Hansen (software dev.)
- Johannes Roth (software dev., design, docs)
- Jonas Dippel (software dev.)
- Lukas Muttenthaler (software dev., design, docs, general responsibility)
- Martin N. Hebart (design)
- Oliver Contier (docs)
- Philipp Kaniuth (design, docs)
- Roman Leipe (sofware dev., docs),
sorted alphabetically.
This is a joint open-source project between the Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, and the Machine Learning Group at Technische Universtität Berlin. Correspondence and requests for contributing should be adressed to Lukas Muttenthaler. Feel free to contact us if you want to become a contributor or have any suggestions/feedback. For the latter, you could also just post an issue or engange in discussions. We'll try to respond as fast as we can.