Expand description
The vision
module groups functions and models related to
computer vision.
Modules
- AlexNet. https://arxiv.org/abs/1404.5997
- The CIFAR-10 dataset.
- ConvMixer implementation.
- A simple dataset structure shared by various computer vision datasets.
- DenseNet implementation.
- DINOv2: Learning Robust Visual Features without Supervision https://github.com/facebookresearch/dinov2 The weights can be extracted from pre-trained Python models using
python src/vision/export_dinov2.py
. - EfficientNet implementation.
- Utility functions to manipulate images.
- Helper functions for ImageNet like datasets.
- InceptionV3.
- The MNIST hand-written digit dataset.
- MobileNet V2 implementation. https://ai.googleblog.com/2018/04/mobilenetv2-next-generation-of-on.htmla
- ResNet implementation.
- SqueezeNet implementation.
- VGG models.