Catalyst is a PyTorch framework for Deep Learning Research and Development.
It focuses on reproducibility, rapid experimentation, and codebase reuse so you can create something new rather than write yet another train loop.
Break the cycle - use the Catalyst!
Get Started View Documentation
Deep learning – introduction, backpropagation algorithm. Optimization methods. Neural Network in numpy.
Regularization methods and deep learning frameworks. Pytorch basics & extras.
CNN. Model Zoo. Convolutional kernels. ResNet. Simple Noise Attack.
Object Detection. (One, Two)-Stage methods. Anchors. Image Segmentation. Up-scaling. FCN, U-net, FPN. DeepMask.
Metric Learning. Contrastive and Triplet Loss. Samplers. Cross Entropy Loss modifications. SphereFace, CosFace, ArcFace.
AutoEncoders. Denoise, Sparse, Variational. Generative Models. Autoregressive models.
Generative Adversarial Networks. VAE-GAN. AAE. Energy based model.
Pretrained Transformers. BERT. GPT. Data Augmentation in Texts. Domain Adaptation.
Reinforcement Learning. DQN Algorithm. DDPG Algorithm. Wolpertinger.
Pipeine for Image Super-Resolution task that based on a frequently cited paper, ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (Wang Xintao et al.), published in 2018.