Revisiting Self-Supervised Visual Representation Learning
This codebase allows to reproduce core experiments from the paper "Revisiting Self-Supervised Visual Representation Learning" by A. Kolesnikov, X. Zhai, and L. Beyer. It contains …
Keywords: self-supervised representation learning, machine learning, computer vision, visual representation learning, TensorFlow
RLStructures is a library to facilitate the implementation of new reinforcement learning algorithms. It includes a library, a tutorial, and different RL algorithms provided as …
Keywords: reinforcement learning, Python
An improved version of Differentiable Architecture Search (DARTS) with added regularization that lowers solution curvature and improves generalization properties.
Keywords: neural architecture search, differentiable architecture search, DARTS, machine learning, Python, PyTorch
Official implementation of ROLAND for creating dynamic GNN from any static GNN. ROLAND is adaptive and scalable.
Keywords: graph neural networks, GNN, dynamic graphs, machine learning, Python, PyTorch, PyG
Official PyTorch implementation of the RotNet models proposed in the paper ""Unsupervised Representation Learning by Predicting Image Rotations" by S. Gidaris, P. Singh, and N. …
Keywords: RotNet, visual representation learning, unsupervised learning, self-supervised learning, computer vision, image understanding, machine learning
RouteRAG is a novel approach that trains language models with reinforcement learning to dynamically decide when to reason, which type of retrieval to use (passage, …
Keywords: retrieval-augmented generation, RAG, reinforcement learning, Python, PyTorch
The reference implementation for Self-Attention graph pooling (SAGPool) for graph classification and regression.
Keywords: graph classification, geometric deep learning, self-attention, graph pooling, PyTorch, Python
Segment Anything Model 2 (SAM 2) is a foundation model towards solving promptable visual segmentation in images and videos. We extend SAM to video by …
Keywords: computer vision, image segmentation, deep learning, SAM, foundational model, segment anything model, semantic segmentation, META, PyTorch, Python
The reference implementation, configuration files and pretrained models for the paper "SCAN: Learning to Classify Images without Labels". A new SOTA method for unsupervised classification …
Keywords: SCAN, computer vision, unsupervised image classification, self supervised learning, deep learning, neural networks
The official implementation of Scattering GCN model that combines graph convolutional networks with geometric scattering in order to tackle the over-smoothing problem in graph neural …
Keywords: scattering GCN, over-smoothing, graph signal processing, geometric scattering, graph neural networks, graph machine learning, geometric deep learning, Python, PyTorch