Projects sharing features with MARTA GAN

This is an educational curriculum for building and training neural networks using PyTorch. It serves as a deep learning training guide and resource, providing a structured series of lessons on tensor computation and architecture development. The course uses an interactive learning model that synchronizes academic theory with practice. It pairs theoretical lecture slides with exercise-driven notebooks, requiring students to implement model logic within predefined templates to validate their conceptual understanding. The curriculum covers a broad range of deep learning capabilities, including

This project is an implementation of the ALBERT language model architecture, providing a framework for training and evaluating transformer-based text classifiers and similarity models. It specifically includes pre-trained assets and tools optimized for generating semantic embeddings and representations of Chinese text. The framework distinguishes itself through tools for converting heavy language model checkpoints into lightweight formats to enable low-latency inference on mobile devices. It utilizes specific weight reduction techniques, including cross-parameter sharing and factorized embedd

TensorFlow Similarity is a Python framework designed for training neural networks to learn high-dimensional vector representations and perform similarity-based retrieval. It provides a comprehensive toolkit for metric learning, enabling the development of systems that group similar items together in vector space and identify them through distance-based comparisons. The library distinguishes itself by integrating specialized training techniques, such as contrastive and triplet-based learning, with robust data management tools that ensure stable model convergence. It supports self-supervised re

This project is a PyTorch vision transformer framework designed for self-supervised learning. It implements a model that trains visual representations using a momentum teacher and self-distillation without the need for labeled data. The library functions as an image feature extractor and visual attention visualizer, allowing for the generation of high-dimensional vectors and the rendering of self-attention maps as heatmaps or videos to analyze model focus. It provides comprehensive tools for downstream vision evaluation, including linear probe classification, k-nearest neighbor categorizatio

This is a PyTorch library and framework for self-supervised vision learning. It provides an implementation of masked autoencoders and vision transformers designed to learn image representations by reconstructing masked image patches from unlabeled data. The project features a distributed training pipeline that scales workloads across multiple GPU nodes. This infrastructure includes multi-node orchestration and gradient accumulation to manage large batch sizes and coordinate resource requests across clusters. The toolkit covers a complete workflow from self-supervised masked pre-training to d

PyTorch implementation of SwAV https//arxiv.org/abs/2006.09882

VISSL is FAIR's library of extensible, modular and scalable components for SOTA Self-Supervised Learning with images.

Self-Supervised Learning by Cross-Modal Audio-Video Clustering (NeurIPS 2020)

This repository contains the code of BadEncoder, which injects backdoors into a pre-trained image encoder such that the downstream classifiers built based on the backdoored image encoder for different downstream tasks simultaneously inherit the backdoor behavior. Here is an overview of our…

Toward Multimodal Image-to-Image Translation

CycleGAN is a generative adversarial network framework designed for unpaired image-to-image translation. It enables the conversion of images between two distinct visual domains using datasets that do not require direct one-to-one matching examples. The project implements a deep learning style transfer tool capable of artistic style transfer, object transfiguration, and domain-to-domain conversion. It uses a dual-generator architecture and cycle-consistency loss to ensure that images translated to a target domain and back recover their original state. The framework covers core machine learnin

Lightly is a self-supervised learning framework and computer vision data curation tool designed to manage large image datasets and train models on unlabeled data. It functions as a PyTorch vision library and dataset management SDK, providing tools to convert raw images into high-dimensional vectors for similarity search, visualization, and feature extraction. The project implements a variety of self-supervised architectures, including MoCo, SimCLR, VICReg, Barlow Twins, and masked image modeling. It distinguishes itself by combining these learning frameworks with active learning capabilities,

Tools for training and using unsupervised autoencoders and supervised deep learning classifiers for hyperspectral data.

We have a reimplementation of the UNIT method that is more performant. It is avaiable at Imaginaire

pix2pixHD is a conditional generative adversarial network designed to transform semantic label maps into high-resolution photorealistic images. It functions as a high-resolution image synthesizer and an image-to-image translation model capable of producing synthetic images at 2048x1024 resolution. The system includes a semantic image editor that allows for the modification of high-resolution visuals by updating the underlying semantic label maps. This enables interactive image editing and the generation of photorealistic images based on source images or discrete label maps. The framework pro

Pytorch implementation for high-resolution (e.g., 2048x1024) photorealistic video-to-video translation. It can be used for turning semantic label maps into photo-realistic videos, synthesizing people talking from edge maps, or generating human motions from poses. The core of video-to-video…

Multimodal Unsupervised Image-to-Image Translation

pix2pix is a framework for image-to-image translation using conditional generative adversarial networks. It functions as a supervised trainer and visual domain mapper designed to learn a mapping between input and output images for style and domain transfer. The system utilizes a U-Net encoder-decoder architecture combined with a PatchGAN local discriminator to enforce high-frequency local consistency. It employs L1 loss regularization to ensure generated outputs remain structurally close to the ground truth. The project covers a broad range of computer vision capabilities, including semantic

Pytorch implementation for the paper "Structural-analogy from a Single Image Pair"

Contrastive unpaired image-to-image translation, faster and lighter training than cyclegan (ECCV 2020, in PyTorch)