4 repository-uri
Awesome GitHub RepositoriesStochastic Depth Regularization
Regularization techniques that randomly drop network paths during training to prevent overfitting.
Distinct from Overfitting Reduction Techniques: Specifically implements stochastic depth as a path-dropping regularization, distinct from general data-driven overfitting reduction.
Explore 4 awesome GitHub repositories matching artificial intelligence & ml · Stochastic Depth Regularization. Refine with filters or upvote what's useful.
Awesome Stochastic Depth Regularization GitHub Repositories
Găsește cele mai bune repo-uri cu AI.Vom căuta cele mai potrivite repository-uri folosind AI.
- facebookresearch/convnext
facebookresearch/ConvNeXt
6,388Vezi pe GitHubCode release for ConvNeXt model
Applies stochastic depth regularization by randomly dropping residual blocks during training.
Python - mosaicml/composer
mosaicml/composer
5,485Vezi pe GitHubComposer este un framework de antrenare distribuită PyTorch conceput pentru scalarea modelelor de mari dimensiuni pe clustere GPU multi-nod. Acesta funcționează ca un antrenor de modele lingvistice mari (LLM), un optimizator de modele distribuite și un manager al ciclului de viață al antrenării. Proiectul se diferențiază ca o bibliotecă de regularizare pentru deep learning, oferind tehnici de optimizare specializate precum Sharpness Aware Minimization, MixUp și CutMix pentru a îmbunătăți generalizarea modelului. De asemenea, distinge fluxul de antrenare prin utilizarea warmup-ului pentru lungimea secvenței, înghețarea progresivă a straturilor și checkpointing-ul stării sharded pentru recuperarea modelelor la scară largă. Framework-ul acoperă o suprafață largă de capabilități, inclusiv orchestrarea antrenării distribuite, gestionarea hardware-ului cu precizie mixtă și streaming-ul de date cloud-native. Oferă, de asemenea, instrumente extinse de monitorizare și observabilitate pentru diagnosticarea memoriei GPU, detectarea divergenței antrenării și urmărirea throughput-ului. Proiectul include un launcher CLI pentru automatizarea execuției joburilor de antrenare multi-GPU pe mai multe noduri.
Provides stochastic depth regularization to randomly drop network paths during training and prevent overfitting in deep models.
Python - ai-dawang/plugnplay-modules
ai-dawang/PlugNPlay-Modules
4,968Vezi pe GitHubPlugNPlay-Modules is a collection of reusable PyTorch computer vision modules and deep learning architectural components. It provides a library of standardized building blocks for constructing neural networks, focusing on attention mechanisms, signal processing layers, and feature fusion modules. The project is distinguished by its extensive variety of attention primitives, covering spatial, channel, and temporal weighting, as well as specialized variants like deformable, frequency-enhanced, and linear-complexity attention. It also implements advanced signal processing tools within the neural
Implements stochastic depth to randomly drop sample paths during training to improve model generalization.
Python - lightly-ai/lightly
lightly-ai/lightly
3,684Vezi pe GitHubLightly 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,
Injects dropout into transformer blocks via stochastic depth regularization to prevent overfitting during training.
Pythoncomputer-visioncontrastive-learningcontributions-welcome
