{"categories":[{"label":"Artificial Intelligence","url":"https://skillfed.io/packages/category/scientific-engineering-artificial-intelligence/8"}],"enrichment":{"capability":"Provides PyTorch implementations of Vision Transformer (ViT) and related transformer-based vision models for image classification and self-supervised learning tasks.","skillfed_tags":["transformer-vision","self-supervised-learning","knowledge-distillation"],"use_cases":["Train a vision transformer from scratch on a custom image classification dataset using the base ViT or SimpleViT variant","Distill knowledge from a pretrained ResNet or other CNN teacher into a smaller, faster ViT student model","Implement masked image modeling or masked autoencoder pretraining for self-supervised representation learning","Experiment with variable-resolution image batching using NaViT for faster training on mixed-size datasets","Deploy efficient vision models on resource-constrained devices using MobileViT or LeViT variants"],"what_it_does":"vit-pytorch is a collection of PyTorch implementations of Vision Transformer architectures and related vision models. It provides the core ViT model alongside numerous variants\u2014SimpleViT, NaViT, CaiT, Token-to-Token ViT, LeViT, MobileViT, and others\u2014each addressing different efficiency or accuracy trade-offs. The package also includes support for self-supervised learning techniques like masked autoencoders and masked image modeling, as well as knowledge distillation from convolutional networks.\n\nThe library is designed for researchers and practitioners building image classification systems with transformers. It depends on torch for computation, torchvision for standard vision utilities, and einops for flexible tensor reshaping. Models are instantiated with configurable parameters (patch size, embedding dimension, depth, attention heads, dropout rates) and accept batches of images as input, returning class predictions or intermediate representations.","worth_installing":"Yes. The package is actively maintained, has low install friction, carries permissive MIT licensing, and offers a comprehensive suite of transformer vision models backed by recent research. It is well-suited for both research prototyping and production image classification systems. No known vulnerabilities."},"id":"vit-pytorch","links":{"html":"https://skillfed.io/packages/vit-pytorch","md":"https://skillfed.io/packages/vit-pytorch.md","pypi":"https://pypi.org/project/vit-pytorch/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2026-08-02","license_spdx":null,"license_treatment":"permissive","name":"vit-pytorch","python_support":"supports_current","summary":"Vision Transformer (ViT) - Pytorch"},"popularity":{"monthly_downloads":184548,"position":10033,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"1.24.2"}