huggingface/pytorch-image-models

Health v2 · maintenance only

Adoption not part of the score

37222 stars · 5216 forks observed · 2026-10-17

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-28, confidence 0.97

timm (PyTorch Image Models) is a Python library that provides a large collection of state‑of‑the‑art vision models, utilities, and training scripts for PyTorch. It includes over 700 pretrained backbones such as ResNet, EfficientNet, ViT, and MobileNet, along with data loaders, augmentations, and optimizers.

Use cases

• load a pretrained vision model for image classification
• fine‑tune a backbone on a custom dataset
• benchmark different image encoders on ImageNet
• extract intermediate features for downstream tasks
• run distributed training of large vision models
• evaluate model performance with built‑in scripts
• experiment with new transformer‑based vision architectures

When to choose

• You need a ready‑to‑use collection of SOTA vision models with pretrained weights.
• Your workflow is based on PyTorch and you want seamless integration with its ecosystem.
• You require utilities for data loading, augmentation, and training scripts out of the box.

When to avoid

• You are working in a non‑Python environment or need models for TensorFlow/JAX.
• Your project requires a lightweight library without the extensive model zoo.
• You need commercial licensing beyond Apache‑2.0.

Facets

library · maturity active

computer-vision image-processing machine-learning deep-learning artificial-intelligence computer-vision deep-learning python windows pretrained-models vision-backbones model-zoo training-scripts distributed-training linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="huggingface/pytorch-image-models")