Res2Net/Res2Net-PretrainedModels

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1114 stars · 216 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Official PyTorch implementation of Res2Net, a multi-scale CNN backbone architecture published in TPAMI, with ImageNet-pretrained model weights. The Res2Net block adds hierarchical residual-like connections within a single residual block and can be plugged into backbones like ResNet, ResNeXt, and DLA.

Use cases

• load pretrained res2net backbone for image classification
• use res2net as backbone for object detection in mmdetection
• improve semantic segmentation with a multi-scale cnn backbone
• replace resnet backbone with res2net in detectron2
• fine-tune imagenet pretrained res2net on my own dataset
• compare res2net vs hrnet for detection performance

When to choose

• you need a strong multi-scale CNN backbone for detection or segmentation
• you want ImageNet-pretrained weights that transfer well to downstream vision tasks
• you use PyTorch, MMDetection, Detectron2, or Jittor and want drop-in Res2Net support

When to avoid

• you need a maintained library with active development or a license
• you want transformer-based backbones like ViT or Swin
• you need training code in this repo (training lives in a separate repository)

Facets

library · maturity maintenance

machine-learning deep-learning image-processing computer-vision deep-learning computer-vision machine-learning python res2net pretrained-models backbone-network pytorch imagenet multi-scale-features model-zoo gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Res2Net/Res2Net-PretrainedModels")