pytorchcv

What it is and what it does

pytorchcv is a model zoo for PyTorch that bundles implementations of dozens of computer vision architectures—from classic networks like AlexNet, VGG, and ResNet to modern designs like EfficientNet, MobileNetV3, and HRNet. Models are pretrained on standard datasets including ImageNet-1K, CIFAR-10/100, COCO, and others, and load with ordinary normalization applied automatically. It covers classification, semantic segmentation, object detection, and pose estimation tasks.

The package depends on torch, torchvision, numpy, and requests. It is intended for researchers and practitioners who want quick access to reference architectures without reimplementing them. Training and evaluation scripts are maintained separately in the imgclsmob repository. The package is classified as Alpha and has not seen active development recently, though the repository remains accessible.

Use it for:

• Load a pretrained ResNet or EfficientNet for image classification without writing model code from scratch.
• Benchmark multiple architectures on a custom dataset using the same normalization and interface.
• Extract features from intermediate layers of pretrained models for transfer learning tasks.
• Evaluate segmentation models like PSPNet or DeepLabv3 on Pascal VOC or ADE20K without reimplementing them.
• Prototype mobile-efficient models like MobileNetV3 or ShuffleNet for deployment scenarios.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides a collection of pretrained computer vision models for PyTorch, including image classification, segmentation, detection, and pose estimation architectures.

Yes, if you need quick access to a broad range of pretrained PyTorch vision models and can tolerate aging maintenance. The package has no known vulnerabilities and low install friction. However, verify the license terms in the repository first, and confirm that model loading and normalization behavior match your requirements—the lack of recent updates and unclear license status warrant a brief review before production use.

Install

pytorchcv on PyPI

pip install pytorchcv

uv add pytorchcv

poetry add pytorchcv

Installing pytorchcv

Before you install

Low install friction with a pure Python wheel. Maintenance status is aging—last commit was 2026-01-13 and the package has received no releases in 213 days, though the repository remains active and not archived.

License in practice

License status is unclear; no SPDX identifier or raw license text is available in the package metadata. Verify the actual license terms in the repository before use in proprietary or restricted contexts.

Quickstart

pip install pytorchcv

import pytorchcv
model = pytorchcv.get_model('resnet50', pretrained=True)
# model is ready for inference or feature extraction

Requires PyTorch and torchvision to be installed; pretrained models download on first use and may require network access and disk space.

Verify before relying

• Whether all pretrained models load automatically or require manual download/caching during first use.
• Specific PyTorch and torchvision version compatibility constraints beyond the stated Python >=3.10 requirement.
• Whether the package includes training/evaluation scripts or only inference-ready models.

Package facts

Evidence: pytorchcv-0.0.74-py3-none-any.whl

Keywords: machine-learning, deep-learning, neuralnetwork, image-classification, pytorch, imagenet, cifar, svhn, vgg, resnet, pyramidnet, diracnet, densenet, condensenet, wrn, drn, dpn, darknet, fishnet, espnetv2, xdensnet, squeezenet, squeezenext, shufflenet, menet, mobilenet, igcv3, mnasnet, darts, xception, inception, polynet, nasnet, pnasnet, ror, proxylessnas, dianet, efficientnet, mixnet, image-segmentation, voc, ade20k, cityscapes, coco, pspnet, deeplabv3, fcn

Tags

More Image Recognition packages

Further reading