kenshohara/3D-ResNets-PyTorch
A PyTorch implementation of 3D ResNets for video-based human action recognition.
Not currently ranked — collecting fresh signals.
star history
Implements 3D convolutional residual networks (3D ResNets) for recognizing human actions in video sequences. The code supports training and testing on standard video action recognition datasets including Kinetics-700, Moments in Time, and STAIR-Actions, and includes support for distributed training, R(2+1)D model variants, and pretrained models.
Frequently asked
- What is kenshohara/3D-ResNets-PyTorch?
- A PyTorch implementation of 3D ResNets for video-based human action recognition.
- Is 3D-ResNets-PyTorch open source?
- Yes — kenshohara/3D-ResNets-PyTorch is open source, released under the MIT license.
- What language is 3D-ResNets-PyTorch written in?
- kenshohara/3D-ResNets-PyTorch is primarily written in Python.
- How popular is 3D-ResNets-PyTorch?
- kenshohara/3D-ResNets-PyTorch has 4k stars on GitHub.
- Where can I find 3D-ResNets-PyTorch?
- kenshohara/3D-ResNets-PyTorch is on GitHub at https://github.com/kenshohara/3D-ResNets-PyTorch.