ML Research Wiki / Benchmarks / Action Recognition / Sports-1M

Sports-1M

Action Recognition Benchmark

Performance Over Time

📊 Showing 8 results | 📏 Metric: Video hit@1

Top Performing Models

Rank Model Paper Video hit@1 Date Code
1 ip-CSN-152 (RGB) Video Classification with Channel-Separated Convolutional Networks 92.80 2019-04-04 📦 open-mmlab/mmaction2 📦 facebookresearch/R2Plus1D 📦 facebookresearch/VMZ
2 ip-CSN-101 (RGB) Video Classification with Channel-Separated Convolutional Networks 92.60 2019-04-04 📦 open-mmlab/mmaction2 📦 facebookresearch/R2Plus1D 📦 facebookresearch/VMZ
3 R[2+1]D-Two-Stream-32frame A Closer Look at Spatiotemporal Convolutions for Action Recognition 91.90 2017-11-30 📦 microsoft/computervision-recipes 📦 open-mmlab/mmaction2 📦 facebookresearch/VMZ
4 R[2+1]D-RGB-32frame A Closer Look at Spatiotemporal Convolutions for Action Recognition 91.50 2017-11-30 📦 microsoft/computervision-recipes 📦 open-mmlab/mmaction2 📦 facebookresearch/VMZ
5 Conv pooling Beyond Short Snippets: Deep Networks for Video Classification 90.40 2015-03-31 📦 shobrook/sequitur
6 R[2+1]D-Flow-32frame A Closer Look at Spatiotemporal Convolutions for Action Recognition 88.70 2017-11-30 📦 microsoft/computervision-recipes 📦 open-mmlab/mmaction2 📦 facebookresearch/VMZ
7 P3D Learning Spatio-Temporal Representation with Pseudo-3D Residual Networks 87.40 2017-11-28 📦 qijiezhao/pseudo-3d-pytorch 📦 ZhaofanQiu/pseudo-3d-residual-networks
8 C3D Learning Spatiotemporal Features with 3D Convolutional Networks 85.50 2014-12-02 📦 open-mmlab/mmaction2 📦 facebookarchive/C3D 📦 mindspore-ai/models

All Papers (8)

open-mmlab/mmaction2 facebookresearch/R2Plus1D
open-mmlab/mmaction2 facebookresearch/R2Plus1D
microsoft/computervision-recipes open-mmlab/mmaction2
microsoft/computervision-recipes open-mmlab/mmaction2
shobrook/sequitur
microsoft/computervision-recipes open-mmlab/mmaction2
qijiezhao/pseudo-3d-pytorch ZhaofanQiu/pseudo-3d-residual-networks
open-mmlab/mmaction2 facebookarchive/C3D