Self-Supervised MultiModal Versatile Networks

Benchmark Model Rank Results
audio-classification-on-audiosetMMV#38Test mAP: 0.309
self-supervised-action-recognition-on-hmdb51-1MMV#2Top-1 Accuracy: 70.1
self-supervised-action-recognition-on-kinetics-2MMV#5Top-1 Accuracy: 55.5
self-supervised-action-recognition-on-ucf101MMV TSM-50x2#73-fold Accuracy: 95.2Pre-Training Dataset: Audioset + Howto100M
self-supervised-action-recognition-on-ucf101-1MMV#73-fold Accuracy: 91.5