Self-Supervised Audio-Visual Representation Learning with Relaxed Cross-Modal Synchronicity

Open paper
Benchmark Model Rank Results
audio-classification-on-dcaseCrissCross (AudioSet)#1Top-1 Accuracy: 97PRE-TRAINING DATASET: AudioSet
audio-classification-on-dcaseCrissCross (Kinetics-400)#2Top-1 Accuracy: 96PRE-TRAINING DATASET: Kinetics-400
audio-classification-on-dcaseCrissCross (Kinetics-Sound)#5Top-1 Accuracy: 93PRE-TRAINING DATASET: Kinetics-Sound
self-supervised-action-recognition-on-hmdb51CrissCross (AudioSet)#11Top-1 Accuracy: 66.8Pre-Training Dataset: AudioSetFrozen: false
self-supervised-action-recognition-on-hmdb51CrissCross (Kinetics400)#18Top-1 Accuracy: 64.7Pre-Training Dataset: Kinetics400Frozen: false
self-supervised-action-recognition-on-hmdb51CrissCross (Kinetics-Sound)#25Top-1 Accuracy: 60.5Pre-Training Dataset: Kinetics-SoundFrozen: false
self-supervised-action-recognition-on-ucf101CrissCross (AudioSet)#153-fold Accuracy: 92.4Pre-Training Dataset: AudioSetFrozen: false
self-supervised-action-recognition-on-ucf101CrissCross (Kinetics400)#183-fold Accuracy: 91.5Pre-Training Dataset: Kinetics400Frozen: false
self-supervised-action-recognition-on-ucf101CrissCross (Kinetics-Sound)#243-fold Accuracy: 88.3Pre-Training Dataset: Kinetics-Sound…

Rank counts only results with a code link.