List your company

Active Speaker Detection Inference Efficiency Profiling

1.25VRAM (GB)LR-ASD
0.84523.57766.319.0424Dec 17, 2025

Evaluation results

MethodLinks
LR-ASDVideo frames=10002025.121.2538.6325,890
TalkNetVideo frames=10002025.121.8948.6520,556
LR-ASDVideo frames=20002025.122.578.6425,431
GateFusionVideo frames=10002025.123.3577.212,953
TalkNetVideo frames=20002025.123.52102.5219,508
GateFusionVideo frames=20002025.124.96174.3711,470
LR-ASDVideo frames=40002025.124.99156.0925,627
LoCoNetVideo frames=10002025.125.02135.457,383
TalkNetVideo frames=40002025.126.96213.8818,702
LR-ASDVideo frames=60002025.127.48236.3225,390
GateFusionVideo frames=40002025.128.17423.379,448
LoCoNetVideo frames=20002025.1210.19274.977,273
TalkNetVideo frames=60002025.1210.32328.0118,292
GateFusionVideo frames=60002025.1211.37749.568,005