GTC ON-DEMAND

 
SEARCH SESSIONS
SEARCH SESSIONS

Search All
 
Refine Results:
 
Year(s)

SOCIAL MEDIA

EMAIL SUBSCRIPTION

 
 

GTC ON-DEMAND

Artificial Intelligence and Deep Learning
Presentation
Media
Multi-Label ConvNets vs Dedicated ConvNets per Label
Abstract:
When training multi-label model sizes of 100,000+, do you fine-tune one per label? Or use ResNet 100, replace softmax with sigmoid and train? The first has terrible run time but high acc, the second has great run time but terrible acc. We'll show a new architecture that has the best of both worlds.
 
Topics:
Artificial Intelligence and Deep Learning, AI Startup, Computer Vision
Type:
Talk
Event:
GTC Silicon Valley
Year:
2017
Session ID:
S7619
Download:
Share: