vitalwarley / research

2 stars 0 forks source link

Experimentos futuros #85

Open vitalwarley opened 1 month ago

vitalwarley commented 1 month ago
vitalwarley commented 1 month ago
vitalwarley commented 1 month ago

Otimizar a perda contrastiva via $L{align}$ e $L{uniformity}$, como exposto em Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere.

image