search

JasonWayne / deep-learning-essay

0 stars 0 forks source link

GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks #234

Open JasonWayne opened 3 years ago

JasonWayne commented 3 years ago

One sentence

自动平衡多目标模型中回传的梯度

Paper

https://arxiv.org/abs/1711.02257

Code

Motivation

Novelties & Key contribution

Method

Experiments

Rethinking

Reference