MHLR: Moving Haar Learning Rate Scheduler for Large-scale Face Recognition Training with One GPU.
Xueyuan GongYain-Whar SiZheng ZhangXiaochen YuanKe WangXinyuan ZhangCong LinXiaoxiang LiuPublished in: CoRR (2024)
Keyphrases
- learning rate
- multilayer neural networks
- face recognition
- training speed
- training algorithm
- adaptive learning rate
- learning algorithm
- convergence rate
- hidden layer
- real time
- feed forward neural networks
- error function
- training process
- rapid convergence
- feedforward neural networks
- convergence speed
- convergence theorem
- real world
- weight vector
- activation function
- face images
- supervised learning
- feature extraction
- delta bar delta
- bp neural network
- dynamic programming
- artificial neural networks
- feature selection
- bp neural network algorithm