mindspore.experimental.optim.lr_scheduler.StepLR
- class mindspore.experimental.optim.lr_scheduler.StepLR(optimizer, step_size, gamma=0.1, last_epoch=- 1)[源代码]
每 step_size 个epoch按 gamma 衰减每个参数组的学习率。StepLR 对于学习率的衰减可能与外部对于学习率的改变同时发生。
警告
这是一个实验性的动态学习率接口,需要和 mindspore.experimental.optim 下的接口配合使用。
- 参数:
optimizer (
mindspore.experimental.optim.Optimizer
) - 优化器实例。step_size (int) - 学习率衰减的周期。
gamma (float,可选) - 学习率衰减的乘法因子。默认值:
0.1
。last_epoch (int,可选) - 当前scheduler的 step() 方法的执行次数。默认值:
-1
。
- 支持平台:
Ascend
GPU
CPU
样例:
>>> import mindspore >>> from mindspore import nn >>> from mindspore.experimental import optim >>> # Define the network structure of LeNet5. Refer to >>> # https://gitee.com/mindspore/docs/blob/r2.4.1/docs/mindspore/code/lenet.py >>> net = LeNet5() >>> loss_fn = nn.SoftmaxCrossEntropyWithLogits(sparse=True) >>> optimizer = optim.Adam(net.trainable_params(), lr=0.05) >>> # Assuming optimizer uses lr = 0.05 for all groups >>> # lr = 0.05 if epoch < 2 >>> # lr = 0.005 if 2 <= epoch < 4 >>> # lr = 0.0005 if 4 <= epoch < 6 >>> scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=2, gamma=0.1) >>> def forward_fn(data, label): ... logits = net(data) ... loss = loss_fn(logits, label) ... return loss, logits >>> grad_fn = mindspore.value_and_grad(forward_fn, None, optimizer.parameters, has_aux=True) >>> def train_step(data, label): ... (loss, _), grads = grad_fn(data, label) ... optimizer(grads) ... return loss >>> for epoch in range(6): ... # Create the dataset taking MNIST as an example. Refer to ... # https://gitee.com/mindspore/docs/blob/r2.4.1/docs/mindspore/code/mnist.py ... for data, label in create_dataset(): ... train_step(data, label) ... scheduler.step() ... current_lr = scheduler.get_last_lr()