mindformers.wrapper.MFTrainOneStepCell

class mindformers.wrapper.MFTrainOneStepCell(network, optimizer, use_clip_grad=False, max_grad_norm=1.0, scale_sense=1.0, **kwargs)

TrainOneStep For MindFormer. Network training with loss scaling, grad clip, gradient accumulation, exponential moving average and so on.

This is a training step with loss scaling. It takes a network, an optimizer and a scale update Cell(or a Tensor) as args. The loss scale value can be updated in both host side or device side. If you want to update it on host side, using a value of Tensor type as scale_sense, otherwise, using a Cell instance for updating loss scale as scale_sense.

参数

• network (Cell) – The training network. The network only supports single output.

• optimizer (Cell) – Optimizer for updating the network parameters.

• use_clip_grad (bool) – Whether to use the gradient clipping function. Default: False.

• max_grad_norm (float) – Maximum gradient value. Default: 1.0.

• scale_sense (Union[Tensor, Cell]) – If this value is a Cell, it will be called by MFTrainOneStepCell to update loss scale. If this value is a Tensor, the loss scale can be modified by set_sense_scale, the shape should be \(()\) or \((1,)\).

Inputs:

• (*inputs) (Tuple(Tensor)) - Tuple of input tensors with shape \((N, \ldots)\).

Outputs:

Tuple of 3 Tensor, the loss, overflow flag and current loss scale value.

• loss (Tensor) - A scalar, the loss value.

• overflow (Tensor) - A scalar, whether overflow occur or not, the type is bool.

• loss scale (Tensor) - The loss scale value, the shape is \(()\) or \((1,)\).

引发

• TypeError – If scale_sense is neither Cell nor Tensor.

• ValueError – If shape of scale_sense is neither (1,) nor ().