softmax_with_cross_entropy

paddle.nn.functional.loss. softmax_with_cross_entropy ( logits, label, soft_label=False, ignore_index=- 100, numeric_stable_mode=True, return_softmax=False, axis=- 1 ) [source]