在Pytorch中可以利用[torch.cuda.memory_allocated()](https://pytorch.org/docs/stable/generated/torch.cuda.memory_allocated.html?highlight=torch%20cuda%20memory_allocated#torch.cuda.memory_allocated)、[torch.cuda.max_memory_allocated()](https://pytorch.org/docs/stable/generated/torch.cuda.max_memory_allocated.html?highlight=torch%20cuda%20memory_allocated#torch.cuda.max_memory_allocated)和[torch.cuda.memory_reserved()](https://pytorch.org/docs/stable/generated/torch.cuda.memory_reserved.html?highlight=torch%20cuda%20memory_reserved#torch.cuda.memory_reserved)对模型进行显存占用分析,具体可参看[PyTorch显存机制分析](https://zhuanlan.zhihu.com/p/424512257). 但是,我在paddle中没有找到对应的api,请问paddle中有相应的api吗? 谢谢!