报错：
If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
当reserved memory is >> allocated memory，进行如下设置，可解决此bug：
代码如下：
import os
os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "max_split_size_mb:128"
文章出处登录后可见！
已经登录？立即刷新
