DennisJ
DennisJ
[issue](https://github.com/pytorch/pytorch/issues/17869) 貌似到现在还没有支持
您好，按照如上配置之前是好用的，但今天我尝试下载https://github.com/microsoft/onnxruntime/releases/download/v1.13.1/onnxruntime-win-x64-cuda-1.13.1.zip 这个文件时又无法下载了
感谢回复！
me too,观察train.py一直在被运行，然后爆显存
Same question,it seems like that the train.py is runned by many times,and the gpu memory is filled. did you solve the question?thanks!
@ThanhPham1987 thank you for your reply!🥸
num_workers设置为1
遇到了同样的问题，预训练模型训练后无法opt。推理模型下载下来可以直接opt。请问有解决吗 @balcklive