search

kenziyuliu / MS-G3D

[CVPR 2020 Oral] PyTorch implementation of "Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition"
https://arxiv.org/abs/2003.14111
MIT License
430 stars 96 forks source link

Vs code运行代码直接终止，并且没有报错 #45

Closed MosasaurYi closed 2 years ago

MosasaurYi commented 3 years ago

你好，请问我复现该代码的时候，使用VS code 和anaconda来跑，环境都没问题，在跑的时候，进入第一次训练，然后程序就直接终止了了，也没有任何的报错，请问这怎么解决呢？

Anirudh257 commented 3 years ago

@MosasaurYi Can you post the screenshot of your error? I was able to run successfully using the python virtualenvironment.

MosasaurYi commented 3 years ago

here,it stop directly without any error ------------------ 原始邮件 ------------------ 发件人: "kenziyuliu/MS-G3D" @.>; 发送时间: 2021年7月19日(星期一) 晚上7:30 @.>; @.**@.>; 主题: Re: [kenziyuliu/MS-G3D] Vs code运行代码直接终止，并且没有报错 (#45)

@MosasaurYi Can you post the screenshot of your error? I was able to run successfully using the python virtualenvironment.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

MosasaurYi commented 3 years ago

@MosasaurYi Can you post the screenshot of your error? I was able to run successfully using the python virtualenvironment.

I find my code will stop in “loss.backward()”，and without error，do you know why，thanks

Anirudh257 commented 3 years ago

Decrease your batch-size and see if your GPU memory is getting full.

MosasaurYi commented 3 years ago

Decrease your batch-size and see if your GPU memory is getting full.

it is all ok,when i debug it,i find the tensor of loss was empty and the error “RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED”，is there any problem with the version of cuda and cudann,cuda9.2 and cudann 7.6.5for cuda92 in my computer,What version are you using?Thanks for reply

Anirudh257 commented 3 years ago

I am using python3.5, cuda 10, cudnn 7 on ubuntu.

kenziyuliu commented 2 years ago

Hi there, sorry for the late reply. Have you resolved the issue? This looks like a problem relating to your local environment setup.