an error in model_stage1.train(epoch) by the test 10X_data
Opened this issue · 7 comments
Hello,authors:
I tested the test data_10X which from author and got the same error with my data. The error is the follow:
I am so confused that why the test data also came out the same problem, I would appreciate it if you could reply to me soon. Thanks!
@fengweimin-maker
Did you solve this problem? I also got the same error using example data.
Thanks.
@YingxinLin
Could you help us to fix this error? Many Thanks
Would you like to test the data downloaded from here: http://www.maths.usyd.edu.au/u/yingxinl/wwwnb/scJoint/data_10x.zip.
Also I am wondering what is your pytorch and python version.
Best wishes,
Yingxin
Hi @YingxinLin
My python version is Python 3.6.13, pytorch version is 1.0.0.
I will gave a try.
Best,
Xiaowen
Hi @YingxinLin My python version is Python 3.6.13, pytorch version is 1.0.0. I will gave a try. Best, Xiaowen
Did you submit a task on cluster server? I finall found it came out such an error by submitting tasks,but it can work on a computer node.
Hi @fengweimin-maker
Thank you for your suggestion. I submitted tasks to run the example data. Because if I run it on command line, I can not use the GPU, then I got the error "no CUDA-capable device is detected".
If you cannot use GPU from your current device, you can try set self.use_cuda = False in the config.py to disable the GPU usage in scJoint.
Yingxin
@YingxinLin
Sorry about the late reply.
First, I tried the data downloaded from here: http://www.maths.usyd.edu.au/u/yingxinl/wwwnb/scJoint/data_10x.zip.
I still got the same error.
Second, I have GPU in our cluster, however, I can only use GPU when I submit the job to HPC. so I can not run "python main.py" on command line when I set "self.use_cuda = True".
Third, like fengweimin-maker's mentioned, It works when I run "python main.py" on command line, when I set "self.use_cuda = False". However, if I submit a job using *.sbatch file which still use CPU, I got the same error again.
Thanks.
Best,
Xiaowen