Can I package the Torch-GPU with PyInstaller

[zhanghongyong123456]

I have a question. I have a Python project that is implemented on Torch-GPU. Can I package it with PyInstaller and run it independently on other computers, or what do I need to configure to run it on other computers.

I have tried a demo, which can be used normally on this machine, but this bug appears on other computers, I hope to get your reply

[rokm]

You probably need to install Visual C++ Redistributable on the target computers.

Or, try using PyInstaller develop branch, which has improved collection of those VC runtime DLLs.

[zhanghongyong123456]

I have tried both methods and still got the same error. Could you please give me another method? Thank you very much

[rokm]

Then you're on your own; try opening the caffe2_detectron_ops_gpu.dll with dependency walker on the target computer and see what DLL it is missing.

[zhanghongyong123456]

I have tried both methods and still got the same error. Could you please give me another method? Thank you very much