wkcn / MobulaOP

A Simple & Flexible Cross Framework Operators Toolkit
MIT License
164 stars 21 forks source link

[Fix Bug] Crash when training model on multiple GPUs #63

Closed wkcn closed 4 years ago

wkcn commented 4 years ago

This PR will fix the bug which crash when training model on multiple GPUs. #62

The reason is that NDArrays are released before the function runs.

However, this PR will allocate the memory before the function runs rather than during the function runs. I will try to find a better solution.

coveralls commented 4 years ago

Pull Request Test Coverage Report for Build 687


Totals Coverage Status
Change from base Build 680: 0.0%
Covered Lines: 1526
Relevant Lines: 1787

💛 - Coveralls