Closed Vaibhavdixit02 closed 1 year ago
This isn't sufficient, the C wrappers of other optimizers are hardcoded to give out Float64s so we can't make this handling completely automatic. The loss function would need to do a Float32 conversion despite this. But this is probably good to have even otherwise so merging this.
Codecov Report
0.00% <0.00%> (ø)
0.00% <ø> (ø)
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more