Slimmable Network using Mixed Precision

@JiahuiYu Thank you for your great work, I have some questions about slimmable network using mixed precision.

How to use Mixed Precision when training the Slimmable Network ?

How to make the convolution below half-precision?

slimmable_networks/models/slimmable_ops.py

Lines 50 to 53 in 5dc14d0

    
           y = nn.functional.conv2d( 
        
               input, weight, bias, self.stride, self.padding, 
        
               self.dilation, self.groups) 
        
           return y

	y = nn.functional.conv2d(
	input, weight, bias, self.stride, self.padding,
	self.dilation, self.groups)
	return y