Is there any support for automatic downcasting float32 model to float16

Questions
#1

Hi all,

I looking for automatic downcasting the float32 model to float16 after loading in tvm. If there is support for this please point me to the documentation or example to get information how to use.

thanks