Design
Specification
API
API Mapping
Migration Guide
FAQ
RELEASE NOTES
Converts input tensor dtype to float16.
Tensor, converted to the float16 dtype.
Ascend GPU CPU
Ascend
GPU
CPU
Examples
>>> input_x = Tensor(np.ones([2,2]), mindspore.int32) >>> output = input_x.half() >>> print(output.dtype) Float16