Float16 values are unconditionally output in scientific notation

[KStocky]

Describe the bug
As the title says. If a float16 value is appears in an assert message it is formatted in scientific notation no matter what the user wants. Although the value printed is correct, it is still not always what the user wants.

To Reproduce
Steps to reproduce the behaviour including the following:

1. Write a test which asserts that two differing float16_t values are equal
2. Run the test
3. Observe that the float16 values are output in scientific notation

Expected behavior
I would prefer no scientific notation unless I wanted it formatted as such

Additional context
This comes from the operator << in the float16 library. The fix could be just to fix that output.