*   `tf.keras`:

    *   `Model.fit` major improvements:
        *   You can now use custom training logic with `Model.fit` by overriding
            `Model.train_step`.
        *   Easily write state-of-the-art training loops without worrying about
            all of the features `Model.fit` handles for you (distribution
            strategies, callbacks, data formats, looping logic, etc)

*  basis, without warranty of any kind, either expressed, implied, or  *
*  statutory, including, without limitation, warranties that the       *
*  Covered Software is free of defects, merchantable, fit for a        *
*  particular purpose or non-infringing. The entire risk as to the     *
*  quality and performance of the Covered Software is with You.        *

<ol>
<li>
When converting between <a href="#Numeric_types">integer types</a>, if the value is a signed integer, it is
sign extended to implicit infinite precision; otherwise it is zero extended.
It is then truncated to fit in the result type's size.
For example, if <code>v := uint16(0x10F0)</code>, then <code>uint32(int8(v)) == 0xFFFFFFF0</code>.
The conversion always yields a valid value; there is no indication of overflow.
</li>
<li>