0. Definitions.

  "This License" refers to version 3 of the GNU Affero General Public License.

  "Copyright" also means copyright-like laws that apply to other kinds of
works, such as semiconductor masks.

  "The Program" refers to any copyrightable work licensed under this
License.  Each licensee is addressed as "you".  "Licensees" and
"recipients" may be individuals or organizations.

        [`tf.keras.layers.MultiHeadAttention`](https://www.tensorflow.org/api_docs/python/tf/keras/layers/MultiHeadAttention).
        *   Implicit masks for `query`, `key` and `value` inputs will
            automatically be used to compute a correct attention mask for the
            layer. These padding masks will be combined with any
            `attention_mask` passed in directly when calling the layer. This can
            be used with

		// In that case (handled below), we emit a bound check (and Spectre mask)
		// and then the prove pass will remove the bounds check.
		// In theory the prove pass could potentially remove certain
		// Spectre masks, but it's very delicate and probably better
		// to be conservative and leave them all in.
		return idx
	}

	bNext := s.f.NewBlock(ssa.BlockPlain)
	bPanic := s.f.NewBlock(ssa.BlockExit)

`end = begin + 1 = 2`. To disambiguate from the original spec `2:4` we
also set the appropriate bit in `shrink_axis_mask`.

2. `2:4` is contributes 2, 4, 1 to begin, end, and stride. All masks have
zero bits contributed.

3. None is a synonym for `tf.newaxis`. This means insert a dimension of size 1
dimension in the final shape. Dummy values are contributed to begin,