`tf.sparse.cross`/`tf.ragged.cross` directly.
    *   Added additional `standardize` and `split` modes to `TextVectorization`:
        *   `standardize="lower"` will lowercase inputs.
        *   `standardize="string_punctuation"` will remove all punctuation.
        *   `split="character"` will split on every unicode character.
    *   Added an `output_mode` argument to the `Discretization` and `Hashing`

border-top-left-radius: 0;\n  border-bottom-left-radius: 0;\n}\n\n.dropdown-toggle-split {\n  padding-right: 0.5625rem;\n  padding-left: 0.5625rem;\n}\n\n.dropdown-toggle-split::after,\n.dropup .dropdown-toggle-split::after,\n.dropright .dropdown-toggle-split::after {\n  margin-left: 0;\n}\n\n.dropleft .dropdown-toggle-split::before {\n  margin-right: 0;\n}\n\n.btn-sm + .dropdown-toggle-split, .btn-group-sm > .btn + .dropdown-toggle-split {\n  padding-right: 0.375rem;\n  padding-left: 0.375rem;\n}\n\n.btn-lg...

border-top-left-radius: 0;\n  border-bottom-left-radius: 0;\n}\n\n.dropdown-toggle-split {\n  padding-right: 0.5625rem;\n  padding-left: 0.5625rem;\n}\n.dropdown-toggle-split::after, .dropup .dropdown-toggle-split::after, .dropend .dropdown-toggle-split::after {\n  margin-left: 0;\n}\n.dropstart .dropdown-toggle-split::before {\n  margin-right: 0;\n}\n\n.btn-sm + .dropdown-toggle-split, .btn-group-sm > .btn + .dropdown-toggle-split {\n  padding-right: 0.375rem;\n  padding-left: 0.375rem;\n}\n\n.btn-lg + ...

draws 12 bits at a time from r, interprets them in // little-endian, and rejects values higher than q, until it drew 256 // values. (The rejection rate is approximately 19%.) // // To do this from a bytes stream, it draws three bytes at a time, and // splits them into two uint16 appropriately masked. // // r₀ r₁ r₂ // |- - - - - - - -|- - - - - - - -|- - - - - - - -| // // Uint16(r₀ || r₁) // |- - - - - - - - - - - - - - - -| // |- - - - - - - - - - - -| // d₁ // // Uint16(r₁ || r₂) // |- - - - - - -...