*   Add tf.keras.layers.AbstractRNNCell as the preferred implementation of
        RNN cell for TF v2. User can use it to implement RNN cell with custom
        behavior.
    *   Adding `clear_losses` API to be able to clear losses at the end of
        forward pass in a custom training loop in eager.
    *   Add support for passing list of lists to the `metrics` param in Keras
        `compile`.

mini-xs) .control-sidebar{z-index:9999}.sidebar-collapse .form-control-sidebar,.sidebar-collapse .form-control-sidebar~.input-group-append,.sidebar-collapse .sidebar-search-results{display:none}[data-widget=sidebar-search] input[type=search]::-ms-clear,[data-widget=sidebar-search] input[type=search]::-ms-reveal{display:none;width:0;height:0}[data-widget=sidebar-search] input[type=search]::-webkit-search-cancel-button,[data-widget=sidebar-search] input[type=search]::-webkit-search-decoration,[dat...