the same across browsers. Vendor-specific rules for pseudo\n// elements cannot be mixed. As such, there are no shared styles for focus or\n// active states on prefixed selectors.\n\n.custom-range {\n  width: 100%;\n  height: add($custom-range-thumb-height, $custom-range-thumb-focus-box-shadow-width * 2);\n  padding: 0; // Need to reset padding\n  background-color: transparent;\n  appearance: none;\n\n  &:focus {\n    outline: none;\n\n    // Pseudo-elements must be split across multiple rulesets to...

the same across browsers. Vendor-specific rules for pseudo\n// elements cannot be mixed. As such, there are no shared styles for focus or\n// active states on prefixed selectors.\n\n.custom-range {\n  width: 100%;\n  height: add($custom-range-thumb-height, $custom-range-thumb-focus-box-shadow-width * 2);\n  padding: 0; // Need to reset padding\n  background-color: transparent;\n  appearance: none;\n\n  &:focus {\n    outline: none;\n\n    // Pseudo-elements must be split across multiple rulesets to...

*   Clean up `BatchNormalization` layer's `trainable` property to act like
    standard python state when it's used inside `tf.functions` (frozen at
    tracing time), instead of acting like a pseudo-variable whose updates *kind
    of sometimes* get reflected in already-traced `tf.function` traces.
*   Add the `Conv1DTranspose` layer.
*   Refine the semantics of `SensitivitySpecificityBase` derived metrics. See