int64_t bias_len = bias_ty.getDimSize(0);
  if (feature_dim != ShapedType::kDynamic && bias_len != ShapedType::kDynamic &&
      feature_dim != bias_len) {
    return op.emitOpError()
           << "requires channel dimension and feature dimension to match; "
              "found "
           << feature_dim << " and " << bias_len << ", respectively";
  }
  return success();
}

        input_shape,
        filter_shape,
        self._input_saved_model_path,
        bias_fn,
        activation_fn,
        has_batch_norm,
        strides,
        dilations,
        'SAME',
        has_func_alias,
    )
    # TODO: b/331809306 - Investigate why these test fail then re-enable.
    if has_batch_norm and (bias_fn or not input_shape_dynamic):
      return

        """
        self.filters = np.random.uniform(low=-1.0, high=1.0, size=weight_shape)

        if bias_fn is not None:
          self.bias = np.random.uniform(
              low=-1.0, high=1.0, size=weight_shape[-1]
          )

      def has_reshape(self) -> bool:
        return self.bias_fn() and self.bias_size != self.filters.shape[-1]

      @def_function.function

```
offset = max_diag_len - diag_len(d) ; if (`align` in {RIGHT_LEFT, RIGHT_RIGHT}
                                           and `d >= 0`) or
                                         (`align` in {LEFT_RIGHT, RIGHT_RIGHT}
                                           and `d <= 0`)
         0                          ; otherwise
```
where `diag_len(d) = min(cols - max(d, 0), rows + min(d, 0))`.