num_bits=8,
            narrow_range=False,
        )
        return {'output': out}

    bias = None
    if has_bias:
      bias_shape = shapes[1][-1]
      if bias_shape is not None:
        bias = array_ops.constant(
            np.random.uniform(size=[shapes[1][-1]]), dtype=dtypes.float32
        )
    model = MatmulModel(bias)
    x = array_ops.constant(

        filter_quantized_element_type.getZeroPoint());
  }

  SmallVector<int64_t, 1> bias_shape = {filter_shape[0]};
  auto bias_type =
      RankedTensorType::getChecked(loc, bias_shape, bias_quantized_type);

  auto bias_value_type = RankedTensorType::getChecked(
      loc, std::move(bias_shape), rewriter.getI32Type());
  auto bias_value = DenseIntElementsAttr::get(

}

std::optional<RankedTensorType> InferWindowOutputShape(
    const ShapedType& base_shape, const xla::Window& window,
    Type element_type) {
  if (window.dimensions_size() != base_shape.getRank()) {
    llvm::errs() << "Window has dimension " << window.dimensions_size()
                 << " but base shape has dimension " << base_shape.getRank()
                 << "\n";
    return std::nullopt;
  }