}

func.func @depthwiseConv2D(tensor<256x32x32x3xf32>, tensor<3x3x3x4xf32>, tensor<256x3x32x32xf32>) -> (tensor<256x30x30x12xf32>, tensor<256x12x30x30xf32>, tensor<256x30x30x12xf32>, tensor<256x30x30x12xf32>) {

        /*depth_multiplier=*/rewriter.getI32IntegerAttr(multiplier));
  }

 private:
  /// Legalize the given filter by converting it from TensorFlow filter data
  /// format to TFLite DepthwiseConv2D op filter data format and return Value
  /// for the converted filter.  TensorFlow filter data format is
  /// [filter_height, filter_width, in_channels, channel_multiplier] and TFLite

        collisions across summary types.

*   When running on GPU (with cuDNN version 7.6.3 or
    later),`tf.nn.depthwise_conv2d` backprop to `filter` (and therefore also
    `tf.keras.layers.DepthwiseConv2D`) now operate deterministically (and
    `tf.errors.UnimplementedError` is no longer thrown) when op-determinism has
    been enabled via `tf.config.experimental.enable_op_determinism`. This closes