// CHECK-SAME:      [[ARG:%.+]]: tensor<2x12x21x7xf16>
// CHECK:           [[CONV32:%.+]] = mhlo.convert %arg0 : (tensor<2x12x21x7xf16>) -> tensor<2x12x21x7xf32>
// CHECK:           [[ZERO:%.+]] = mhlo.constant dense<0.000000e+00> : tensor<f32>
// CHECK:           [[DIVIDEND:%.+]] = "mhlo.reduce_window"([[CONV32]], [[ZERO]])
// CHECK-SAME:        window_dimensions = dense<[1, 2, 2, 1]>

// CHECK:           %[[CONV:.*]] = "tf.Conv2D"(%[[SLICED_ARG0]], %[[ARG1]])
// CHECK-SAME:      explicit_paddings = [0, 0, 4, 0, 0, 2, 0, 0]
// CHECK-SAME:      (tensor<128x5x4x64xf32>, tensor<3x2x64x4xf32>) -> tensor<128x4x3x4xf32>
// CHECK:           return %[[CONV]] : tensor<128x4x3x4xf32>
// CHECK:         }

};

using ConvertConv2DDynamic =
    ConvertConvDynamic<TF::Conv2DOp, /*num_spatial_dims=*/2>;

// Converts the TensorFlow conv op in template to the generic HLO conv op by
// converting TensorFlow op attributes to HLO op attributes.
//
// Sample result for Conv2D:
//
//   %conv = "mhlo.convolution"(%input, %filter) {
//     strides = [1, 2],
//     paddings = [[1, 0], [1, 1]],
//     ...
//   }
//