module {
  func.func @simpleTest(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>, %arg2: tensor<1xf32>, %arg3: tensor<1xf32>) -> tensor<2x1xf32> {
    %0 = func.call @func_0_GPU_FLOAT(%arg0, %arg1, %arg2) {tac.interface_name = "func_0"} : (tensor<1xf32>, tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    %1 = func.call @func_1_GPU_FLOAT(%arg0, %arg3) {tac.interface_name = "func_1"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>

    %4 = "tf.LayerNorm"(%a1, %a2, %a3, %a4) {_tfl_quant_trait = "fully_quantizable", device = ""} : (tensor<128x128xf32>, tensor<1xf32>, tensor<1xf32>, tensor<1xi32>) -> tensor<128x128xf32>
   "tfl.yield"(%4) : (tensor<128x128xf32>) -> ()
  }) {_tfl_quant_trait = "fully_quantizable", device = ""} : (tensor<128x128xf32>, tensor<1xf32>, tensor<1xf32>, tensor<1xi32>) -> tensor<128x128xf32>

// CHECK-SAME: (tensor<1x2x2x3xi8>, tensor<2x2x3x2xi8>, tensor<2xi32>, tensor<f32>, tensor<i32>, tensor<2xf32>, tensor<2xi32>, tensor<2xf32>, tensor<2xi32>, tensor<f32>, tensor<i32>) -> tensor<*xf32>
// CHECK: return %[[conv_quant]]

// CHECK-LABEL: func private @quantized_conv2d_with_bias_and_relu6_float_output_fn_0

  func.func private @internal_calculate_quant_params(%input : tensor<*xf32>) -> (tensor<1xf32>, tensor<1xi32>) {
    %zero = "tf.Const"() {value = dense<0.0> : tensor<1xf32>} : () -> tensor<1xf32>
    %shape = "tf.Const"() {value = dense<[-1]> : tensor<1xi32>} : () -> tensor<1xi32>
    %dim = "tf.Const"() { value = dense<0> : tensor<1xi64> } : () -> tensor<1xi64>

  %0 = "tfr.cast"(%arg0) : (tensor<*xf32>) -> !tfr.tensor
  %1 = "tfr.cast"(%0) : (!tfr.tensor) -> tensor<*xi32>
  %2 = "tfr.cast"(%0) : (!tfr.tensor) -> tensor<2xi32>
  func.return %1, %2 : tensor<*xi32>, tensor<2xi32>

// CHECK: %[[tf_cast_unranked:.*]] = "tf.Cast"(%arg0) <{Truncate = false}> : (tensor<*xf32>) -> tensor<*xi32>

  return %0 : tensor<*xi32>
}
func.func private @_batched(%arg0: tensor<1xi32>) -> tensor<1xi32> {
  return %arg0 : tensor<1xi32>
}

}

// -----

module attributes {tf_saved_model.semantics} {

// Test hoisting write side-effect ops.

  %b = arith.constant dense<0.0> : tensor<16xf32>
  %conv_3d = "tfl.conv_3d"(%arg0, %w, %b) {dilation_d_factor = 1 : i32, dilation_h_factor = 1 : i32, dilation_w_factor = 1 : i32, fused_activation_function = "NONE", padding = "SAME", stride_d = 1 : i32, stride_h = 1 : i32, stride_w = 1 : i32} : (tensor<1x32x32x32x8xf32>, tensor<1x1x1x8x16xf32>, tensor<16xf32>) -> tensor<1x32x32x32x16xf32>
  func.return %conv_3d : tensor<1x32x32x32x16xf32>

// CHECK:   %[[MUL_0:.*]] = "tf.Mul"(%arg1, %arg0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
// CHECK:   return %[[MUL_0]] : tensor<1xf32>
// CHECK: }

// CHECK: %[[v5:.*]] = "tf.Gather"(%3, %[[cst_1]]) <{validate_indices = true}> : (tensor<3xi32>, tensor<2xi32>) -> tensor<2xi32>
// CHECK: %[[v6:.*]] = "tf.Gather"(%4, %[[cst_2]]) <{validate_indices = true}> : (tensor<4xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: %[[v7:.*]] = "tf.Concat"(%[[cst_3]], %5, %6) : (tensor<i32>, tensor<2xi32>, tensor<1xi32>) -> tensor<3xi32>

// PerTensor-LABEL: QuantizeGatherWeightOnly
func.func @QuantizeGatherWeightOnly(%arg0: tensor<3xi32>) -> tensor<3x3x3x3xf32> {
  %w = arith.constant dense<1.270000e+02> : tensor<64x3x3x3xf32>
  %emb = "tfl.gather"(%w, %arg0) {axis = 0 : i32, batch_dims = 0 : i32} : (tensor<64x3x3x3xf32>, tensor<3xi32>) -> tensor<3x3x3x3xf32>