%4 = "tf.Transpose"(%3, %cst_1) {_tpu_replicate = "cluster", device = ""} : (tensor<1x3x3x4xbf16>, tensor<4xi32>) -> tensor<1x3x4x3xbf16>

  %fq = "tf.FakeQuantWithMinMaxVars"(%in, %mini, %maxi) {num_bits = 3, narrow_range = false} : (tensor<3x3x3x16xf32>, tensor<f32>, tensor<f32>) -> tensor<3x3x3x16xf32>
  %rst = "tf.Conv2D"(%arg, %fq) {T = "tfdtype$DT_FLOAT", data_format = "NHWC", dilations = [1, 2, 3, 1], padding = "SAME", strides = [1, 4, 5, 1]} : (tensor<256x32x32x3xf32>, tensor<3x3x3x16xf32>) -> tensor<256x8x7x16xf32>

// IntPerLayer-DAG: "tf.DumpTensor"(%[[output0_unquantized]]) <{enabled = true, file_name = "unquantized_tensor_data.pb", func_name = "multiple_conv2d", log_dir_path = "/tmp/dumps/composite_conv2d_with_bias_and_relu6_fn_2", node_name = "Conv2D"}>

  // cuDNN prefers NCHW data format for spatial convolutions in f16 before
  // compute capability 7.0 (NVIDIA Tensor Cores).

  // CHECK: "tf.Conv2D"(%[[INPUT_TRANSPOSE:[0-9]*]], %arg1)
  // CHECK-SAME: data_format = "NCHW"
  %0 = "tf.Conv2D"(%input, %filter)
       {
         data_format = "NHWC",
         padding = "VALID",
         strides = [1, 1, 1, 1]

    }
  }
  attr {
    key: "_class"
    value {
      list {
        s: "loc:@MobilenetV1/Conv2d_0/weights"
      }
    }
  }
}
node {
  name: "MobilenetV1/MobilenetV1/Conv2d_0/Conv2D"
  op: "Conv2D"
  input: "input"
  input: "MobilenetV1/Conv2d_0/weights/read"
  attr {
    key: "T"
    value {
      type: DT_FLOAT
    }
  }
  attr {
    key: "data_format"
    value {

 ReadVariableOp 2
 VarHandleOp 435
 _Retval 2
cluster 0 size 2178
 Add 17
 AddN 72
 ArgMax 1
 AssignAddVariableOp 1
 AssignSubVariableOp 106
 BiasAdd 1
 BiasAddGrad 1
 Cast 3
 Const 357
 Conv2D 53
 Conv2DBackpropFilter 53
 Conv2DBackpropInput 52
 DivNoNan 1
 Equal 1
 FusedBatchNorm 53
 FusedBatchNormGrad 53
 Identity 2
 MatMul 3
 MaxPool 1
 MaxPoolGrad 1
 Mean 1
 Mul 164
 Pad 1

// CHECK-NOT: func private @internal_matmul_fn
// CHECK: func private @quantized_matmul_fn
// CHECK-SAME: tf_quant.quantized_ops = ["MatMul"]
// CHECK: func private @quantized_conv2d_fn
// CHECK-SAME: tf_quant.quantized_ops = ["Conv2D"]
// CHECK: func private @quantized_depthwise_conv2d_fn
// CHECK-SAME: tf_quant.quantized_ops = ["DepthwiseConv2D"]

// UQ-CHECK: func private @quantized_conv2d_fn
// UQ-CHECK: func private @quantized_depthwise_conv2d_fn

    save_options = None
    if has_func_alias:
      save_options = tensorflow.saved_model.SaveOptions(
          function_aliases={FUNC_ALIAS: model.conv2d}
      )
    saved_model_save.save(
        model,
        saved_model_path,
        signatures=model.conv2d.get_concrete_function(
            tensor_spec.TensorSpec(
                shape=input_shape, dtype=dtypes.float32, name='input_tensor'
            )

 ReadVariableOp 2
 Switch 1
 _Arg 435
 _Retval 2
cluster 0 size 1910
 Add 16
 AddN 71
 ArgMax 1
 AssignAddVariableOp 1
 BiasAdd 1
 BiasAddGrad 1
 Cast 115
 Const 407
 Conv2D 53
 Conv2DBackpropFilter 53
 Conv2DBackpropInput 52
 Equal 1
 FusedBatchNormGradV2 53
 FusedBatchNormV2 53
 MatMul 3
 MaxPool 1
 MaxPoolGrad 1
 Mean 1
 Mul 218
 Pad 2
 ReadVariableOp 538

def LiftConvWithBiasDynamic : Pat<
  (StableHLO_AddOp:$res
    (StableHLO_ConvolutionOp:$conv_0 $lhs, $rhs, $window_strides, $padding,
        $lhs_dilation, $rhs_dilation, $window_reversal, $dimension_numbers,
        $feature_group_count, $batch_group_count, $precision_config),
    (StableHLO_DynamicBroadcastInDimOp
      $bias,
      (Shape_ShapeOfOp $conv_1), $_, $_, $_)),
  (LiftAsTFXlaCallModule<"composite_conv_with_bias_dynamic_fn">