func.return %1 : tensor<3x2xi32>
}

// CHECK: %cst = arith.constant dense<[1, 0]> : tensor<2xi32>
// CHECK: [1, 3, 5], [2, 4, 6]
// CHECK: %0 = tfl.add %arg0, %cst_0 {fused_activation_function = "NONE"} : tensor<2x3xi32>
// CHECK: %1 = "tfl.transpose"(%0, %cst) : (tensor<2x3xi32>, tensor<2xi32>) -> tensor<3x2xi32>
// CHECK: return %1 : tensor<3x2xi32>

// -----

// CHECK-LABEL: pushTposeAfterSubSimple

  // CHECK: return %1 : tensor<1x750xf32>
}

// CHECK-LABEL: @unfold_splat_constant_integer
func.func @unfold_splat_constant_integer() -> tensor<1x750xi32> {
  %cst = mhlo.constant dense<1> : tensor<1x750xi32>
  func.return %cst : tensor<1x750xi32>

  // CHECK-DAG: %0 = mhlo.constant dense<1> : tensor<i32>

  ^bb0(%arg2: tensor<5xf32>, %arg3: tensor<5xf32>):
    // CHECK:      %1 = "tfl.abs"(%arg2) {fused_activation_function = "NONE"} : (tensor<5xf32>) -> tensor<5xf32>
    // CHECK-NEXT: %2 = "vhlo.add_v1"(%1, %arg2) : (tensor<5xf32>, tensor<5xf32>) -> tensor<5xf32>
    // CHECK-NEXT: %3 = "tfl.abs"(%2) {fused_activation_function = "NONE"} : (tensor<5xf32>) -> tensor<5xf32>

    return %0 : tensor<*xf32>
  }

// CHECK-LABEL: func @depthwise_conv
// CHECK-DAG: %[[CONST_0:.*]] = arith.constant dense<0.000000e+00> : tensor<2xf32>
// CHECK-DAG: %[[CONST_1:.*]] = arith.constant dense<3.000000e+00> : tensor<2x3x1x1536xf32>
// CHECK: %0 = "quantfork.qcast"(%[[CONST_1]]) : (tensor<2x3x1x1536xf32>) -> tensor<2x3x1x1536x!quant.uniform<i8<-127:127>:f32:3, {0.023622047244094488,

    return %0 : tensor<*xf32>
  }

// CHECK-LABEL: func @conv2d
// CHECK-DAG: %[[CONST_0:.*]] = arith.constant dense<0.000000e+00> : tensor<2xf32>
// CHECK-DAG: %[[CONST_1:.*]] = arith.constant dense<3.000000e+00> : tensor<2x3x3x512xf32>
// CHECK: %0 = "quantfork.qcast"(%[[CONST_1]]) : (tensor<2x3x3x512xf32>) -> tensor<2x3x3x512x!quant.uniform<i8<-127:127>:f32, 0.023622047244094488>>

plugins {
    `java-library`
}

// tag::api[]
tasks.named("check")                    // <1>
tasks.register("myTask1")               // <2>

tasks.named<JavaCompile>("compileJava") // <3>
tasks.register<Copy>("myCopy1")         // <4>

tasks.named("assemble") {               // <5>
    dependsOn(":myTask1")
}
tasks.register("myTask2") {             // <6>
    description = "Some meaningful words"
}

// RUN: json_to_flatbuffer %p/test_schema.fbs %s | flatbuffer_translate --tflite-flatbuffer-to-mlir -o - | FileCheck %s

// CHECK: %0 = "tfl.pseudo_const"() <{value = dense<2> : tensor<2xi32>}> : () -> tensor<2xi32>
// CHECK: %1 = "tfl.reshape"(%arg0, %0) : (tensor<1x4xf32>, tensor<2xi32>) -> tensor<2x2xf32>

{
  "version": 3,
  "operator_codes": [
    {
      "builtin_code": "RESHAPE"
    }
  ],
  "subgraphs": [
    {
      "tensors": [

  func.return %2 : tensor<1x6x6x16xf32>

// CHECK: %0 = "tfl.dequantize"(%arg0)
// CHECK: %1 = "tfl.dequantize"(%arg1)
// CHECK: %2 = "tfl.maximum"(%0, %1)
// CHECK: %3 = "tfl.quantize"(%2)
// CHECK: %4 = "tfl.dequantize"(%3)
// CHECK: return %4 : tensor<1x6x6x16xf32>
}

// CHECK-LABEL: QuantizeMinimum

// CHECK-LABEL: avg_pool2d_3
// CHECK: %cst = arith.constant dense<[0, 2, 3, 1]> : tensor<4xi32>
// CHECK: %0 = "tfl.transpose"(%arg0, %cst) : (tensor<1x1x1x8xf32>, tensor<4xi32>) -> tensor<1x1x8x1xf32>
// CHECK{LITERAL}: %cst_0 = arith.constant dense<[[0, 0], [0, 0], [1, 1], [0, 0]]> : tensor<4x2xi32>

  func.return %1 : tensor<2x1xf32>

// CHECK:  %cst = arith.constant
// CHECK:  %0 = "tf.Reshape"(%arg0, %cst)
// CHECK-SAME: tensor<2x1xf32>
// CHECK:  %1 = "tf.FakeQuantWithMinMaxVars"(%0, %arg1, %arg2)
// CHECK:  return %1
}

// CHECK-LABEL: fakeQuantFollowedByReshapes