// RUN: tf-mlir-translate -mlir-tf-to-hlo-text %s -tf-input-shapes=: -tf-xla-emit-use-tuple-args -tf-xla-emit-return-tuple | FileCheck %s
// RUN: tf-mlir-translate -mlir-tf-to-hlo-text %s -tf-input-shapes=: | FileCheck -check-prefix=NO_TUPLES %s
// RUN: tf-mlir-translate -mlir-tf-to-hlo-text-via-builder %s -tf-input-shapes=: | FileCheck -check-prefix=NO_TUPLES %s

module attributes {tf.versions = {producer = 179 : i32}} {
  func.func @main() -> (tensor<0xi32>, tensor<0xi32>) {

// RUN: tf-opt "-xla-legalize-tf=use-tf2xla-fallback=true device-type=XLA_CPU_JIT" -verify-diagnostics %s | FileCheck --check-prefix SUPPORTED_FALLBACK_DEVICE %s
// RUN: tf-opt "-xla-legalize-tf=use-tf2xla-fallback=true" %s | FileCheck --check-prefix UNSPECIFIED_FALLBACK_DEVICE %s
// RUN: tf-opt "-xla-legalize-tf=use-tf2xla-fallback=true device-type=INVALID_DEVICE_TYPE" %s | FileCheck --check-prefix UNSUPPORTED_FALLBACK_DEVICE %s

// We run this test four times:

// RUN: odml-to-stablehlo-opt %s --tf-stablehlo=skip-partitioned-calls=true | FileCheck %s --check-prefix=CHECK-SKIP
// RUN: odml-to-stablehlo-opt %s --tf-stablehlo=skip-partitioned-calls=false | FileCheck %s --check-prefix=CHECK-NOSKIP

module {
  func.func @partitioned_call(%arg0: tensor<1x2x2x3xf32>) -> (tensor<1x2x2x3xf32>) {
    %0 = "tf.StatefulPartitionedCall"(%arg0) <{
      config = "", config_proto = "", executor_type = "", f = @some_func
    }> {

// RUN: tf-mlir-translate -mlir-tf-to-hlo-text %s -tf-input-shapes=8,16,16,64:64 -tf-xla-emit-use-tuple-args -tf-xla-emit-return-tuple | FileCheck %s
// RUN: tf-mlir-translate -mlir-tf-to-hlo-text-via-builder %s -tf-input-shapes=8,16,16,64:64 | FileCheck %s

module attributes {tf.versions = {producer = 179 : i32}} {

// RUN: tf-opt -tf-shape-inference -inline='default-pipeline=''' %s | FileCheck %s --dump-input=always
// RUN: tf-opt -tf-standard-pipeline=enable-inliner %s | FileCheck %s --dump-input=always

// Tests function with argument has no resource subtype but caller operand has a
// resource subtype, and after shape inference, function argument is refined and
// no `tf.Cast` ops are generated.

# RUN: tf-mlir-translate -graphdef-to-mlir -tf-enable-shape-inference-on-import=false %s -tf-input-arrays=p,x -tf-input-shapes=512,1024: -tf-input-data-types="DT_INT32,DT_RESOURCE(512,1024:DT_INT32)" -tf-output-arrays=p,x -o - | FileCheck %s -check-prefix=CHECK-SHAPE

// RUN: tf-opt -tf-saved-model-mark-initialized-variables-test %s | FileCheck %s
// RUN: tf-opt -tf-saved-model-mark-initialized-variables-invalid-session-test %s | FileCheck %s --check-prefix=INVALID


module attributes {tf_saved_model.semantics, tf_saved_model.under_construction} {

// RUN: tf-opt %s -tf-drop-while-shape-invariant | FileCheck %s
// RUN: tf-opt %s -tf-drop-while-shape-invariant-in-device-cluster | FileCheck -check-prefix=IN-CLUSTER %s


func.func @while_cond(%arg0: tensor<*xf32>) -> tensor<i1> {
  %0 = "tf.Const"() {value = dense<true> : tensor<i1>} : () -> tensor<i1>
  func.return %0 : tensor<i1>
}

func.func @while_body(%arg0: tensor<*xf32>) -> (tensor<*xf32>) {

// RUN: tf-quant-opt %s -quant-insert-quantized-functions='quantization-method=drq' | FileCheck %s
// RUN: tf-quant-opt %s -quant-insert-quantized-functions='quantization-method=drq target-opset=UNIFORM_QUANTIZED' | FileCheck --check-prefix=UQ-CHECK %s

// Empty module
module {
  func.func @simple_fn(%arg0: tensor<*xf32>) -> tensor<*xf32> {
    func.return %arg0 : tensor<*xf32>
  }
}

// CHECK-NOT: func private @internal_calculate_quant_params

// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s -o - | flatbuffer_to_string - | FileCheck %s
// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s -o - | flatbuffer_translate -tflite-flatbuffer-to-mlir - -o - | FileCheck --check-prefix=IMPORT %s

func.func @main(tensor<4xf32>) -> tensor<4xf32> {
^bb0(%arg0: tensor<4xf32>):
// CHECK: {
// CHECK-NEXT:   version: 3,
// CHECK-NEXT:     operator_codes: [ {
// CHECK-NEXT:       deprecated_builtin_code: 80,