-   x86 and arm64
-   CPU-only, or with NVIDIA CUDA support (Linux only), or with TPUs

## How to Test Your Changes to TensorFlow

You may check how your changes will affect TensorFlow by:

1. Creating a PR and observing the presubmit test results
2. Running the CI scripts locally, as explained below
3. **Google employees only**: Google employees can use an internal-only tool

        "Encountered while processing output ", i, " from function '", fn_name,
        "'");
    TF_RETURN_WITH_CONTEXT_IF_ERROR(ValidateNonRefOutput(node, idx),
                                    "Encountered while creating function '",
                                    fn_name, "'");
    output_tensors->emplace_back(node, idx);
  }
  return absl::OkStatus();
}

  EXPECT_TRUE(absl::StrContains(s, expected))
      << "'" << s << "' does not contain '" << expected << "'";
}

// Returns the GPU device name if there is one (with arbitrary tie breaking if
// there are more than one), or "" otherwise.
string GPUDeviceName(TF_Session* session) {
  std::unique_ptr<TF_Status, decltype(&TF_DeleteStatus)> status(
      TF_NewStatus(), TF_DeleteStatus);

    std::vector<TensorHandlePtr> components, absl::Span<const int64_t> shape,
    TF_Status* status) {
  if (components.empty()) {
    TF_SetStatus(status, TF_INTERNAL,
                 "No components are provide for creating a ParallelTensor");
    return nullptr;
  }
  TFE_TensorHandleGetStatus(components[0].get(), status);
  if (!status->status.ok()) {
    return nullptr;
  }

                           TF_Status* status) {
#if defined(IS_MOBILE_PLATFORM) || defined(IS_SLIM_BUILD)
  status->status = tensorflow::errors::Unimplemented(
      "Creating while loops is not supported on mobile. File a bug at "
      "https://github.com/tensorflow/tensorflow/issues if this feature is "
      "important to you");
  return EmptyWhileParams();
#else

#ifdef __cplusplus
extern "C" {
#endif

// Resets `op_to_reset` with `op_or_function_name` and `raw_device_name`. This
// is for performance optimization by reusing an exiting unused op rather than
// creating a new op every time. If `raw_device_name` is `NULL` or empty, it
// does not set the device name. If it's not `NULL`, then it attempts to parse
// and set the device name. It's effectively `TFE_OpSetDevice`, but it is faster

// GraphDef type. These functions must be deleted by calling TF_DeleteFunction.
//
// If `mutate_proto_func` is non-NULL, run it over each FunctionDef proto,
// before creating a TF_Function out of the possibly mutated proto.
static std::vector<UniqueFuncPtr> CreateFunctionsFromTextProto(
    const char* text_proto,
    std::function<void(FunctionDef*)>* mutate_proto_func, TF_Status* status) {

# Release 2.17.0

## TensorFlow

<INSERT SMALL BLURB ABOUT RELEASE FOCUS AREA AND POTENTIAL TOOLCHAIN CHANGES>

### Breaking Changes

* <DOCUMENT BREAKING CHANGES HERE>
* <THIS SECTION SHOULD CONTAIN API, ABI AND BEHAVIORAL BREAKING CHANGES>

### Known Caveats

* <CAVEATS REGARDING THE RELEASE (BUT NOT BREAKING CHANGES).>
* <ADDING/BUMPING DEPENDENCIES SHOULD GO HERE>
* <KNOWN LACK OF SUPPORT ON SOME PLATFORM, SHOULD GO HERE>

TensorFlow coding style.

#### General guidelines and philosophy for contribution

*   Include unit tests when you contribute new features, as they help to a)
    prove that your code works correctly, and b) guard against future breaking
    changes to lower the maintenance cost.
*   Bug fixes also generally require unit tests, because the presence of bugs
    usually indicates insufficient test coverage.

# BEGIN TF REMOTE BUILD EXECUTION OPTIONS
# Options when using remote execution
# WARNING: THESE OPTIONS WONT WORK IF YOU DO NOT HAVE PROPER AUTHENTICATION AND PERMISSIONS

# Allow creation of resultstore URLs for any bazel invocation
build:resultstore --google_default_credentials
build:resultstore --bes_backend=buildeventservice.googleapis.com
build:resultstore --bes_instance_name="tensorflow-testing"