steps:
      - name: Stop old running containers (if any)
        shell: bash
        run: |
          running_containers=$(docker ps -q) && \
          if [[ $running_containers == "" ]]; then
            echo "No running containers";
          else
            echo "Running container(s) found" && \
            docker stop $running_containers;
          fi
          docker container prune -f

research](https://scholar.google.com/scholar?q=gpu+side+channel) shows that it
might be possible to use side channel attacks on the GPU to leak data from other
running models or processes in the same system. GPUs can also have
implementation bugs that might allow attackers to leave malicious code running
and leak or tamper with applications from other users. Please report
vulnerabilities to the vendor of the affected hardware accelerator.

TF_CAPI_EXPORT void TFE_ContextSetLogDevicePlacement(TFE_Context* ctx,
                                                     unsigned char enable,
                                                     TF_Status* status);

// Enables running eager ops as function.
TF_CAPI_EXPORT void TFE_ContextSetRunEagerOpAsFunction(TFE_Context* ctx,
                                                       unsigned char enable,

  echo 'sure they are present on your system and try again.'
  exit 1
fi

# "TFCI_MACOS_BAZEL_TEST_DIR_PATH" specifies the directory that Bazel should use
# when running tests. Each test will be executed in a separate subdirectory
# inside this directory. TF Mac builds need ~150 GB of disk space to be able to
# run all the tests. Since TFCI Mac VMs execute Bazel test commands in a

class EagerContext;
class CustomDevice;
class CustomDeviceOpHandler;
class Device;

// LINT.IfChange
// Note: Keep in sync with exported copy of enum in eager/c_api.h.
enum ContextDevicePlacementPolicy {
  // Running operations with input tensors on the wrong device will fail.
  DEVICE_PLACEMENT_EXPLICIT = 0,
  // Copy the tensor to the right device but log a warning.
  DEVICE_PLACEMENT_WARN = 1,

        *   If you experience issues with oneDNN optimizations on, we recommend
            turning them off.
    *   To explicitly enable or disable oneDNN optimizations, set the
        environment variable `TF_ENABLE_ONEDNN_OPTS` to `1` (enable) or `0`
        (disable) before running TensorFlow. (The variable is checked during
        `import tensorflow`.) To fall back to default settings, unset the

    TF_Status status;
    // Let this custom device choose the device to pin this op on if it
    // implements the pinning function.
    if (device_.shall_pin_to_this_device != nullptr) {
      return device_.shall_pin_to_this_device(tensorflow::wrap(op), &status);
    }
    return errors::Unimplemented("No custom device pinning implementation.");
  }

 private:
  TFE_Context* context_;
  TFE_CustomDevice device_;
  void* info_;

  (*config.mutable_device_count())["CPU"] = num_cpu_devices;

  // TODO(b/113217601): This is needed for EagerContext::runner_ to use a
  // threadpool, so that we avoid the possibility of running the runner_ in the
  // threadpool of GPU event mgr, as that can trigger more callbacks to be
  // scheduled on that same threadpool, causing a deadlock in cases where the

    $env:PATH = [Environment]::GetEnvironmentVariable('PATH', 'Machine'); \
    gcloud components install docker-credential-gcr kubectl gsutil;

# Install cygwin and packages
# Running a seperate ps1 file since when running inside a Dockerfile, it does
# not work.
COPY install/install_cygwin.ps1 c:/
RUN c:/install_cygwin.ps1; \
    $old_path = [Environment]::GetEnvironmentVariable(\"PATH\", \"Machine\"); \

  // source tensors. The output gradients are used if not empty and not
  // null. The result is populated with one tensor per target element.
  // When running backward functions, builds zeros-like tensors for
  // incoming grads which are nullptrs, unless `build_default_zeros_grads`
  // is set to false.
  Status ComputeGradient(