licenses = ["notice"],
)

# This is only for plugins
cc_library(
    name = "filesystem_interface",
    hdrs = ["filesystem_interface.h"],
    visibility = ["//visibility:public"],
    deps = [
        "//tensorflow/c:tf_file_statistics",
        "//tensorflow/c:tf_status",
    ],
)

# Core TensorFlow depends on this, will be included in main library
cc_library(

struct TFE_Op;

namespace tensorflow {

class ImmediateExecutionContext;
class AbstractOpAttrs;

// Abstract interface to an operation.
class ImmediateExecutionOperation : public AbstractOperation {
 public:
  virtual void Clear() = 0;

  // Returns the inputs of this op.
  virtual absl::Span<ImmediateExecutionTensorHandle* const> GetInputs()
      const = 0;
  virtual Status SetInput(size_t index,

  // for computing gradient so we could safely drop it.
  outputs[0] = temp_outputs[1];
  temp_outputs[0]->Unref();
  return absl::OkStatus();
}

class CppGradients
    : public ::testing::TestWithParam<std::tuple<const char*, bool, bool>> {
 protected:
  void SetUp() override {
    TF_StatusPtr status(TF_NewStatus());
    TF_SetTracingImplementation(std::get<0>(GetParam()), status.get());

# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

# Bring in all of the public TensorFlow interface into this
# module.

# pylint: disable=g-bad-import-order
from tensorflow.python import pywrap_tensorflow  # pylint: disable=unused-import

/// \brief An LRU cache of string keys and arbitrary values, with configurable
/// max item age (in seconds) and max entries.
///
/// This class is thread safe.
template <typename T>
class ExpiringLRUCache {
 public:
  /// A `max_age` of 0 means that nothing is cached. A `max_entries` of 0 means
  /// that there is no limit on the number of entries in the cache (however, if
  /// `max_age` is also 0, the cache will not be populated).

}
AbstractTensorHandle* TapeTensor::GetHandle() const { return handle_; }

AbstractTensorHandle* TapeTensor::ZerosLike() const { return nullptr; }

class TapeVSpace
    : public eager::VSpace<AbstractTensorHandle, GradientFunction, TapeTensor> {
 public:
  explicit TapeVSpace(AbstractContext* ctx) : ctx_(ctx) {}
  ~TapeVSpace() override {}

  // Returns the number of elements in the gradient tensor.

# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Bring in all of the public TensorFlow interface into this module."""

import distutils as _distutils
import importlib
import inspect as _inspect
import os as _os
import site as _site
import sys as _sys

#include "tensorflow/core/framework/op_def.pb.h"
#include "tensorflow/core/framework/tensor.pb.h"
#include "tensorflow/core/platform/logging.h"
#include "tensorflow/core/platform/strcat.h"
#include "tensorflow/core/public/session_options.h"

using tensorflow::GraphDef;
using tensorflow::NodeDef;

static void BoolDeallocator(void* data, size_t, void* arg) {
  delete[] static_cast<bool*>(data);
}

*  In the `keras` PyPI package, accessible symbols are now restricted to symbols that are intended to be public. This may affect your code if you were using `import keras` and you used `keras` functions that were not public APIs, but were accessible in earlier versions with direct imports. In those cases, please use the following guideline:

mitigate the risk in this scenario is to sandbox the model execution.

### Hardware attacks

Physical GPUs or TPUs can also be the target of attacks. [Published
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