Merge master into r1.0 to fetch all the fixes.

configure

@@ -4,7 +4,7 @@ set -e
 set -o pipefail
 
 # Find out the absolute path to where ./configure resides
-pushd `dirname $0` #> /dev/null
+pushd `dirname $0` > /dev/null
 SOURCE_BASE_DIR=`pwd -P`
 popd > /dev/null
 
@@ -145,7 +145,7 @@ while [ "$TF_NEED_CUDA" == "" ]; do
 done
 
 export TF_NEED_CUDA
-export TF_NEED_SYCL
+export TF_NEED_OPENCL
 if [[ "$TF_NEED_CUDA" == "0" ]] && [[ "$TF_NEED_OPENCL" == "0" ]]; then
   echo "Configuration finished"
   bazel_clean_and_fetch
@@ -465,7 +465,6 @@ while true; do
   COMPUTECPP_TOOLKIT_PATH=""
 done
 
-export TF_NEED_OPENCL
 # end of if "$TF_NEED_OPENCL" == "1"
 fi
 

tensorflow/BUILD

@@ -133,6 +133,7 @@ filegroup(
         "//tensorflow/contrib/losses:all_files",
         "//tensorflow/contrib/metrics:all_files",
         "//tensorflow/contrib/ndlstm:all_files",
+        "//tensorflow/contrib/nn:all_files",
         "//tensorflow/contrib/opt:all_files",
         "//tensorflow/contrib/rnn:all_files",
         "//tensorflow/contrib/seq2seq:all_files",

tensorflow/contrib/BUILD

@@ -38,6 +38,7 @@ py_library(
         "//tensorflow/contrib/losses:losses_py",
         "//tensorflow/contrib/metrics:metrics_py",
         "//tensorflow/contrib/ndlstm",
+        "//tensorflow/contrib/nn:nn_py",
         "//tensorflow/contrib/opt:opt_py",
         "//tensorflow/contrib/quantization:quantization_py",
         "//tensorflow/contrib/rnn:rnn_py",
@@ -47,8 +48,8 @@ py_library(
         "//tensorflow/contrib/solvers:solvers_py",
         "//tensorflow/contrib/specs",
         "//tensorflow/contrib/stat_summarizer:stat_summarizer_py",
-        "//tensorflow/contrib/tensor_forest:tensor_forest_py",
-        "//tensorflow/contrib/tensor_forest/hybrid:ops_lib",
+        "//tensorflow/contrib/tensor_forest:init_py",
+        "//tensorflow/contrib/tensor_forest/hybrid:ops_lib",  # XXX: no ref but need for pip
         "//tensorflow/contrib/tensorboard",
         "//tensorflow/contrib/testing:testing_py",
         "//tensorflow/contrib/tfprof",

tensorflow/contrib/__init__.py

@@ -42,6 +42,7 @@
 from tensorflow.contrib import lookup
 from tensorflow.contrib import losses
 from tensorflow.contrib import metrics
+from tensorflow.contrib import nn
 from tensorflow.contrib import opt
 from tensorflow.contrib import quantization
 from tensorflow.contrib import rnn

tensorflow/contrib/android/java/org/tensorflow/contrib/android/TensorFlowInferenceInterface.java

@@ -76,23 +76,31 @@ public TensorFlowInferenceInterface() {
    */
   public native int runInference(String[] outputNames);
 
+  /**
+   * Whether to collect and log stats to logcat during inference via StepStats and StatSummarizer.
+   * This should only be enabled when needed, as it will add overhead.
+   */
+  public native void enableStatLogging(boolean enabled);
+
+  /** Returns the last stat summary string if logging is enabled. */
+  public native String getStatString();
+
   /**
    * Cleans up the native variables associated with this Object. initializeTensorFlow() can then
    * be called again to initialize a new session.
-   *
    */
   public native void close();
 
   // Methods for creating a native Tensor and filling it with values.
   public native void fillNodeFloat(String inputName, int[] dims, float[] values);
-
   public native void fillNodeInt(String inputName, int[] dims, int[] values);
-
   public native void fillNodeDouble(String inputName, int[] dims, double[] values);
+  public native void fillNodeByte(String inputName, int[] dims, byte[] values);
 
   public native void readNodeFloat(String outputName, float[] values);
   public native void readNodeInt(String outputName, int[] values);
   public native void readNodeDouble(String outputName, double[] values);
+  public native void readNodeByte(String outputName, byte[] values);
 
   /**
    * Canary method solely for determining if the tensorflow_inference native library should be

tensorflow/contrib/android/jni/tensorflow_inference_jni.cc

@@ -53,6 +53,9 @@ struct SessionVariables {
   int num_runs = 0;
   int64 timing_total_us = 0;
 
+  bool log_stats = false;
+  StatSummarizer* summarizer = nullptr;
+
   InputMap input_tensors;
   std::vector<std::string> output_tensor_names;
   std::vector<tensorflow::Tensor> output_tensors;
@@ -129,6 +132,10 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(initializeTensorFlow)(
   LOG(INFO) << "GraphDef loaded from " << model_str << " with "
             << tensorflow_graph.node_size() << " nodes.";
 
+  // Whether or not stat logging is currently enabled, the StatSummarizer must
+  // be initialized here with the GraphDef while it is available.
+  vars->summarizer = new StatSummarizer(tensorflow_graph);
+
   LOG(INFO) << "Creating TensorFlow graph from GraphDef.";
   tensorflow::Status s = session->Create(tensorflow_graph);
 
@@ -193,8 +200,28 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(runInference)(
   }
 
   vars->output_tensors.clear();
-  s = vars->session->Run(input_tensors, vars->output_tensor_names, {},
-                         &(vars->output_tensors));
+
+  if (vars->log_stats) {
+    RunOptions run_options;
+    run_options.set_trace_level(RunOptions::FULL_TRACE);
+    RunMetadata run_metadata;
+
+    s = vars->session->Run(run_options, input_tensors,
+                           vars->output_tensor_names, {},
+                           &(vars->output_tensors), &run_metadata);
+
+    assert(run_metadata.has_step_stats());
+    const StepStats& step_stats = run_metadata.step_stats();
+    vars->summarizer->ProcessStepStats(step_stats);
+
+    // Print the full output string, not just the abbreviated one returned by
+    // getStatString().
+    vars->summarizer->PrintStepStats();
+  } else {
+    s = vars->session->Run(input_tensors, vars->output_tensor_names, {},
+                           &(vars->output_tensors));
+  }
+
   end_time = CurrentWallTimeUs();
   const int64 elapsed_time_inf = end_time - start_time;
   vars->timing_total_us += elapsed_time_inf;
@@ -208,6 +235,24 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(runInference)(
   return s.code();
 }
 
+JNIEXPORT void JNICALL TENSORFLOW_METHOD(enableStatLogging)(
+    JNIEnv* env, jobject thiz, jboolean enableStatLogging) {
+  SessionVariables* vars = GetSessionVars(env, thiz);
+  vars->log_stats = enableStatLogging;
+}
+
+JNIEXPORT jstring JNICALL TENSORFLOW_METHOD(getStatString)(JNIEnv* env,
+                                                           jobject thiz) {
+  // Return an abbreviated stat string suitable for displaying on screen.
+  SessionVariables* vars = GetSessionVars(env, thiz);
+  std::stringstream ss;
+  ss << vars->summarizer->GetStatsByMetric("Top 10 CPU",
+                                           StatSummarizer::BY_TIME, 10);
+  ss << vars->summarizer->GetStatsByNodeType();
+  ss << vars->summarizer->ShortSummary();
+  return env->NewStringUTF(ss.str().c_str());
+}
+
 JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz) {
   SessionVariables* vars = GetSessionVars(env, thiz);
 
@@ -216,6 +261,8 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz) {
     LOG(ERROR) << "Error closing session: " << s;
   }
 
+  delete vars->summarizer;
+
   mutex_lock l(mutex_);
   std::map<int64, SessionVariables*>& sessions = *GetSessionsSingleton();
   sessions.erase(vars->id);
@@ -225,7 +272,7 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz) {
 }
 
 // TODO(andrewharp): Use memcpy to fill/read nodes.
-#define FILL_NODE_METHOD(DTYPE, JAVA_DTYPE, TENSOR_DTYPE)                  \
+#define FILL_NODE_METHOD(DTYPE, JAVA_DTYPE, CTYPE, TENSOR_DTYPE)           \
   FILL_NODE_SIGNATURE(DTYPE, JAVA_DTYPE) {                                 \
     SessionVariables* vars = GetSessionVars(env, thiz);                    \
     jboolean iCopied = JNI_FALSE;                                          \
@@ -237,7 +284,7 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz) {
     }                                                                      \
     env->ReleaseIntArrayElements(dims, dim_vals, JNI_ABORT);               \
     tensorflow::Tensor input_tensor(TENSOR_DTYPE, shape);                  \
-    auto tensor_mapped = input_tensor.flat<JAVA_DTYPE>();                  \
+    auto tensor_mapped = input_tensor.flat<CTYPE>();                       \
     j##JAVA_DTYPE* values = env->Get##DTYPE##ArrayElements(arr, &iCopied); \
     j##JAVA_DTYPE* value_ptr = values;                                     \
     const int array_size = env->GetArrayLength(arr);                       \
@@ -253,14 +300,14 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz) {
     vars->input_tensors[input_name] = input_pair;                          \
   }
 
-#define READ_NODE_METHOD(DTYPE, JAVA_DTYPE)                                \
+#define READ_NODE_METHOD(DTYPE, JAVA_DTYPE, CTYPE)                         \
   READ_NODE_SIGNATURE(DTYPE, JAVA_DTYPE) {                                 \
     SessionVariables* vars = GetSessionVars(env, thiz);                    \
     Tensor* t = GetTensor(env, thiz, node_name_jstring);                   \
     if (t == nullptr) {                                                    \
       return -1;                                                           \
     }                                                                      \
-    auto tensor_mapped = t->flat<JAVA_DTYPE>();                            \
+    auto tensor_mapped = t->flat<CTYPE>();                                 \
     jboolean iCopied = JNI_FALSE;                                          \
     j##JAVA_DTYPE* values = env->Get##DTYPE##ArrayElements(arr, &iCopied); \
     j##JAVA_DTYPE* value_ptr = values;                                     \
@@ -273,10 +320,12 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz) {
     return 0;                                                              \
   }
 
-FILL_NODE_METHOD(Float, float, tensorflow::DT_FLOAT)
-FILL_NODE_METHOD(Int, int, tensorflow::DT_INT32)
-FILL_NODE_METHOD(Double, double, tensorflow::DT_DOUBLE)
+FILL_NODE_METHOD(Float, float, float, tensorflow::DT_FLOAT)
+FILL_NODE_METHOD(Int, int, int, tensorflow::DT_INT32)
+FILL_NODE_METHOD(Double, double, double, tensorflow::DT_DOUBLE)
+FILL_NODE_METHOD(Byte, byte, uint8_t, tensorflow::DT_UINT8)
 
-READ_NODE_METHOD(Float, float)
-READ_NODE_METHOD(Int, int)
-READ_NODE_METHOD(Double, double)
+READ_NODE_METHOD(Float, float, float)
+READ_NODE_METHOD(Int, int, int)
+READ_NODE_METHOD(Double, double, double)
+READ_NODE_METHOD(Byte, byte, uint8_t)

tensorflow/contrib/android/jni/tensorflow_inference_jni.h

@@ -48,15 +48,23 @@ JNIEXPORT jint JNICALL TENSORFLOW_METHOD(initializeTensorFlow)(
 JNIEXPORT jint JNICALL TENSORFLOW_METHOD(runInference)(
     JNIEnv* env, jobject thiz, jobjectArray output_name_strings);
 
+JNIEXPORT void JNICALL TENSORFLOW_METHOD(enableStatLogging)(
+    JNIEnv* env, jobject thiz, jboolean enableStatLogging);
+
+JNIEXPORT jstring JNICALL TENSORFLOW_METHOD(getStatString)(JNIEnv* env,
+                                                           jobject thiz);
+
 JNIEXPORT jint JNICALL TENSORFLOW_METHOD(close)(JNIEnv* env, jobject thiz);
 
 FILL_NODE_SIGNATURE(Float, float);
 FILL_NODE_SIGNATURE(Int, int);
 FILL_NODE_SIGNATURE(Double, double);
+FILL_NODE_SIGNATURE(Byte, byte);
 
 READ_NODE_SIGNATURE(Float, float);
 READ_NODE_SIGNATURE(Int, int);
 READ_NODE_SIGNATURE(Double, double);
+READ_NODE_SIGNATURE(Byte, byte);
 
 #ifdef __cplusplus
 }  // extern "C"

tensorflow/contrib/bayesflow/BUILD

@@ -26,6 +26,8 @@ py_library(
         "//tensorflow/python:training",
         "//tensorflow/python:util",
         "//tensorflow/python:variable_scope",
+        "//third_party/py/numpy",
+        "@six_archive//:six",
     ],
 )
 
@@ -35,9 +37,16 @@ cuda_py_test(
     srcs = ["python/kernel_tests/entropy_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//third_party/py/numpy",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/contrib/layers:layers_py",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:math_ops",
+        "//tensorflow/python:nn_ops",
         "//tensorflow/python:platform_test",
+        "//tensorflow/python:variables",
     ],
 )
 
@@ -47,9 +56,17 @@ cuda_py_test(
     srcs = ["python/kernel_tests/stochastic_variables_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//third_party/py/numpy",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/python:array_ops",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:math_ops",
         "//tensorflow/python:platform_test",
+        "//tensorflow/python:random_ops",
+        "//tensorflow/python:variable_scope",
+        "//tensorflow/python:variables",
     ],
 )
 
@@ -59,8 +76,13 @@ cuda_py_test(
     srcs = ["python/kernel_tests/monte_carlo_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/contrib/layers:layers_py",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:math_ops",
         "//tensorflow/python:platform_test",
     ],
 )
@@ -71,9 +93,13 @@ cuda_py_test(
     srcs = ["python/kernel_tests/special_math_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//third_party/py/numpy",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:gradients",
         "//tensorflow/python:platform_test",
+        "//tensorflow/python:variables",
     ],
 )
 
@@ -83,8 +109,14 @@ cuda_py_test(
     srcs = ["python/kernel_tests/stochastic_graph_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/python:array_ops",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:control_flow_ops",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:gradients",
+        "//tensorflow/python:math_ops",
         "//tensorflow/python:platform_test",
     ],
 )
@@ -95,9 +127,15 @@ cuda_py_test(
     srcs = ["python/kernel_tests/variational_inference_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/contrib/layers:layers_py",
+        "//tensorflow/python:array_ops",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:math_ops",
         "//tensorflow/python:platform_test",
+        "//tensorflow/python:variables",
     ],
 )
 
@@ -107,7 +145,11 @@ cuda_py_test(
     srcs = ["python/kernel_tests/stochastic_tensor_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//third_party/py/numpy",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/python:array_ops",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
         "//tensorflow/python:platform_test",
     ],
@@ -119,9 +161,15 @@ cuda_py_test(
     srcs = ["python/kernel_tests/stochastic_gradient_estimators_test.py"],
     additional_deps = [
         ":bayesflow_py",
-        "//tensorflow:tensorflow_py",
+        "//third_party/py/numpy",
+        "//tensorflow/contrib/distributions:distributions_py",
+        "//tensorflow/python:client_testlib",
+        "//tensorflow/python:framework_for_generated_wrappers",
         "//tensorflow/python:framework_test_lib",
+        "//tensorflow/python:gradients",
+        "//tensorflow/python:math_ops",
         "//tensorflow/python:platform_test",
+        "//tensorflow/python:variables",
     ],
 )