Java Code Examples for org.tensorflow.Tensors#create()

@Test
public void testDummyLoadOfTensorFlowGraphWithPrefix() throws Exception {
  final String prefix = "test";
  final Path graphFile = createADummyTFGraph();
  try (final TensorFlowGraphModel model =
          TensorFlowGraphModel.create(graphFile.toUri(), null, prefix);
      final Session session = model.instance();
      final Tensor<Double> double3 = Tensors.create(3.0D)) {
    List<Tensor<?>> result = null;
    try {
      result =
          session
              .runner()
              .fetch(prefix + "/" + mulResult)
              .feed(prefix + "/" + inputOpName, double3)
              .run();
      assertEquals(result.get(0).doubleValue(), 6.0D, Double.MIN_VALUE);
    } finally {
      if (result != null) {
        result.forEach(Tensor::close);
      }
    }
  }
}
 

@Test
public void testBooleanTensor() {
  final boolean[] booleanValue = {true, true, false, true, false};
  final Tensor<Boolean> tensor = Tensors.create(booleanValue);
  final JTensor jt = JTensor.create(tensor);
  assertEquals(DataType.BOOL, jt.dataType());
  assertEquals(1, jt.numDimensions());
  assertArrayEquals(shape, jt.shape());
  assertArrayEquals(booleanValue, jt.booleanValue());
  testException(jt, JTensor::stringValue);
  testException(jt, JTensor::intValue);
  testException(jt, JTensor::longValue);
  testException(jt, JTensor::floatValue);
  testException(jt, JTensor::doubleValue);
}
 

private static Tuple2<String, Float> classify(
        String review, SavedModelBundle model, WordIndex wordIndex
) {
    try (Tensor<Float> input = Tensors.create(wordIndex.createTensorInput(review));
         Tensor<?> output = model.session().runner()
                                 .feed("embedding_input:0", input)
                                 .fetch("dense_1/Sigmoid:0").run().get(0)
    ) {
        float[][] result = new float[1][1];
        output.copyTo(result);
        return tuple2(review, result[0][0]);
    }
}
 

/**
 * Creates a simple TensorFlow graph that multiplies Double on input by 2.0, result is available
 * via multiply operation.
 */
private Path createADummyTFGraph() throws IOException {
  final Path graphFile;
  try (final Graph graph = new Graph();
      final Tensor<Double> t = Tensors.create(2.0D)) {
    final Output<Double> input =
        graph
            .opBuilder("Placeholder", inputOpName)
            .setAttr("dtype", t.dataType())
            .build()
            .output(0);

    final Output<Double> two =
        graph
            .opBuilder("Const", "two")
            .setAttr("dtype", t.dataType())
            .setAttr("value", t)
            .build()
            .output(0);

    graph.opBuilder("Mul", mulResult).addInput(two).addInput(input).build();

    graphFile = Files.createTempFile("tf-graph", ".bin");
    Files.write(graphFile, graph.toGraphDef());
  }
  return graphFile;
}
 

private static Graph createDummyGraph() {
  final Tensor<Double> t2 = Tensors.create(2.0);
  final Tensor<Double> t3 = Tensors.create(3.0);

  final Graph graph = new Graph();
  final Output<Double> input =
      graph.opBuilder("Placeholder", "input").setAttr("dtype", DataType.DOUBLE).build().output(0);

  final Output<Double> two =
      graph
          .opBuilder("Const", "two")
          .setAttr("dtype", t2.dataType())
          .setAttr("value", t2)
          .build()
          .output(0);

  final Output<Double> three =
      graph
          .opBuilder("Const", "three")
          .setAttr("dtype", t3.dataType())
          .setAttr("value", t3)
          .build()
          .output(0);

  graph.opBuilder("Mul", mul2).addInput(input).addInput(two).build();

  graph.opBuilder("Mul", mul3).addInput(input).addInput(three).build();

  return graph;
}
 

@Test
public void doubleTensorSerializable() throws IOException {
  final double[] doubleValue = {1, 2, 3, 4, 5};
  final Tensor<Double> tensor = Tensors.create(doubleValue);
  final JTensor jt = JTensor.create(tensor);

  new ObjectOutputStream(new ByteArrayOutputStream()).writeObject(jt);
}
 

@Test
public void floatTensorSerializable() throws IOException {
  final float[] floatValue = {1, 2, 3, 4, 5};
  final Tensor<Float> tensor = Tensors.create(floatValue);
  final JTensor jt = JTensor.create(tensor);

  new ObjectOutputStream(new ByteArrayOutputStream()).writeObject(jt);
}
 

@Test
public void longTensorSerializable() throws IOException {
  final long[] longValue = {1, 2, 3, 4, 5};
  final Tensor<Long> tensor = Tensors.create(longValue);
  final JTensor jt = JTensor.create(tensor);

  new ObjectOutputStream(new ByteArrayOutputStream()).writeObject(jt);
}
 

@Test
public void intTensorSerializable() throws IOException {
  final int[] intValue = {1, 2, 3, 4, 5};
  final Tensor<Integer> tensor = Tensors.create(intValue);
  final JTensor jt = JTensor.create(tensor);

  new ObjectOutputStream(new ByteArrayOutputStream()).writeObject(jt);
}
 

@Test
public void multidimensionalStringTensorSerializable() throws IOException {
  final byte[][][][] byteArray = toByteArray(STRING_ARRAY_3DIMENSIONS);
  final Tensor<String> tensor = Tensors.create(byteArray);
  final JTensor jt = JTensor.create(tensor);
  new ObjectOutputStream(new ByteArrayOutputStream()).writeObject(jt);
}
 

/**
 * TensorFlow Example prediction function.
 *
 * @deprecated Use {@link #example(Function, String...)}
 * @param outTensorExtractor Function to extract the output value from JTensor's
 * @param fetchOps operations to fetch.
 */
@Deprecated
static <InputT, ValueT> TensorFlowPredictFn<InputT, List<Example>, ValueT> exampleBatch(
    final Function<Map<String, JTensor>, ValueT> outTensorExtractor, final String... fetchOps) {
  final BiFunction<TensorFlowModel, List<Example>, ValueT> predictFn =
      (model, examples) -> {
        final byte[][] bytes = examples.stream().map(Example::toByteArray).toArray(byte[][]::new);

        try (final Tensor<String> t = Tensors.create(bytes)) {
          final Session.Runner runner =
              model.instance().session().runner().feed("input_example_tensor", t);
          final Map<String, JTensor> result = TensorFlowExtras.runAndExtract(runner, fetchOps);

          return outTensorExtractor.apply(result);
        }
      };

  return (model, vectors) -> {
    final List<CompletableFuture<Prediction<InputT, ValueT>>> predictions =
        vectors
            .stream()
            .map(
                vector ->
                    CompletableFuture.supplyAsync(() -> predictFn.apply(model, vector.value()))
                        .thenApply(v -> Prediction.create(vector.input(), v)))
            .collect(Collectors.toList());

    return CompletableFutures.allAsList(predictions);
  };
}
 

@Test
public void testFloatTensor() {
  final float[] floatValue = {1, 2, 3, 4, 5};
  final Tensor<Float> tensor = Tensors.create(floatValue);
  final JTensor jt = JTensor.create(tensor);
  assertEquals(DataType.FLOAT, jt.dataType());
  assertEquals(1, jt.numDimensions());
  assertArrayEquals(shape, jt.shape());
  assertArrayEquals(floatValue, jt.floatValue(), 0.0f);
  testException(jt, JTensor::stringValue);
  testException(jt, JTensor::intValue);
  testException(jt, JTensor::longValue);
  testException(jt, JTensor::doubleValue);
}
 

@Test
public void testLongTensor() {
  final long[] longValue = {1, 2, 3, 4, 5};
  final Tensor<Long> tensor = Tensors.create(longValue);
  final JTensor jt = JTensor.create(tensor);
  assertEquals(DataType.INT64, jt.dataType());
  assertEquals(1, jt.numDimensions());
  assertArrayEquals(shape, jt.shape());
  assertArrayEquals(longValue, jt.longValue());
  testException(jt, JTensor::stringValue);
  testException(jt, JTensor::intValue);
  testException(jt, JTensor::floatValue);
  testException(jt, JTensor::doubleValue);
}
 

@Test
public void testIntTensor() {
  final int[] intValue = {1, 2, 3, 4, 5};
  final Tensor<Integer> tensor = Tensors.create(intValue);
  final JTensor jt = JTensor.create(tensor);
  assertEquals(DataType.INT32, jt.dataType());
  assertEquals(1, jt.numDimensions());
  assertArrayEquals(shape, jt.shape());
  assertArrayEquals(intValue, jt.intValue());
  testException(jt, JTensor::stringValue);
  testException(jt, JTensor::longValue);
  testException(jt, JTensor::floatValue);
  testException(jt, JTensor::doubleValue);
}
 

@Test
public void testStringTensor3DimensionsDescending() {
  final byte[][][][] byteArray = toByteArray(STRING_ARRAY_3DIMENSIONS_DESCENDING);
  final Tensor<String> tensor = Tensors.create(byteArray);
  final JTensor jt = JTensor.create(tensor);
  testMultidimensionalStringTensor(jt, STRING_ARRAY_3DIMENSIONS_DESCENDING, new long[] {3, 2, 1});
}
 

@Test
public void testStringTensor3DimensionsAscending() {
  final byte[][][][] byteArray = toByteArray(STRING_ARRAY_3DIMENSIONS_ASCENDING);
  final Tensor<String> tensor = Tensors.create(byteArray);
  final JTensor jt = JTensor.create(tensor);
  testMultidimensionalStringTensor(jt, STRING_ARRAY_3DIMENSIONS_ASCENDING, new long[] {1, 2, 3});
}
 

/**
 * Running neural network
 *
 * @param image cropped, centralized face
 * @return describing of a face based on 128 float features
 */
private FaceFeatures passImageThroughNeuralNetwork(BufferedImage image, int faceType) {
    FaceFeatures features;

    try (Session session = new Session(graph)) {
        Tensor<Float> feedImage = Tensors.create(imageToMultiDimensionalArray(image));

        long timeResponse = System.currentTimeMillis();

        Tensor<Float> response = session.runner()
                .feed("input", feedImage)
                .feed("phase_train", Tensor.create(false))
                .fetch("embeddings")
                .run().get(0)
                .expect(Float.class);

        FileUtils.timeSpent(timeResponse, "RESPONSE");

        final long[] shape = response.shape();

        //first dimension should return 1 as for image with normal size
        //second dimension should give 128 characteristics of face
        if (shape[0] != 1 || shape[1] != 128) {
            throw new IllegalStateException("illegal output values: 1 = " + shape[0] + " 2 = " + shape[1]);
        }

        float[][] featuresHolder = new float[1][128];
        response.copyTo(featuresHolder);

        features = new FaceFeatures(featuresHolder[0], faceType);

        response.close();
    }

    return features;
}
 

@Test
public void testStringTensor1Dimension() {
  final byte[][] byteArray = toByteArray(STRING_ARRAY_1DIMENSION);
  final Tensor<String> tensor = Tensors.create(byteArray);
  final JTensor jt = JTensor.create(tensor);
  testMultidimensionalStringTensor(jt, STRING_ARRAY_1DIMENSION, new long[] {3});
}
 

@Test
public void testStringTensor() {
  final String stringValue = "world";
  final Tensor<String> tensor = Tensors.create(stringValue);
  final JTensor jt = JTensor.create(tensor);
  assertEquals(DataType.STRING, jt.dataType());
  assertEquals(0, jt.numDimensions());
  assertArrayEquals(new long[0], jt.shape());
  assertEquals(stringValue, jt.stringValue());
  testException(jt, JTensor::intValue);
  testException(jt, JTensor::longValue);
  testException(jt, JTensor::floatValue);
  testException(jt, JTensor::doubleValue);
}
 

/**
 * See <a href="https://github.com/tensorflow/tensorflow/issues/24331#issuecomment-447523402">GitHub issue</a>
 */
private static Tensor<?> makeImageStringTensor(File imageFile) throws IOException {
    var content = FileUtils.readFileToByteArray(imageFile);
    byte[][] data = { content };
    return Tensors.create(data);
}