Python tensorrt.init_libnvinfer_plugins方法代码示例

# 需要导入模块: import tensorrt [as 别名]
# 或者: from tensorrt import init_libnvinfer_plugins [as 别名]
def _load_plugins(self):
        ctypes.CDLL("models/ssd_mobilenet/libflattenconcat.so")
        trt.init_libnvinfer_plugins(self.trt_logger, '') 

# 需要导入模块: import tensorrt [as 别名]
# 或者: from tensorrt import init_libnvinfer_plugins [as 别名]
def _load_plugins(self):
        if trt.__version__[0] < '7':
            ctypes.CDLL("ssd/libflattenconcat.so")
        trt.init_libnvinfer_plugins(self.trt_logger, '') 

# 需要导入模块: import tensorrt [as 别名]
# 或者: from tensorrt import init_libnvinfer_plugins [as 别名]
def main():
    parser = argparse.ArgumentParser()
    parser.add_argument('model', type=str, choices=list(MODEL_SPECS.keys()))
    args = parser.parse_args()

    # initialize
    if trt.__version__[0] < '7':
        ctypes.CDLL(LIB_FILE)
    TRT_LOGGER = trt.Logger(trt.Logger.INFO)
    trt.init_libnvinfer_plugins(TRT_LOGGER, '')

    # compile the model into TensorRT engine
    model = args.model
    spec = MODEL_SPECS[model]
    dynamic_graph = add_plugin(
        gs.DynamicGraph(spec['input_pb']),
        model,
        spec)
    _ = uff.from_tensorflow(
        dynamic_graph.as_graph_def(),
        output_nodes=['NMS'],
        output_filename=spec['tmp_uff'],
        text=True,
        debug_mode=DEBUG_UFF)
    with trt.Builder(TRT_LOGGER) as builder, builder.create_network() as network, trt.UffParser() as parser:
        builder.max_workspace_size = 1 << 28
        builder.max_batch_size = 1
        builder.fp16_mode = True

        parser.register_input('Input', INPUT_DIMS)
        parser.register_output('MarkOutput_0')
        parser.parse(spec['tmp_uff'], network)
        engine = builder.build_cuda_engine(network)

        buf = engine.serialize()
        with open(spec['output_bin'], 'wb') as f:
            f.write(buf) 

# 需要导入模块: import tensorrt [as 别名]
# 或者: from tensorrt import init_libnvinfer_plugins [as 别名]
def __init__(self, trt_engine_path, uff_model_path, trt_engine_datatype=trt.DataType.FLOAT, batch_size=1):
        """Initializes TensorRT objects needed for model inference.

        Args:
            trt_engine_path (str): path where TensorRT engine should be stored
            uff_model_path (str): path of .uff model
            trt_engine_datatype (trt.DataType):
                requested precision of TensorRT engine used for inference
            batch_size (int): batch size for which engine
                should be optimized for
        """

        # We first load all custom plugins shipped with TensorRT,
        # some of them will be needed during inference
        trt.init_libnvinfer_plugins(TRT_LOGGER, '')

        # Initialize runtime needed for loading TensorRT engine from file
        self.trt_runtime = trt.Runtime(TRT_LOGGER)
        # TRT engine placeholder
        self.trt_engine = None

        # Display requested engine settings to stdout
        print("TensorRT inference engine settings:")
        print("  * Inference precision - {}".format(trt_engine_datatype))
        print("  * Max batch size - {}\n".format(batch_size))

        # If engine is not cached, we need to build it
        if not os.path.exists(trt_engine_path):
            # This function uses supplied .uff file
            # alongside with UffParser to build TensorRT
            # engine. For more details, check implmentation
            self.trt_engine = engine_utils.build_engine(
                uff_model_path, TRT_LOGGER,
                trt_engine_datatype=trt_engine_datatype,
                batch_size=batch_size)
            # Save the engine to file
            engine_utils.save_engine(self.trt_engine, trt_engine_path)

        # If we get here, the file with engine exists, so we can load it
        if not self.trt_engine:
            print("Loading cached TensorRT engine from {}".format(
                trt_engine_path))
            self.trt_engine = engine_utils.load_engine(
                self.trt_runtime, trt_engine_path)

        # This allocates memory for network inputs/outputs on both CPU and GPU
        self.inputs, self.outputs, self.bindings, self.stream = \
            engine_utils.allocate_buffers(self.trt_engine)

        # Execution context is needed for inference
        self.context = self.trt_engine.create_execution_context()

        # Allocate memory for multiple usage [e.g. multiple batch inference]
        input_volume = trt.volume(model_utils.ModelData.INPUT_SHAPE)
        self.numpy_array = np.zeros((self.trt_engine.max_batch_size, input_volume))