[TensorRT] ERROR: Network must have at least one output

使用TensorRT转换ONNX模型时报错

def build_engine():
    with trt.Builder(TRT_LOGGER) as builder, builder.create_network() as network, trt.OnnxParser(network, TRT_LOGGER) as parser:
        builder.max_workspace_size = 1 << 30
        with open(ONNX_MODEL, "rb") as model:
            parser.parse(model.read())
        return builder.build_cuda_engine(network)

• Pytorch==1.8.0
• TensorRT==7.2.3.4

[TensorRT] ERROR: Network must have at least one output 
[TensorRT] ERROR: Network validation failed.

https://forums.developer.nvidia.com/t/tensorrt-error-network-must-have-at-least-one-output-tensorrt-error-network-validation-failed/156296/3
这个issue中有提到报错的原因是TensorRT的OnnxParser解析模型失败。

百度之后发现，#328中@cwentland0提到，在TensorRT-7.0中，OnnxParser只支持full-dimensions mode，也就是需要固定Batch Size并传一个explicitBatch flag到create_network方法中。

官方文档的Builder中有提到具体的实现方法：

pass a value of 1 << NetworkDefinitionCreationFlag.EXPLICIT_BATCH to create_network()。

不过具体操作中trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH并不是一个int值，因此需要进行强制类型转换：

1 << int(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH)

转换之后可以成功解析模型