protocol-buffers - 无法将 ONNX 模型解析为 TensorRT

Question

我正在使用jetson nano

我试图转换 onnx 模型https://github.com/onnx/models/tree/master/vision/body_analysis/emotion_ferplus

遇到这个错误：

https://user-images.githubusercontent.com/28679735/86281506-a75e5380-bbab-11ea-8608-9bf8e2f50cc6.png

附加信息：

https://user-images.githubusercontent.com/28679735/86281617-d674c500-bbab-11ea-8bbe-16f6d3db7203.png

Answer 1

创建模型后，使用以下代码：

TRT_LOGGER = trt.Logger(trt.Logger.WARNING)
EXPLICIT_BATCH = 1 << (int)(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH)
with trt.Builder(TRT_LOGGER) as builder, builder.create_network(EXPLICIT_BATCH) as network, trt.OnnxParser(network, TRT_LOGGER) as parser:
        with open("modelfile.onnx", 'rb') as model:
            if not parser.parse(model.read()):
                for error in range(parser.num_errors):
                    print(parser.get_error(error))
        engine = builder.build_cuda_engine(network)

您可以直接使用引擎，也可以保存并稍后重复使用。

with open("output.engine", "wb") as f:
            f.write(engine.serialize())