使用PaddleSeg推理（infer.py）出现的问题

[LarryHYQ]

2022-01-16 13:13:51 [INFO] Use GPU 2022-01-16 13:13:51 [INFO] Use TRT 2022-01-16 13:13:51 [INFO] Use manual set dynamic shape Traceback (most recent call last): File "deploy/python/infer.py", line 393, in main(args) File "deploy/python/infer.py", line 380, in main predictor = Predictor(args) File "deploy/python/infer.py", line 137, in init self.predictor = create_predictor(self.pred_cfg) ValueError: (InvalidArgument) Pass tensorrt_subgraph_pass has not been registered. Please use the paddle inference library compiled with tensorrt or disable the tensorrt engine in inference configuration! [Hint: Expected Has(pass_type) == true, but received Has(pass_type):0 != true:1.] (at /paddle/paddle/fluid/framework/ir/pass.h:236)

（infer代码没有改动，命令行输入的是：!python deploy/python/infer.py --config output_quant/segformer_quant/deploy.yaml --image_path ./../../data/data_2022_baseline/JPEGImages --batch_size 1 --save_dir output_quant/result_quant --use_trt True --precision int8）

[juncaipeng]

请参考文档安装支持trt的whl包：https://github.com/PaddlePaddle/PaddleSeg/blob/release/2.3/docs/deployment/inference/python_inference.md

然后下载trt包，进行配置。