jetson nx jetpack4.6 使用trt_fp16推理出现报错

gxwyxajh88 commented 2 years ago

问题确认 Search before asking

[x] 我已经查询历史issue，没有报过同样bug。I have searched the issues and found no similar bug report.

bug描述 Describe the Bug

--run_mode=paddle 不开启trt的情况下可以正常运行输出，用命令“python3 tools/export_model.py -c configs/ppyoloe/ppyoloe_crn_s_300e_coco.yml -o weights=https://paddledet.bj.bcebos.com/models/ppyoloe_crn_s_300e_coco.pdparams trt=True”导出trt以后进行fp16推理出现异常，输出信息如下： ----------- Running Arguments ----------- action_file: None batch_size: 1 camera_id: -1 cpu_threads: 1 device: gpu enable_mkldnn: False enable_mkldnn_bfloat16: False image_dir: None image_file: demo/000000014439_640x640.jpg model_dir: output_inference/ppyoloe_crn_s_300e_coco output_dir: output random_pad: False reid_batch_size: 50 reid_model_dir: None run_benchmark: False run_mode: trt_fp16 save_images: False save_mot_txt_per_img: False save_mot_txts: False save_results: False scaled: False threshold: 0.5 tracker_config: None trt_calib_mode: False trt_max_shape: 1280 trt_min_shape: 1 trt_opt_shape: 640 use_dark: True use_gpu: False video_file: None window_size: 50

----------- Model Configuration ----------- Model Arch: YOLO Transform Order: --transform op: Resize --transform op: NormalizeImage --transform op: Permute

E0704 17:15:40.951094 9144 helper.h:111] Invalid parameter: NMS topK (10000) exceeds limit (4096) E0704 17:15:40.951248 9144 helper.h:111] Invalid parameter: NMS topK (10000) exceeds limit (4096) E0704 17:15:40.953336 9144 helper.h:111] Invalid parameter: NMS topK (10000) exceeds limit (4096) E0704 17:15:44.247506 9144 helper.h:111] Invalid parameter: NMS topK (10000) exceeds limit (4096) E0704 17:15:45.962250 9144 helper.h:111] Invalid parameter: NMS topK (10000) exceeds limit (4096) E0704 17:15:46.186007 9144 helper.h:111] Invalid parameter: NMS topK (10000) exceeds limit (4096)

复现环境 Environment

PaddlePaddle: 2.3.0 (飞浆官网下载的 paddlepaddle_gpu-2.3.0-cp36-cp36m-linux_aarch64.whl) PaddleDetection: 2.4 Python: 3.6 CUDA: 10.2 CUDNN: 8.2.1.32 TensorRT: 8.0.1.6

是否愿意提交PR Are you willing to submit a PR?

[ ] Yes I'd like to help by submitting a PR!

OPPOA113 commented 1 year ago

GTX 3090 paddledet 2.5.0 paddlepaddle-gpu 2.4.2.post112 运行yolox-s，出现同样的问题： GPU 模式，正常，GPU-trtfp32错误！！

bbilixzc commented 6 months ago

使用PaddleInference将PaddleDetection YOLOX系列的模型部署到Jetson Nano和Jetson Xavier NX，不使用TensorRT推理正常，开启TensorRT出现相似的问题。 b345b427ba14ec94310ce4994c911ee

GTX 3090 paddledet 2.5.0 paddlepaddle-gpu 2.4.2.post112 运行yolox-s，出现同样的问题： GPU 模式，正常，GPU-trtfp32错误！！

PaddlePaddle / PaddleDetection