PaddlePaddle / Paddle-Lite

PaddlePaddle High Performance Deep Learning Inference Engine for Mobile and Edge (飞桨高性能深度学习端侧推理引擎)
https://www.paddlepaddle.org.cn/lite
Apache License 2.0
6.9k stars 1.6k forks source link

Amlogic A311D no device specified #10008

Closed zhangzaixiu closed 1 year ago

zhangzaixiu commented 1 year ago

Amlogic A311D上运行图像分类示例程序run.sh运行成功 用mobilenetv1_full_api.py YOLOv3-MobileNetV3 训练的识别人摔倒的目标检测模型,报如下错误 image 用mobilenetv1_light_api.py YOLOv3-MobileNetV3 训练的识别人摔倒的目标检测模型,报如下错误 image

paddle-bot[bot] commented 1 year ago

您好,我们已经收到了您的问题,会安排技术人员尽快解答您的问题,请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时,您也可以通过查看官网文档常见问题历史Issue来寻求解答。祝您生活愉快~

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the APIFAQ and Github Issue to get the answer.Have a nice day!

hong19860320 commented 1 year ago

@yingshengBD 跟一下这个问题吧

yingshengBD commented 1 year ago

1,首先这个示例程序,默认使用full api。light api是搭配 opt工具使用的,这里就不展开了 2,你的full的截图,没有截到错误信息 3,如果要启用311d NPU的算力,是需要做模型全量化的。yolov3的neck和head相对后续版本更复杂、性能也更差,全量化可能会有部分op量化信息缺失,这里先建议使用其他检测模型,推荐 yolov5、ppyoloe(yolov6 7 8也能支持,但是综合精度和性能考虑还是推荐yolov5、ppyoloe)

zhangzaixiu commented 1 year ago

1,首先这个示例程序,默认使用full api。light api是搭配 opt工具使用的,这里就不展开了 2,你的full的截图,没有截到错误信息 3,如果要启用311d NPU的算力,是需要做模型全量化的。yolov3的neck和head相对后续版本更复杂、性能也更差,全量化可能会有部分op量化信息缺失,这里先建议使用其他检测模型,推荐 yolov5、ppyoloe(yolov6 7 8也能支持,但是综合精度和性能考虑还是推荐yolov5、ppyoloe)

1用full的时候用的是没有opt转化的模型 用light的时候用的是已经opt转化的模型 2 full的有错误信息 如第一个截图所示 3 即使用推荐的demo,也是提示这个错误,yolo3虽然比不得yolo5,但是既然兼容,应该跑的通才对 image

yingshengBD commented 1 year ago

image 这里面m没有错误信息 都是info和warning 不影响

zhangzaixiu commented 1 year ago

Reference in new

image 这里面m没有错误信息 都是info和warning 不影响

这种确实也没运行出结果来,无论用文档给的模型还是自己训练的模型,该怎么处理