search

dog-qiuqiu / Yolo-Fastest

:zap: Based on yolo's ultra-lightweight universal target detection algorithm, the calculation amount is only 250mflops, the ncnn model size is only 666kb, the Raspberry Pi 3b can run up to 15fps+, and the mobile terminal can run up to 178fps+
Other
1.98k stars 427 forks source link

TensorRT supported? #8

Open CaeTism opened 3 years ago

CaeTism commented 3 years ago

Amazing work! I test the model of yolo-fastest-xl on a jetson nano, the FPS is more than 10 which is quite good. I wonder if the TensorRT is introduced according to jkjung's work(https://github.com/jkjung-avt/tensorrt_demos), which could we achieve a real-time detection on edge devices. Many thanks!

xinyuabcd commented 3 years ago

@CaeTism Hello，the FPS is just 1.5 when I test the model of yolo-fastest on the jetson nano. How did you do it ? Thanks very much.

CaeTism commented 3 years ago

@CaeTism Hello，the FPS is just 1.5 when I test the model of yolo-fastest on the jetson nano. How did you do it ? Thanks very much.

I test a local video in /usr/share/visionworks/sources/data/pedestrians.mp4 with a yolo-fastest-xl model.

MuhammadAsadJaved commented 3 years ago

@CaeTism Have you managed to run with TensorRT?

clementzhao commented 3 years ago

I tested xl on jetson nano with only 14 fps. And have some error while translating to onnx format. Does anyone successful run it with tensorRT?

yueyihua commented 3 years ago

@clementzhao @CaeTism @dog-qiuqiu @MuhammadAsadJaved TensorRT can support？