haotian/duan8
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

yolov5 m/l/x speed test

Browse Source
This commit is contained in:
wang-xinyu 2020-08-03 19:20:03 +08:00
parent ee6dd8788c
commit 4897bb32ee
3 changed files with 7 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@ -87,7 +87,6 @@ Some tricky operations encountered in these models, already solved, but might ha
|-|-|:-:|:-:|:-:|:-:| |-|-|:-:|:-:|:-:|:-:|
| YOLOv3-tiny | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 333 | | YOLOv3-tiny | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 333 |
| YOLOv3(darknet53) | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 39.2 | | YOLOv3(darknet53) | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 39.2 |
| YOLOv3-spp(darknet53) | Xeon E5-2620/GTX1080 | 1 | FP32 | 256x416 | 94 |
| YOLOv3-spp(darknet53) | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 38.5 | | YOLOv3-spp(darknet53) | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 38.5 |
| YOLOv4(CSPDarknet53) | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 35.7 | | YOLOv4(CSPDarknet53) | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 35.7 |
| YOLOv4(CSPDarknet53) | Xeon E5-2620/GTX1080 | 4 | FP16 | 608x608 | 40.9 | | YOLOv4(CSPDarknet53) | Xeon E5-2620/GTX1080 | 4 | FP16 | 608x608 | 40.9 |
@ -95,6 +94,9 @@ Some tricky operations encountered in these models, already solved, but might ha
| YOLOv5-s | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 142 | | YOLOv5-s | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 142 |
| YOLOv5-s | Xeon E5-2620/GTX1080 | 4 | FP16 | 608x608 | 173 | | YOLOv5-s | Xeon E5-2620/GTX1080 | 4 | FP16 | 608x608 | 173 |
| YOLOv5-s | Xeon E5-2620/GTX1080 | 8 | FP16 | 608x608 | 190 | | YOLOv5-s | Xeon E5-2620/GTX1080 | 8 | FP16 | 608x608 | 190 |
| YOLOv5-m | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 71 |
| YOLOv5-l | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 40 |
| YOLOv5-x | Xeon E5-2620/GTX1080 | 1 | FP16 | 608x608 | 27 |
| RetinaFace(resnet50) | TX2 | 1 | FP16 | 384x640 | 15 | | RetinaFace(resnet50) | TX2 | 1 | FP16 | 384x640 | 15 |
| RetinaFace(resnet50) | Xeon E5-2620/GTX1080 | 1 | FP32 | 928x1600 | 15 | | RetinaFace(resnet50) | Xeon E5-2620/GTX1080 | 1 | FP32 | 928x1600 | 15 |
| ArcFace(LResNet50E-IR) | Xeon E5-2620/GTX1080 | 1 | FP32 | 112x112 | 333 | | ArcFace(LResNet50E-IR) | Xeon E5-2620/GTX1080 | 1 | FP32 | 112x112 | 333 |

5
yolov5/README.md
View File

@ -2,7 +2,7 @@
The Pytorch implementation is [ultralytics/yolov5](https://github.com/ultralytics/yolov5). The Pytorch implementation is [ultralytics/yolov5](https://github.com/ultralytics/yolov5).
I made a copy of [yolov5s.pt(google drive)](https://drive.google.com/drive/folders/1Yaamfa-t_V3ImxYRBESqGzy7k4Arlt95?usp=sharing). Just in case the yolov5 model updated. I made a copy of [yolov5-s/m/l/x.pt(google drive)](https://drive.google.com/drive/folders/1Yaamfa-t_V3ImxYRBESqGzy7k4Arlt95?usp=sharing). Just in case the yolov5 model updated.
## Config ## Config
@ -25,6 +25,7 @@ git clone https://github.com/ultralytics/yolov5.git
// download its weights 'yolov5s.pt' // download its weights 'yolov5s.pt'
cd yolov5 cd yolov5
cp ../tensorrtx/yolov5/gen_wts.py . cp ../tensorrtx/yolov5/gen_wts.py .
// ensure the file name is yolov5s.pt and yolov5s.wts in gen_wts.py
python gen_wts.py python gen_wts.py
// a file 'yolov5s.wts' will be generated. // a file 'yolov5s.wts' will be generated.
@ -32,7 +33,7 @@ python gen_wts.py
mv yolov5s.wts ../tensorrtx/yolov5/ mv yolov5s.wts ../tensorrtx/yolov5/
cd ../tensorrtx/yolov5 cd ../tensorrtx/yolov5
ensure the macro NET in yolov5.cpp is s // ensure the macro NET in yolov5.cpp is s
mkdir build mkdir build
cd build cd build
cmake .. cmake ..