YOLOv5-Seg

YOLOv5

This version of YOLOv5 has been converted to run on the Axera NPU using w8a16 quantization.

This model has been optimized with the following LoRA:

Compatible with Pulsar2 version: 3.4

Convert tools links:

For those who are interested in model conversion, you can try to export axmodel through

• The repo of ax-samples, which you can get the how to build the ax_yolov5s_seg

• The repo of axcl-samples, which you can get the how to build the axcl_yolov5s_seg

• Pulsar2 Link, How to Convert ONNX to axmodel

Support Platform

• AX650
  • M4N-Dock(爱芯派Pro)
  • M.2 Accelerator card
• AX630C
  • 爱芯派2
  • Module-LLM
  • LLM630 Compute Kit

How to use

Download all files from this repository to the device

root@ax650 ~/yolov5-seg # tree -L 2
.
├── ax650
│   └── yolov5s-seg.axmodel
├── ax_aarch64
│   └── ax_yolov5s_seg
├── config.json
├── football.jpg
├── README.md
├── yolov5_seg_config.json
├── yolov5s-seg-cut.onnx
├── yolov5s-seg.onnx
└── yolov5s_seg_out.jpg

3 directories, 10 files

Inference

Input image:

Inference with AX650 Host, such as M4N-Dock(爱芯派Pro)

root@ax650 ~/yolov5-seg # ./ax_yolov5s_seg -m yolov5s-seg.axmodel -i football.jpg
--------------------------------------
model file : yolov5s-seg.axmodel
image file : football.jpg
img_h, img_w : 640 640
--------------------------------------
Engine creating handle is done.
Engine creating context is done.
Engine get io info is done.
Engine alloc io is done.
Engine push input is done.
--------------------------------------
post process cost time:9.19 ms
--------------------------------------
Repeat 1 times, avg time 9.55 ms, max_time 9.55 ms, min_time 9.55 ms
--------------------------------------
detection num: 6
 0:  90%, [ 747,  224, 1140, 1147], person
 0:  89%, [1356,  337, 1622, 1035], person
 0:  88%, [   3,  364,  308, 1094], person
 0:  81%, [ 491,  479,  668, 1015], person
32:  78%, [ 777,  887,  827,  942], sports ball
 0:  59%, [1840,  690, 1905,  812], person
--------------------------------------

Output image: