YOLOV5-SWIFT

This repository is based on ultralytics/yolov5. Seeed has modified it to make it more suitable for AIoT hardware devices launched by Seeed.

Model test results on the coco dataset

Model	Input_size	FLOPs	Params	Size（M）	[email protected]	[email protected]:0.95
yolov5n6-xiao	192×192	0.07G	0.4M	0.5	0.28	0.15

Test speed in different situations

Equipment	Computing backend	Input	specification	time(ms)
Intel	[email protected]	192×192	pt	15
Himax6537	400MHz	192×192	tflite	700

Quick start

Install

Python>=3.7.0 is required with all requirements.txt installed including PyTorch>=1.7:

$ git clone https://github.com/Seeed-Studio/yolov5-swift
$ cd YOLOv5-swift
$ pip install -r requirements.txt

Inference with detect.py

detect.py runs inference on a variety of sources, downloading models automatically from the YOLOv5n6-xiao and saving results to runs/detect.

$ python detect.py --source 0  # webcam
                            file.jpg  # image 
                            file.mp4  # video
                            path/  # directory
                            path/*.jpg  # glob
                            'https://youtu.be/NUsoVlDFqZg'  # YouTube
                            'rtsp://example.com/media.mp4'  # RTSP, RTMP, HTTP stream

Training

$ python train.py --data coco.yaml --cfg yolov5n6-xiao.yaml --weights yolov5n6-xiao.pt --batch-size 128    

Export TFlite

The trained pt format model can be exported to the int8 type tflite model by the following command

$ python export.py --data coco.yaml --cfg yolov5n6-xiao.yaml --weights yolov5n6-xiao.pt --imgsz 192 --int8   

Export UF2

UF2 is a file format, developed by Microsoft. Seeed uses this format to convert .tflite to .uf2, allowing tflite files to be stored on the AIoT devices launched by Seeed.

Currently Seeed's devices support up to 4 models, each model (.tflite) is less than 1M .

You can specify the model to be placed in the corresponding index with -t.

$ python uf2conv.py -f GROVEAI -t 1 -c xxx.tflite -o xxx.uf2 # Place the model to index 1

$ python uf2conv.py -f GROVEAI -t 1 xxx.tflite -o xxx.uf2 # Place the model to index 1 & flash it

References

[1] Official YOLOV5 repository

[2] Official UF2 repository