Use your yolov5 predictions as supervisely annotations
-
clone this repo
-
install dependencies:
pip install -r requirements.txt
-
set meta.json file
-
set
WEIGHTS
,DEVICE
,SOURCE_DIR
,SAVE_DIR
,IMAGE_SIZE
,CONF_THRESHOLD
,IOU_THRESHOLD
,META_JSON_PATH
params in yolov5_to_supervisely.py. -
run:
python yolov5_to_supervisely.py
- upload output folder into supervisely via
Import Via Apps/Import Images in Supervisely Format
thanks to yavuzKomecoglu for the inspiration