Skip to content

how to upload a Yolov5 dataset to Label-Studio through Fiftyone library

Notifications You must be signed in to change notification settings

fatemeh-mohseni-AI/fiftyone_labelstudio

Repository files navigation

** my env: ubuntu 20.04 python 3.8.10 **

just install label-studio version 1.7.3 on version 1.8.0 faces an unrecognized weired error shown as below :

<< raise ValidationError(f'{ext} extension is not supported') >>

so install noted version . after first run django will return an error due to limitation on upload file numbers and is does not related to label-studio . it is a Django limitation . ~~ i changed directly the line of code from the line :

<< venv/lib/python3.8/site-packages/django/http/multipartparser.py", line 242 >> and change << settings.DATA_UPLOAD_MAX_NUMBER_FILES >> to number 1000 . so it compare number of uploaded files with 1000 not the django setting .

run label-studio first , then run label_studio_upload.py .

if you faced runtime-error just do following change : <<venv/lib/python3.8/site-packages/label_studio_sdk/client.py>>

~~change the line TIMEOUT= (10,50) line 16 to the : TIMEOUT = (100.0, 5000.0)

NOTE : the tree is as follows :

dir/ | ------dataset.yaml ------images | ---train ---test ------labels | ---train ---test

and dataset.yaml content looks like :

                                                | |
                                                \ /

train: ./images/train/ val: ./images/val/

number of classes

nc: 5

class names

names: ["clean_water_pot", "dirty_water_pot", "feeding_line", "feeding_pot", "watering_line"]

About

how to upload a Yolov5 dataset to Label-Studio through Fiftyone library

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published