Setup:

Python version>=3.6 is required

1. Clone the repo : https://github.com/voxel51/fiftyone.git

2. Setup the environment by running the pip install -r requirements.txt command.

3. Now run the command pip install fiftyone.

4. To start the instance run the command fiftyone quickstart.

5. Now create the dataset.yaml file containing the path for the dataset and the respective classes which is similar to the custom_data.yaml so rename it to dataset.yaml and place it in the dataset directory.

6. Now run the command to view the dataset into fiftyone app fiftyone app view
   --dataset-dir $DATASET_DIR
   --type fiftyone.types.YOLOv5Dataset

7. Where $DATASET_DIR is replaced with the path for the dataset directory.

8. Now in the app we can check the number of images for each label by clicking on the right side show written along with histogram icon .

9. Also you can check the id, file_path and images are correctly labeled or not from the right side ground_truth instance.

#Referrence

#Voxel-fityone doc

Troubleshooting steps:

If encountered with the mongodb error or any database error run the following command: pip install fiftyone-db-ubuntu1604