What ı should to do when ı made some mistakes on creating DataSet (DS) ?
Hello Data Collectors/ML/DL or AI engineers!
This causerie may help you when you are creating your Data Set for ML or DL. Because so many junior AI engineer (as like me) making some mistakes when collecting Data Set. What is the mistakes of collecting Data Set ?
For instance;
- Mixing the Sort
- Mixing information in .json or .coco file
- Mismatch of Photo and .json File annotations
and another mistakes. Maybe you can ask me or send me an email for about this mistakes and ı can give you some support specially.
Anyway, first of all the solution, DO NOT USE ANOTHER CONFİGURE PROGRAM!
For example don’t use rename program or another thing because there are some reason for my attention.
- You don’t know how it works this type of programs.
- You don’t know are these programs safe or not ?
- Generally this type of programs listing your images or json files complicated and renaming also complicated.
What you should to do ?
You need to find or write some codes in Python or another programming language for sorting and renaming your files. And if you made some mistakes which ı wrote upside than you need to find easiest way for configure them. For example if you gonna change image names inside of json files it will more complicated situation. Than you can change the name of images from image folder. And you can use json file as a references because json files will include older name of images.
