Hey there! So, I’ve been using picstork to train models for my college project to detect Water Bodies and I’m a bit stuck on one thing: How many annotations do you think are enough to really train a model effectively?
1 Like
This totally depends upon the type of objects and images . Do share some samples
1 Like
Hello @Aditi , Absolutely it depends upon your use case and your accuracy requirements .Ask yourself a few question’s before you train your model . How accurate i want my model ? , Do i need to setup my task in a way which only count’s the water bodies or do i want to get count or water-bodies have some kind of defects making all the water bodies look different from each other this would require more number of examples i mean more number of annotations in order to get a accurate model.
1 Like