I use this web(jetson-inference/pytorch-collect.md at master · dusty-nv/jetson-inference · GitHub)for training. How do I allocate the amount of photos of train, test and val?
Hi,
Based on the follow document, it’s recommended to collect at least 100 images for training.
In general, we reserve 1/5 for testing and validation.
Thanks.
Hi,
Suppose yes.
This is a general frameworks for collecting training dataset.
It can be used for both the default resnet18 and googlenet.
Thanks.