Google Collaborator: Full example of Kaggle Seedlings + DPN92 on

Part 1 (2018) Beginner (2018)
1

Following numerous requests from the crowd, regarding porting our #Applied #Medical-AI labs to #google #collaborator, I am happy to publish the first fully working #PyTorch version :slight_smile:

It has everything you need, including installing #Kaggle CLI, downloading the Datasets and running a DP292 (Dual Path Network) on the Kaggle Seedlings data set.

Comments are welcomed,
https://colab.research.google.com/drive/1y0pgDW_0r4tPSk6URgWc3UekejIKBxDd

lab-0-DPN92-PyTorch-SeedLings-Google-Collab_ipynb_-_Colaboratory
lab-0-DPN92-PyTorch-SeedLings-Google-Collab_ipynb_-_Colaboratory.png1366×1218 260 KB

2

Why do you tell people to set up a tree structure for data, while the code uses only the train set and separates it into train and valid sets on its own?

It apparently confuses people.