Strengths4
- End-to-end execution from data loading to submission generation.
- Proper handling of class imbalance via weighted loss.
- Use of stratified split for reliable validation.
- Model checkpointing based on validation F1-score.
Competition Solution · Jupyter Notebook
This repository contains a Jupyter notebook that implements a soil image classification model using a pretrained ResNet18 with a custom classification head and dropout, trained on the Annam.ai Kaggle competition dataset, and generates a submission file for the competition.
Merit describes the code at this commit. Later pushes count once the owner rescans. How merit is scored
Each field's share of the project's AI work, as scoring counts it. It ranks in the field it is most about.
Read from the project's code at its scored commit.
The owner keeps this project's limitations private.