Skip to content

Latest commit

Β 

History

History
43 lines (34 loc) Β· 1.52 KB

File metadata and controls

43 lines (34 loc) Β· 1.52 KB

Important Files Not on GitHub

Due to GitHub's file size limitations, some important files are stored locally only:

πŸ“Š Large Dataset Files (Not on GitHub)

  • dataset/train.csv (~50MB) - Main training dataset
  • dataset/test.csv (~25MB) - Test dataset for predictions

How to get them: These are the original competition datasets. Download from the source or contact repository owner.

πŸ–ΌοΈ Downloaded Images (Not on GitHub)

  • image_cache/*.jpg (2,183 files, ~500MB total) - Product images for computer vision
  • Only image_cache/image_features.json is included (metadata)

How to get them: Run the download script:

python scripts/download_computer_vision_images.py

βœ… What IS on GitHub

  • βœ… All Python code and models
  • βœ… Best model predictions (results/test_out_meta_learning.csv)
  • βœ… Image metadata (image_cache/image_features.json)
  • βœ… Sample datasets for testing
  • βœ… Complete documentation
  • βœ… All requirements and setup files

πŸš€ Quick Setup After Cloning

git clone https://github.com/abbinavv/AmazonML.git
cd AmazonML
python -m venv .venv
source .venv/bin/activate  # (.venv\Scripts\activate on Windows)
pip install -r requirements.txt

# Optional: Download images for computer vision
python scripts/download_computer_vision_images.py

πŸ“ˆ Best Results Available

The repository includes the best model predictions (50.12% SMAPE) in:

  • results/test_out_meta_learning.csv - Meta-learning model (best)
  • results/test_out_neural_enhanced.csv - Neural enhanced model