Implementation of 'Variational Autoencoders for Collaborative Filtering' paper [https://arxiv.org/abs/1802.05814] in Pytorch
- Data Processing
- Denoising AutoEncoder Model in the paper
- Validation for the model
- Model Ablation Studies: a. Explicit Use of l2 norm for model weights b. Effect of dropout
Dependency: Python3
Will upload the code when the results are comparable
Final: Model with explicit l2 loss no_l2 : Model without explicit l2 loss no_drop: Model without dropout. It overfits as expected.
- Clean up the code
- Add VAE model
- Add test matrix in the code
Note: Data preparation is mostly based on original implementation source with some changes.