python src/main.py --mode 'train' --data_path {dataset path}python main.py --mode 'test_all_point' --data_path {dataset path}RUN draw_plot.ipynb- Qualitative Result: Anomaly score plot for all moment
- Quantative Result: AUROC, AUPRC, Best-F1 Score
[Hyperparameter]
| Name | Description |
|---|---|
| Epochs | 10 |
| Batch Size | 256 |
| Learning Rate | 1e-4 |
| lambda_energy | 0.1 |
| lambda_cov | 0.005 |
| number of gaussian components | 5 |
[Paper] Deep Autoencoding Gaussian Mixture Model for Unsupervised Anomaly Detection (ICLR,2018)
[Youtube Review] 발표자: 고려대학교 산업경영공학과 DSBA 연구실 이윤승(https://github.com/yun-ss97)
Reference: [code]