README.md

Documentation

Here we detail every step performed inside the repository.

• 01_dataset_creation:
  • scraping, filtering, annotation, quality review, and constructing COCO-style dataset.
• 02_data-analysis:
  • brief exploratory data analysis.
• 03_training:
  • Facere, Facere+ model architectures, training, experiment tracking.
• 04_inference:
  • Converting trained models to ONNX format, inference on both Facere and Facere+ and other SoTA models such as; Attribute Mask R-CNN and Fashionformer.
• 05_evaluation:
  • Evaluation metrics and results.
• 06_visualization:
  • Visualizing model predictions (bounding boxes and segmentation masks), analysis on False Positives, various visualizations including Reliability Diagrams.
• FAQ:
  • Other details regarding pipeline, dataset, further work, and closing remarks.

---

Next: Dataset Creation