Skip to content

Latest commit

 

History

History
27 lines (22 loc) · 1.04 KB

README.md

File metadata and controls

27 lines (22 loc) · 1.04 KB

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.