Skip to content

Latest commit

 

History

History
17 lines (12 loc) · 909 Bytes

README.md

File metadata and controls

17 lines (12 loc) · 909 Bytes
Raw

Analysis & Data Presentation Laboratory- 094295

HW2-face-mask-detection

Python code for face-mask object detection & classification task using Torchvision fasterRCNN.

The repository includes a project report under Project_Report.pdf file.

In order to evaluate the trained model on a new dataset, run python predict.py >PATH_TO_FOLDER>. This will save a prediction.csv file with the predictions, in the same directory.

The repository also includes:

  • basic_experiment dir with the basic model we have trained- for more information please look at the report.
  • loading dataset script- dataset.py
  • training script- train.py
  • main script- main.py that runs the training process using functions from my_utils dir, and parameters from cfg dir.
  • EDA script- EDA.py

In addition, the repository includes example images, example images script and example prediction script.