Week 1

Feb 1, 23

1. fmepy
   1. Create a project board
   2. Check the progress and other stats from Vannary's run
   3. Check the commit made by VS
   4. Talk about review-nb and its membership thing
2. ocr-in-python
   1. This is something which is on low-priority
   2. Going from the perspective of object detection will not help, as object-detection also has a notion of associated text with the box
   3. EasyOCR works better than pytessaract
3. CCTV stuff
   1. creating the training pipeline with timm
   2. weighted loss functions? Implementation created does not work as expected
4. ORSANCO
   1. datetime issues in the frontend
   2. gte, lte were used for the query on the backend
5. EPA
   1. No new updates on the EPA front
6. CHAMA
   1. No new updates on the EPA front

Feb 2, 23

Feb 3, 23

1. Next steps for Vannary on the CCTV front
   1. Train the model from scratch -- do not use pretraining
2. Presentation for SD1
   1. Some insights about the data
   2. Some stats