Progress Reports

Progress report 07/21 - 08/21

this report is about the applied techniques to improve the performance of Mask RCNN (tensorflow matterport version) on deepfashion2 dataset

Progress report 10/21 

the impact of reducing the classes of the dataset to 5.

Progress Report 11/21 and 12/21

- Reading the paper of DETR 

- Searching about the most effective cloud GPU (price and instance memory)

- Cross Validation of mask rcnn (matterport TF 1.4)

Progress Report 09/21

-implementation of coco evaluation metrics

- inspecting the subsets of deep fashion2

- extracting good data

- training for days

Progress report of 02/22

after starting to work on stile.ai

Progress report 03/22

after training on mask rcnn detecton2 (pytorch), extracting results and evaluation 

progress report 04/22

-stile.ai diagram using AWS infrastructure

-5 folds cross valdation using mask rcnn detectron2 (pytorch) 

-applying detection inference on Gif and Videos

progress report 05/22

dockerization and deployment