Random Deep Learning experiments

• See training code and inference results in notebooks
• Implemented models are in models

Object Detection and Classification on Cats-and-Dogs dataset

No postprocessing added to Unet-based models

Model	cat AP	dog AP	mAP
Unet	51.62	67.31	59.47
Unet + SEBlocks	95.19	86.73	90.96
RetinaNet	46.38	68.65	57.52

Result samples from SEUnet

Image generation on Facades dataset

Model	pix2pix	sagan
Samples

Deep Fashion on Viton dataset

Reproduced models and pipeline from VITON: An Image-based Virtual Try-on Network

Result samples: