Deep Metric Learning | Contrastive, N-Pair, Triplet Loss, ArcFace, CosFace | Kaggle | Transformer скачать в хорошем качестве

Deep Metric Learning | Contrastive, N-Pair, Triplet Loss, ArcFace, CosFace | Kaggle | Transformer

Deep Metric Learning от Kha Vu, Microsoft. Transformer | Прогрессивное ML комьюнити Must have лекция по Metric Learning: ▪️ What is Metric Learning, and why do we need it? ▪️ Contrastive approaches ▪️ When can they produce good results, and what are their limitations? ▪️Contrastive, Triplet, and N-Pair losses ▪️Angular margin approaches ▪️In which cases are they better than Contrastive approaches? ▪️ArcFace, CosFace, and other SOTA methods ▪️A survey of how Metric Learning is used to solve real-world and Kaggle problems • Сайт проекта Transformer — https://transformer.community • Курс INTRODUCTION TO MACHINE LEARNING & NEURAL NETWORKS — https://transformer.community/introdu... • Телеграм-канал Мишин Лернинг — https://t.me/mishin_learning Подписывайтесь на YouTube-канал "Transformer", чтобы не пропустить следующее видео! #transformer