I love Machine Learning / NLP , especially for analyzing scientific texts
4x faster pinned CPU -> GPU data transfer than Pytorch pinned CPU tensors, and 110x faster GPU -> CPU transfer. Augment parameter size by hosting on CPU. Use non sparse optimizers for sparse training
Sitcom Audience Reaction in Real Life
A book recommender for rare books, with 2829853 books in our system
Fast CPU -> GPU for Pytorch
We propose a model for optimal decision-making on arbitrary, complex modalities, with applications to healthcare, robotics, HCI, and general artificial intelligence.
Ask us a medical question, and our AI will help you find related questions and answers, and generate advice specifically tailored for your query.