Optimizing the latency for transformer inference by removing redundant operations

How do model size and training data size affect validation loss and...

Different training metrics wrt to model quality, throughput, compute, memory

Exposing char LM as an API with different decoding strategies

Building a character level language model that learns patterns in data with...

Breaking down one of the most important architectures in history. Transformers

Implemented popular attention variants in pytorch Deep Learning RNN

Developing the utilies to analyze and track gradients in a architecture Deep Learning RNN

Covering all architectural advancements from RNN to Sequence-2-Sequence Model Deep Learning

Classifying sentiments of text using Seq2Vec Architecture Deep Learning RNN