Home

Awesome

Awesome Transformer Awesome

A curated list of NLP resources focused on Transformer.

What is Transformer?

Transformer is a novel neural network architecture which has shown great promise in many natural language processing(NLP) tasks such as machine translation, language modeling, question answering and reading comprehension.

<img src="https://github.com/basicv8vc/awesome-transformer/blob/master/imgs/transformer.png" width = "500" height = "700" alt="The Transformer - model architecture " align=center />

More info here.

Table of Contents

Implementations

repositoryframeworkorganization
tensor2tensorTensorFlow(official)Google
allennlpPyTorchAllen AI
annotated-transformerPyTorchHarvard NLP
attention-is-all-you-need-pytorchPyTorchjadore801120
gluon-nlpMXNetdmlc
attention-is-all-you-need-kerasKerasLsdefine
attention_is_all_you_needChainersoskek

Theory

Lectures

lecturecourseschool
Transformers and Self-Attention For Generative ModelsCS224n: Natural Language Processing with Deep LearningStanford
Self Attention/CNNsCS 11-731 Machine Translation and Sequence-to-Sequence ModelsCMU
AttentionCS 11-747 Neural Networks for NLPCMU
AttentionCSC 421/2516 Neural Networks and Deep LearningToronto