This repository contains implementations of various transformer models for different natural language processing and computer vision tasks.
Paper: BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
- Fill-mask language model pretraining for downstream tasks โ
- Sequence classification โ
- Token classification ๐
- Next sentence prediction ๐
Paper: Language Models are Unsupervised Multitask Learners
- Semi-supervised training for sequence generation ๐
Paper: An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- Image inpainting ๐
- Image classification ๐