What’s that now?
Long story, but the most developed family here is the transformer or Sparse Transformer etc for particularly developed examples and explanations of this sub-field. The best illustrated blog post is Jay Alammar’s Illustrated Transformer.
These networks are absolutely massive (heh) in natural language processing right now.
A key point about these networks is that they can be made extremely large but still remain trainable. This leads to interesting scaling laws.
A good paper read is Yannic Kilcher’s.
HuggingFace distributes and documents and implements a lot of Transformer/attention NLP models and seem to be the most active neural NLP project. Certainly too active to explain what they are up to in between pumping out all the code.
The library currently contains PyTorch and Tensorflow implementations, pre-trained model weights, usage scripts and conversion utilities for the following models:
- BERT (from Google) released with the paper BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding by Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova.
- GPT (from OpenAI) released with the paper Improving Language Understanding by Generative Pre-Training by Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever.
- GPT-2 (from OpenAI) released with the paper Language Models are Unsupervised Multitask Learners by Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever.
- Transformer-XL (from Google/CMU) released with the paper Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context by Zihang Dai, Zhilin Yang, Yiming Yang, Jaime Carbonell, Quoc V. Le, and Ruslan Salakhutdinov.
- XLNet (from Google/CMU) released with the paper XLNet: Generalized Autoregressive Pretraining for Language Understanding by Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, and Quoc V. Le.
- XLM (from Facebook) released together with the paper Cross-lingual Language Model Pretraining by Guillaume Lample and Alexis Conneau.
- RoBERTa (from Facebook), released together with the paper a Robustly Optimized BERT Pretraining Approach by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov.
- DistilBERT (from HuggingFace) released together with the paper DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter by Victor Sanh, Lysandre Debut, and Thomas Wolf. The same method has been applied to compress GPT2 into DistilGPT2.
- [very long list excised]
GPT-Neo is the code name for a series of transformer-based language models loosely styled around the GPT architecture that we plan to train and open source. Our primary goal is to replicate a GPT-3 sized model and open source it to the public, for free.
This guide to pruning multihead attention NN should probably go somewhere useful if i actually end up doing NLP like all the recruiters seem to want.