// the find
ashishpatel26/Treasure-of-Transformers
💁 Awesome Treasure of Transformers Models for Natural Language processing contains papers, videos, blogs, official repo along with colab Notebooks. 🛫☑️
A link collection of transformer model papers, blogs, YouTube videos, and Colab notebooks spanning 2017-2021. It's aimed at ML practitioners and students who want a single page to find canonical resources for BERT, GPT variants, and the long tail of transformer architecture research from that era.
The paper+blog+video+Colab four-column format is genuinely useful — you get both the theory and a runnable example without hunting. Coverage of the 2019-2020 efficiency variants (Longformer, Linformer, Reformer, Performer) is solid and those are still practically relevant. Having GPT-Neo, RAG, and CodeBERT alongside the classics gives it reasonable breadth beyond vanilla BERT fine-tuning tutorials.
The list stops cold around 2021, so nothing post-GPT-3 era is here — no LLaMA, no Mistral, no instruction-tuned models, no RLHF architectures. A substantial number of Colab and video links are dead or point to unrelated notebooks. The single GPT-Neo notebook is the only actual code in the repo; everything else is pointers, so if those external resources disappear this becomes an empty table. No curation signal either — a marginal 2020 paper gets the same row as BERT.