Transformers (stub)

Purpose: a transformer-centric view of modern model families (LLMs, vision transformers, multimodal), plus key architectural variants.

Original architecture: Vaswani et al., 2017 (Attention Is All You Need)
Masked-language modeling (encoder-only): BERT
Autoregressive next-token prediction (decoder-only): GPT-3
Transformers for vision: Vision Transformer (ViT)
Open-access multilingual LLM example: BLOOM