← All notes

Summary

Flamingo: a Visual Language Model for Few-Shot Learning (2022) was a landmark few-shot visual language model. It set an interleaved image-text architecture that later open and closed systems still follow: a frozen vision encoder, a large language model, and a way to insert visual tokens into the text stream.

What it introduced: Flamingo's gated cross-attention layers for mixing visual features into a pretrained language model, trained so the model can be prompted with interleaved images and text in a few-shot style.

What it led to: later open multimodal assistants, including LLaVA's visual instruction-tuning recipe (a full note in batch B). Live notes: ao-evals-multimodal-agents. See also the CLIP pointer in batch B, which is a common vision-encoder ancestor.

Where to read it: https://arxiv.org/abs/2204.14198 (v2, last revised 15 Nov 2022). NeurIPS 2022.

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

@misc{iil_note_r7en1qfh,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{Flamingo: a Visual Language Model for Few-Shot Learning}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2204-14198-flamingo/},
  note = {Permanent ID iil:note:r7en1qfh. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2204-14198-flamingo.md.