gemma-3-1b-it
google/gemma-3-1b-it · GPL-3.0
A 1B model that reads images and speaks 140+ languages with a 128K context.
この解説はまだ日本語がありません。英語の原文を表示しています。
Gemma 3 1B brings multimodal input and an unusually long context to the smallest tier, handy for multilingual, document-heavy or image-aware edge apps.
スコア
導入の要点
pip install transformers 深読み
gemma-3-1b-it is Google's smallest Gemma 3 instruct model: 1B parameters, multimodal (text + image), a 128K context window and support for 140+ languages — a remarkably capable edge model.
It is gated on Hugging Face: you sign in and accept Google's usage terms before downloading, which is the standard flow for the whole Gemma family.
For developers it is the cheapest way into the Gemma 3 architecture — the same 128K context and multilingual coverage as its bigger siblings, at a size that runs on a phone or a single small GPU.