CodeGemma

Google · April 2024

activeOpen Sourcedecoder onlycode
Parameters2B / 7B
Context Window8K tokens
Variants2B, 7B

Why It Matters

Google's lightweight code model built on Gemma, showing that even 2B-parameter models could be useful for code completion.

Description

Google's lightweight code model built on the Gemma architecture, available in 2B and 7B parameter sizes. The 2B variant is specifically optimized for fast code completion in IDEs (suggesting the next few tokens as you type), while the 7B variant handles more complex tasks like code generation and explanation.

Key Innovations

Code Gen
Code GenAbility to write, debug, and understand programming code across multiple languages.

Family Tree

Built On

Lineage

GemmaCodeGemma