DeepSeek-Coder

DeepSeek · November 2023

activeOpen Sourcedecoder onlycode
Parameters1.3B - 33B
Context Window16K tokens
Variants1.3B, 6.7B, 33B, Instruct

Description

DeepSeek's code-specialized model family, trained from scratch on 2 trillion tokens of code and natural language. Available in sizes from 1.3B to 33B parameters, it was one of the first Chinese-developed open-source code models to rival Western alternatives on coding benchmarks.

Key Innovations

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

Family Tree

Built On

Lineage

DeepSeek V1DeepSeek-Coder

Successors (1)

External Links