# Trillion-token training

Poro is a 34-billion-parameter multilingual large language model focused on Finnish, English, and code processing, open-sourced under Apache 2.0 license.

StarCoderBase-1B is a 1-billion-parameter code generation model trained on over 80 programming languages, supporting fill-in-the-middle generation and multi-query attention mechanisms.

OpenLLaMA is an open-source reproduction of Meta AI's LLaMA large language model, offering pretrained models with 3B, 7B, and 13B parameter scales

OpenLLaMA is an open-source reproduction of Meta AI's LLaMA large language model, offering pre-trained models with 3B, 7B, and 13B parameter scales

StarCoderBase is a large-scale code generation model with 15.5 billion parameters, trained on over 80 programming languages, supporting code completion and generation tasks.