The goal of the OpenLLM-Ro is to bring together the Romanian community that builds open Romanian models and to collect these models in a single place.

We value:

• using public and open corpora
• open-source training and evaluation code.

In this organization, you can find RoLLM models, based on different underlying models and in different flavours (i.e., foundational, instruct, or chat variants). There are currently two collections:

• RoLlama2: Romanian models based on Llama2
• RoMistral: Romanian models based on Mistral
• RoGemma: Romanian models based on Gemma
• RoLlama3: Romanian models based on Llama3

See details in https://arxiv.org/abs/2405.07703.

We encourage the community to engage in discussions (to provide feedback, ask questions, or make improvement suggestions) in Hugging Face or GitHub.

We will also organize physical meetings (announced in advance) to brainstorm ideas, roadmap, and other technical aspects.