mii-llm/maestrale-chat-v0.2-alpha-sft
mii-llm/maestrale-chat-v0.2-alpha-sft is a 7 billion parameter Mistral-7b based language model, developed by @efederici and @mferraretto, specifically optimized for the Italian language. It underwent continued pre-training on a curated large-scale, high-quality Italian corpus and was fine-tuned with approximately 270k Italian conversations/instructions. This model is designed for chat-based applications requiring strong Italian language capabilities.
Loading preview...
Maestrale Chat v0.2 Alpha
Maestrale Chat v0.2 Alpha is a 7 billion parameter language model built upon the Mistral-7b architecture, developed by @efederici and @mferraretto. Its primary distinction lies in its extensive specialization for the Italian language.
Key Capabilities
- Italian Language Proficiency: The model has undergone significant continued pre-training using a large, high-quality corpus specifically curated for Italian.
- Instruction Following: It has been fine-tuned with approximately 270,000 Italian conversations and instructions, enhancing its ability to understand and respond to user prompts in a chat format.
- ChatML Format: The model is designed to work with the ChatML prompt format, including a system prompt for safe and helpful assistance.
Intended Uses & Limitations
This model is currently an alpha version, indicating ongoing development. It is not yet fully aligned, and the developers are actively working on alignment data and evaluation metrics. Users should be aware of its alpha status and potential limitations in terms of alignment and safety compared to more mature models. It is best suited for experimental applications and development focusing on Italian language generation and conversational AI.