TheBloke/Vigogne-Instruct-13B-HF
TheBloke/Vigogne-Instruct-13B-HF is a 13 billion parameter LLaMA-based instruction-following language model specifically fine-tuned for French language tasks. Developed by bofenghuang, this model excels at understanding and generating responses to instructions provided in French. It is primarily intended for research use, offering a specialized solution for French natural language processing applications.
Loading preview...
Vigogne-Instruct-13B-HF: A French Instruction-Following LLaMA Model
Vigogne-Instruct-13B-HF is a 13 billion parameter LLaMA-based model, fine-tuned by bofenghuang to specialize in following instructions in French. This model is a merged LoRA version of the original Vigogne Instruct 13B, provided in fp16 Hugging Face format by TheBloke for broader accessibility.
Key Capabilities
- French Instruction Following: Specifically optimized for understanding and responding to prompts and instructions given in French.
- LLaMA-13B Base: Built upon the robust LLaMA-13B architecture, providing a strong foundation for language generation.
- Research-Oriented: Intended and licensed for research use only, aligning with the original Stanford Alpaca and Vigogne project guidelines.
Good For
- French NLP Research: Ideal for academic and research projects focusing on French language models and instruction tuning.
- Experimentation: Useful for developers and researchers looking to experiment with French-specific LLMs.
Limitations
As an ongoing development, Vigogne may generate harmful, biased, incorrect, or unhelpful content. Its use is restricted to non-commercial research purposes due to its CC BY NC 4.0 dataset license.