metadata
language:
- fi
license: apache-2.0
tags:
- finnish
- llama
inference: false
pipeline_tag: text-generation
Llama-3b V2 for Finnish
Training for the new V2 model in progress. 10% of the training already exceeds the previous Finnish-NLP/llama-3b-finnish and Finnish-NLP/llama-7b-finnish models in the FIN-bench evaluation.