p208p2002/llama-3-zhtw-8B

The p208p2002/llama-3-zhtw-8B is an 8 billion parameter Llama 3-based language model developed by p208p2002, fine-tuned with 800M additional tokens for Traditional Chinese (zhtw) language capabilities. It maintains the original Llama 3's English MMLU performance due to its continued pre-training on FineWeb, while also incorporating Chinese and code datasets. This model is designed for applications requiring strong English language understanding alongside Traditional Chinese processing, offering a balanced performance profile.