Llama3-FFM
The first Traditional Chinese language model to achieve a TMMLU+ score that surpasses GPT-4.0 sets a new record for Traditional Chinese models!
Llama3-FFM is trained on Meta-Llama3 with an excellent Traditional Chinese corpus, featuring a native 128K multilingual extended vocabulary, powerful reasoning capabilities, and diverse application scenarios. In addition, the model has a Tool Call function, which can quickly connect to external information to expand comprehensive commercial applications.
Llama3-FFM
★ First Traditional Chinese model benchmark surpasses GPT-4.0 ★
★ Offers FFM large language model 70B/8B options ★
★ The Llama series has been fully upgraded, with significant improvements in all aspects of performance. ★
★ Supports Tool Call functionality for quick integration with external applications ★
★ Features a native 128K multilingual expanded vocabulary ★