The Yi series of models is made with 6-bit and 34-bit LLMs, which are both open source. There’s a sequence length of 4K, in a 34-billion parameter large training model that outperforms the 70 billion ...
Some results have been hidden because they may be inaccessible to you