从我以前炼的一个模型（<a target="_blank" rel="ugc" href="https://huggingface.co/meishi/AiMeishi_cnXianxia/tree/main">https://huggingface.co/meishi/AiMeishi_cnXianxia/tree/main</a>）中提取出来的Lora，效果可能比大模型稍微差一点，但从体积上看性价比颇高。The Lora model, extracted from a model I trained previously (<a target="_blank" rel="ugc" href="https://huggingface.co/meishi/AiMeishi_cnXianxia/tree/main">https://huggingface.co/meishi/AiMeishi_cnXianxia/tree/main</a>), may have slightly lower performance than the large model, but it is quite cost-effective in terms of size.

Files with SHA256: 85fd49c684b9f6ad7769d1b68b6c34929f53a6d9982bd45c4981a37608d04ed2