search

jianzhnie / LLamaTuner

Easy and Efficient Finetuning LLMs. (Supported LLama, LLama2, LLama3, Qwen, Baichuan, GLM , Falcon) 大模型高效量化训练+部署.
https://jianzhnie.github.io/llmtech/
Apache License 2.0
547 stars 60 forks source link

llama-2-13b的模型用单卡跑lora就会报错 #95

Open Enoch202 opened 8 months ago

Enoch202 commented 8 months ago

用单卡跑就会显示“NotImplementedError: Cannot copy out of meta tensor； no data!” 用双卡跑就正常，但是占用显存很大。两张48G的A6000，每张都会占用35G左右

jianzhnie commented 7 months ago

试一下用微软的 deepspeed 呢