Qingyi Zhong
Qingyi Zhong
使用官方提供的finetune.sh (ZeRO2+QLora)微调Qwen1.5-0.5B-Chat模型后,在推理过程中遇到相同的问题,用ZeRO3 + Lora微调Qwen1.5-0.5B-Chat模型并没有遇到此问题,请问是什么原因?QAQ...
> me too, it solved? Not yet, I'm still waiting for a solution or workaround.😵💫😵💫😵💫
> Hi @Harsha-Nori , Thank you for the prompt response!!!😃😃😃 I've tested the code you provided, and it runs without any errors. **Result:** ``` { "id": "chatcmpl-7qE02z3mUo1ezoq3VUhPYtDRn3lga", "object": "chat.completion.chunk", "created":...