Closed lonelydancer closed 1 week ago
MiniCPM-2B-sft-int4 lora微调后,无法merge权重。
raise ValueError("Cannot merge LORA layers when the model is gptq quantized")
https://github.com/yangjianxin1/Firefly/blob/master/script/merge_lora.py
No response
- OS: [e.g. Ubuntu 20.04] - Pytorch: [e.g. torch 2.0.0] - CUDA: [e.g. CUDA 11.8] - Device: [e.g. A10, RTX3090]
请问需要自己用bf16版本做微调merge后做量化? 如果自己用gptq做量化,数据集应该用什么呢?
可以使用bf16进行训练后再做量化,具体的量化操作可以参考这个github地址:https://github.com/LDLINGLINGLING/MiniCPM/tree/ollama_manual_installation 的readme文件中gptq量化部分,量化的数据集可以是通用数据集加上你自己的训练数据集。
Is there an existing issue ? / 是否已有相关的 issue ?
Describe the bug / 描述这个 bug
MiniCPM-2B-sft-int4 lora微调后,无法merge权重。
raise ValueError("Cannot merge LORA layers when the model is gptq quantized")
To Reproduce / 如何复现
https://github.com/yangjianxin1/Firefly/blob/master/script/merge_lora.py
Expected behavior / 期望的结果
No response
Screenshots / 截图
No response
Environment / 环境
Additional context / 其他信息
No response
请问需要自己用bf16版本做微调merge后做量化? 如果自己用gptq做量化,数据集应该用什么呢?