-
Notifications
You must be signed in to change notification settings - Fork 15
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
PULSE-7b模型,GPTQ-for-PULSE量化 #10
Comments
@genggui001 |
这样的形式组成jsonl文件就好; |
感谢您的回复! |
您好,@hanrui1sensetime 想再咨询您一个问题,就是量化完成后生成的bin文件,如果想通过model = AutoModelForCausalLM.from_pretrained()这个官方接口加载使用的话,需要怎么生成一下量化后对应的这些文件呢? 有没有相关的教程可以学习一下? |
Hi, @zhaofeng3012 |
按照文档,对PULSE-7b模型,使用GPTQ-for-PULSE进行模型量化:
CUDA_VISIBLE_DEVICES=0 python bloom.py ${MODEL_DIR} custom --wbits 4 --act-order --groupsize 128 --save pulse7b-4bit-128g.bin --calib_data ${CALIB_DATA_PATH}
命令参数${MODEL_DIR}指的应该是PULSE-7bv5的模型文件夹路径吧?
--calib_data ${CALIB_DATA_PATH}是什么意思?要设置什么路径呢?
@hanrui1sensetime 求解答
另外,生成的pulse7b-4bit-128g.bin要怎么使用?在设置的地方替换原来的PULSE-7bv5的模型文件夹路径就可以了么?
The text was updated successfully, but these errors were encountered: