Linly 在运行llama-7b微服务时，出现错误,

eivnm1vs 于 4个月前发布在其他

关注(0)|答案(2)|浏览(157)

微信图片126.png

Linly

来源：https://github.com/CVI-SZU/Linly/issues/76

2条答案

按热度按时间

jucafojl1#

欢迎来到bitsandbytes。对于错误报告，请运行
python -m bitsandbytes

并将此信息与您的错误跟踪一起提交给：https://github.com/TimDettmers/bitsandbytes/issues

bin /root/miniconda3/envs/linly_llama/lib/python3.8/site-packages/bitsandbytes/libbitsandbytes_cpu.so
/root/miniconda3/envs/linly_llama/lib/python3.8/site-packages/bitsandbytes/cextension.py:33: UserWarning: 已安装的bitsandbytes版本未编译支持GPU。8位优化器、8位乘法和GPU量化不可用。
warn("已安装的bitsandbytes版本未编译支持GPU。"
CUDA SETUP: 加载二进制文件 /root/miniconda3/envs/linly_llama/lib/python3.8/site-packages/bitsandbytes/libbitsandbytes_cpu.so...
Traceback (most recent call last):
File "llama_server.py", line 86, in
init_model()
File "llama_server.py", line 45, in init_model
args.tokenizer = Tokenizer(model_path=args.spm_model_path)
File "/data/lihu/llama_inference/model/tokenize.py", line 16, in init
assert os.path.isfile(model_path), model_path
AssertionError: ./root/lihu/ChatFlow-7B/tokenizer.model