LLAMA-Factory安装教程（办理报错cannot allocate memory in static TLS bl ...

步骤一： 下载底子镜像

1. # 配置docker DNS
2. vi /etc/docker/daemon.json

复制代码

  # daemon.json文件中
  { "insecure-registries": ["https://swr.cn-east-317.qdrgznjszx.com"], "registry-mirrors": ["https://docker.mirrors.ustc.edu.cn"] }

1. systemctl restart docker.service
2. docker pull swr.cn-east-317.qdrgznjszx.com/donggang/llama-factory-ascend910b:cann8-py310-torch2.2.0-ubuntu18.04
3. mkdir /root/llama_factory_model

复制代码

步骤二：新建底子容器 

1. docker create -it -u root --ipc=host --net=host --name=llama-factory   -e LANG="C.UTF-8"\
2.         --device=/dev/davinci0 \
3.         --device=/dev/davinci1 \
4.         --device=/dev/davinci2 \
5.         --device=/dev/davinci3 \
6.         --device=/dev/davinci4 \
7.         --device=/dev/davinci5 \
8.         --device=/dev/davinci6 \
9.         --device=/dev/davinci7 \
10.         --device=/dev/davinci_manager \
11.         --device=/dev/devmm_svm \
12.         --device=/dev/hisi_hdc \
13.         -v /usr/local/Ascend/driver:/usr/local/Ascend/driver \
14.         -v /usr/local/Ascend/add-ons/:/usr/local/Ascend/add-ons/ \
15.         -v /usr/local/sbin/npu-smi:/usr/local/sbin/npu-smi \
16.         -v /mnt/:/mnt/ \
17.         -v /root/llama_factory_model:/root/llama_factory_model \
18.         -v /var/log/npu:/usr/slog swr.cn-east-317.qdrgznjszx.com/donggang/llama-factory-ascend910b:cann8-py310-torch2.2.0-ubuntu18.04 \
19.         /bin/bash \

复制代码

步骤三：安装llamafactory

1. docker start llama-factory
2. docker exec -it llama-factory bash
4. # 安装llama-factory
5. wget https://codeload.github.com/hiyouga/LLaMA-Factory/zip/refs/heads/main -O LLaMA-Factory.zip
6. unzip LLaMA-Factory.zip
7. mv LLaMA-Factory-main LLaMA-Factory
9. cd LLaMA-Factory
10. pip install -e ".[torch-npu,metrics]"
11. apt install libsndfile1
13. # 激活昇腾环境变量（建议加入 ~/.bashrc中）
14. source /usr/local/Ascend/ascend-toolkit/set_env.sh
16. #使用以下指令对 LLaMA-Factory × 昇腾的安装进行校验
17. llamafactory-cli env

复制代码

1. # 运行llamafactory webui（访问本机7860端口）
2. nohup llamafactory-cli webui> llama_factory_output.log 2>&1 &
3. # 查看llamafactory运行日志
4. tail -f /home/HwHiAiUser/LLaMA-Factory/llama_factory_output.log

复制代码

办理报错

题目描述

RuntimeError: Failed to import transformers.generation.utils because of the following error (look up to see its traceback):
/usr/local/python3.10.13/lib/python3.10/site-packages/sklearn/utils/../../scikit_learn.libs/libgomp-d22c30c5.so.1.0.0: cannot allocate memory in static TLS block
办理思路

1. vim ~/.bashrc
3. #文档末尾添加
4. export LD_PRELOAD=/usr/local/python3.10.13/lib/python3.10/site-packages/sklearn/utils/../../scikit_learn.libs/libgomp-d22c30c5.so.1.0.0
6. source ~/.bashrc

复制代码

免责声明：如果侵犯了您的权益，请联系站长，我们会及时删除侵权内容，谢谢合作！更多信息从访问主页：qidao123.com:ToB企服之家，中国第一个企服评测及商务社交产业平台。