ai - llama factory的安装和使用

访问量: 8

refer to: https://github.com/hiyouga/LlamaFactory

git clone 

cd ..

先查看显卡的硬件：$ nvidia-smi 

Sun Mar 29 13:19:54 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 576.88                 Driver Version: 576.88         CUDA Version: 12.9     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                  Driver-Model | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 3080      WDDM  |   00000000:04:00.0  On |                  N/A |
| 50%   44C    P8             17W /  320W |    3156MiB /  20480MiB |     12%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

pip install -e . -i https://mirrors.aliyun.com/pypi/simple/

pip install -r .\requirements\metrics.txt -i https://mirrors.aliyun.com/pypi/simple/

对于windows，还需要：

pip uninstall torch torchvision torchaudio

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126

python -c "import torch; print(torch.cuda.is_available())"

需要返回True

windows用户：安装 bitsandbytes 

具体需要下载的版本号参考：https://www.doubao.com/thread/wa26102ff5387b63e

我的是3080, 20G, 所以使用了这个：（最新的）

这里比较折腾。 参考：https://www.doubao.com/thread/w1fee0eccea95cfa9

pip uninstall -y numpy torch bitsandbytes
pip install torch==2.4.0 torchvision==0.19.0 torchaudio==2.4.0 --index-url https://download.pytorch.org/whl/cu121
pip install C:\Users\你的用户名\Desktop\bitsandbytes-0.41.2.post2-py3-none-win_amd64.whl

验证：
python -c "import numpy; import torch; import bitsandbytes; print('环境正常！NumPy:',numpy.__version__,'CUDA:',torch.version.cuda)"


订阅/RSS Feed

Subscribe