https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

mkdir -p ~/miniconda3
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda3/miniconda.sh
bash miniconda.sh -b -u -p ~/miniconda3
rm -rf ~/miniconda3/miniconda.sh
#初始化
~/miniconda3/bin/conda init bash
~/miniconda3/bin/conda init zsh
#环境进入和退出
source ~/miniconda3/bin/activate
conda env list
conda deactivate
source deactivate your_env_name
#禁用自动激活 base 环境
conda config --set auto_activate_base false
#激活环境
source activate your_env_name
conda remove --name glm130b --all
#软件包
conda list
conda search package_name# 查询包
conda install package_name
conda install package_name=1.5.0
conda update package_name
conda remove package_name
#国内加速
sudo conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
sudo conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
sudo conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
sudo conda config --set show_channel_urls yes


#pip
sudo dnf install -y python-pip
sudo apt-get install -y python-pip
#更改默认源为清华大学源	
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple pip -U
pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
#官方默认源软件
#pip config set global.index-url  https://pypi.org/simple 
#git
install -y git
#AI相关包安装
pip install vllm torch torchvision
#pytorch
pip install torch --index-url https://download.pytorch.org/whl/cu118
#vLLM 安装
pip install vllm
pip install git+https://github.com/vllm-project/vllm.git
#Hugging Face CLI 安装
pip install --upgrade huggingface_hub
huggingface-cli login
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
#国内源
export HF_ENDPOINT=https://hf-mirror.com
export CUDA_HOME=/usr/local/cuda
export PATH=$PATH:$CUDA_HOME/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CUDA_HOME/lib64
#检查 CUDA：
python -c "import torch; print(torch.cuda.is_available())"
#检查 vLLM：
python -c "from vllm import LLM"
#检查 HF CLI：
huggingface-cli --version

huggingface-cli download deepseek-ai/DeepSeek-R1-Distill-Qwen-7B \
    --local-dir /path/to/model     --local-dir-use-symlinks False
#国内站
pip install modelscope -i https://mirrors.ustc.edu.cn/pypi/web/simple
export PATH="$HOME/.local/bin:$PATH"
modelscope download --model deepseek-ai/$Model_Name --local_dir $Path/$Model_Name
modelscope download --model Tencent-Hunyuan/HY-WorldPlay --local_dir /mnt/HY-WorldPlay

python -m vllm.entrypoints.openai.api_server --model /path/to/model --tensor-parallel-size 1 --max-num-batched-tokens 8192 --max-num-seqs 4 --port 8000

curl http://localhost:8000/v1/chat/completions     -H "Content-Type: application/json"     -d '{
        "model": "deepseek-7b",
        "messages": [{"role": "user", "content": "你好"}]
    }'

#启动服务
python -m vllm.entrypoints.openai.api_server \
  --model /mnt/Kimi-Linear-48B-A3B-Instruct \
  --served-model-name Kimi-Linear-48B-A3B-Instruct \
  --tensor-parallel-size 8 \
  --dtype bfloat16 \
  --max-model-len 16384 \
  --trust-remote-code \
  --gpu-memory-utilization 0.88 \
  --max-num-seqs 1 \
  --host 0.0.0.0 \
  --port 8000
#验证vllm
python - << 'EOF'
from transformers import AutoTokenizer
tok = AutoTokenizer.from_pretrained(
    "/mnt/Kimi-Linear-48B-A3B-Instruct",
    trust_remote_code=True
)
print(tok("你好，世界"))
EOF


#webui  OPENAI_API
docker run -d \
  --name open-webui \
  --restart always \
  -p 3000:8080 \
  -e OPENAI_API_BASE_URL=http://192.168.1.160:8000/v1 \
  -e OPENAI_API_KEY=sk-local \
  -e DEFAULT_MODEL=/mnt/Kimi-Linear-48B-A3B-Instruct \
  -v open-webui:/app/backend/data \
  ghcr.io/open-webui/open-webui:main



#miho
SGLANG_USE_MODELSCOPE=true 
python3 -m sglang.launch_server \
    --model-path /mnt/MiMo-V2-Flash \
    --served-model-name mimo-v2-flash \
    --pp-size 1 \
    --dp-size 2 \
    --enable-dp-attention \
    --tp-size 8 \
    --moe-a2a-backend deepep \
    --page-size 1 \
    --host 0.0.0.0 \
    --port 8000 \
    --trust-remote-code \
    --mem-fraction-static 0.75 \
    --max-running-requests 128 \
    --chunked-prefill-size 16384 \
    --reasoning-parser qwen3 \
    --tool-call-parser mimo \
    --context-length 262144 \
    --attention-backend fa3 \
    --speculative-algorithm EAGLE \
    --speculative-num-steps 3 \
    --speculative-eagle-topk 1 \
    --speculative-num-draft-tokens 4 \
    --enable-mtp