Leonurus-free

chatgpt-next-web

docker run -d -p 3000:3000 \
   -e CUSTOM_MODELS=-all,+Qwen1.5-110B-Chat-AWQ \
   -e BASE_URL=http://192.168.100.142:8200 \
   yidadaa/chatgpt-next-web
   
docker run -d -p 3001:3000 -e CUSTOM_MODELS=-all,+qwen2.5-72b-instruct-awq -e BASE_URL=http://192.168.100.160:8200 --name chatgpt-next-web --restart=always yidadaa/chatgpt-next-web

docker run -d -p 3002:3000 --name chatgpt-next-web-custom --restart=always yidadaa/chatgpt-next-web

stirling-pdf

docker run -d \
  -p 8080:8080 \
  -v /home/free/python_project/Stirling-PDF/trainingData:/usr/share/tessdata \
  -v /home/free/python_project/Stirling-PDF/extraConfigs:/configs \
  -v /home/free/python_project/Stirling-PDF/logs:/logs \
  -v /home/free/python_project/Stirling-PDF/customFiles:/customFiles \
  -e DOCKER_ENABLE_SECURITY=true \
  -e INSTALL_BOOK_AND_ADVANCED_HTML_OPS=true \
  -e LANGS=zh_CN \
  -e http_proxy=socks5://192.168.20.21:1080 \
  -e https_proxy=socks5://192.168.20.21:1080 \
  --name stirling-pdf-nologin \
  frooodle/s-pdf:latest-fat
  
docker run -d \
  -p 8080:8080 \
  -v /home/free/python_project/Stirling-PDF/trainingData:/usr/share/tessdata \
  -v /home/free/python_project/Stirling-PDF/extraConfigs:/configs \
  -v /home/free/python_project/Stirling-PDF/logs:/logs \
  -v /home/free/python_project/Stirling-PDF/customFiles:/customFiles \
  -e DOCKER_ENABLE_SECURITY=true \
  -e INSTALL_BOOK_AND_ADVANCED_HTML_OPS=true \
  -e SECURITY_ENABLE_LOGIN=true \
  -e SECURITY_INITIALLOGIN_USERNAME=admin \
  -e SECURITY_INITIALLOGIN_PASSWORD=123 \
  -e LANGS=zh_CN \
  --name stirling-pdf-login \
  frooodle/s-pdf:latest-fat
  

docker run -d -p 3004:8080 -v E:/temp/Stirling-PDF/trainingData:/usr/share/tessdata -v E:/temp/Stirling-PDF/extraConfigs:/configs -v E:/temp/Stirling-PDF/logs:/logs -v E:/temp/Stirling-PDF/customFiles:/customFiles -e DOCKER_ENABLE_SECURITY=true -e INSTALL_BOOK_AND_ADVANCED_HTML_OPS=true -e LANGS=zh_CN --name stirling-pdf-nologin --restart=always frooodle/s-pdf:0.27.0-fat


docker run -d \
  --name stirling-pdf \
  -p 8980:8080 \
  --restart=always \
  -v "/home/ubuntu/luo/.Stirling-PDF/trainingData:/usr/share/tessdata" \
  -v "/home/ubuntu/luo/.Stirling-PDF/extraConfigs:/configs" \
  -v "/home/ubuntu/luo/.Stirling-PDF/customFiles:/customFiles/" \
  -v "/home/ubuntu/luo/.Stirling-PDF/logs:/logs/" \
  -v "/home/ubuntu/luo/.Stirling-PDF/pipeline:/pipeline/" \
  -e DOCKER_ENABLE_SECURITY=false \
  -e LANGS=zh_CN \
  stirling-pdf:0.45.6-fat
  
docker run -d \
  --name stirling-pdf \
  -p 8980:8080 \
  --restart=always \
  -v "/data/.Stirling-PDF/trainingData:/usr/share/tessdata" \
  -v "/data/.Stirling-PDF/extraConfigs:/configs" \
  -v "/data/.Stirling-PDF/customFiles:/customFiles/" \
  -v "/data/.Stirling-PDF/logs:/logs/" \
  -v "/data/.Stirling-PDF/pipeline:/pipeline/" \
  -e DOCKER_ENABLE_SECURITY=false \
  -e LANGS=zh_CN \
  stirling-pdf:0.45.6-fat
  
docker run -d --name stirling-pdf -p 8980:8080 -v "E:/temp/Stirling-PDF/trainingData:/usr/share/tessdata" -v "E:/temp/Stirling-PDF/extraConfigs:/configs" -v "E:/temp/Stirling-PDF/customFiles:/customFiles/" -v "E:/temp/Stirling-PDF/logs:/logs/" -v "E:/temp/Stirling-PDF/pipeline:/pipeline/" -e DOCKER_ENABLE_SECURITY=false stirling-pdf:0.45.6-fat

xinference

docker run -d -v F:/.xinference:/root/.xinference \
  -v F:/.xinference/.cache/huggingface:/root/.cache/huggingface \
  -v F:/.xinference/.cache/modelscope:/root/.cache/modelscope \
  -v F:/.xinference/logs:/workspace/xinference/logs \
  -e XINFERENCE_MODEL_SRC=modelscope \

  -p 9997:9997 \
  --gpus all \
  --name xinference \
  --restart=always \
  xprobe/xinference:v0.15.2 \
  xinference-local -H 0.0.0.0 \
  --log-level debug
  
docker run -d -v F:/.xinference:/root/.xinference -v F:/.xinference/.cache/huggingface:/root/.cache/huggingface -v F:/.xinference/.cache/modelscope:/root/.cache/modelscope -v F:/.xinference/logs:/workspace/xinference/logs -p 9997:9997 --gpus all --name xinference --restart=always xprobe/xinference:v0.15.2 xinference-local -H 0.0.0.0 --log-level debug

lobe-chat

docker run -d -p 3210:3210 \
  -e OPENAI_API_KEY=sk-123 \
  -e OPENAI_PROXY_URL=http://192.168.100.143:8090/v1 \
  --name lobe-chat \
  lobehub/lobe-chat:v1.19.33
  
docker run -d -p 3210:3210 -e OPENAI_API_KEY=sk-123 -e OPENAI_PROXY_URL=http://192.168.100.143:8090/v1 --name lobe-chat lobehub/lobe-chat:v1.19.33

vllm

docker pull vllm/vllm-openai:latest  
docker run -d
	--runtime nvidia \
	--gpus all \
	--name vllm_Qwen2.5-7B-Instruct \
    -v F:/.vllm/model:/root/model \
    -v F:/.vllm/vllm:/root/.cache/vllm \
    -v F:/.vllm/huggingface:/root/.cache/huggingface \
    --env "HUGGING_FACE_HUB_TOKEN=<secret>" \
    -p 8888:8000 \
    --ipc=host \
    vllm/vllm-openai:v0.6.3.post1 \   # 以上是docker配置，以下是vllm配置
    --model /root/model/Qwen2.5-7B-Instruct \
    --gpu-memory-utilization 0.8 \
    --tensor-parallel-size 2 \
    --max-model-len 8129 \
    --served-model-name Qwen2.5-7B-Instruct 
    
    
docker run -d --runtime nvidia --gpus "device=1" --name vllm_Qwen2.5-7B-Instruct -v F:/.vllm/model:/root/model -v F:/.vllm/vllm:/root/.cache/vllm -v F:/.vllm/huggingface:/root/.cache/huggingface --env "HUGGING_FACE_HUB_TOKEN=<secret>" -p 8888:8000 --ipc=host vllm/vllm-openai:v0.6.6.post1 --model /root/model/Qwen2.5-7B-Instruct --gpu-memory-utilization 0.8 --tensor-parallel-size 1 --max-model-len 16384 --served-model-name Qwen2.5-7B-Instruct 

docker run -d --runtime nvidia --gpus all --name vllm_Qwen2.5-7B-Instruct-lora -v F:/.vllm/model:/root/model -v F:/.vllm/vllm:/root/.cache/vllm -v F:/.vllm/huggingface:/root/.cache/huggingface --env "HUGGING_FACE_HUB_TOKEN=<secret>" -p 8888:8000 --ipc=host vllm/vllm-openai:v0.6.3.post1 --model /root/model/outputs-Qwen2.5-7B-Instruct-lora --gpu-memory-utilization 0.8 --tensor-parallel-size 2 --max-model-len 8129 --served-model-name Qwen2.5-7B-Instruct-lora 
    
 
 # 采用vllm.entrypoints.openai.api_server启动
 
CUDA_VISIBLE_DEVICES=2 python -m vllm.entrypoints.openai.api_server \
    --model /home/ubuntu/luo/Corrective_fintune_v1/models/Qwen/outputs-Qwen2.5-7B-Instruct-lora \
    --tokenizer /home/ubuntu/luo/Corrective_fintune_v1/models/Qwen/outputs-Qwen2.5-7B-Instruct-lora \
    --max-model-len 2048 \
    --gpu-memory-utilization 1 \
    --enforce-eager \
    --dtype half \
    --port 8888
    


# DeepSeek-R1-Distill-Qwen-7B

docker run -d --runtime nvidia --gpus "device=1" --name vllm-DeepSeek-R1-Distill-Qwen-7B -v F:/.vllm/model:/root/model -v F:/.vllm/vllm:/root/.cache/vllm -v F:/.vllm/huggingface:/root/.cache/huggingface --env "HUGGING_FACE_HUB_TOKEN=<secret>" -p 8888:8000 --ipc=host vllm/vllm-openai:v0.6.6.post1 --model /root/model/DeepSeek-R1-Distill-Qwen-7B --gpu-memory-utilization 0.8 --tensor-parallel-size 1 --max-model-len 16384 --served-model-name DeepSeek-R1-Distill-Qwen-7B

# Qwen/Qwen2.5-VL-72B-Instruct-AWQ
docker run -d \
	--runtime nvidia \
	--gpus '"device=0,1,2"' \
	--name Qwen2.5-VL-72B-Instruct-AWQ \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	vllm/vllm-openai:v0.8.4 \
    --model /root/model/Qwen2.5-VL-72B-Instruct-AWQ \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 3 \
    --max-model-len 32768 \
    --served-model-name Qwen2.5-VL-72B-Instruct-AWQ
   
# ChineseErrorCorrector3-4B
docker run -d \
	--runtime nvidia \
	--gpus '"device=3,4"' \
	--name ChineseErrorCorrector3-4B \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	vllm/vllm-openai:v0.10.1.1 \
    --model /root/model/ChineseErrorCorrector3-4B \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name ChineseErrorCorrector3-4B
    
docker run -d --runtime nvidia  --gpus '"device=0,1"' --name Qwen2.5-VL-72B-Instruct-AWQ -v /home/ubuntu/luo/model_files:/root/model -v /home/ubuntu/luo/.vllm/vllm:/root/.cache/vllm -v /home/ubuntu/luo/.vllm/huggingface:/root/.cache/huggingface --env "HUGGING_FACE_HUB_TOKEN=<secret>" -p 8888:8000 --ipc=host vllm/vllm-openai:v0.8.4 --model /root/model/Qwen2.5-VL-72B-Instruct-AWQ --gpu-memory-utilization 0.9 --tensor-parallel-size 2 --max-model-len 32768 --served-model-name Qwen2.5-VL-72B-Instruct-AWQ


# GLM-4-32B-0414
# Deploy with docker on Linux:
docker run -d
	--runtime nvidia \
	--gpus '"device=0,1,2"' \
	--name vllm_GLM-4-32B-0414 \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	vllm/vllm-openai:v0.8.4 \
    --model /root/model/GLM-4-32B-0414 \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 3 \
    --max-model-len 32768 \
    --served-model-name GLM-4-32B-0414
    
docker run -d --runtime nvidia  --gpus '"device=0,1"' --name vllm_GLM-4-32B-0414 -v /home/ubuntu/luo/model_files:/root/model -v /home/ubuntu/luo/.vllm/vllm:/root/.cache/vllm -v /home/ubuntu/luo/.vllm/huggingface:/root/.cache/huggingface --env "HUGGING_FACE_HUB_TOKEN=<secret>" -p 8888:8000 --ipc=host vllm/vllm-openai:v0.8.4 --model /root/model/GLM-4-32B-0414 --gpu-memory-utilization 0.9 --tensor-parallel-size 2 --max-model-len 32768 --served-model-name GLM-4-32B-0414

# Qwen2.5-Coder-32B-Instruct-AWQ
docker run -d \
	--runtime nvidia \
	--gpus '"device=2,3"' \
	--name Qwen2.5-Coder-32B-Instruct-AWQ \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen2.5-Coder-32B-Instruct-AWQ/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen2.5-Coder-32B-Instruct-AWQ/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8882:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen2.5-Coder-32B-Instruct-AWQ \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name Qwen2.5-Coder-32B-Instruct-AWQ \
    --dtype auto 

# Qwen3-32B-AWQ
docker run -d \
	--runtime nvidia \
	--gpus '"device=0,1"' \
	--name Qwen3-32B-AWQ \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-32B-AWQ/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-32B-AWQ/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-32B-AWQ \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name Qwen3-32B-AWQ \
    --dtype auto 
    
# Qwen3-32B-AWQ tool
docker run -d \
	--runtime nvidia \
	--gpus '"device=0,1"' \
	--name Qwen3-32B-AWQ-tool \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-32B-AWQ/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-32B-AWQ/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-32B-AWQ \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name Qwen3-32B-AWQ \
    --reasoning-parser qwen3 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --dtype auto 
    
# Qwen3-30B-A3B-Instruct-2507-FP8
docker run -d \
	--runtime nvidia \
	--gpus '"device=0,1"' \
	--name Qwen3-30B-A3B-Instruct-2507-FP8 \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-30B-A3B-Instruct-2507-FP8/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-30B-A3B-Instruct-2507-FP8/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-30B-A3B-Instruct-2507-FP8 \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name Qwen3-30B-A3B-Instruct-2507-FP8 \
    --reasoning-parser qwen3 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --dtype auto 
 
    
# Qwen3-32B-AWQ-reasoning
docker run -d \
	--runtime nvidia \
	--gpus '"device=0,1"' \
	--name Qwen3-32B-AWQ-reasoning \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-32B-AWQ/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-32B-AWQ/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-32B-AWQ \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name Qwen3-32B-AWQ \
    --enable-reasoning \
    --reasoning-parser deepseek_r1 \
    --dtype auto 

# Qwen3-14B-AWQ
docker run -d \
	--runtime nvidia \
	--gpus '"device=4"' \
	--name Qwen3-14B-AWQ \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-14B-AWQ/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-14B-AWQ/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8887:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-14B-AWQ \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --tensor-parallel-size 1 \
    --max-model-len 32768 \
    --served-model-name Qwen3-14B-AWQ \
    --dtype auto 
    
# Qwen3-4B-Thinking-2507
docker run -d \
	--runtime nvidia \
	--gpus '"device=4"' \
	--name Qwen3-4B-Thinking-2507 \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-4B-Thinking-2507/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-4B-Thinking-2507/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8887:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-4B-Thinking-2507 \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --tensor-parallel-size 1 \
    --max-model-len 32768 \
    --served-model-name Qwen3-4B-Thinking-2507 \
    --dtype auto 
    
# Qwen3-32B
docker run -d \
	--runtime nvidia \
	--gpus '"device=0,1,2,3"' \
	--name Qwen3-32B \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen3-32B/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen3-32B/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8888:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen3-32B \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --tensor-parallel-size 4 \
    --max-model-len 32768 \
    --served-model-name Qwen3-32B \
    --enable-reasoning \
    --reasoning-parser deepseek_r1 \
    --dtype auto 

    
# Qwen2.5-VL-32B-Instruct-AWQ
docker run -d \
	--runtime nvidia \
	--gpus '"device=5,6"' \
	--name Qwen2.5-VL-32B-Instruct-AWQ \
    -v /data/model_files:/root/model \
    -v /data/.vllm/Qwen2.5-VL-32B-Instruct-AWQ/vllm:/root/.cache/vllm \
    -v /data/.vllm/Qwen2.5-VL-32B-Instruct-AWQ/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8886:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/Qwen2.5-VL-32B-Instruct-AWQ \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --tensor-parallel-size 2 \
    --max-model-len 32768 \
    --served-model-name Qwen2.5-VL-32B-Instruct-AWQ \
    --dtype auto 
    
# bge-reranker-v2-m3
docker run -d \
	--runtime nvidia \
	--gpus '"device=7"' \
	--name bge-reranker-v2-m3 \
    -v /data/model_files:/root/model \
    -v /data/.vllm/bge-reranker-v2-m3/vllm:/root/.cache/vllm \
    -v /data/.vllm/bge-reranker-v2-m3/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8885:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/bge-reranker-v2-m3 \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 1 \
    --max-model-len 8192 \
    --served-model-name bge-reranker-v2-m3 
    
# bge-large-zh-v1.5
docker run -d \
	--runtime nvidia \
	--gpus '"device=7"' \
	--name bge-large-zh-v1.5 \
    -v /data/model_files:/root/model \
    -v /data/.vllm/bge-large-zh-v1.5/vllm:/root/.cache/vllm \
    -v /data/.vllm/bge-large-zh-v1.5/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8884:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.10.0 \
    --model /root/model/bge-large-zh-v1.5 \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 1 \
    --served-model-name bge-large-zh-v1.5 \
    --task embed
 
# Qwen3-Embedding-0.6B
docker run -d \
	--runtime nvidia \
	--gpus '"device=4"' \
	--name Qwen3-Embedding-0.6B \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/Qwen3-Embedding-0.6B/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/Qwen3-Embedding-0.6B/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8884:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.13.0-x86_64 \
    --model /root/model/Qwen3-Embedding-0.6B \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 1 \
    --served-model-name Qwen3-Embedding-0.6B
    
    
# Qwen3-VL-30B-A3B-Instruct
docker run -d \
	--runtime nvidia \
	--gpus '"device=0"' \
	--name Qwen3-VL-30B-A3B-Instruct \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/Qwen3-VL-30B-A3B-Instruct/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/Qwen3-VL-30B-A3B-Instruct/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8886:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.13.0-x86_64 \
    --model /root/model/Qwen3-VL-30B-A3B-Instruct \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 1 \
    --max-model-len 32768 \
    --served-model-name Qwen3-VL-30B-A3B-Instruct \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --dtype auto 
    
# Qwen3-VL-30B-A3B-Instruct-FP8
docker run -d \
	--runtime nvidia \
	--gpus '"device=0"' \
	--name Qwen3-VL-30B-A3B-Instruct-FP8 \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/Qwen3-VL-30B-A3B-Instruct-FP8/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/Qwen3-VL-30B-A3B-Instruct-FP8/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8886:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:v0.13.0-x86_64 \
    --model /root/model/Qwen3-VL-30B-A3B-Instruct-FP8 \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 1 \
    --max-model-len 32768 \
    --served-model-name Qwen3-VL-30B-A3B-Instruct-FP8 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --dtype auto 

注意事项： tensor_parallel_size只支持 2、4、6、8。

vLLM 支持将模型生成的思考内容解析为结构化消息：

vllm serve Qwen/Qwen3-8B --enable-reasoning --reasoning-parser deepseek_r1

--enable-reasoning \
--reasoning-parser deepseek_r1 \

彻底禁用nouveau驱动！ sudo vim /etc/modprobe.d/blacklist.conf blacklist nouveau

重启后，看是否成功，命令窗输入下面指令，无回复内容，则成功！

lsmod

grep nouveau

Qwen3.5-35B-A3B-AWQ-4bit

docker run -d \
	--runtime nvidia \
	--gpus '"device=0"' \
	--name Qwen3.5-35B-A3B-AWQ-4bit \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/Qwen3.5-35B-A3B-AWQ-4bit/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/Qwen3.5-35B-A3B-AWQ-4bit/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8886:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:latest \
    --model /root/model/Qwen3___5-35B-A3B-AWQ-4bit \
    --gpu-memory-utilization 0.9 \
    --tensor-parallel-size 1 \
    --max-model-len 262144 \
    --served-model-name Qwen3.5-35B-A3B-AWQ-4bit \
    --enable-auto-tool-choice \
    --tool-call-parser qwen3_coder \
    --reasoning-parser qwen3
    

vllm启动embeding

export CUDA_VISIBLE_DEVICES=3  
vllm serve /qwen/Qwen/Qwen3-Embedding-4B \
  --served-model-name Qwen3-Embedding-4B  \
  --port 8010 \
  --enforce-eager \
  --gpu-memory-utilization 0.45 \
  --task embed \
  --max_model_len 10000

http://192.168.100.100:8010/v1/embeddings

docker run -d \
	--runtime nvidia \
	--gpus '"device=4"' \
	--name Qwen3-Embedding-4B \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/Qwen3-Embedding-4B/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/Qwen3-Embedding-4B/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8884:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:latest \
    --model /root/model/Qwen3-Embedding-4B \
    --gpu-memory-utilization 0.4 \
    --tensor-parallel-size 1 \
    --max_model_len 8192 \
    --served-model-name Qwen3-Embedding-4B \
    --hf-overrides '{"is_matryoshka": true, "matryoshka_dimensions": [256,512,768,1024,2048,4096]}'

vllm启动reranker

export CUDA_VISIBLE_DEVICES=3
vllm serve /qwen/Qwen/Qwen3-Reranker-4B \
  --served-model-name Qwen3-Reranker-4B \
  --enforce-eager \
  --gpu-memory-utilization 0.45 \
  --port 8011 \
  --task score \
  --max_model_len 10000
  
  
docker run -d \
	--runtime nvidia \
	--gpus '"device=4"' \
	--name Qwen3-Reranker-4B \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/Qwen3-Reranker-4B/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/Qwen3-Reranker-4B/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8885:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm/vllm-openai:latest \
    --model /root/model/Qwen3-Reranker-4B \
    --gpu-memory-utilization 0.4 \
    --tensor-parallel-size 1 \
    --max_model_len 8192 \
    --served-model-name Qwen3-Reranker-4B \
    --task score
    
    
docker run -d \
  --runtime nvidia \
  --gpus '"device=4"' \
  --name Qwen3-Reranker-4B \
  -v /home/ubuntu/luo/model_files:/root/model \
  -v /home/ubuntu/luo/.vllm/Qwen3-Reranker-4B/vllm:/root/.cache/vllm \
  -v /home/ubuntu/luo/.vllm/Qwen3-Reranker-4B/huggingface:/root/.cache/huggingface \
  --env "HUGGING_FACE_HUB_TOKEN=<secret>" \
  -p 8885:8000 \
  --ipc=host \
  --restart=unless-stopped \
  vllm/vllm-openai:latest \
  /root/model/Qwen3-Reranker-4B \
  --trust-remote-code \
  --gpu-memory-utilization 0.4 \
  --tensor-parallel-size 1 \
  --max-model-len 8192 \
  --served-model-name Qwen3-Reranker-4B \
  --hf_overrides '{
    "architectures": ["Qwen3ForSequenceClassification"],
    "classifier_from_token": ["no", "yes"],
    "is_original_qwen3_reranker": true
  }'

whisper

docker run -d \
	--runtime nvidia \
	--gpus '"device=4"' \
	--name whisper-large-v3-turbo \
    -v /home/ubuntu/luo/model_files:/root/model \
    -v /home/ubuntu/luo/.vllm/whisper-large-v3-turbo/vllm:/root/.cache/vllm \
    -v /home/ubuntu/luo/.vllm/whisper-large-v3-turbo/huggingface:/root/.cache/huggingface \
 	--env "HUGGING_FACE_HUB_TOKEN=<secret>" \
	-p 8883:8000 \
	--ipc=host \
	--restart=unless-stopped \
	vllm-whisper:v0.13.0 \
    --model /root/model/whisper-large-v3-turbo \
    --enforce-eager \
    --gpu-memory-utilization 0.15 \
    --tensor-parallel-size 1 \
    --served-model-name whisper-large-v3-turbo

压力测试

运行全套性能测试：
python run_benchmarks.py --llm_url "http://192.168.100.166:8888/v1" --api_key "sk-123" --model "Qwen3-30B-A3B-Instruct-2507-FP8" --use_long_context

运行单次并发测试：
python llm_benchmark.py --llm_url "http://192.168.100.166:8888/v1" --api_key "sk-123" --model "Qwen3-32B-AWQ" --num_requests 100 --concurrency 50 --output_tokens 25000

python run_benchmarks.py --llm_url "http://192.168.100.166:8887/v1" --api_key "sk-123" --model "Qwen3-14B-AWQ" --use_long_context


python run_benchmarks.py --llm_url "http://192.168.100.166:8886/v1" --api_key "sk-123" --model "Qwen2.5-VL-32B-Instruct-AWQ" --use_long_context

Whisper-vllm

# Dockerfile
FROM vllm/vllm-openai:v0.13.0

# 安装音频相关依赖
RUN apt-get update && apt-get install -y \
    libglib2.0-0 \
    libsndfile1 \
    && rm -rf /var/lib/apt/lists/*

# 安装vLLM音频支持包
RUN uv pip install --system vllm[audio]==0.13.0

构建命令

# 构建镜像（指定镜像名称和版本）
docker build -t vllm-whisper:v0.13.0 .

# 或者使用完整标签
docker build -t vllm-whisper:v0.13.0 -f Dockerfile .

postman

curl http://192.168.100.50:8000/v1/chat/completions -H "Content-Type: application/json" -d '{
    "model": "/home/data/models/Qwen1___5-7B-Chat",
    "messages": [
    {"role": "system", "content": "你是一个有用的助手."},
    {"role": "user", "content": "告诉我关于尼米兹号航母的的一些知识."}
    ]
}'

omniparse

# 构建镜像
docker build -t omniparse:v2 .

# 启动镜像
docker run -d --runtime nvidia --gpus "device=1" -p 18001:8001 -v F:/.omniparse:/tmp -v F:/pythonProject/omniparse/omniparseV2/.env:/app/.env -v F:/pythonProject/omniparse/omniparseV2/models:/app/models --name omniparse omniparse:v2

# 100.100
docker run -d --runtime nvidia --gpus "device=4" -p 18801:8001 -v /home/ubuntu/luo/.omniparse/tmp:/tmp -v /home/ubuntu/luo/.omniparse/.env:/app/.env -v /home/ubuntu/luo/.omniparse/models:/app/models --name omniparse omniparse:20250422

mysql

docker run \
	-d \
   	--name mysql_9.1.0 \
   	-e MYSQL_ROOT_PASSWORD=1qaz2wsx \
   	-p 33306:3306 \
   	-v F:/.mysql/conf:/etc/mysql/conf.d \
   	-v F:/.mysql/data:/var/lib/mysql \
   	-v F:/.mysql/logs:/var/log/mysql \
   	--restart=always \
   	mysql:9.1.0
   	
# 注意mysql9.1.0中删除了以下服务器选项和变量：
–mysql本机密码服务器选项
–mysql本机密码代理用户服务器选项
–default_authentication_plugin服务器系统变量
9版本推荐使用caching_sha2_password远程密码连接
   	
docker run -d --name mysql_8.0 -e MYSQL_ROOT_PASSWORD=1qaz2wsx -p 33306:3306 -v F:/.mysql/conf:/etc/mysql/conf.d -v F:/.mysql/data:/var/lib/mysql -v F:/.mysql/logs:/var/log/mysql --restart=always mysql:8.0

docker exec -it mysql_8.0 bin/bash

docker run -d --name mysql_8.0 -e MYSQL_ROOT_PASSWORD=1qaz2wsx -p 33306:3306 -v /home/ubuntu/luo/.mysql/conf:/etc/mysql/conf.d -v /home/ubuntu/luo/.mysql/data:/var/lib/mysql -v /home/ubuntu/luo/.mysql/logs:/var/log/mysql --restart=always mysql:8.0

docker run -d --name mysql_8.0 -e MYSQL_ROOT_PASSWORD=1qaz2wsx -p 33306:3306 -v /data/.mysql/conf:/etc/mysql/conf.d -v /data/.mysql/data:/var/lib/mysql -v /data/.mysql/logs:/var/log/mysql --restart=always mysql:8.0

dev_env

docker run -d \
  --privileged
  --runtime nvidia --gpus all \
  -p 98080:8080 \ 
  -p 90022:22 \
  -p 97000:7000 \
  -p 98000:8000 \
  -v F:/pythonProject:/home/pythonProject \
  --restart=always \
  --name dev_env \
  dev_env_cuda12.5:latest

docker run -d --privileged --runtime nvidia --gpus "device=1" -p 18080:8080 -p 10022:22 -p 17000:7000 -p 18000:8000 -v F:/pythonProject:/home/pythonProject --restart=always --name dev_env dev_env_conda_ssh:12.6.3-cudnn-devel-ubuntu22.04

# A6000

docker run -d --gpus '"device=0,1,2"' -p 18080:8080 -p 10022:22 -p 17000:7000 -p 18000:8000 -v /home/ubuntu/luo:/home/luo --restart=always --name dev_env_A6000 dev_env_conda_ssh:12.6.3-cudnn-devel-ubuntu22.04

# 3090

docker run -d --gpus '"device=3,4"' -p 18081:8081 -p 10023:22 -p 17001:7001 -p 18001:8001 -v /home/ubuntu/luo:/home/luo --restart=always --name dev_env_3090 dev_env_conda_ssh:12.6.3-cudnn-devel-ubuntu22.04

# 本机 使用3090GPU 满足后端，前端
docker run -d --privileged --runtime nvidia --gpus "device=1" -p 15080:8080 -p 15022:22 -p 15700:7000 -p 15701:7001 -p 15702:7002 -p 15703:7003 -p 15704:7004 -p 15705:7005 -p 15666:5666 -p 15800:8000 -p 15801:8001 -p 15802:8002 -p 15803:8003 -p 15804:8004 -p 15805:8005 -v F:/pythonProject:/home/pythonProject -v F:/frontend_project:/home/frontend_project -v F:/paperCode:/home/paperCode -v F:/python_full_project:/home/python_full_project -v F:/projects:/home/projects --restart=unless-stopped --name dev_env_v2 dev_env_conda_ssh:12.6.3-cudnn-devel-ubuntu22.04

自己在容器中安装ssh

DataTransfer

docker build -t datatransfer:v1 .

docker run -d --restart=always --name DataTransfer datatransfer:v1

NotesSynchronization

docker build -t notes_synchronization:v1 .

docker run -d -v D:/MyNotes:/app/MyNotes --restart=always --name notes_synchronization notes_synchronization:v1

onlyoffice

sudo docker run -i -t -d -p 80:80 --restart=always \
    -v F:/.onlyoffice/DocumentServer/logs:/var/log/onlyoffice  \
    -v F:/.onlyoffice/DocumentServer/data:/var/www/onlyoffice/Data  \
    -v F:/.onlyoffice/DocumentServer/lib:/var/lib/onlyoffice \
    -v F:/.onlyoffice/DocumentServer/db:/var/lib/postgresql -e JWT_SECRET=false onlyoffice/documentserver:8.3.0
    

docker run -i -t -d -p 80:80 --restart=always -v F:/.onlyoffice/DocumentServer/logs:/var/log/onlyoffice  -v F:/.onlyoffice/DocumentServer/data:/var/www/onlyoffice/Data  -v F:/.onlyoffice/DocumentServer/lib:/var/lib/onlyoffice -v F:/.onlyoffice/DocumentServer/db:/var/lib/postgresql -e JWT_SECRET=false onlyoffice/documentserver:8.3.0

open-webui

docker run -d -p 3003:8080 -e OPENAI_API_KEY=sk-123 -e OPENAI_API_BASE_URL=http://192.168.100.160:8200/v1 -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

docker run -d -p 30031:8080 -e OPENAI_API_KEY=sk-123 -e OPENAI_API_BASE_URL=http://192.168.100.100:8888/v1 -v open-webui:/app/backend/data --name open-webui-vl --restart always ghcr.io/open-webui/open-webui:main

docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui --restart always --env=OPENAI_API_BASE_URL=http://192.168.:8000/v1 --env=OPENAI_API_KEY=token-abc123 --env=ENABLE_OLLAMA_API=false --env=ENABLE_RAG_WEB_SEARCH=true --env=RAG_WEB_SEARCH_ENGINE=duckduckgo  ghcr.io/open-webui/open-webui:main

docker run -d -p 3003:8080 -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:v0.6.26

docker run -d --name open-webui -p 3003:8080 -v open-webui-data:/app/backend/data -e HF_HUB_OFFLINE=1 --restart unless-stopped open-webui:0_6_43_custom

docker run -d \
    --name open-webui \
    -p 3003:8080 \
    -e ENABLE_OLLAMA_API=false \
    -e OPENAI_API_BASE_URL=http://192.168.100.143:8200 \
    -e OPENAI_API_KEY=sk-123 \
    -e CORS_ALLOW_ORIGIN='*' \
    -e FORWARDED_ALLOW_IPS='*' \
    -e SCARF_NO_ANALYTICS=true \
    -e DO_NOT_TRACK=true \
    -e ANONYMIZED_TELEMETRY=false \
    -e RAG_EMBEDDING_ENGINE=openai \
    -e RAG_EMBEDDING_MODEL=Qwen3-Embedding-0.6B \
    -e RAG_OPENAI_API_BASE_URL=http://192.168.100.100:8884 \
    -e RAG_OPENAI_API_KEY=sk-123 \
    -e ENABLE_COMMUNITY_SHARING=false \
    -v /home/ubuntu/luo/.open-webui:/app/backend/data \
    open-webui:0_6_43_custom
    
docker run -d \
    --name open-webui \
    -p 3003:8080 \
    -e ENABLE_OLLAMA_API=false \
    -e CORS_ALLOW_ORIGIN='http://192.168.100.100:3003' \
    -e FORWARDED_ALLOW_IPS='*' \
    -e SCARF_NO_ANALYTICS=true \
    -e DO_NOT_TRACK=true \
    -e ANONYMIZED_TELEMETRY=false \
    -v /home/ubuntu/luo/.open-webui:/app/backend/data \
    open-webui:0_7_2_custom
    
docker run -d \
    --name open-webui-1 \
    -p 3004:8080 \
    -e ENABLE_OLLAMA_API=false \
    -e CORS_ALLOW_ORIGIN='*' \
    -e FORWARDED_ALLOW_IPS='*' \
    -e SCARF_NO_ANALYTICS=true \
    -e DO_NOT_TRACK=true \
    -e ANONYMIZED_TELEMETRY=false \
    -v /home/ubuntu/luo/.open-webui:/app/backend/data \
    open-webui:0_6_43_custom
    
docker build -t open-webui:0_7_2_custom .
docker save -o open-webui.tar open-webui:0_7_2_custom

open-webui+casdoor

name: openwebui-production

services:
  # ============================================
  # Open WebUI 主服务
  # ============================================
  openwebui:
    image: open-webui:0_7_2_custom
    container_name: openwebui
    restart: unless-stopped
    stop_grace_period: 30s
    depends_on:
      open_webui_postgres:
        condition: service_healthy
      open_webui_redis:
        condition: service_started
      open_webui_qdrant:
        condition: service_started
    ports:
      - "${WEBUI_PORT:-23303}:8080"
    environment:
      # 数据库配置 (PostgreSQL)
      - DATABASE_URL=postgresql://${POSTGRES_USER:-openwebui}:${POSTGRES_PASSWORD:-1qaz2wsx}@open_webui_postgres:5432/${POSTGRES_DB:-openwebui}
      
      # Redis 缓存配置
      - REDIS_URL=redis://:${REDIS_PASSWORD:-1qaz2wsx}@open_webui_redis:6379/0
      
      # 向量数据库配置 (Qdrant)
      - VECTOR_DB=qdrant
      - QDRANT_URI=http://open_webui_qdrant:6333
      
      # WebSocket 配置
      - ENABLE_WEBSOCKET_SUPPORT=true
      - WEBSOCKET_MANAGER=redis
      - WEBSOCKET_REDIS_URL=redis://:${REDIS_PASSWORD:-1qaz2wsx}@open_webui_redis:6379/1
      
      # WebUI 配置
      - ENV=prod
      - ENABLE_OLLAMA_API=false
      - CORS_ALLOW_ORIGIN=http://192.168.100.100:23303
      - FORWARDED_ALLOW_IPS='*'
      - SCARF_NO_ANALYTICS=true
      - DO_NOT_TRACK=true
      - ANONYMIZED_TELEMETRY=false
      - WEBUI_URL=http://192.168.100.100:23303
      - WEBUI_SECRET_KEY=caj6D3Jzif1DfBXCP8aMQE4h6R/Gbt0lg5vN/A1eUAk=
      
      # 数据库连接池优化
      - DATABASE_POOL_SIZE=50
      - DATABASE_POOL_MAX_OVERFLOW=20
      - DATABASE_POOL_TIMEOUT=60
      - DATABASE_POOL_RECYCLE=3600
      
      # 性能优化配置
      - THREAD_POOL_SIZE=1000
      - CHAT_RESPONSE_STREAM_DELTA_CHUNK_SIZE=7
      - ENABLE_REALTIME_CHAT_SAVE=false
      - ENABLE_COMPRESSION_MIDDLEWARE=true
      
      # SSO
      - OAUTH_CLIENT_ID=34c0ccfa111882b676cf
      - OAUTH_CLIENT_SECRET=de3b07d7365fb356630835e189aaeeebd76ae2b7
      - OPENID_PROVIDER_URL=http://192.168.100.100:18104/.well-known/openid-configuration
      - OPENID_REDIRECT_URI=http://192.168.100.100:23303/oauth/oidc/callback
      - WEBUI_AUTH_SIGNOUT_REDIRECT_URL=http://192.168.100.100:23303/auth
      - OAUTH_SCOPES=openid email profile
      - OAUTH_PROVIDER_NAME=沧澜SSO
      - ENABLE_OAUTH_SIGNUP=true
      - OAUTH_MERGE_ACCOUNTS_BY_EMAIL=true
      - OAUTH_ALLOWED_DOMAINS=*
      - ENABLE_OAUTH_PERSISTENT_CONFIG=false
      
      # 登录界面
      - ENABLE_PASSWORD_AUTH=false
      - ENABLE_LOGIN_FORM=false
      - ENABLE_SIGNUP=false
      
    volumes:
      - /home/ubuntu/luo/projectDockerCompose/open_webui/openwebui:/app/backend/data
    networks:
      - open_webui_network
    healthcheck:
      test: ["CMD-SHELL", "unset http_proxy https_proxy; curl --silent --fail http://localhost:8080/health | jq -ne 'input.status == true' || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
    deploy:
      resources:
        limits:
          cpus: '4'
          memory: 8G
    security_opt:
      - no-new-privileges:true

  # ============================================
  # PostgreSQL 数据库服务
  # ============================================
  open_webui_postgres:
    image: postgres:18.1-bookworm
    container_name: openwebui-postgres
    restart: unless-stopped
    stop_grace_period: 30s
    ports:
      - "${POSTGRESQL_PORT:-25432}:5432"
    environment:
      - POSTGRES_USER=${POSTGRES_USER:-openwebui}
      - POSTGRES_PASSWORD=${POSTGRES_PASSWORD:-1qaz2wsx}
      - POSTGRES_DB=${POSTGRES_DB:-openwebui}
      - TZ=Asia/Shanghai
    volumes:
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_postgres/data:/var/lib/postgresql
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_postgres/config:/etc/postgresql
    networks:
      - open_webui_network
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-openwebui} -d ${POSTGRES_DB:-openwebui}"]
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 30s
    deploy:
      resources:
        limits:
          cpus: '4'
          memory: 16G
        reservations:
          cpus: '2'
          memory: 8G
    security_opt:
      - no-new-privileges:true
    shm_size: 256MB

  # ============================================
  # Redis 缓存服务
  # ============================================
  open_webui_redis:
    image: redis:7.4.2
    ports:
      - "${DOCKER_MAP_REDIS_PORT:-26379}:6379"
    container_name: open_webui_redis
    restart: unless-stopped
    stop_grace_period: 30s
    environment:
      - TZ=Asia/Shanghai
    volumes:
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_redis/conf:/etc/redis  # 单独存放配置文件
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_redis/data:/data       # 单独存放数据文件
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_redis/logs:/var/log/redis  # 日志目录
    command: redis-server /etc/redis/redis.conf --appendonly yes
    networks:
      - open_webui_network
    healthcheck:
      test: ["CMD", "redis-cli", "-a", "${REDIS_PASSWORD:-1qaz2wsx}", "ping"]
      interval: 30s
      timeout: 5s
      retries: 5
      start_period: 10s
    deploy:
      resources:
        limits:
          cpus: '2'
          memory: 2G
        reservations:
          cpus: '1'
          memory: 512M
    security_opt:
      - no-new-privileges:true

  # ============================================
  # Qdrant 向量数据库服务
  # ============================================
  open_webui_qdrant:
    image: qdrant/qdrant:v1.16.3-gpu-nvidia
    container_name: openwebui-qdrant
    runtime: nvidia  # 如果使用 GPU，确保启用
    restart: unless-stopped
    stop_grace_period: 30s
    environment:
      - TZ=Asia/Shanghai
    volumes:
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_qdrant/storage:/qdrant/storage
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_qdrant/config/custom_config.yaml:/qdrant/config/production.yaml
    networks:
      - open_webui_network
    ports:
      - "${QDRANT_PORT:-26333}:6333"
      - "${QDRANT_GRPC_PORT:-26334}:6334"
    healthcheck:
      test: ["CMD", "bash", "-c", "exec 3<>/dev/tcp/127.0.0.1/6333"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    deploy:
      resources:
        limits:
          cpus: '4'
          memory: 8G
        reservations:
          cpus: '2'
          memory: 4G
    security_opt:
      - no-new-privileges:true
      
  # ============================================
  # pipelines服务
  # ============================================
  open_webui_pipelines:
    image: ghcr.io/open-webui/pipelines:main
    container_name: openwebui-pipelines
    stop_grace_period: 30s
    volumes:
      - /home/ubuntu/luo/projectDockerCompose/open_webui/open_webui_pipelines:/app/pipelines
    restart: unless-stopped
    environment:
      - TZ=Asia/Shanghai
      - PIPELINES_API_KEY=1qaz2wsx
    networks:
      - open_webui_network
    ports:
      - "29099:9099"
    healthcheck:
      test: ["CMD", "bash", "-c", "exec 3<>/dev/tcp/127.0.0.1/9099"]
      interval: 30s
      timeout: 10s
      retries: 5
      start_period: 30s

# ============================================
# 网络配置
# ============================================
networks:
  open_webui_network:
    name: open_webui_network
    driver: bridge

open_webui数据迁移

安装 pgloader
pgloader sqlite://webui.db postgresql://openwebui:1qaz2wsx@192.168.100.100:25432/openwebui
或者直接使用navicat

FileCodeBox

docker run -d --restart=always -p 12345:12345 -v E:/.fcb-data/FileCodeBox/:/app/data --name filecodebox lanol/filecodebox:beta

docker run -d --restart=always -p 12345:12345 -v /home/ubuntu/luo/.fcb-data/FileCodeBox/:/app/data --name filecodebox lanol/filecodebox:beta

初始：
访问 /#/admin
输入管理员密码 FileCodeBox2023

redis

wget http://download.redis.io/redis-stable/redis.conf

bind 127.0.0.1 # 这行要注释掉，解除本地连接限制
protected-mode no # 默认yes，如果设置为yes，则只允许在本机的回环连接，其他机器无法连接。
daemonize no # 默认no 为不守护进程模式，docker部署不需要改为yes，docker run -d本身就是后台启动，不然会冲突
requirepass 123456 # 设置密码
appendonly yes # 持久化
databases 32 # 数据库个数（可选）

docker run --restart=always --privileged=true --name redis_7.4 -p 36379:6379 -v F:/.redis/redis.conf:/etc/redis/redis.conf -v F:/.redis:/data -d redis:7.4.2 redis-server /etc/redis/redis.conf --appendonly yes

docker run --restart=always --privileged=true --name redis_7.4 -p 36379:6379 -v /home/ubuntu/luo/.redis/redis.conf:/etc/redis/redis.conf -v /home/ubuntu/luo/.redis:/data -d redis:7.4.2 redis-server /etc/redis/redis.conf --appendonly yes

docker run --restart=always --privileged=true --name redis_7.4 -p 36379:6379 -v /data/.redis/redis.conf:/etc/redis/redis.conf -v /data/.redis:/data -d redis:7.4.2 redis-server /etc/redis/redis.conf --appendonly yes

fastapi-sso

# 部署

docker build -t fastapi-sso:v1 .

# 启动

docker run -d -p 19999:9999 --restart=always -v .env:/home/.env --name fastapi-sso fastapi-sso:v1

第二版

docker build -t fastapi-sso:20250705 .

docker-compose up -d

luo-vben5

# 部署
docker build -t luo_vben5:v1 .
# 启动 
docker run -d -p 15666:5666 --restart=always -v F:/frontend_project/luo-vben5/apps/web-antd/docker_env/.env.development:/app/apps/web-antd/.env.development --name luo_vben5 luo_vben5:v1

docker save -o luo-vben5.tar luo_vben5:v1

Prompt Optimizer

# 运行容器（默认配置）
docker run -d -p 80:80 --restart unless-stopped --name prompt-optimizer linshen/prompt-optimizer

# 运行容器（配置API密钥）
docker run -d -p 180:80 \
  -e VITE_CUSTOM_API_KEY=sk-123 \
  -e VITE_CUSTOM_API_BASE_URL=http://192.168.100.160:8200/v1 \
  -e VITE_CUSTOM_API_MODEL=qwen2.5-72b-instruct-awq \
  --restart unless-stopped \
  --name prompt-optimizer \
  linshen/prompt-optimizer
  
docker run -d -p 180:80 -e VITE_CUSTOM_API_KEY=sk-123 -e VITE_CUSTOM_API_BASE_URL=http://192.168.100.160:8200/v1 -e VITE_CUSTOM_API_MODEL=qwen2.5-72b-instruct-awq --restart unless-stopped --name prompt-optimizer linshen/prompt-optimizer

docker run -d -p 180:80 --restart unless-stopped --name prompt-optimizer linshen/prompt-optimizer

mtranserver

compose.yml

services:
  mtranserver:
    image: xxnuo/mtranserver:2.1.1
    container_name: mtranserver
    restart: unless-stopped
    ports:
      - "18989:8989"
    volumes:
      - ./models:/app/models
    environment:
      - CORE_API_TOKEN=1qaz2wsx

docker run -d -p 18989:8989 --restart unless-stopped --name mtranserver -v /home/ubuntu/luo/.mtranserver/models:/app/models -e CORE_API_TOKEN=1qaz2wsx xxnuo/mtranserver:4.0.32

docker run -d -p 18989:8989 --restart unless-stopped --name mtranserver -v /data/.mtranserver/models:/app/models -e CORE_API_TOKEN=1qaz2wsx xxnuo/mtranserver:2.1.1

请求方式：

Postman 请求配置

1. 请求方法：POST
2. 请求 URL：http://localhost:18989/translate
3. Headers：
   • Content-Type: application/json
   • Authorization: 1qaz2wsx （替换为你的实际 token）
4. Body (选择 raw 并设置为 JSON 格式)：

{
    "from": "en",
    "to": "zh",
    "text": "Hello, world!"
}

预期响应

成功请求后，你将会收到类似以下的响应：

{
    "result": "你好，世界！"
}

注意事项

• 确保你的本地服务 http://localhost:18989 正在运行。
• 如果接口需要认证，请将 your_token 替换为有效的授权令牌。
• 如果接口有跨域限制，你可能需要在服务端配置 CORS。

使用

下面表格内的 localhost 可以替换为你的服务器地址或 Docker 容器名。

下面表格内的 8989 端口可以替换为你在 compose.yml 文件中设置的端口值。

如果未设置 CORE_API_TOKEN 或者设置为空，翻译插件使用无密码的 API。

如果设置了 CORE_API_TOKEN，翻译插件使用有密码的 API。

下面表格中的 your_token 替换为你在 config.ini 文件中设置的 CORE_API_TOKEN 值。

翻译插件接口：

注：

• 沉浸式翻译 在设置页面，开发者模式中启用Beta特性，即可在翻译服务中看到自定义 API 设置(官方图文教程)。然后将自定义 API 设置的每秒最大请求数拉高以充分发挥服务器性能准备体验飞一般的感觉。我设置的是每秒最大请求数为5000，每次请求最大段落数为10。你可以根据自己服务器配置设置。
• 简约翻译 在设置页面，接口设置中滚动到下面，即可看到自定义接口 Custom。同理，设置最大请求并发数量、每次请求间隔时间以充分发挥服务器性能。我设置的是最大请求并发数量为100，每次请求间隔时间为1。你可以根据自己服务器配置设置。

接下来按下表的设置方法设置插件的自定义接口地址。注意第一次请求会慢一些，因为需要加载模型。以后的请求会很快。

名称	URL	插件设置
沉浸式翻译无密码	http://localhost:8989/imme	自定义API 设置 - API URL
沉浸式翻译有密码	http://localhost:8989/imme?token=your_token	同上，需要更改 URL 尾部的 your_token 为你的 CORE_API_TOKEN 值
简约翻译无密码	http://localhost:8989/kiss	接口设置 - Custom - URL
简约翻译有密码	http://localhost:8989/kiss	同上，需要 KEY 填 your_token
划词翻译自定义翻译源无密码	http://localhost:8989/hcfy	设置-其他-自定义翻译源-接口地址
划词翻译自定义翻译源有密码	http://localhost:8989/hcfy?token=your_token	设置-其他-自定义翻译源-接口地址

普通用户参照表格内容设置好插件使用的接口地址就可以使用了。

TranslationV4

# 部署

docker build -t translation:v4 .

# 启动

docker run -d -p 17001:7000 --restart=always --network my_network_1 -v F:/pythonProject/TranslationV4/docker_env/.env:/app/.env -v F:/pythonProject/TranslationV4/InputFile:/app/InputFile -v F:/pythonProject/TranslationV4/OutputFile:/app/OutputFile --name translation_v4 translation:v4

docker save -o translation_4.tar translation:v4

翻译项目docker compose

docker-compose up -d

services:
  frontend:
    image: luo_vben5:20250422
    container_name: translation_luo_vben5_docker_compose
    restart: always
    ports:
      - "15666:5666"
    volumes:
      - "F:/projectDockerCompose/translation/frontend/.env.development:/app/apps/web-antd/.env.development"
    networks:
      - my_network

  backend:
    image: translation:20250422
    container_name: translation_v4_docker_compose
    restart: always
    ports:
      - "17001:7000"
    volumes:
      - "F:/projectDockerCompose/translation/backend/.env:/app/.env"
      - "F:/projectDockerCompose/translation/backend/InputFile:/app/InputFile"
      - "F:/projectDockerCompose/translation/backend/OutputFile:/app/OutputFile"
      - "F:/projectDockerCompose/translation/backend/logs:/app/logs"
    networks:
      - my_network
    depends_on:
      - redis
      - mysql

  redis:
    image: redis:7.4.2
    container_name: translation_redis_7.4_docker_compose
    restart: always
    privileged: true
    ports:
      - "16379:6379"
    volumes:
      - "F:/projectDockerCompose/translation/redis/conf:/etc/redis"  # 单独存放配置文件
      - "F:/projectDockerCompose/translation/redis/data:/data"       # 单独存放数据文件
    command: redis-server /etc/redis/redis.conf --appendonly yes
    networks:
      - my_network

  mysql:
    image: mysql:8.0
    container_name: translation_mysql_8.0_docker_compose
    restart: always
    environment:
      MYSQL_ROOT_PASSWORD: "1qaz2wsx"
    ports:
      - "13306:3306"
    volumes:
      - "F:/projectDockerCompose/translation/mysql/conf:/etc/mysql/conf.d"  # 配置目录
      - "F:/projectDockerCompose/translation/mysql/data:/var/lib/mysql"     # 数据目录
      - "F:/projectDockerCompose/translation/mysql/logs:/var/log/mysql"     # 日志目录
    networks:
      - my_network

networks:
  my_network:
    name: translation_network

Umi-OCR_runtime_linux

docker run -d --name umi-ocr \
    -e HEADLESS=true \
    -p 1224:1224 \
    umi-ocr-paddle
    
docker run -d --name umi-ocr --restart=always -e HEADLESS=true -p 12298:1224 umi-ocr-paddle:latest

etherpad-lite

docker-compose up -d

Perplexica/Vane

docker-compose up -d

docker run -d -p 23000:3000 -e SEARXNG_API_URL=http://192.168.100.100:6789 -v /home/ubuntu/luo/.vane:/home/vane/data --name vane itzcrazykns1337/vane:slim-latest


docker run -d -p 23000:3000 -v /home/ubuntu/luo/.vane:/home/vane/data --name vane itzcrazykns1337/vane:latest

comfyUI

# Nvidia GPU Base Images
# For NVIDIA GPU with stable CUDA version
FROM nvidia/cuda:12.6.3-cudnn-runtime-ubuntu24.04 AS base

# For NVIDIA GPU with latest CUDA version
# FROM nvidia/cuda:12.8.1-cudnn-runtime-ubuntu24.04 AS base 

# AMD GPU Base Images
# For AMD GPU with stable ROCm version
# FROM rocm/dev-ubuntu-24.04:6.2.4-complete AS base 

# For AMD GPU with latest ROCm version
# FROM rocm/dev-ubuntu-24.04:6.3.4-complete AS base 

# Environment variables
ENV DEBIAN_FRONTEND=noninteractive

# Install necessary dependencies and Python 3.12
RUN apt-get update \
    && apt-get install -y \
    git \
    software-properties-common \
    curl \
    python3.12 \
    python3.12-dev \
    python3.12-venv \
    python3-setuptools \
    wget \
    ffmpeg \
    libsm6 \
    libxext6 \
    libgl1 \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

# Set the working directory
WORKDIR /app

# Clone the ComfyUI repository and set up virtual environment
RUN git clone https://github.com/comfyanonymous/ComfyUI.git /app/comfyui \
    && python3.12 -m venv /app/venv \
    && /app/venv/bin/pip install --upgrade pip \
    && /app/venv/bin/pip install pyyaml \
    && /app/venv/bin/pip install -r /app/comfyui/requirements.txt

# Clone ComfyUI-Manager and install its dependencies
RUN git clone https://github.com/ltdrdata/ComfyUI-Manager.git /app/temp/ComfyUI-Manager \
    && mv /app/temp/* /app/comfyui/custom_nodes/ \
    && rm -rf /app/temp \
    && /app/venv/bin/pip install -r /app/comfyui/custom_nodes/ComfyUI-Manager/requirements.txt

    # NVIDIA GPU PyTorch Installation

    # Install PyTorch with CUDA 12.6 support (stable version)
# RUN /app/venv/bin/pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu126

    # Install PyTorch with CUDA 12.8 support (latest version)
# RUN /app/venv/bin/pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

    # AMD GPU PyTorch Installation

    # Install PyTorch with ROCm 6.2 support (stable version)
# RUN /app/venv/bin/pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.2.4

    # Install PyTorch with ROCm 6.3 support (latest version)
# RUN /app/venv/bin/pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.3


# Expose the backend port
EXPOSE 8188

# Set the entrypoint to run the app
CMD ["/bin/bash", "-c", "source /app/venv/bin/activate && python3 /app/comfyui/main.py --listen 0.0.0.0 --port 8188"]

docker build -t comfyui:0.3.40 .

docker run -it -d --name comfyui_0340 -p 8188:8188 -v F:/.comfyui/input:/app/comfyui/input -v F:/.comfyui/models:/app/comfyui/models -v F:/.comfyui/output:/app/comfyui/output -v F:/.comfyui/user:/app/comfyui/user -e NVIDIA_VISIBLE_DEVICES=all -e NVIDIA_DRIVER_CAPABILITIES=all --restart=always --runtime nvidia comfyui:0.3.40

docker run -it -d --name comfyui_0340 -p 8188:8188 -v /home/ubuntu/luo/.comfyui/input:/app/comfyui/input -v /home/ubuntu/luo/.comfyui/models:/app/comfyui/models -v /home/ubuntu/luo/.comfyui/output:/app/comfyui/output -v /home/ubuntu/luo/.comfyui/user:/app/comfyui/user -e NVIDIA_VISIBLE_DEVICES=all -e NVIDIA_DRIVER_CAPABILITIES=all --restart=always --runtime nvidia comfyui:0.3.40

docker run -it -d --name comfyui_0340 -p 8188:8188 -v /data/.comfyui/input:/app/comfyui/input -v /data/.comfyui/models:/app/comfyui/models -v /data/.comfyui/output:/app/comfyui/output -v /data/.comfyui/user:/app/comfyui/user -e NVIDIA_VISIBLE_DEVICES="7" -e NVIDIA_DRIVER_CAPABILITIES=all --restart=always --runtime nvidia comfyui:0.3.40

新的命令

mkdir -p \
  storage \
  storage-models/models \
  storage-models/hf-hub \
  storage-models/torch-hub \
  storage-user/input \
  storage-user/output \
  storage-user/workflows

docker run -it -d \
  --name comfyui \
  --runtime nvidia \
  --gpus all \
  -p 8188:8188 \
  -v "$(pwd)"/storage:/root \
  -v "$(pwd)"/storage-models/models:/root/ComfyUI/models \
  -v "$(pwd)"/storage-models/hf-hub:/root/.cache/huggingface/hub \
  -v "$(pwd)"/storage-models/torch-hub:/root/.cache/torch/hub \
  -v "$(pwd)"/storage-user/input:/root/ComfyUI/input \
  -v "$(pwd)"/storage-user/output:/root/ComfyUI/output \
  -v "$(pwd)"/storage-user/workflows:/root/ComfyUI/user/default/workflows \
  -e CLI_ARGS="--disable-xformers" \
  yanwk/comfyui-boot:cu128-megapak

n8n

docker volume create n8n_data
docker run -it -d --name n8n --restart=always -p 5678:5678 -v n8n_data:/home/node/.n8n -e N8N_SECURE_COOKIE=false docker.n8n.io/n8nio/n8n

videolingo

docker build -t videolingo .

docker run -d -p 8501:8501 --gpus all videolingo:latest

searxng

docker run -d -p 6789:8080 --name searxng -v "F:/.searxng/searxng:/etc/searxng" -e "BASE_URL=http://localhost:6789/" -e "INSTANCE_NAME=my-instance" searxng/searxng:2025.5.21-156d1eb

去文件夹F:/.searxng/searxng找配置文件settings.yml，然后进行修改。

docker run -d -p 6789:8080 --restart=always --name searxng -v "/home/ubuntu/luo/.searxng:/etc/searxng" searxng/searxng:2025.6.3-eb36de8

docker run --name searxng -d \
    --restart=always \
    -p 6789:8080 \
    -v "/home/ubuntu/luo/.searxng/config/:/etc/searxng/" \
    -v "/home/ubuntu/luo/.searxng/data/:/var/cache/searxng/" \
	searxng/searxng:2026.3.13-3c1f68c59

portainer

docker run -d -p 9000:9000 -v /var/run/docker.sock:/var/run/docker.sock -v /home/luo/.dockerData/portainer:/data --restart=always --name portainer portainer/portainer-ce:2.30.1

drawio

docker run -d -v "/home/ubuntu/luo/.drawiodata/letsencrypt-log:/var/log/letsencrypt/" -v "/home/ubuntu/luo/.drawiodata/letsencrypt-etc:/etc/letsencrypt/" -v "/home/ubuntu/luo/.drawiodata/letsencrypt-lib:/var/lib/letsencrypt" -e LETS_ENCRYPT_ENABLED=true -e PUBLIC_DNS=drawio.example.com --restart=always --name=drawio -p 11180:80 -p 11443:8443 jgraph/drawio:27.0.9

docker run -d --restart=always --name=drawio -p 11188:8080 -p 18443:8443 jgraph/drawio:27.0.9

打开地址：
http://192.168.100.100:11188/?lang=zh

-e LANG=zh_CN.UTF-8 -e LANGUAGE=zh_CN:zh -e LC_ALL=zh_CN.UTF-8

IOPaint

docker run -d -p 8380:8080 --runtime nvidia -v F:/.IOPaint/models:/root/.cache --restart=unless-stopped --name IOPaint thr3a/iopaint:20250409

docker run -d -p 8380:8080 --runtime nvidia --name IOPaint thr3a/iopaint:20250409

docker run -d \
-p 8380:8080 \
--runtime nvidia \
-v /home/ubuntu/luo/.IOPaint/models:/root/.cache/ \
--restart=unless-stopped \
--name IOPaint \
thr3a/iopaint:20250409 \
HF_ENDPOINT=https://hf-mirror.com

wewe-rss

version: '3.9'

services:
  db:
    image: mysql:8.3.0
    command: --mysql-native-password=ON
    environment:
      # 请修改为自己的密码
      MYSQL_ROOT_PASSWORD: 1qaz2wsx
      TZ: 'Asia/Shanghai'
      MYSQL_DATABASE: 'wewe-rss'
    # ports:
    #   - 13306:3306
    volumes:
      - db_data:/var/lib/mysql
    healthcheck:
      test: ['CMD', 'mysqladmin', 'ping', '-h', 'localhost']
      timeout: 45s
      interval: 10s
      retries: 10

  app:
    image: cooderl/wewe-rss:v2.6.1
    ports:
      - 14000:4000
    depends_on:
      db:
        condition: service_healthy
    environment:
      # 数据库连接地址
      - DATABASE_URL=mysql://root:1qaz2wsx@db:3306/wewe-rss?schema=public&connect_timeout=30&pool_timeout=30&socket_timeout=30
      # 服务接口请求授权码
      - AUTH_CODE=1qaz2wsx
      # 提取全文内容模式
      - FEED_MODE=fulltext
      # 定时更新订阅源Cron表达式
      - CRON_EXPRESSION=35 5,17 * * *
      # 服务接口请求限制，每分钟请求次数
      - MAX_REQUEST_PER_MINUTE=60
      # 外网访问时，需设置为服务器的公网 IP 或者域名地址
      # - SERVER_ORIGIN_URL=http://localhost:4000

networks:
  wewe-rss:

volumes:
  db_data:

docker-compose up -d

ConvertX

# docker-compose.yml
services:
  convertx:
    image: c4illin/convertx:pr-337
    container_name: convertx
    restart: unless-stopped
    ports:
      - "8981:3000"
    environment:
      - JWT_SECRET=aLongAndSecretStringUsedToSignTheJSONWebToken1234 # will use randomUUID() if unset
      - ALLOW_UNAUTHENTICATED=true
      - HTTP_ALLOWED=true
      - LANGUAGE=zh
    volumes:
      - ./data:/app/data

docker-compose up -d

docker run -d \
-p 8981:3000 \
-v /home/ubuntu/luo/.convertx/data:/app/data \
-e LANGUAGE=zh \
-e JWT_SECRET=aLongAndSecretStringUsedToSignTheJSONWebToken1234 \
-e HTTP_ALLOWED=true \
-e ALLOW_UNAUTHENTICATED=true \
--restart=unless-stopped \
--name convertx \
convertx:v337

docker run -d \
-p 8981:3000 \
-v /data/.convertx/data:/app/data \
-e LANGUAGE=zh \
-e JWT_SECRET=aLongAndSecretStringUsedToSignTheJSONWebToken1234 \
-e HTTP_ALLOWED=true \
-e ALLOW_UNAUTHENTICATED=true \
--restart=unless-stopped \
--name convertx \
convertx:v337

rabbitmq

docker run -id --restart=always --name=rabbitmq_4.1.1 -v F:/.rabbitmq:/var/lib/rabbitmq -p 35673:15672 -p 35672:5672 -e RABBITMQ_DEFAULT_USER=admin -e RABBITMQ_DEFAULT_PASS=1qaz2wsx rabbitmq:4.1.1-management

 docker network connect my_network_1 rabbitmq_4.1.1

oneline

docker build -t oneline:2.3.0 . 

docker run -p 23001:3000 \
  --restart=unless-stopped \
  --name=oneline \
  -e NEXT_PUBLIC_API_ENDPOINT=http://192.168.100.166:8888/v1/chat/completions \
  -e NEXT_PUBLIC_API_MODEL=Qwen3-32B-AWQ \
  -e NEXT_PUBLIC_API_KEY=sk-123 \
  -e NEXT_PUBLIC_SEARXNG_URL=http://192.168.100.100:6789 \
  -e NEXT_PUBLIC_SEARXNG_ENABLED=true \
  -e NEXT_PUBLIC_ALLOW_USER_CONFIG=false \
  -e NEXT_PUBLIC_ACCESS_PASSWORD= \
  -d oneline:2.3.0
  
docker run -p 23001:3000 --restart=unless-stopped --name=oneline -e NEXT_PUBLIC_API_ENDPOINT=http://192.168.100.143:8200/v1/chat/completions -e NEXT_PUBLIC_API_MODEL=Qwen3-235B-A22B-Instruct-2507-FP8 -e NEXT_PUBLIC_API_KEY=sk-123 -e NEXT_PUBLIC_SEARXNG_URL=http://192.168.100.100:6789 -e NEXT_PUBLIC_SEARXNG_ENABLED=true -e NEXT_PUBLIC_ALLOW_USER_CONFIG=false -e NEXT_PUBLIC_ACCESS_PASSWORD= -d oneline:2.3.0

# windows
docker run -p 23001:3000 --restart=unless-stopped --name=oneline -v "${PWD}\.env.local:/app/.env.local" -d oneline:2.3.0

# Linux
docker run -p 23001:3000 --restart=unless-stopped --name=oneline -v "$(pwd)/.env.local:/app/.env.local" -d oneline:2.3.0

环境变量说明

环境变量	说明	默认值
NEXT_PUBLIC_API_ENDPOINT	外部API端点	-
NEXT_PUBLIC_API_MODEL	API模型名称	gemini-2.0-flash-exp-search
NEXT_PUBLIC_API_KEY	API密钥	-
NEXT_PUBLIC_SEARXNG_URL	SearXNG搜索服务URL	https://sousuo.emoe.top
NEXT_PUBLIC_SEARXNG_ENABLED	是否启用SearXNG	false
NEXT_PUBLIC_ALLOW_USER_CONFIG	是否允许用户在前端修改配置	true
NEXT_PUBLIC_ACCESS_PASSWORD	访问密码	-

注意事项：

• 环境变量配置的优先级高于前端用户配置
• 当NEXT_PUBLIC_ALLOW_USER_CONFIG设置为false时，用户将无法在前端修改API设置
• 当设置了NEXT_PUBLIC_ACCESS_PASSWORD时，用户需要输入正确的密码才能访问API设置
• 推荐使用SearXNG搜索增强功能，这样可以避免配置API密钥

drawdb

git clone https://github.com/drawdb-io/drawdb

docker build -t drawdb .

docker run -d -p 23002:80 --restart=unless-stopped --name drawdb drawdb
# 访问地址
http://192.168.100.123:23002/editor

drawnix

docker run pubuzhixing/drawnix:v0.2.0

docker run -d -p 23003:80 --restart=unless-stopped --name drawnix pubuzhixing/drawnix:v0.2.0

http://192.168.100.123:23003/

minio

docker run -p 9091:9000 -p 9090:9090 \
     --name minio \
     -d --restart=unless-stopped \
     -e "MINIO_ROOT_USER=root" \
     -e "MINIO_ROOT_PASSWORD=1qaz2wsx" \
     -v /data/.minio/data:/data \
     -v /data/.minio/config:/root/.minio \
     minio/minio:RELEASE.2025-07-23T15-54-02Z server \
     /data --console-address ":9090" --address ":9000"
     
docker run -p 9091:9000 -p 9090:9090 --name minio -d --restart=unless-stopped -e "MINIO_ROOT_USER=root" -e "MINIO_ROOT_PASSWORD=1qaz2wsx" -v F:/.minio/data:/data -v F:/.minio/config:/root/.minio minio/minio:RELEASE.2025-07-23T15-54-02Z server /data --console-address ":9090" --address ":9000"
     

openai-whisper-asr-webservice

docker run -d --runtime nvidia --gpus "device=1" -p 9000:9000 \
  -e ASR_MODEL=base \
  -e ASR_ENGINE=openai_whisper \
  onerahmet/openai-whisper-asr-webservice:latest-gpu
  
docker run -d --runtime nvidia --gpus "device=1" -p 9000:9000 -e ASR_MODEL=base -e ASR_ENGINE=whisperx onerahmet/openai-whisper-asr-webservice:latest-gpu

docker run -d --runtime nvidia --gpus "device=1" --name whisper_engine -p 9007:9000 -e ASR_MODEL=medium.en -e ASR_ENGINE=whisperx -e ASR_MODEL_PATH=/data/whisper -e HF_TOKEN=hf_xxxxxxxxxxxxxxxxxxxxxx -v F:/.whisperx/cache:/data/whisper onerahmet/openai-whisper-asr-webservice:latest-gpu

# 100.100
docker run -d --runtime nvidia --gpus "device=4" --name whisper_engine -p 9007:9000 -e ASR_MODEL=medium.en -e ASR_ENGINE=whisperx -e HF_TOKEN=hf_xxxxxxxxxxxxxxxxxxxxxx -v /home/ubuntu/luo/.whisper-asr/cache:/root/.cache/ -v /home/ubuntu/luo/.whisper-asr/swagger-ui-assets:/swagger-ui-assets onerahmet/openai-whisper-asr-webservice:latest-gpu

manga-image-translator

docker build -t image-translator:v1 .

docker run -d \
    --name image_translator_gpu \
    -p 15003:5003 \
    --ipc=host \
    --gpus "device=0" \
    -v /home/ubuntu/luo/.image-translator/models:/app/models \
    -v /home/ubuntu/luo/.image-translator/result:/app/result \
    -e CUSTOM_OPENAI_API_KEY='sk-123' \
    -e CUSTOM_OPENAI_API_BASE='http://192.168.100.143:8200/v1' \
    -e CUSTOM_OPENAI_MODEL='Qwen3-235B-A22B-Instruct-2507-FP8' \
    -w /app/server \
    --entrypoint python \
    image-translator:v1 \
    main.py --verbose --start-instance --host=0.0.0.0 --port=5003 --use-gpu
    
docker run -d --name image_translator_gpu -p 15003:5003 --ipc=host --gpus "device=1" -v F:/pythonProject/manga-image-translator/models:/app/models -v F:/pythonProject/manga-image-translator/result:/app/result -e CUSTOM_OPENAI_API_KEY='sk-123' -e CUSTOM_OPENAI_API_BASE='http://192.168.100.143:8200/v1' -e CUSTOM_OPENAI_MODEL='Qwen3-235B-A22B-Instruct-2507-FP8' -w /app/server --entrypoint python image-translator:v1 smain.py --verbose --start-instance --host=0.0.0.0 --port=5003 --use-gpu

whisperlivekit

docker build -t wlk .

docker run -d --gpus "device=1" -p 8800:8000 --name wlk-production -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models/huggingface/hub:/root/.cache/huggingface/hub -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models/torch:/root/.cache/torch -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models/whisper:/root/.cache/whisper -e WHISPER_CACHE_DIR=/root/.cache/whisper -e HF_HOME=/root/.cache/huggingface -e TORCH_HOME=/root/.cache/torch wlk:diarization --model base --diarization --diarization-backend sortformer --target-language zh-CN

# 说话人识别
docker run -d --gpus "device=1" -p 8800:8000 --name wlk -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models:/whisperlivekit_models -e WHISPER_CACHE_DIR=/whisperlivekit_models/whisper -e HF_HOME=/whisperlivekit_models/huggingface -e TORCH_HOME=/whisperlivekit_models/torch -e XDG_CACHE_HOME=/whisperlivekit_models wlk:latest --model large-v3 --diarization --target-language zh 

docker run -d \
  --gpus "device=1" \
  -p 8800:8000 \
  --name wlk-production \
  --restart unless-stopped \
  -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models/huggingface/hub:/root/.cache/huggingface/hub \
  -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models/torch:/root/.cache/torch \
  -v F:/python_full_project/WhisperLiveKitProject/WhisperLiveKit/whisperlivekit_models/whisper:/root/.cache/whisper \
  -e WHISPER_CACHE_DIR=/root/.cache/whisper \
  -e HF_HOME=/root/.cache/huggingface \
  -e TORCH_HOME=/root/.cache/torch \
  wlk:diarization \
  --model large-v3 \
  --language auto \
  --diarization \
  --diarization-backend sortformer \
  --target-language zh-CN

方案11（使用 generateBundle 钩子直接注入代码）

原来采用虚拟模块的方式的代码需要修改么？

next-ai-draw-io

docker run -d -p 3008:3000 \
  --name next-ai-draw-io \
  -d --restart=unless-stopped \
  -e AI_PROVIDER=openai \
  -e AI_MODEL=Qwen3-235B-A22B-Instruct-2507-FP8 \
  -e OPENAI_BASE_URL=http://192.168.100.143:8200/v1 \
  -e OPENAI_API_KEY=sk-123 \
  ghcr.io/dayuanjiang/next-ai-draw-io:0.4.7

我是一个大一在校大学生，专业是软件工程。平时除了上课闲暇时间就喜欢逛逛论坛，看一些有趣的技术分享，平时也喜欢和大佬讨论C++、python编程技术。最近在google中搜到l站的llm和vps的相关帖子，作为爱好者想进来交流一下，希望能和大家一起学习。

我是在计算机行业工作5年的小菜鸡，目前做的工作是python软件开发和NLP算法，平时对软件开发和NLP算法探索感兴趣。最近在google中搜到l站的llm和vps的相关帖子，作为爱好者想进来交流一下，希望能和大家一起学习。

linuxdo站我一开始接触是在网上搜各种代码出错解决办法和教程，linuxdo总是能给出好的解决方法，因此觉得linuxdo生态特别好。之前一直就曾想申请账号，但是由于种种原因在注册的时候没有成功。希望这次贵站能够同意我的申请。

casbin/casdoor

docker-compose.yml

services:
  casdoor:
    image: casbin/casdoor::2.277.0
    container_name: casdoor
    restart: unless-stopped
    ports:
      - "8000:8000"
    environment:
      - GIN_MODE=release
    volumes:
      - ./conf:/conf
      - ./logs:/logs
    networks:
      - casdoor-network

networks:
  casdoor-network:
    driver: bridge

//div[@class=’btn-web-link’]/a/@href

//main[@id=’page-content’]

https://www.rand.org/pubs/external_publications/EP71199.html

elasticsearch

version: '3.8'

services:
  elasticsearch:
    image: elasticsearch:8.19.11
    container_name: elasticsearch
    restart: unless-stopped
    # 容器拥有root权限
    privileged: true
    # 在 Linux 里使用 ulimit 命令可以对进程的资源进行限制，这里设置为无限
    ulimits:
      memlock:
        soft: -1
        hard: -1
    environment:
      # 该环境变量设置 ElasticSearch 的最小和最大内存使用都是 1G
      - ES_JAVA_OPTS=-Xms1024m -Xmx1024m
      # 该环境变量设置成 0.0.0.0 表示允许任意客户端机器的连接访问
      - http.host=0.0.0.0
      - node.name=elastic
      - cluster.name=cluster_elasticsearch
      # 该环境变量设置为 single-node 表示部署的 ElasticSearch 为单节点
      - discovery.type=single-node
    volumes:
      - ./elasticsearch/elasticsearch-data:/usr/share/elasticsearch/data
      - ./elasticsearch/elasticsearch-config:/usr/share/elasticsearch/config
      - ./elasticsearch/elasticsearch-plugins:/usr/share/elasticsearch/plugins
      - ./elasticsearch/elasticsearch-logs:/usr/share/elasticsearch/logs
    ports:
      - "9200:9200"
    networks:
      - es-network

  kibana:
    image: kibana:8.19.11
    container_name: kibana
    restart: unless-stopped
    environment:
      - ELASTICSEARCH_HOSTS=http://elasticsearch:9200
    ports:
      - "5601:5601"
    depends_on:
      - elasticsearch
    networks:
      - es-network
    volumes:
      - ./kibana/kibana-data:/usr/share/kibana/data
      - ./kibana/kibana-config:/usr/share/kibana/config

networks:
  es-network:
    driver: bridge