Docker images for multiple LLM applications
Note: Select the appropriate image based on your system (x64 CPU/NVIDIA GPU support required)
Docker Images for Apps
FramePack
Windows requirement: NVIDIA drivers + CUDA. Install CUDA to Enable GPU Acceleration for LLM Apps
# FramePack CUDA 12.8
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/framepack:1.0.0-gpu-cu128
# FramePack CUDA 12.6
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/framepack:1.0.0-gpu-cu126
# FramePack CUDA 12.4
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/framepack:1.0.0-gpu-cu124
# FramePack CUDA 12.1
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/framepack:1.0.0-gpu-cu121
More architecture support coming soon (under testing).
Demo
docker run -d --gpus all --name framepack-local -v "D:/work/lmd_data_root/apps/framepack/hf_download:/root/FramePack/hf_download" -p 7860:7860 -e HF_ENDPOINT=https://hf-mirror.com crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/framepack:1.0.0-gpu-cu128
Note that adding --gpus all is necessary for the container to access the host machine's GPU. Use the method of mounting the host directory to specify the model directory. If you cannot access https://huggingface.co/
normally, add HF_ENDPOINT=https://hf-mirror.com
as a proxy. Other parameters are more common.
MaxKB
This application can run in both Windows and macOS environments and has no special requirements for the graphics card.
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/3rd-party/1panel-maxkb:v1.10.8-lts
Other App images
More application images are currently under development and testing, coming soon.
Base Images
# Python 3.10 and CUDA 12.8
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/python-torch-cuda:1.0.0-cp310-torch2.7.0-cu128
# Python 3.10 and CUDA 12.6
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/python-torch-cuda:1.0.0-cp310-torch2.7.0-cu126
# Python 3.10 and CUDA 12.4
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/python-torch-cuda:1.0.0-cp310-torch2.6.0-cu124
# Python 3.10 and CUDA 12.1
crpi-ru0fy7yg2nq6feyc.cn-shanghai.personal.cr.aliyuncs.com/lmdown/python-torch-cuda:1.0.0-cp310-torch2.5.0-cu121