salernosimone/hunyuanworld-mirror repository overview
HunyuanWorld-Mirror
HunyuanWorld-Mirror is a versatile feed-forward model for comprehensive 3D geometric prediction. It integrates diverse geometric priors (camera poses, calibrated intrinsics, depth maps) and simultaneously generates various 3D representations (point clouds, multi-view depths, camera parameters, surface normals, 3D Gaussians) in a single forward pass.
Quickstart
sudo docker run -it --rm -v /path/to/images:/HunyuanWorld-Mirror/examples salernosimone/hunyuanworld-mirror bash
# inside container
cd /HunyuanWorld-Mirror
python infer.py --input_path /path/to/your/input --output_path /path/to/your/output --save_colmap --save_gs
Dockerfile
It is based on Python 3.11, CUDA 12.4, PyTorch 2.4.1.
FROM salernosimone/torch:py311-cu124
RUN git clone https://github.com/Tencent-Hunyuan/HunyuanWorld-Mirror.git && \
cd HunyuanWorld-Mirror && \
TORCH_CUDA_ARCH_LIST="7.5 8.0 8.6 8.9 9.0" pip install git+https://github.com/nerfstudio-project/gsplat.git && \
pip install -r requirements.txt
RUN micromamba install -y "huggingface_hub[cli]" && hf download tencent/HunyuanWorld-Mirror --local-dir /HunyuanWorld-Mirror/ckpts
RUN apt update && apt install -y libgl1 libglib2.0-0 ffmpeg && rm -rf /var/lib/apt/lists/*
RUN sed -i 's|tencent/HunyuanWorld-Mirror|ckpts|g' /HunyuanWorld-Mirror/infer.py
Tag summary
Content type
Image
Digest
sha256:596554376…
Size
14.4 GB
Last updated
about 1 year ago
docker pull salernosimone/hunyuanworld-mirror