-
Notifications
You must be signed in to change notification settings - Fork 64
/
Copy pathDockerfile
39 lines (27 loc) · 1.12 KB
/
Dockerfile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
FROM pytorch/pytorch:2.1.2-cuda11.8-cudnn8-devel
ARG DEBIAN_FRONTEND=noninteractive
ARG FORCE_CUDA=1
ENV FORCE_CUDA=${FORCE_CUDA}
ENV NVIDIA_VISIBLE_DEVICES \
${NVIDIA_VISIBLE_DEVICES:-all}
ENV NVIDIA_DRIVER_CAPABILITIES \
${NVIDIA_DRIVER_CAPABILITIES:+$NVIDIA_DRIVER_CAPABILITIES,}graphics
RUN apt-get update -y && apt-get install -y git wget
WORKDIR /
RUN git clone --recursive https://github.com/NVlabs/InstantSplat.git &&\
cd InstantSplat &&\
git submodule update --init --recursive &&\
cd submodules/dust3r/ &&\
mkdir -p checkpoints/ &&\
wget https://download.europe.naverlabs.com/ComputerVision/DUSt3R/DUSt3R_ViTLarge_BaseDecoder_512_dpt.pth -P checkpoints/
WORKDIR /InstantSplat
RUN pip3 install -r requirements.txt
ARG TORCH_CUDA_ARCH_LIST="3.5;5.0;6.0;6.1;7.0;7.5;8.0;8.6+PTX"
RUN pip3 install submodules/simple-knn
RUN pip3 install submodules/diff-gaussian-rasterization
RUN pip3 install submodules/fused-ssim
RUN cd croco/models/curope/ &&\
python3 setup.py build_ext --inplace
RUN pip3 install plyfile
RUN apt-get install -y libgl1-mesa-dev libglib2.0-0
ENV CUDA_VISIBLE_DEVICES=0