[Docker] [Experiments] NVIDIA Kaolin Wisp配置

Installing Kaolin Wisp

Reference: Kaolin Wisp installation

1. Docker (Python 3.8 environment)

docker run --gpus all --ipc=host --net=host --pid=host --name vqad -v /public/data0/MULT:/MULT -v /etc/localtime:/etc/localtime -v /public/data0/data0:/data0 -v /public/data1:/data1 -v /raid:/raid -it nvcr.io/nvidia/pytorch:21.08-py3-pt1.10dvgo /bin/bash

docker start vqad

docker exec -it vqad bash

pip install --upgrade pip

2. Install OpenEXR (Linux) (root)

apt-get update

apt-get install libopenexr-dev

3. Install Kaolin

Reference: Kaolin installation

git clone --recursive https://github.com/NVIDIAGameWorks/kaolin

cd kaolin

python setup.py develop

kaolinsetup

Testing

Run a quick test of your installation and version:

python -c “import kaolin; print(kaolin.version)”

quicktest

For an exhaustive check, install testing dependencies and run tests as follows: (Linux)

pip install -r tools/ci_requirements.txt
export CI=‘true’
pytest tests/python/

4. Install the rest of the dependencies

requirements.txt

cd …/code/nerf/kaolin-wisp/

pip install -r requirements.txt -i https://pypi.douban.com/simple

5. Installing Wisp

python setup.py develop

wissetup

Testing Your Installation

1. Download some sample data

RTMV dataset
Project: RTMV: A Ray-Traced Multi-View Synthetic Dataset for Novel View Synthesis

2. Train a NeRF using NGLOD

WISP_HEADLESS=1 python app/main.py --config configs/nglod_nerf.yaml --dataset-path /path/to/V8 --dataset-num-workers 4

3. Errors

AttributeError: partially initialized module ‘cv2’ has no attribute ‘_registerMatType’ (most likely due to a circular import)

_registerMatType
安装低版本的opencv_python:
Reference:AttributeError: partially initialized module ‘cv2’ has no attribute ‘_registerMatType’

pip uninstall opencv-python
pip uninstall opencv-contrib-python
pip install opencv_python==3.4.10.37

TypeError: Descriptors cannot not be created directly.

Descriptors
安装低版本的protobuf:
Reference: TypeError: Descriptors cannot not be created directly.

pip uninstall protobuf
pip install protobuf==3.20.*

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值