NICE-SLAM code reproduction and analysis (to solve the problem of slow download when using yaml file to configure the conda environment)

code run

Code address : https://github.com/cvg/nice-slam/tree/master

Environment configuration

1. Download the compressed package and open the environment.yaml file
   

2. Add the following code to the yaml file

  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/  
  - https://mirrors.ustc.edu.cn/anaconda/pkgs/free/   

It seems that you have to comment out the channels: inside - pytorch

3. According to the environment configuration tutorial in github, run in the base environment:

sudo apt-get install libopenexr-dev
conda env create -f environment.yaml

The previous command will create a nice-slam environment and activate the environment

conda activate nice-slam

After being in this environment, it seems that a requirements.txt file needs to be written to add the ones that need to be installed by pip (it is best to install with Tsinghua source)

Visualizing NICE-SLAM Results

1. Download the data set
   We are here to test the visualization effect of Apartment
   
   According to github, our domestic players should choose which extra data set in the download script to download manually, as follows (very conscientious, this cloud disk does not download without speed limit)
   

2. Unzip the downloaded compressed package to the folder corresponding to ouput, and run:

python visualizer.py configs/Apartment/apartment.yaml --output output/vis/Apartment

You can see the following effect:

3. After the operation is completed, the effect is still good

run demo

1. The method of downloading the dataset
   is the same as above, after downloading, put it in the ./Datasets created under the project, and unzip it

2. run

python -W ignore run.py configs/Demo/demo.yaml

Assi~~~

My GTX3070 has been sent, the game is over~