Exercises for 3D Computer Vision

Requirements

Python 3

Installation

Unix

For the installation I recommend a python venv.

git clone git@github.com:vislearn/3dcv-students.git
cd 3dcv/
python3 -m venv .
source bin/activate
pip install -r requirements.txt

Windows

Download Python here. Make sure that the installer adds Python to your PATH.

In PowerShell run

git clone git@github.com:vislearn/3dcv-students.git

(if you don't use git, just download and extract the zip file)

cd .\3dcv\
pip install torch torchvision -f https://download.pytorch.org/whl/torch_stable.html
pip install -r .\requirements.txt

Usage

In order to edit the notebooks run

jupyter-lab

to start the local webserver. If the notebook home page does not show up immediately, open http://localhost:8888 in your browser. Now, just open one of the task notebooks and start editing.

Project Organization

├── 1.0-tl-scientific-python.ipynb          <- The Notebooks, containing your tasks
│
├── ...
│
├── LICENSE                                 <- The License
│
├── README.md                               <- The top-level README with installation instructions
│
├── data                                    <- The necessary data files, e.g. datasets
│
├── models                                  <- Trained and serialized models, model predictions, or model summaries
│
├── requirements.txt                        <- The requirements file to setup the environment
│
├── task.pdf                               <- Detailed information about your tasks
│
├── vll                                     <- Boilerplate source code provided by us
│   │
│   ├── data                                <- Scripts to download or generate data
│   │
│   ├── model                               <- Scripts defining the neural network models
│   │
│   ├── utils                               <- Scripts utilities used during data generation or training
│   │
│   ├── validate                            <- Scripts to validate models
│   │
│   └── visualize                           <- Scripts to create exploratory and results oriented visualizations