________                  ___________.__            .___            
\_____  \_______   ____   \_   _____/|__| ____    __| _/___________ 
 /   |   \_  __ \_/ __ \   |    __)  |  |/    \  / __ |/ __ \_  __ \
/    |    \  | \/\  ___/   |     \   |  |   |  \/ /_/ \  ___/|  | \/
\_______  /__|    \___  >  \___  /   |__|___|  /\____ |\___  >__|   
        \/            \/       \/            \/      \/    \/      

If you have already mined in Minecraft, you may have found yourself stressing out of fear of walking past valuable ores. With OreFinder, this time is over! OreFinder implements the one-stage object detector YOLOV5(m) v6 [1], trained on a labelled dataset of 3k Minecraft gameplay screenshots.

Project undertaken as part of the INFO8010 course given by Pr. Louppe at ULiège. Final grade : 15/20. Note : No copying (even partial) of this code within the scope of the INFO8010 course will be tolerated.

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh # Accept everything, say 'yes'...
conda init # Run it in a new terminal
conda create -n info8010
conda activate info8010
conda info --envs #check your conda envs
which python3 # check that the python version is the one from conda (something like /home/username/miniconda3/envs/info8010/bin/python3) 
conda install ipykernel
sudo apt install graphviz

When running the notebook, make sure to select the kernel info8010.

Official Implementation : https://github.com/ultralytics/yolov5/tree/master/models

YOLOV5 explained:

• https://iq.opengenus.org/yolov5/
• https://towardsai.net/p/l/yolov5m-implementation-from-scratch-with-pytorch
• https://sh-tsang.medium.com/brief-review-yolov5-for-object-detection-84cc6c6a0e3a
• ultralytics/yolov5#3181

• YOLOv5: https://universe.roboflow.com/yolo-minecraft/minecraft-ogpjp/dataset/10

• Course page: https://github.com/glouppe/info8010-deep-learning