stepbystep.txt

# Git installieren
sudo apt install git

# Tensorflow Models Repository clonen:
git clone https://github.com/tensorflow/models.git

# Eigenes git clonen
git clone https://github.com/93Dennis/cultural-metadata-ai.git
cd models/research/delf/delf/python/training

# Install Python
sudo apt install python3-pip

# Install curl
sudo apt install curl

# Im Installationsscript Zeile 111 anpassen (/research/delf/delf/python/training/install_delf.sh):
cd ../../../../object_detection/packages/tf1

# Install object detection
pip install tensorflow-object-detection-api

# modules einbinden [ACHTUNG PFAD ANPASSEN!]
PYTHONPATH=$PYTHONPATH:/[path]/models/research/delf:/[path]/photo-geolocalization
export PYTHONPATH

# Installationsscript ausführen (https://github.com/tensorflow/models/blob/master/research/delf/INSTALL_INSTRUCTIONS.md): 
# From models/research/delf/delf/python/training
bash install_delf.sh

# R50-DELG-GLDv2-clean Model in parameters Ordner laden
# From cultural-metadata-ai/delg/parameters
wget https://storage.googleapis.com/delf/r50delg_gldv2clean_20200914.tar.gz
tar -xvzf r50delg_gldv2clean_20200914.tar.gz