MSCOCO dataset: n_samples = 13500
For geting dataset:
bash getData.sh
We recommend anaconda/miniconda.
You can setup your environment(x86_64) by entering:
conda env create -f environment.yml
If you want to use pip, enter following in your venv:
pip install -r requirements.txt
python train.py --p /path/to/training/data --a /path/to/annontations
If you are on Ubuntu or similar distribution juse python
with python3
It also supports some more flags which you can see by entering
python train.py --help
python scene_captioning.py