- Create a Conda environment. For this you need to have Conda or Miniconda installed.
conda create --name name_env python=3.11 ipython- Activate the environment.
conda activate name_env- Install the dependencies.
pip install -r requirements.txt- Train.
python src/main.py4b. Alternatively, to train in CALCULA (UPC)
srun -A veu -p veu --mem=16G -c 2 --gres=gpu:4 python src/main.py
To use this code, it is needed the .wav files in three different folders: train valid and test. The code will automatically create a pandas dataframe for every dataset. If it is needed to change the data once we have already trained, for example for using data with spectral subtraction, it will be required to delete the previous dataframes.
The OCEAN project is focused on enhancing operator awareness in navigation, to reduce the frequency of severe accidents like collision and grounding, to mitigate ship-strike risks to marine mammals, and to mitigate the risk presented by floating obstacles to ships.
The OCEAN project will contribute to an improved understanding of accident root causes, and will strive to reduce the resulting human, environmental and economic losses through socio-technical innovations supporting ship navigators.
The OCEAN consortium, coordinated by Western Norway University of Applied Sciences, includes 13 partner organizations across 7 different European countries from the industry, academia, NGOs and end users
The project was awarded funding by Horizon Europe and launched in October 2022, and it is due to run until 2025.