Create the ndviforecasting environment with the following dependencies:
conda create -n ndviforecasting python==3.8.8
conda activate ndviforecasting
pip install torch==1.9.0+cu102 torchvision==0.10.0+cu102 torchaudio===0.9.0 h5py -f https://download.pytorch.org/whl/torch_stable.html
pip install git+https://github.com/catalyst-team/catalyst.git
conda install pandas matplotlib seaborn scikit-learn
pip install wandb
And voila!
Then, clone the git repository:
https://github.com/waldnerf/NDVI_forecasting.git
cd NDVI_forecasting
If you want to save the performance of your models on wandb, you will need to link your account by either doing:
>> wandb login
or
python
import wandb
wandb.login()
and follow the instructions.
You can also run the code in Colab
Below is the architecture of the folder. Code in run_model.py is used to train a model, and evaluate_model.py serves for model evaluation. The different models that are available can be found in nn/models
.
├── README.md
├── run_model.py
├── run_model.ipynb
├── evaluate_model.py
├── data
├── nn
│ ├── loss
│ └── models
└── utils