https://www.kaggle.com/competitions/predict-electricity-consumption/overview
https://gist.github.com/ericavonb/3c79e5035567c8ef3267
- Create a virtual environment using
python -m venv env
- Install dependencies
pip install -r requirements.txt
- If you want to train using GPU
- While using cuda 11.3 write
pip install -r requirements-cuda113.txt
instead - Make sure your
torch
installations matchescuda
installed on your server - Do not use
cuda
version10.2
(error in multiplication) (minimal example to test)
- While using cuda 11.3 write
- If you want to train using GPU
- Enlist virtual environment as usable jupyter kernel
python -m ipykernel --user --name <name-of-kernel>
- Run JupyterServer and your Jupyter IDE
- Make sure the notebook uses your
<name-of-kernel>
kernel for execution
LSTM code is in lstm branch