Skip to content

Kairixir/ml-tartu-electricity-prediciton

Repository files navigation

ml-tartu-electricity-prediciton

https://www.kaggle.com/competitions/predict-electricity-consumption/overview

Commit guidelines

https://gist.github.com/ericavonb/3c79e5035567c8ef3267

Install instructions

  1. Create a virtual environment using python -m venv env
  2. Install dependencies pip install -r requirements.txt
    • If you want to train using GPU
      1. While using cuda 11.3 write pip install -r requirements-cuda113.txt instead
      2. Make sure your torch installations matches cuda installed on your server
      3. Do not use cuda version 10.2 (error in multiplication) (minimal example to test)
  3. Enlist virtual environment as usable jupyter kernel python -m ipykernel --user --name <name-of-kernel>
  4. Run JupyterServer and your Jupyter IDE
  5. Make sure the notebook uses your <name-of-kernel> kernel for execution

Important

LSTM code is in lstm branch