Skip to content

mbacvanski/AnalogET

Repository files navigation

Getting Started

Install uv for package management

https://docs.astral.sh/uv/getting-started/installation/

Install and activate virtual environment

uv sync
source .venv/bin/activate

Train model

model_energy_train.py creates the training and validation data and saves them to the data folder. It also saved the best model weights and the weights from the last epoch to the data folder.

uv run model_energy_train.py

Visualization notebook

These notebooks use the pretrained weights stored in data/model_best.npz.

make_plots.ipynb contains various plots, including sample energy and visible neuron trajectories and training/validation loss/accuracy.

About

Code for the Analog Energy Transformer

Resources

Stars

Watchers

Forks

Contributors 2

  •  
  •