🎶 https://youtu.be/Y3ywicffOj4 😴

structure

• neuralnet.py defines network, training, dreaming, neuron-selector
• train.py trains initial NN
• dream.py starts dreaming for a given model
• datagen.py functions for generating data
• AnalyzeDream.ipynb to load and analyze dreaming data by hand

setting up on the cluster

module load anaconda/3/2021.11
virtualenv --python=python3.9 venv
source venv/bin/activate
pip install -r requirements.txt

for data generation, configure configs/datagen.yaml, and run

sbatch startup/datagen

or

sbatch startup/datagen_parallel

for training, configure configs/train.yaml, and run

sbatch startup/train

models and config files are saved

for dreaming, configure configs/dream.yaml, and run

sbatch startup/dream

random

useful commands:

• du -h filename (size of file)