Learning Curve Fitting

Learning Curve Fitting Paper Experiments

Running out experiments

First, we need to download the LCDB

wget ...

Then (optionally) download our image:

singularity build learningcurvefitting.sif docker://taylanot/learningcurvefitting:latest

Then one can run the experiments with the image:

singularity exec learningcurvefitting.sif python3 main.py -n name-of-your-experiment -F model_name

Above code will save the experiment in a unique file with all the run information. If you want to change configuration from the command line just do:

singularity exec learningcurvefitting.sif python3 main.py -n name-of-your-experiment -F experiments with conf.x=y

where x and y are related parts that you want to change. For setting the seed following command will work.

singularity exec learningcurvefitting.sif python3 main.py -n name-of-your-experiment -F experiments with seed=24