Algorithm Outputs

The output directory contains the following:

Output Directory Structure
pyt_save/ PyTorch implementations only. A directory containingeverything needed to restore the trained agent and valuefunctions. (Details for PyTorch saves below.)
tf1_save/ Tensorflow implementations only. A directory containingeverything needed to restore the trained agent and valuefunctions. (Details for Tensorflow saves below.)
config.json A dict containing an as-complete-as-possible description of the args and kwargs you used to launch the trainingfunction. If you passed in something which can’t be serialized to JSON, it should get handled gracefully by thelogger, and the config file will represent it with a string.
Note: this is meant for record-keeping only. Launching an experiment from a config file is not currently supported.
progress.txt A tab-separated value file containing records of the metricsrecorded by the logger throughout training. eg, Epoch,AverageEpRet, etc.
vars.pkl A pickle file containing anything about the algorithm state which should get stored. Currently, all algorithms only use this to save a copy of the environment.

PyTorch Save Directory

The pyt_save directory contains:

Pyt_Save Directory Structure
model.pt A file created with torch.save, essentially just a pickled PyTorch nn.Module. Loading it will restorea trained agent as an ActorCritic object with an actmethod.

Save Directory Location

Experiment results will, by default, be saved in the same directory as the Spinning Up package, in a folder called data:

spinningup/
data/
        ...
    docs/
        ...
    spinup/
        ...
    LICENSE
    setup.py