search

PG642 / multi-sample-factory

High throughput reinforcement learning on clusters
MIT License
2 stars 0 forks source link

readme

multi-sample-factory

High throughput reinforcement learning on clusters

Installation (for member of the organization only)

  1. Download the .whl-file for PyTorch in your LiDO3 work directory from the OneDrive link in Discord.
  2. Clone tis repo with git clone https://github.com/PG642/multi-sample-factory.git.
  3. Execute the installation script with 
    cd /work/USERNAME/multi-sample-factory
./install-multi-sample-factory.sh
  4. Now you can try sample-factory with RoboLeague by executing multi_sample_factory_examples/train_rocket_league_env.py in any SLURM script. An example script is uploaded in the same OneDrive link as before. Remember to fill in your email address and your username for the output parameter.

Fork of sample-factory