Configuration

Overriding Configurations

You can override configurations directly from the command line by appending key=value pairs to your command. For example, if you want to change the learning rate, which is defined as follows in YAML:

# configurations/experiment/xxx.yaml
experiment:
  training:
    lr: 1e-3

you can override it with the following command:

python -m main ... experiment.training.lr=5e-3

Dataset-Dependent Configurations

In many cases, algorithm and experiment configurations are dependent on the dataset. For instance, in the video_generation experiment using the flowm_video algorithm, configurations such as batch size, learning rate, and model size vary based on the dataset. To manage this variability, we utilize Hydra's Specializing Configuration feature.

This is implemented in the configurations/shortcode folder, enabling dataset-specific customization of experiment and algorithm settings. YAMLs are organized there based on dataset and experiment.

Other configurations

Auxiliary configuration files are described below:

• ckpt_map: Mapping between the model name and the location that it is stored. You can make one for each server, if they are stored in different locations.

• secrets: Used to store all information not intended to be pushed to github, such as the wandb api key.

• cluster: Cluster settings for submitting experiments through slurm. Setting cluster=xxx in the command line arguments should automatically send it through slurm instead.