mlff_builder_driver.py Command Help

Command: $SCHRODINGER/run mlff_builder_gui_dir/mlff_builder_driver.py

usage: $SCHRODINGER/run mlff_builder_gui_dir/mlff_builder_driver.py
       [-h] -training_cycle_size TRAINING_CYCLE_SIZE -batch_size BATCH_SIZE
       -training_cycles TRAINING_CYCLES [-gpu] [-use_dispersion_energy]
       -forcefield
       {MLFF:official:Hybrid_MPNICE,MLFF:official:Hybrid_MPNICE_I,MLFF:official:Hybrid_MPNICE_O,MLFF:official:Inorganic_MPNICE,MLFF:official:Organic_Crystals_MPNICE,MLFF:official:Organic_Crystals_MPNICE_v1.1,MLFF:official:Organic_MPNICE,MLFF:official:UMA_sm_oc20,MLFF:official:UMA_sm_odac,MLFF:official:UMA_sm_omat,MLFF:official:UMA_sm_omc,MLFF:official:UMA_sm_omol}
       -codebase {qe,jaguar} [-seed SEED] [-weight_decay WEIGHT_DECAY]
       [-learning_rate LEARNING_RATE]
       input_file

MLFF builder driver. Copyright Schrodinger, LLC. All rights reserved.

positional arguments:
  input_file            File to read structures from for training.

options:
  -h, -help             Show this help message and exit.
  -training_cycle_size TRAINING_CYCLE_SIZE
                        Train cycle size. (default: None)
  -batch_size BATCH_SIZE
                        Training batch size. (default: None)
  -training_cycles TRAINING_CYCLES
                        Maximum number of validation cycles. (default: None)
  -gpu
  -use_dispersion_energy
                        Whether to use Grimme dispersion energy. (default:
                        False)
  -forcefield {MLFF:official:Hybrid_MPNICE,MLFF:official:Hybrid_MPNICE_I,MLFF:official:Hybrid_MPNICE_O,MLFF:official:Inorganic_MPNICE,MLFF:official:Organic_Crystals_MPNICE,MLFF:official:Organic_Crystals_MPNICE_v1.1,MLFF:official:Organic_MPNICE,MLFF:official:UMA_sm_oc20,MLFF:official:UMA_sm_odac,MLFF:official:UMA_sm_omat,MLFF:official:UMA_sm_omc,MLFF:official:UMA_sm_omol}
                        Force field to use in energy evaluations. (default:
                        None)
  -codebase {qe,jaguar}
                        Codebase used to generate the reference data.
                        (default: None)
  -seed SEED            Seed for random number generator. (default: 1234)
  -weight_decay WEIGHT_DECAY
                        Weight decay (if negative, decay towards initial
                        parameter values). (default: 0.0001)
  -learning_rate LEARNING_RATE
                        Learning rate for optimizer. (default: 0.001)