Advanced options

The following guide gives an overview over advanced capabilities of MALA, such as performance improvements and more in-depth hyperparameter searches. Parallelization and hardware acceleration will also be discussed. When using the former, i.e., when computing across multiple CPUs, it is useful to employ the MALA parallel-safe printing function for all output.

import mala

mala.printout("This message will only be printed once, even "
               "if multiple CPUS are used")

• Improved training performance
• Storing data with OpenPMD
• Improved data conversion
• Improved hyperparameter optimization
• Using MALA in production