databases

This is summary documentation for hippynn.databases module. Click here for full documentation.

API Summary for sub-modules:

• SNAPJson
• database
• h5_pyanitools
• ondisk
• restarter

Links to Full Documentation:

Organized datasets for training and prediction.

Note

Databases constructed from disk (i.e. anything besides the base Database class) will load floating point data in the format (float32 or float64) specified via the torch.get_default_dtype() function. Use torch.set_default_dtype() to control this behavior.

Modules

SNAPJson	Load database from SNAP format.
database	Base database functionality from dictionary of numpy arrays
h5_pyanitools	Read Databases in the pyanitools H5 format.
ondisk	Dataset stored as NPY files in directory or as NPZ dictionary.
restarter	Classes for allowing reload of database.