3.14. Saving / Loading

3.14.1. Saving / loading the state of the network

To save or load the network state you can use the following methods:

ANNarchy.save(filename, populations=True, projections=True, net_id=0)[source]

Save the current network state (parameters and variables) to a file.

  • If the extension is ‘.npz’, the data will be saved and compressed using np.savez_compressed (recommended).
  • If the extension is ‘.mat’, the data will be saved as a Matlab 7.2 file. Scipy must be installed.
  • If the extension ends with ‘.gz’, the data will be pickled into a binary file and compressed using gzip.
  • Otherwise, the data will be pickled into a simple binary text file using cPickle.
Parameters:
  • filename – filename, may contain relative or absolute path.
  • populations – if True, population data will be saved (by default True)
  • projections – if True, projection data will be saved (by default True)

Warning

The ‘.mat’ data will not be loadable by ANNarchy, it is only for external analysis purpose.

Example:

save('results/init.npz')

save('results/init.data')

save('results/init.txt.gz')

save('1000_trials.mat')
ANNarchy.load(filename, populations=True, projections=True, net_id=0)[source]

Loads a saved state of the network.

Warning: Matlab data can not be loaded.

Parameters:
  • filename – the filename with relative or absolute path.
  • populations – if True, population data will be loaded (by default True)
  • projections – if True, projection data will be loaded (by default True)

Example:

load('results/network.npz')

Please note that these functions are only usable after the call to ANNarchy.compile().

3.14.2. Saving / loading the parameters of the network

ANNarchy.save_parameters(filename, net_id=0)[source]

Saves the global parameters of a network (flag population for neurons, projection for synapses) to a JSON file.

Parameters:
  • filename – path to the JSON file.
  • net_id – ID of the network (default: 0, the global network).
ANNarchy.load_parameters(filename, global_only=True, verbose=False, net_id=0)[source]

Loads the global parameters of a network (flag population for neurons, projection for synapses) from a JSON file.

Parameters:
  • filename – path to the JSON file.
  • global_only – True if only global parameters (flags population and projection) should be loaded, the other values are ignored. (default: True)
  • verbose – True if the old and new values of the parameters should be printed (default: False).
  • net_id – ID of the network (default: 0, the global network).
Returns:

a dictionary of additional parameters not related to populations or projections (keyword network in the JSON file).

It is advised to generate the JSON file first with save_parameters() and later edit it manually.

A strong restriction is that population/projection names cannot change between saving and loading. By default, they take names such as pop0 or proj2, we advise setting explicitly a name in their constructor for readability.

If you add a parameter name to the JSON file but it does not exist in te neuron/synapse, it will be silently skipped. Enable verbose=True to see which parameters are effectively changed.

If you set global_only to True, you will be able to set values for non-global parameters (e.g. synapse-specific), but a single value will be loaded for all. The JSON file cannot contain arrays.

If you want to save/load the value of variables after a simulation, please refer to save() or load().