def save_model(self, net):
"""Save the model.
This function saves some or all of the following:
- model parameters;
- optimizer state;
- criterion state;
- training history;
- custom modules;
- entire model object.
"""
kwargs_module, kwargs_other = _check_f_arguments(
self.__class__.__name__, **self._f_kwargs())
for key, val in kwargs_module.items():
if val is None:
continue
f = self._format_target(net, val, -1)
key = key[:-1] # remove trailing '_'
self._save_params(f, net, 'f_' + key, key + " state")
f_history = kwargs_other.get('f_history')
if f_history is not None:
f = self.f_history_
self._save_params(f, net, "f_history", "history")
f_pickle = kwargs_other.get('f_pickle')
if f_pickle:
f_pickle = self._format_target(net, f_pickle, -1)
with open_file_like(f_pickle, 'wb') as f:
pickle.dump(net, f)
def save_model(self, net):
"""Save the model.
This function saves some or all of the following:
- model parameters;
- optimizer state;
- training history;
- entire model object.
"""
if self.f_params is not None:
f = self._format_target(net, self.f_params, -1)
self._save_params(f, net, "f_params", "model parameters")
if self.f_optimizer is not None:
f = self._format_target(net, self.f_optimizer, -1)
self._save_params(f, net, "f_optimizer", "optimizer state")
if self.f_history is not None:
f = self.f_history_
self._save_params(f, net, "f_history", "history")
if self.f_pickle:
f_pickle = self._format_target(net, self.f_pickle, -1)
with open_file_like(f_pickle, 'wb') as f:
pickle.dump(net, f)
def save_history(self, f):
"""Saves the history of ``NeuralNet`` as a json file. In order
to use this feature, the history must only contain JSON encodable
Python data structures. Numpy and PyTorch types should not
be in the history.
Parameters
----------
f : file-like object or str
Examples
--------
>>> before = NeuralNetClassifier(mymodule)
>>> before.fit(X, y, epoch=2) # Train for 2 epochs
>>> before.save_params('path/to/params')
>>> before.save_history('path/to/history.json')
>>> after = NeuralNetClassifier(mymodule).initialize()
>>> after.load_params('path/to/params')
>>> after.load_history('path/to/history.json')
>>> after.fit(X, y, epoch=2) # Train for another 2 epochs
"""
with open_file_like(f, 'w') as fp:
json.dump(self.history.to_list(), fp)
def from_file(cls, f):
"""Load the history of a ``NeuralNet`` from a json file.
Parameters
----------
f : file-like object or str
"""
with open_file_like(f, 'r') as fp:
return cls(json.load(fp))
def load_history(self, f):
"""Load the history of a ``NeuralNet`` from a json file. See
``save_history`` for examples.
Parameters
----------
f : file-like object or str
"""
with open_file_like(f, 'r') as fp:
self.history = History(json.load(fp))
def to_file(self, f):
"""Saves the history as a json file. In order to use this feature,
the history must only contain JSON encodable Python data structures.
Numpy and PyTorch types should not be in the history.
Parameters
----------
f : file-like object or str
"""
with open_file_like(f, 'w') as fp:
json.dump(self.to_list(), fp)
def save_model(self, net):
"""Save the model.
This function saves some or all of the following:
- model parameters;
- training history;
- entire model object.
"""
if self.f_params:
net.save_params(self._format_target(net, self.f_params))
if self.f_history:
net.save_history(self._format_target(net, self.f_history))
if self.f_pickle:
f_pickle = self._format_target(net, self.f_pickle)
with open_file_like(f_pickle, 'wb') as f:
pickle.dump(net, f)
