Helper Module for Deep Learning with pytorch.
This work is made available by a community of people, amoung which the CEA Neurospin BAOBAB laboratory.
- Official source code repo: https://github.com/neurospin/pynet
- HTML documentation (last stable release): http://neurospin.github.io/pynet
- HTML documentation (master release): https://python-network.readthedocs.io/
You can list all available Deep Learning tools by executing in a Python shell:
from pprint import pprint
import pynet
pprint(pynet.get_tools())
The 'get_tools' function returns a dictionary with all available 'networks', 'losses', 'regularizers', and 'metrics'.
Then each network has been embeded in a Deep Learning training interface. Network parameters are set using the NetParameters object. You can list all these interfaces by executing in a Python shell:
from pprint import pprint
import pynet
pprint(pynet.get_interfaces(family=None))
params = pynet.NetParameters(param1=1, param2=2)
params.param3 = 3
The 'get_interfaces' function returns a dictionary with interfaces sorted by family names. You can filter the result by providing the family name or a list of family names of interest.
You can list also all available data fetchers by executing in a Python shell:
from pprint import pprint
import pynet.datasets import get_fetchers
pprint(get_fetchers())
The 'get_fetchers' function returns a dictionary with all the declared fetchers. Finally you may want to look at the data manger class that provides convenient tools to split/stratify you dataset:
from pynet.datasets import DataManager
Make sure you have installed all the package dependencies. Further instructions are available at https://neurospin.github.io/pynet/generated/installation.html