MCPcopy Create free account
hub / github.com/modAL-python/modAL / set_dropout_mode

Function set_dropout_mode

modAL/dropout.py:405–436  ·  view source on GitHub ↗

Function to change the mode of the dropout layers (bool: train_mode -> train or evaluation) Args: model: Pytorch model dropout_layer_indexes: Indexes of the dropout layers which should be activated Choose indices from : list(torch_model.mod

(model, dropout_layer_indexes: list, train_mode: bool)

Source from the content-addressed store, hash-verified

403
404
405def set_dropout_mode(model, dropout_layer_indexes: list, train_mode: bool):
406 """
407 Function to change the mode of the dropout layers (bool: train_mode -> train or evaluation)
408
409 Args:
410 model: Pytorch model
411 dropout_layer_indexes: Indexes of the dropout layers which should be activated
412 Choose indices from : list(torch_model.modules())
413 train_mode: boolean, true <=> train_mode, false <=> evaluation_mode
414 """
415
416 modules = list(model.modules()) # list of all modules in the network.
417
418 if len(dropout_layer_indexes) != 0:
419 for index in dropout_layer_indexes:
420 layer = modules[index]
421 if layer.__class__.__name__.startswith('Dropout'):
422 if True == train_mode:
423 layer.train()
424 elif False == train_mode:
425 layer.eval()
426 else:
427 raise KeyError(
428 "The passed index: {} is not a Dropout layer".format(index))
429
430 else:
431 for module in modules:
432 if module.__class__.__name__.startswith('Dropout'):
433 if True == train_mode:
434 module.train()
435 elif False == train_mode:
436 module.eval()

Callers 1

get_predictionsFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected

Used in the wild real call sites across dependent graphs

searching dependent graphs…