Mercurial > ift6266

comparison deep/stacked_dae/stacked_dae.py @ 185:b9ea8e2d071a

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Enlevé ce qui concernait la réutilisation de résultats de préentraînement (trop compliqué pour peu de bénéfice: c'est le finetuning qui est vraiment long
author fsavard
date Fri, 26 Feb 2010 17:45:52 -0500
parents 1f5937e9e530
children d364a130b221
comparison
equal deleted inserted replaced
172:4d3d3627df3e 185:b9ea8e2d071a
142 142
143 class SdA(object): 143 class SdA(object):
144 def __init__(self, train_set_x, train_set_y, batch_size, n_ins, 144 def __init__(self, train_set_x, train_set_y, batch_size, n_ins,
145 hidden_layers_sizes, n_outs, 145 hidden_layers_sizes, n_outs,
146 corruption_levels, rng, pretrain_lr, finetune_lr, input_divider=1.0): 146 corruption_levels, rng, pretrain_lr, finetune_lr, input_divider=1.0):
147 # Just to make sure those are not modified somewhere else afterwards
148 hidden_layers_sizes = copy.deepcopy(hidden_layers_sizes)
149 corruption_levels = copy.deepcopy(corruption_levels)
147 update_locals(self, locals()) 150 update_locals(self, locals())
148 151
149 self.layers = [] 152 self.layers = []
150 self.pretrain_functions = [] 153 self.pretrain_functions = []
151 self.params = [] 154 self.params = []
237 # symbolic variable that points to the number of errors made on the 240 # symbolic variable that points to the number of errors made on the
238 # minibatch given by self.x and self.y 241 # minibatch given by self.x and self.y
239 242
240 self.errors = self.logLayer.errors(self.y) 243 self.errors = self.logLayer.errors(self.y)
241 244
242 @classmethod
243 def copy_reusing_lower_layers(cls, obj, num_hidden_layers, new_finetuning_lr=None):
244 assert(num_hidden_layers <= obj.n_layers)
245
246 if not new_finetuning_lr:
247 new_finetuning_lr = obj.finetune_lr
248
249 new_sda = cls(train_set_x= obj.train_set_x, \
250 train_set_y = obj.train_set_y,\
251 batch_size = obj.batch_size, \
252 n_ins= obj.n_ins, \
253 hidden_layers_sizes = obj.hidden_layers_sizes[:num_hidden_layers], \
254 n_outs = obj.n_outs, \
255 corruption_levels = obj.corruption_levels[:num_hidden_layers],\
256 rng = obj.rng,\
257 pretrain_lr = obj.pretrain_lr, \
258 finetune_lr = new_finetuning_lr, \
259 input_divider = obj.input_divider )
260
261 # new_sda.layers contains only the hidden layers actually
262 for i, layer in enumerate(new_sda.layers):
263 original_layer = obj.layers[i]
264 for p1,p2 in zip(layer.params, original_layer.params):
265 p1.value = p2.value.copy()
266
267 return new_sda
268
269 def get_params_copy(self):
270 return copy.deepcopy(self.params)
271
272 def set_params_from_copy(self, copy):
273 # We don't want to replace the var, as the functions have pointers in there
274 # We only want to replace values.
275 for i, p in enumerate(self.params):
276 p.value = copy[i].value
277
278 def get_params_means(self):
279 s = []
280 for p in self.params:
281 s.append(numpy.mean(p.value))
282 return s
283
284 if __name__ == '__main__': 245 if __name__ == '__main__':
285 import sys 246 import sys
286 args = sys.argv[1:] 247 args = sys.argv[1:]
287 248