Mercurial > ift6266

comparison deep/stacked_dae/v_sylvain/nist_sda_retrieve.py @ 306:a78dbbc61f37

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Meilleure souplesse d'execution, un parametre hard-coade est maintenant plus propre
author SylvainPL <sylvain.pannetier.lebeuf@umontreal.ca>
date Wed, 31 Mar 2010 21:02:27 -0400
parents f9b93ae45723
children a76bae0f2388
comparison
equal deleted inserted replaced
305:fe5d428c2acc 306:a78dbbc61f37
82 ## "You have to correct the code (and be patient, P07 is huge !!)\n"+ 82 ## "You have to correct the code (and be patient, P07 is huge !!)\n"+
83 ## "or reduce the number of pretraining epoch to run the code (better idea).\n") 83 ## "or reduce the number of pretraining epoch to run the code (better idea).\n")
84 ## print('\n\tpretraining with P07') 84 ## print('\n\tpretraining with P07')
85 ## optimizer.pretrain(datasets.nist_P07(min_file=0,max_file=nb_file)) 85 ## optimizer.pretrain(datasets.nist_P07(min_file=0,max_file=nb_file))
86 print ('Retrieve pre-train done earlier') 86 print ('Retrieve pre-train done earlier')
87
88 if state['pretrain_choice'] == 0:
89 PATH=PATH_NIST
90 elif state['pretrain_choice'] == 1:
91 PATH=PATH_P07
87 92
88 sys.stdout.flush() 93 sys.stdout.flush()
94 channel.save()
89 95
90 #Set some of the parameters used for the finetuning 96 #Set some of the parameters used for the finetuning
91 if state.has_key('finetune_set'): 97 if state.has_key('finetune_set'):
92 finetune_choice=state['finetune_set'] 98 finetune_choice=state['finetune_set']
93 else: 99 else:
105 111
106 #Decide how the finetune is done 112 #Decide how the finetune is done
107 113
108 if finetune_choice == 0: 114 if finetune_choice == 0:
109 print('\n\n\tfinetune with NIST\n\n') 115 print('\n\n\tfinetune with NIST\n\n')
110 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 116 optimizer.reload_parameters(PATH+'params_pretrain.txt')
111 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1) 117 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1)
112 channel.save() 118 channel.save()
113 if finetune_choice == 1: 119 if finetune_choice == 1:
114 print('\n\n\tfinetune with P07\n\n') 120 print('\n\n\tfinetune with P07\n\n')
115 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 121 optimizer.reload_parameters(PATH+'params_pretrain.txt')
116 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=0) 122 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=0)
117 channel.save() 123 channel.save()
118 if finetune_choice == 2: 124 if finetune_choice == 2:
119 print('\n\n\tfinetune with NIST followed by P07\n\n') 125 print('\n\n\tfinetune with NIST followed by P07\n\n')
120 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 126 optimizer.reload_parameters(PATH+'params_pretrain.txt')
121 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=21) 127 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=21)
122 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=20) 128 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=20)
123 channel.save() 129 channel.save()
124 if finetune_choice == 3: 130 if finetune_choice == 3:
125 print('\n\n\tfinetune with NIST only on the logistic regression on top (but validation on P07).\n\ 131 print('\n\n\tfinetune with NIST only on the logistic regression on top (but validation on P07).\n\
126 All hidden units output are input of the logistic regression\n\n') 132 All hidden units output are input of the logistic regression\n\n')
127 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 133 optimizer.reload_parameters(PATH+'params_pretrain.txt')
128 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1,special=1) 134 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1,special=1)
129 135
130 136
131 if finetune_choice==-1: 137 if finetune_choice==-1:
132 print('\nSERIE OF 3 DIFFERENT FINETUNINGS') 138 print('\nSERIE OF 3 DIFFERENT FINETUNINGS')
133 print('\n\n\tfinetune with NIST\n\n') 139 print('\n\n\tfinetune with NIST\n\n')
134 sys.stdout.flush() 140 sys.stdout.flush()
135 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 141 optimizer.reload_parameters(PATH+'params_pretrain.txt')
136 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1) 142 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1)
137 channel.save() 143 channel.save()
138 print('\n\n\tfinetune with P07\n\n') 144 print('\n\n\tfinetune with P07\n\n')
139 sys.stdout.flush() 145 sys.stdout.flush()
140 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 146 optimizer.reload_parameters(PATH+'params_pretrain.txt')
141 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=0) 147 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=0)
142 channel.save() 148 channel.save()
143 print('\n\n\tfinetune with NIST (done earlier) followed by P07 (written here)\n\n') 149 print('\n\n\tfinetune with NIST (done earlier) followed by P07 (written here)\n\n')
144 sys.stdout.flush() 150 sys.stdout.flush()
145 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_finetune_NIST.txt') 151 optimizer.reload_parameters('params_finetune_NIST.txt')
146 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=20) 152 optimizer.finetune(datasets.nist_P07(min_file=nb_file),datasets.nist_all(),max_finetune_epoch_P07,ind_test=20)
147 channel.save() 153 channel.save()
148 print('\n\n\tfinetune with NIST only on the logistic regression on top.\n\ 154 print('\n\n\tfinetune with NIST only on the logistic regression on top.\n\
149 All hidden units output are input of the logistic regression\n\n') 155 All hidden units output are input of the logistic regression\n\n')
150 sys.stdout.flush() 156 sys.stdout.flush()
151 optimizer.reload_parameters('/u/pannetis/IFT6266/ift6266/deep/stacked_dae/v_sylvain/ift6266h10_db/pannetis_finetuningSDA/1/params_pretrain.txt') 157 optimizer.reload_parameters(PATH+'params_pretrain.txt')
152 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1,special=1) 158 optimizer.finetune(datasets.nist_all(),datasets.nist_P07(min_file=nb_file),max_finetune_epoch_NIST,ind_test=1,special=1)
153 channel.save() 159 channel.save()
154 160
155 channel.save() 161 channel.save()
156 162