mlp_cv_set2.log

load norm file ......
norm file load
parameters input:
fea_file:             ../dataset/fea_cv.pfile
norm_file:            ../dataset/train.norm
targ_file:            ../dataset/lab_cv.pfile
outwts_file:          ./mlp_cv.0.wts
log_file:             ./mlp_cv.0.log
initwts_file:         ./mlp.0.wts
train_sent_range:     0-1574
cv_sent_range:        0-0
fea_dim:              39
fea_context:          11
bunchsize:            1024
train_cache:          102400
init_randem_seed:     27863875
targ_offset:          5
init_randem_weight_max:           0.100000
init_randem_weight_min:           -0.100000
init_randem_bias_max:             0.100000
init_randem_bias_min:             -0.100000
learnrate:                        0.768000
layersizes:                       429,2048,2048,2048,2048,2048,2048,8991,
nmod:                             1
remainder:                        0
discard_prob:                     0.000000
discardLabs:                      
Please check...
Get pfile info over: Training data has 606451 frames, 1575 sentences.
Get chunk info over: Training sentences have 6 chunks, 590651 samples.
Get cv chunk info over: CV sentences have 1 chunks, 105 samples.
starting cross validing:
cv over
total samples:589824 
correct samples: 118650
accuracy: 20.116169
total time cost: 348344.531250ms