timeit_run_statement = """
length = min(E_test.shape[1],template_test.shape[1])
sum(tt.score_template_background_section(template_test[:,:length],bg_test,E_test[:,:length]))"""
num_features = 384
length = 20
E_test = (np.random.rand(length*num_features).reshape(num_features,length) > .6).astype(np.uint8)
bg_test = (np.random.rand(num_features)*.3 + .1).astype(np.float32)
template_test = (np.random.rand(num_features*length).reshape(num_features,length)*.9 + .05).astype(np.float32)
out1 = sum(tt.score_template_background_section(template_test,bg_test,E_test))
log_template = np.log(template_test.T)
log_invtemplate = np.log(1-template_test.T)
E_test_transpose = E_test.T
out2 = tt.score_template_background_section_quantizer(log_template,
log_invtemplate,
bg_test,
E_test_transpose)
import timeit
t = timeit.Timer(timeit_run_statement,timeit_import_statement)
t.repeat(5,10000)
timeit_import_statement2 = """
import numpy as np
root_path = '/home/mark/Template-Speech-Recognition/'
import sys, os, cPickle
sys.path.append(root_path)
import template_speech_rec.test_template as tt
def apply_bms_to_examples(phn_examples,phn_lengths,phn_bgs,bms):
return np.array(tuple(max(tt.score_template_background_section_quantizer(
bms.log_templates[i], bms.log_invtemplates[i],phn_bgs[j],
phn_examples[j][:phn_lengths[j]]) for i in xrange(len(bms.weights)))\
for j in xrange(phn_examples.shape[0])))
20,
1,
threshold=.7,
edge_orientations = edge_orientations,
edge_feature_row_breaks = edge_feature_row_breaks)
E = E.T
bg = np.clip(np.load(tmp_data_path+'avg_E_bgd.npy'),.01,.4).astype(np.float32)
k = ('t','aa')
list_of_syllable_models= alexey_nt_list_dict[k]
syllable_model = list_of_syllable_models[0]
log_template, log_invtemplate = zip(syllable_model.log_templates,syllable_model.log_invtemplates)[0]
tuple( tt.score_template_background_section_quantizer(log_template,
log_invtemplate,
bg,
E[t:t+len(log_template)],return_both_summed=True)
for t in xrange(len(E)-len(log_template)))
alexey_output = dict(
# compute the score over the different syllables
(k,tuple(
# compute the score for each mixture number
tuple(
# Compute the score for each time period
tuple( tt.score_template_background_section_quantizer(log_template,
log_invtemplate,
bg,
E[t:t+len(log_template)],return_both_summed=True)
for t in xrange(len(E)-len(log_template)))
for log_template, log_invtemplate in zip(syllable_model.log_templates,syllable_model.log_invtemplates)
