Mercurial > ift6266

comparison deep/amt/amt.py @ 402:83413ac10913

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Added more stats printing. Now you dont need to parameters which dataset you are testing, it will detect it automatically
author humel
date Wed, 28 Apr 2010 11:28:28 -0400
parents 86d5e583e278
children a11692910312
comparison
equal deleted inserted replaced
401:86d5e583e278 402:83413ac10913
1 # Script usage : python amt.py filname.cvs 1 # Script usage : python amt.py filname.cvs
2 2
3 import csv,numpy,re 3 import csv,numpy,re,decimal
4 4
5 PATH = 'nist/' 5 DATASET_PATH = { 'nist' : '/data/lisa/data/ift6266h10/amt_data/nist/',
6 'p07' : '/data/lisa/data/ift6266h10/amt_data/p07/',
7 'pnist' : '/data/lisa/data/ift6266h10/amt_data/pnist/' }
8
6 CVSFILE = None 9 CVSFILE = None
7 10 #PATH = None
8 answer_labels = [ 'Answer.C'+str(i+1) for i in range(10) ] 11 answer_labels = [ 'answer.c'+str(i+1) for i in range(10) ]
9 img_url = 'Input.image_url' 12 img_url = 'input.image_url'
10 turks_per_batch = 3 13 turks_per_batch = 3
11 14 image_per_batch = 10
12 15
13 def setup_association(): 16 def setup_association():
14 answer_assoc = {} 17 answer_assoc = {}
15 for i in range(0,10): 18 for i in range(0,10):
16 answer_assoc[str(i)]=i 19 answer_assoc[str(i)]=i
35 38
36 errors = numpy.zeros((len(entries),)) 39 errors = numpy.zeros((len(entries),))
37 error_means = numpy.zeros((total_uniq_entries,)) 40 error_means = numpy.zeros((total_uniq_entries,))
38 error_variances = numpy.zeros((total_uniq_entries,)) 41 error_variances = numpy.zeros((total_uniq_entries,))
39 42
40 43 PATH = DATASET_PATH[find_dataset(entries[0])]
41 for i in range(total_uniq_entries): 44 for i in range(total_uniq_entries):
42 for t in range(turks_per_batch): 45 for t in range(turks_per_batch):
43 errors[i*turks_per_batch+t] = get_error(entries[i*turks_per_batch+t]) 46 errors[i*turks_per_batch+t] = get_error(entries[i*turks_per_batch+t],PATH)
44 error_means[i] = errors[i*turks_per_batch:(i+1)*turks_per_batch].mean() 47 error_means[i] = errors[i*turks_per_batch:(i+1)*turks_per_batch].mean()
45 error_variances[i] = errors[i*turks_per_batch:(i+1)*turks_per_batch].var() 48 error_variances[i] = errors[i*turks_per_batch:(i+1)*turks_per_batch].var()
46 49
47 print 'Average test error: ' + str(error_means.mean()*10) +'%' 50 decimal.getcontext().prec = 3
48 print 'Error variance : ' + str(error_variances.mean()*10) +'%' 51 print 'Total entries : ' + str(len(entries))
52 print 'Turks per batch : ' + str(turks_per_batch)
53 print 'Average test error : ' + str(error_means.mean()*image_per_batch) +'%'
54 print 'Error variance : ' + str(error_variances.mean()*image_per_batch) +'%'
49 55
50 56
51 def get_error(entry): 57 def find_dataset(entry):
58 file = parse_filename(entry[img_url])
59 return file.split('_')[0]
60
61 def get_error(entry,PATH):
52 file = parse_filename(entry[img_url]) 62 file = parse_filename(entry[img_url])
53 f = open(PATH+file,'r') 63 f = open(PATH+file,'r')
54 labels = re.sub("\s+", "",f.readline()).strip()[1:-2].split('.') 64 labels = re.sub("\s+", "",f.readline()).strip()[1:-2].split('.')
55 f.close() 65 f.close()
56 test_error = 0 66 test_error = 0