def dump_vectors(qreq_):
"""
Example:
>>> # DISABLE_DOCTEST
>>> from wbia.algo.smk.smk_pipeline import * # NOQA
>>> import wbia
>>> ibs, aid_list = wbia.testdata_aids(defaultdb='PZ_MTEST', a='default:mingt=2,pername=2')
>>> qaids = aid_list[0:2]
>>> daids = aid_list[:]
>>> config = {'nAssign': 1, 'num_words': 8000,
>>> 'sv_on': True}
>>> qreq_ = SMKRequest(ibs, qaids, daids, config)
>>> qreq_.ensure_data()
"""
inva = qreq_.dinva
X = qreq_.dinva.get_annot(qreq_.daids[0])
n_words = inva.wx_list[-1] + 1
n_dims = X.agg_rvecs.shape[1]
n_annots = len(qreq_.daids)
X.agg_rvecs.dtype
vlads = np.zeros((n_annots, n_words, n_dims), dtype=np.float32)
ids_ = list(zip(qreq_.dnids, qreq_.daids))
for count, (nid, aid) in enumerate(ut.ProgIter(ids_, label='vlad')):
# X.rrr()
X = qreq_.dinva.get_annot(aid)
out = vlads[count]
out[X.wx_list] = X.agg_rvecs
# X.to_dense(out=out)
# Flatten out
vlads.shape = (n_annots, n_words * n_dims)
ut.print_object_size(vlads)
fname = 'vlad_%d_d%d_%s' % (n_annots, n_words * n_dims,
qreq_.ibs.get_dbname())
fpath = ut.truepath('~/' + fname + '.mat')
import scipy.io
mdict = {
'vlads': vlads,
'nids': qreq_.dnids,
}
scipy.io.savemat(fpath, mdict)
def hyrule_vocab_test():
from yael.yutils import load_ext
from os.path import join
import sklearn.cluster
dbdir = ut.truepath('/raid/work/Oxford/')
datadir = dbdir + '/smk_data_iccv_2013/data/'
# Files storing descriptors/geometry for Oxford5k dataset
test_sift_fname = join(datadir, 'oxford_sift.uint8')
# test_nf_fname = join(datadir, 'oxford_nsift.uint32')
all_vecs = load_ext(test_sift_fname, ndims=128,
verbose=True).astype(np.float32)
logger.info(ut.print_object_size(all_vecs))
# nfeats_list = load_ext(test_nf_fname, verbose=True)
with ut.embed_on_exception_context:
rng = np.random.RandomState(13421421)
# init_size = int(config['num_words'] * 8)
num_words = int(2**16)
init_size = num_words * 4
# converged after 26043 iterations
minibatch_params = dict(
n_clusters=num_words,
init='k-means++',
# init='random',
init_size=init_size,
n_init=1,
max_iter=100,
batch_size=1000,
tol=0.0,
max_no_improvement=10,
reassignment_ratio=0.01,
)
clusterer = sklearn.cluster.MiniBatchKMeans(compute_labels=False,
random_state=rng,
verbose=1,
**minibatch_params)
clusterer.fit(all_vecs)
words = clusterer.cluster_centers_
logger.info(words.shape)
print('TEST2')
print_ids()
import imp
# Causes utool.write_to to be bad
# because utool doesnt do a rrr itself
imp.reload(utool.util_io)
print_ids()
def docstr_test1():
print_ids()
utool.rrr()
print_ids()
def docstr_test2():
print_ids()
utool.rrr()
print_ids()
if __name__ == '__main__':
utool.print_object_size(utool)
reloading_test1()
#reloading_test2()
reloading_test1()
reloading_test1()
#reloading_test2()
utool.print_object_size(utool)
pass
print('TEST2')
print_ids()
import imp
# Causes utool.write_to to be bad
# because utool doesnt do a rrr itself
imp.reload(utool.util_io)
print_ids()
def docstr_test1():
print_ids()
utool.rrr()
print_ids()
def docstr_test2():
print_ids()
utool.rrr()
print_ids()
if __name__ == '__main__':
utool.print_object_size(utool)
reloading_test1()
#reloading_test2()
reloading_test1()
reloading_test1()
#reloading_test2()
utool.print_object_size(utool)
pass