def pycloud_unpickle(file_name):
# type: (Text) -> Any
"""Unpickle an object from file using cloudpickle."""
from future.utils import PY2
import cloudpickle
with io.open(file_name, 'rb') as f: # pragma: no test
if PY2:
return cloudpickle.load(f)
else:
return cloudpickle.load(f, encoding="latin-1")
def load(cls, model_dir, ngram_featurizer):
# type: (Text, Text) -> NGramFeaturizer
import cloudpickle
if model_dir and ngram_featurizer:
classifier_file = os.path.join(model_dir, ngram_featurizer)
with io.open(classifier_file, 'rb') as f: # pramga: no cover
if PY3:
return cloudpickle.load(f, encoding="latin-1")
else:
return cloudpickle.load(f)
else:
return NGramFeaturizer()
def load(cls, model_dir=None, model_metadata=None, cached_component=None, **kwargs):
# type: (Text, Metadata, Optional[Component], **Any) -> SklearnIntentClassifier
import cloudpickle
if model_dir and model_metadata.get("intent_classifier_sklearn"):
classifier_file = os.path.join(model_dir, model_metadata.get("intent_classifier_sklearn"))
with io.open(classifier_file, 'rb') as f: # pragma: no test
if PY3:
return cloudpickle.load(f, encoding="latin-1")
else:
return cloudpickle.load(f)
else:
return SklearnIntentClassifier()
def load(cls, model_dir, intent_classifier_sklearn):
# type: (Text, Text) -> SklearnIntentClassifier
import cloudpickle
if model_dir and intent_classifier_sklearn:
classifier_file = os.path.join(model_dir, intent_classifier_sklearn)
with io.open(classifier_file, 'rb') as f: # pragma: no test
if PY3:
return cloudpickle.load(f, encoding="latin-1")
else:
return cloudpickle.load(f)
else:
return SklearnIntentClassifier()
def load(cls, model_dir=None, model_metadata=None, cached_component=None, **kwargs):
# type: (Text, Metadata, Optional[Component], **Any) -> NGramFeaturizer
import cloudpickle
if model_dir and model_metadata.get("ngram_featurizer"):
classifier_file = os.path.join(model_dir, model_metadata.get("ngram_featurizer"))
with io.open(classifier_file, 'rb') as f: # pramga: no cover
if PY3:
return cloudpickle.load(f, encoding="latin-1")
else:
return cloudpickle.load(f)
else:
return NGramFeaturizer()
def postprocess(input, output):
vocab = pickle.load(open('vocab.pkl', 'rb'))
entity_vocab = pickle.load(open('entity_vocab.pkl', 'rb'))
entity_inv = {v:k for k,v in entity_vocab.items()}
vocab_inv = {v:k for k,v in vocab.items()}
# this has the shape (timesteps, batches, data), so swap axes to (batches, timesteps, data)
character_probs = numpy.swapaxes(output, 0, 1)
# now extract the guessed entities
predictions = numpy.argmax(output, axis=2)
# find contiguous entity characters across timesteps
non_entity_label = entity_vocab.get('O')
entities = []
for i, query in enumerate(predictions):
previous_label = non_entity_label
entity_string = ""
used_indices = set()
for j, label in enumerate(query):
# find entity start point (expand to space character) and extract the continuous entity
if label != non_entity_label and label != previous_label and j not in used_indices:
entity_start = j
while vocab_inv.get(
numpy.argmax(input[i, entity_start])) not in string.whitespace and entity_start >= 0:
entity_start -= 1
# move start point forward one to get out of whitespace or back to 0 index
entity_start += 1
# now from the start point, extract continuous until whitespace or punctuation
entity_idx = entity_start
while entity_idx < len(query) and \
(
query[entity_idx] == label or
entity_idx == entity_start or
(
entity_idx > entity_start and
vocab_inv.get(numpy.argmax(input[i, entity_idx])) not in string.whitespace + string.punctuation and
vocab_inv.get(numpy.argmax(input[i, entity_idx - 1])) not in string.whitespace + string.punctuation
)
):
entity_string += vocab_inv.get(numpy.argmax(input[i, entity_idx]))
used_indices.add(entity_idx)
entity_idx += 1
# get rid of trailing matched punctuation
if entity_string[-1] in string.punctuation:
entity_string = entity_string[:-1]
# add the entity stripped of whitespace in beginning and end, and reset the string
entities.append((entity_string.strip(), entity_inv.get(label)))
entity_string = ""
previous_label = label
return entities
def _calc_Tx(self, name, x, lambdify=True):
""" Uses Sympy to transform x from the reference frame of a joint
or link to the origin (world) coordinates.
name string: name of the joint or link, or end-effector
x list: the [x,y,z] position of interest in "name"'s reference frame
lambdify boolean: if True returns a function to calculate
the transform. If False returns the Sympy
matrix
"""
# check to see if we have our transformation saved in file
if (os.path.isfile('%s/%s.T' % (self.config_folder, name))):
Tx = cloudpickle.load(open('%s/%s.T' %
(self.config_folder, name), 'rb'))
else:
T = self._calc_T(name=name)
# transform x into world coordinates
Tx = T * sp.Matrix(self.x + [1])
# save to file
cloudpickle.dump(Tx, open('%s/%s.T' %
(self.config_folder, name), 'wb'))
if lambdify is False:
return Tx
return sp.lambdify(self.q + self.x, Tx)
def _calc_T_inv(self, name, x, lambdify=True):
""" Return the inverse transform matrix, which converts from
world coordinates into the robot's end-effector reference frame
name string: name of the joint or link, or end-effector
x list: the [x,y,z] position of interest in "name"'s reference frame
lambdify boolean: if True returns a function to calculate
the transform. If False returns the Sympy
matrix
"""
# check to see if we have our transformation saved in file
if (os.path.isfile('%s/%s.T_inv' % (self.config_folder,
name))):
T_inv = cloudpickle.load(open('%s/%s.T_inv' %
(self.config_folder, name), 'rb'))
else:
T = self._calc_T(name=name)
rotation_inv = T[:3, :3].T
translation_inv = -rotation_inv * T[:3, 3]
T_inv = rotation_inv.row_join(translation_inv).col_join(
sp.Matrix([[0, 0, 0, 1]]))
# save to file
cloudpickle.dump(T_inv, open('%s/%s.T_inv' %
(self.config_folder, name), 'wb'))
if lambdify is False:
return T_inv
return sp.lambdify(self.q + self.x, T_inv)
def create_model(init_config_file=None, vocab={}, label_vocab={}):
# load from a configuration file, or define the model configuration
if init_config_file is not None:
with open(init_config_file, 'rb') as f:
init_config = pickle.load(f)
else:
init_config = {
'input_size': len(vocab),
'hidden_size': 128,
'output_size': len(label_vocab),
'hidden_activation': 'tanh',
'inner_hidden_activation': 'sigmoid',
'activation': 'softmax',
'weights_init': 'uniform',
'weights_interval': 'montreal',
'r_weights_init': 'orthogonal',
'clip_recurrent_grads': 5.,
'noise': 'dropout',
'noise_level': 0.5,
'direction': 'bidirectional',
'cost_function': 'nll',
'cost_args': {'one_hot': True}
}
# instantiate the model!
lstm = LSTM(**init_config)
return lstm
def load(cls, directory: str):
with open(os.path.join(directory, 'rnn.pkl'), 'rb') as f:
params = cloudpickle.load(f)
guesser = RnnGuesser(params['config_num'])
guesser.page_field = params['page_field']
guesser.qanta_id_field = params['qanta_id_field']
guesser.text_field = params['text_field']
guesser.n_classes = params['n_classes']
guesser.gradient_clip = params['gradient_clip']
guesser.n_hidden_units = params['n_hidden_units']
guesser.n_hidden_layers = params['n_hidden_layers']
guesser.nn_dropout = params['nn_dropout']
guesser.use_wiki = params['use_wiki']
guesser.n_wiki_sentences = params['n_wiki_sentences']
guesser.wiki_title_replace_token = params['wiki_title_replace_token']
guesser.lowercase = params['lowercase']
guesser.random_seed = params['random_seed']
guesser.model = RnnModel(
guesser.n_classes,
text_field=guesser.text_field,
init_embeddings=False, emb_dim=300,
n_hidden_layers=guesser.n_hidden_layers,
n_hidden_units=guesser.n_hidden_units
)
guesser.model.load_state_dict(torch.load(
os.path.join(directory, 'rnn.pt'), map_location=lambda storage, loc: storage
))
guesser.model.eval()
if CUDA:
guesser.model = guesser.model.cuda()
return guesser
def test_load_namespace(self):
obj = 1, 2, 3, 4
bio = BytesIO()
cloudpickle.dump(obj, bio)
bio.seek(0)
returned_obj = cloudpickle.load(bio)
self.assertEqual(obj, returned_obj)
def _calc_Mq_g(self, lambdify=True):
""" Uses Sympy to generate the force of gravity in
joint space for the ur5
lambdify boolean: if True returns a function to calculate
the Jacobian. If False returns the Sympy
matrix
"""
# check to see if we have our gravity term saved in file
if os.path.isfile('%s/Mq_g' % self.config_folder):
Mq_g = cloudpickle.load(open('%s/Mq_g' % self.config_folder,
'rb'))
else:
# get the Jacobians for each link's COM
J = [self._calc_J('link%s' % ii, x=[0, 0, 0], lambdify=False)
for ii in range(self.num_links)]
# transform each inertia matrix into joint space and
# sum together the effects of arm segments' inertia on each motor
Mq_g = sp.zeros(self.num_joints, 1)
for ii in range(self.num_joints):
Mq_g += J[ii].T * self._M[ii] * self.gravity
Mq_g = sp.Matrix(Mq_g)
# save to file
cloudpickle.dump(Mq_g, open('%s/Mq_g' % self.config_folder,
'wb'))
if lambdify is False:
return Mq_g
return sp.lambdify(self.q + self.x, Mq_g)
def read(self, filepath): with open(filepath, 'rb') as fo: match = cloudpickle.load(fo) self.for_worker = match.for_worker self.gamename = match.gamename self.map_file = match.map_file self.uuid = match.uuid self.bots = match.bots
def get_state(self):
if self.latest_only:
try:
with open(self.get_snapshot_path(0), "rb") as f:
return cloudpickle.load(f)
except EOFError:
pass
else:
snapshot_files = os.listdir(self.snapshots_folder)
snapshot_files = sorted(
snapshot_files, key=lambda x: int(x.split(".")[0]))[::-1]
for file in snapshot_files:
file_path = os.path.join(self.snapshots_folder, file)
try:
with open(file_path, "rb") as f:
return cloudpickle.load(f)
except EOFError:
pass
def test_cloudpickle_to_file(EN):
f = tempfile.NamedTemporaryFile(delete=False)
p = cloudpickle.CloudPickler(f)
p.dump(EN)
f.close()
loaded_en = cloudpickle.load(open(f.name))
os.unlink(f.name)
doc = loaded_en(unicode('test parse'))
assert len(doc) == 2
def classifier(test_data):
"""
:param test_data: data that needs prediction list or single parameter
:return: predicted class rotten or fresh
"""
# Load the pickle data
model = pickle.load(open('movie_model.pkl','rb'))
vectorizer = pickle.load(open('vectorizer.pkl','rb'))
# Check for the type
if type(test_data) != list:
test_data = list(test_data)
# Transform the test data
transformed = vectorizer.transform(test_data).toarray()
# Predict the class
predicted = model.predict(transformed).tolist()
return predicted
def load(self, PATH):
"""
load trained model froom given path
:param PATH:
:return:
"""
try:
with open(PATH, 'rb') as f:
self.model = cloudpickle.load(f)
except IOError:
return False
def wrapper(args):
kwargs = {k: v for k, v in zip(search.keys(), args)}
result = fun(**kwargs)
try:
with open('/tmp/results.pkl', 'r') as f: data = pickle.load(f)
except (IOError, EOFError):
data = {"kwargs": [], "results": []}
with open('/tmp/results.pkl', 'w') as f:
data["kwargs"].append(kwargs)
data["results"].append(result)
pickle.dump(data, f)
return result
def _load_cache_from_file(self, file_name):
try:
with open(file_name, 'rb') as io:
deterministic_cache = pickle.load(io)
self._deterministic_cache = deterministic_cache.cache
self._cache_valid_for_turns = deterministic_cache.turns
self._logger.debug(
'Loaded cache with %d entries' % len(self._deterministic_cache))
return True
except IOError:
self._logger.debug('Cache file not found. Starting with empty cache')
return False
def load(self, models_dir):
try:
del self.model
tf.keras.backend.clear_session()
self.model = tf.keras.models.load_model(os.path.join(models_dir, "tf_intent_model.hd5"),compile=True)
self.graph = tf.get_default_graph()
print("Tf model loaded")
with open(os.path.join(models_dir, "labels.pkl"), 'rb') as f:
self.label_encoder = cloudpickle.load(f)
print("Labels model loaded")
except IOError:
return False
def read(self, filepath): try: with open(filepath, 'rb') as fo: match = cloudpickle.load(fo) self.for_worker = match.for_worker self.gamename = match.gamename self.map_file = match.map_file self.uuid = match.uuid self.bots = match.bots return True except FileNotFoundError as e: # FIXME logging return False
def load(path):
with open(path, "rb") as f:
model_data, act_params = cloudpickle.load(f)
act = deepq.build_act(**act_params)
sess = tf.Session()
sess.__enter__()
with tempfile.TemporaryDirectory() as td:
arc_path = os.path.join(td, "packed.zip")
with open(arc_path, "wb") as f:
f.write(model_data)
zipfile.ZipFile(arc_path, 'r', zipfile.ZIP_DEFLATED).extractall(td)
load_state(os.path.join(td, "model"))
return ActWrapper(act, act_params)
def string_to_data(query):
vocab = pickle.load(open('vocab.pkl', 'rb'))
# process the raw input data string
data = []
# get the integer encodings
for data_char in query:
data.append(vocab.get(data_char, 0))
# convert the integers to one-hot arrays
data = numpy_one_hot(numpy.asarray(data), n_classes=numpy.amax(vocab.values()) + 1)
# make 3D for model input
seq, dim = data.shape
data = numpy.reshape(data, (1, seq, dim))
return data
def load(cls, directory: str):
with open(os.path.join(directory, 'elmo.pkl'), 'rb') as f:
params = cloudpickle.load(f)
guesser = ElmoGuesser(params['config_num'])
guesser.class_to_i = params['class_to_i']
guesser.i_to_class = params['i_to_class']
guesser.random_seed = params['random_seed']
guesser.dropout = params['dropout']
guesser.model = ElmoModel(len(guesser.i_to_class))
guesser.model.load_state_dict(torch.load(
os.path.join(directory, 'elmo.pt'), map_location=lambda storage, loc: storage
))
guesser.model.eval()
if CUDA:
guesser.model = guesser.model.cuda()
return guesser
def get_data(name, force=False, read=True):
""" remember that the stuff is here
d._rightmove_data__request_object = _GetDataFromUrl()
d._rightmove_data__url
And that weblinks go into df.url. That is what we need to recurse into.
"""
url = urls[name]
filename = get_hash_pickle_name(name, url)
if os.path.exists(filename) and not force:
print("found {}".format(filename))
else:
rightmove_object = rightmove_data(url)
pickle.dump(rightmove_object, open(filename, 'wb'))
if read:
print("reading {}".format(filename))
return pickle.load(open(filename, 'rb'))
def _calc_J(self, name, x, lambdify=True):
""" Uses Sympy to generate the Jacobian for a joint or link
name string: name of the joint or link, or end-effector
lambdify boolean: if True returns a function to calculate
the Jacobian. If False returns the Sympy
matrix
"""
# check to see if we have our Jacobian saved in file
if os.path.isfile('%s/%s.J' % (self.config_folder, name)):
J = cloudpickle.load(open('%s/%s.J' %
(self.config_folder, name), 'rb'))
else:
Tx = self._calc_Tx(name, x=x, lambdify=False)
J = []
# calculate derivative of (x,y,z) wrt to each joint
for ii in range(self.num_joints):
J.append([])
J[ii].append(Tx[0].diff(self.q[ii])) # dx/dq[ii]
J[ii].append(Tx[1].diff(self.q[ii])) # dy/dq[ii]
J[ii].append(Tx[2].diff(self.q[ii])) # dz/dq[ii]
end_point = name.strip('link').strip('joint')
if end_point != 'EE':
end_point = min(int(end_point) + 1, self.num_joints)
# add on the orientation information up to the last joint
for ii in range(end_point):
J[ii] = J[ii] + self.J_orientation[ii]
# fill in the rest of the joints orientation info with 0
for ii in range(end_point, self.num_joints):
J[ii] = J[ii] + [0, 0, 0]
# save to file
cloudpickle.dump(J, open('%s/%s.J' %
(self.config_folder, name), 'wb'))
J = sp.Matrix(J).T # correct the orientation of J
if lambdify is False:
return J
return sp.lambdify(self.q + self.x, J)
def load(self, PATH):
try:
with open(PATH, 'rb') as f:
self.model = cloudpickle.load(f)
except IOError:
return False
def main():
args = argparser.parse_args()
n_pts = int(args.points)
print 'Loading data'
with open(os.path.abspath(args.in_file), 'rb') as f:
data = cloudpickle.load(f)
print 'subtracting mean'
time_arr = data['time_arr'] - np.mean(data['time_arr'])
gyr_arr = data['gyr_arr'] - np.tile( np.mean(data['gyr_arr'], axis=1).reshape((3,1)), (1,data['gyr_arr'].shape[1]) )
acc_arr = data['acc_arr'] - np.tile( np.mean(data['acc_arr'], axis=1).reshape((3,1)), (1,data['acc_arr'].shape[1]) )
# M: number of axes
# N: number of epochs
if acc_arr.shape != gyr_arr.shape:
raise Exception('different sizes')
M, N = gyr_arr.shape
# automate this?
print 'Computing mean dt'
t0 = np.mean(np.diff(time_arr))
fs = np.float64(1.0)/t0
n = np.power(2, np.arange(np.floor(np.log2(N/2.))))
end_log_inc = np.log10(n[-1])
m = shared_from_array( np.unique(np.ceil(np.logspace(0, end_log_inc, n_pts))).astype(np.int64) )
T = m*t0
if (T < 0).any():
print 'T < 0'
set_trace()
# setup input/output shared memory arrays
theta_gyr = shared_from_array( np.cumsum(gyr_arr, axis=1) )
theta_acc = shared_from_array( np.cumsum(acc_arr, axis=1) )
sigma2_gyr = shared_from_array( np.zeros((M, len(m))) )
sigma2_acc = shared_from_array( np.zeros((M, len(m))) )
# shared memory/serialization workaround: define calculation functions here so
# that the shared memory arrays are in scope
def adev_at_tau(i):
"""worker function for parallelization. first part of the Allan deviation
equation.
There is potentially a way to do the Allan deviation calculation without any
for loop whatsoever, but I haven't figured it out yet. It would require 2D
array indexing in NumPy.
"""
k = range(N - 2*m[i])
sigma2_gyr[:,i] = np.sum( np.power( theta_gyr[:,k+2*m[i]] - 2*theta_gyr[:,k+m[i]] + theta_gyr[:,k] , 2 ), axis=1)
sigma2_acc[:,i] = np.sum( np.power( theta_acc[:,k+2*m[i]] - 2*theta_acc[:,k+m[i]] + theta_acc[:,k] , 2 ), axis=1)
def adev_at_tau_wrapper(idxs):
if idxs[0] == 0:
for i in trange(len(idxs)):
adev_at_tau(idxs[i])
else:
for i in idxs:
adev_at_tau(i)
print 'creating procs'
idx_chunks = chunk(range(len(m)), int(args.cores))
procs = [multiprocessing.Process(target=adev_at_tau_wrapper, args=(ichnk,)) for ichnk in idx_chunks]
print '# chunks: ', len(procs)
for proc in procs:
proc.start()
for proc in procs:
proc.join()
div = np.tile(2*np.multiply(np.power(T,2), N-2*m), (M,1))
sigma2_gyr = np.divide(sigma2_gyr, div)
sigma2_acc = np.divide(sigma2_acc, div)
sigma_gyr = np.sqrt(sigma2_gyr)
sigma_acc = np.sqrt(sigma2_acc)
data_dir, in_name = os.path.split(os.path.abspath(args.in_file))
set_name, ext = in_name.split(os.extsep)
out_file_name = os.path.join(data_dir, set_name+'_adev'+os.extsep+ext)
print 'saving to: ', out_file_name
with open(out_file_name, 'wb') as f:
cloudpickle.dump(
{
'T': T,
'sigma2_gyr': sigma2_gyr,
'sigma2_acc': sigma2_acc,
'sigma_gyr': sigma_gyr,
'sigma_acc': sigma_acc,
},
f, -1
)
def load_pickle(filepath):
filestream = open(filepath, "rb")
obj = cloudpickle.load(filestream)
return obj
def load_custom(file_path: str):
with open(file_path, "rb") as file:
return cloudpickle.load(file)
def arl_load(name: str):
"""Load an object from a pre-existing pickle
"""
with open(name, 'rb') as f:
return cloudpickle.load(f)
ncols=3,
sharex=True,
sharey=True,
figsize=(9, 9))
ax = axs.reshape(-1)
normalize = 0
fig.subplots_adjust(hspace=0)
fig.subplots_adjust(wspace=0)
for loop in range(0, 9):
try:
bin_centres, median, per_50, per_16, per_84, per_25, per_75 = np.loadtxt(
'./binned_data/SM_DTM_z' + str(loop) + '.txt',
unpack=True,
comments='#')
ax[loop] = cloudpickle.load(
open('./pkl_hists/SM_DTM_z' + str(loop) + '.pkl', 'rb'))
except IOError:
bin_centres, median, per_50, per_16, per_84, per_25, per_75, normalize = bin_data(
'SM', 'DTM', ax, normalize, loop, 'SM_DTM', nbins=30)
ax[loop].set_xlim([8., 11.98])
ax[loop].set_ylim([-3.98, 2])
plot_params(ax[loop], loop, 'SM', 'DTM')
plot_observations(ax[loop], loop, "DTM_SM")
#ax[loop].errorbar(x_bins,y_median,yerr=(y_mederr),color='k',label='L-Galaxies Median',linewidth=2)
ax[loop].plot(bin_centres,
per_50,
c='k',
def load_checkpoint(filename='checkpoint.pkl'):
with open(filename, 'rb') as fi:
return cloudpickle.load(fi)
def cloudpickle_load(filename, **kwargs):
with open(filename, 'rb') as f:
return cloudpickle.load(f, **kwargs)
def main(args):
use_cuda = torch.cuda.is_available() and True
device = torch.device("cuda" if use_cuda else "cpu")
""" Slice module. """
image = sitk.ReadImage(args.image_path)
if args.mask_path is not None:
mask = sitk.ReadImage(args.mask_path)
else:
mask = None
""" Dummy image """
label = sitk.Image(image.GetSize(), sitk.sitkInt8)
label.SetOrigin(image.GetOrigin())
label.SetDirection(image.GetDirection())
label.SetSpacing(image.GetSpacing())
""" Get the patch size from string."""
matchobj = re.match("([0-9]+)-([0-9]+)-([0-9]+)", args.image_patch_size)
if matchobj is None:
print("[ERROR] Invalid patch size : {}.".fotmat(args.image_patch_size))
sys.exit()
image_patch_size = [int(s) for s in matchobj.groups()]
""" Get the patch size from string."""
matchobj = re.match("([0-9]+)-([0-9]+)-([0-9]+)", args.label_patch_size)
if matchobj is None:
print("[ERROR] Invalid patch size : {}.".fotmat(args.label_patch_size))
sys.exit()
label_patch_size = [int(s) for s in matchobj.groups()]
""" Get the slide size from string."""
if args.slide is not None:
matchobj = re.match("([0-9]+)-([0-9]+)-([0-9]+)", args.slide)
if matchobj is None:
print("[ERROR] Invalid patch size : {}.".fotmat(args.slide))
sys.exit()
slide = [int(s) for s in matchobj.groups()]
else:
slide = None
extractor = extor(image=image,
label=label,
mask=mask,
image_patch_size=image_patch_size,
label_patch_size=label_patch_size,
slide=slide,
phase="segmentation")
extractor.execute()
image_array_list, mask_array_list = extractor.output("Array")
""" Load model. """
with open(args.modelweightfile, 'rb') as f:
model = cloudpickle.load(f)
model = torch.nn.DataParallel(model, device_ids=args.gpuid)
model.eval()
""" Segmentation module. """
segmented_array_list = []
for image_array, mask_array in tqdm(zip(image_array_list, mask_array_list),
desc="Segmenting images...",
ncols=60):
if args.mask_path is not None and (mask_array == 0).all():
segmented_array_list.append(mask_array)
continue
image_array = image_array.transpose(2, 0, 1)
image_array = torch.from_numpy(image_array[np.newaxis, np.newaxis,
...]).to(device,
dtype=torch.float)
segmented_array = model(image_array)
segmented_array = segmented_array.to("cpu").detach().numpy().astype(
np.float)
segmented_array = np.squeeze(segmented_array)
segmented_array = np.argmax(segmented_array, axis=0).astype(np.uint8)
segmented_array = segmented_array.transpose(1, 2, 0)
segmented_array_list.append(segmented_array)
""" Restore module. """
segmented = extractor.restore(segmented_array_list)
createParentPath(args.save_path)
print("Saving image to {}".format(args.save_path))
sitk.WriteImage(segmented, args.save_path, True)
def load_remote(folder: str):
file_path = os.path.join(folder, DefaultRemoteFilename)
with open(file_path, "rb") as file:
return cloudpickle.load(file)
def __call__(self, state, reward):
if not self.rendered:
self.kinematic = cloudpickle.load(
open('./model/swing_kinematic.dll', 'rb'))
self.fig, self.ax = plt.subplots(ncols=1, figsize=(6, 6))
self.ax.set_xlim([-0.3, 0.3])
self.ax.set_ylim([-0.3, 0.15])
self.ax.set_xlabel('X')
self.ax.set_ylabel('Y')
# self.ax[0].set_xlim([-0.3, 0.3])
# self.ax[0].set_ylim([-0.3, 0.3])
# self.ax[0].set_xlabel('X')
# self.ax[0].set_ylabel('Y')
# self.ax[1].set_xlim([-2, 2])
# self.ax[1].set_ylim([-15, 15])
# self.ax[1].set_xlabel('$q_1$')
# self.ax[1].set_ylabel('$\dot{q_1}$')
self.line1, = self.ax.plot([], [],
lw=2,
color='k',
linestyle='-',
marker='o',
ms=10)
self.line2, = self.ax.plot([], [],
lw=4,
color='b',
linestyle='-',
marker='o',
ms=6)
# self.line1, = self.ax[0].plot([], [], lw=1, color='k', linestyle='-', marker='o', ms=5)
# self.line2, = self.ax[0].plot([], [], lw=2, color='b', linestyle='-', marker='o', ms=3)
# self.tau, = self.ax[1].plot([], [], color='b', marker='o', ms=3)
# self.traj = None
# self.time_text = self.ax.text(-0.2, 0.23, '')
# self.reward_text = self.ax.text(-0.2, 0.2, '')
# self.time_text = self.ax[0].text(-0.2, 0.23, '')
# self.reward_text = self.ax[0].text(-0.2, 0.2, '')
self.fig.canvas.draw()
self.rendered = True
if not self.savefig:
plt.show(block=False)
else:
p1, p2, p3, p21, p31 = kinematic(next_state, myParam)
# if self.traj is None:
# self.traj = np.expand_dims(next_state[[0, 3]], axis=0)
# else:
# self.traj = np.vstack([self.traj, next_state[[0, 3]]])
self.line1.set_data([0, p1[0]], [0, p1[1]])
self.line2.set_data([p21[0], p2[0], p3[0], p31[0]],
[p21[1], p2[1], p3[1], p31[1]])
# self.tau.set_data(self.traj[:, 0], self.traj[:, 1])
# self.time_text.set_text('t = %.2f'%(self.step/100.0))
# self.reward_text.set_text('r = %.2f'%reward)
# update canvas
self.ax.draw_artist(self.ax.patch)
#self.ax.draw_artist(self.ax[1].patch)
self.ax.draw_artist(self.line1)
self.ax.draw_artist(self.line2)
# self.ax.draw_artist(self.tau)
# self.ax.draw_artist(self.time_text)
# self.ax.draw_artist(self.reward_text)
# self.ax[0].draw_artist(self.ax[0].patch)
# self.ax[1].draw_artist(self.ax[1].patch)
# self.ax[0].draw_artist(self.line1)
# self.ax[0].draw_artist(self.line2)
# self.ax[1].draw_artist(self.tau)
# self.ax[0].draw_artist(self.time_text)
# self.ax[0].draw_artist(self.reward_text)
self.fig.canvas.flush_events()
self.fig.canvas.update()
plt.axis('off')
if self.savefig:
self.fig.savefig(self.dir_name + '/anim/{}.png'.format(self.step))
self.step += 1
return self.fig
def cli(onto_fp, dcell_fp, costanzo_2010_fp, cpkl_output_fmt, tsv_output):
onto_raw_preds_fp = onto_fp
df = pd.read_csv(onto_raw_preds_fp, sep='\t')
df.columns = ['A', 'B', 'onto_pred', 'true', 'p']
df = df[~df['true'].isna()].reset_index(drop=True)
onto_df = df.sort_values(by=['A', 'B'])
print(onto_df.head())
print('Ontotype prediction data shape:', onto_df.shape)
raw_dcell_preds = dcell_fp
dcell_df = pd.read_csv(raw_dcell_preds, sep='\t', header=None)
dcell_df.columns = ['A', 'B', 'dcell_pred']
dcell_df = dcell_df.sort_values(by=['A', 'B'])
print(dcell_df.head())
print('DCell prediction data shape:', dcell_df.shape)
onto_df = sort_cols(onto_df)
dcell_df = sort_cols(dcell_df)
keys = list(zip(onto_df['A'], onto_df['B']))
keys = [a + ':' + b for (a, b) in keys]
print('Ontotype preds have duplicate preds for gene pairs:',
not len(keys) == len(set(keys)))
keys = list(zip(dcell_df['A'], dcell_df['B']))
keys = [a + ':' + b for (a, b) in keys]
print('DCell preds have duplicate preds for gene pairs:',
not len(keys) == len(set(keys)))
print('Dropping duplicates in DCell preds')
dcell_df = dcell_df.drop_duplicates(['A', 'B'])
keys = list(zip(dcell_df['A'], dcell_df['B']))
keys = [a + ':' + b for (a, b) in keys]
print('DCell preds have duplicate preds for gene pairs:',
not len(keys) == len(set(keys)))
print(
'Performing inner join to merge DCell and Ontotype predictions... (validating that join is 1-1)'
)
merged_df = onto_df.merge(dcell_df,
how='inner',
left_on=['A', 'B'],
right_on=['A', 'B'],
validate='1:1')
print(merged_df.head())
print('** Ontotype: **')
report(merged_df, 'onto_pred')
print('** DCell: **')
report(merged_df, 'dcell_pred')
merged_df.to_csv(tsv_output, sep='\t', index=False)
value_cols = ['onto_pred', 'dcell_pred', 'true', 'p']
#TODO: we need to make this data match the shape of the data with our own processed Gs.
# That is, we need to figure whether a pair is
# A) query - array
# b) array - query (so we swap)
# c) or both array and query, then we add two values to the matrix.
# check for duplicates
# keys = list(zip(merged_df['A'], merged_df['B']))
# keys = [frozenset((a,b)) for (a,b) in keys]
# assert(len(keys) == len(set(keys)))
# print('Merged scores are indexed by unique keys:', len(keys) == len(set(keys)))
with open(costanzo_2010_fp, 'rb') as f:
costanzo10_data = cpkl.load(f)
costanzo10_rows = set(costanzo10_data['rows'])
costanzo10_cols = set(costanzo10_data['cols'])
print(len(costanzo10_rows), len(costanzo10_cols))
needs_duplicates = df['A'].isin(costanzo10_rows) & df['A'].isin(costanzo10_cols) & \
df['B'].isin(costanzo10_rows) & df['B'].isin(costanzo10_cols)
print(sum(needs_duplicates))
# Incorrectly oriented (i.e. A is not in rows, or B is not in cols)
needs_flipping = (~df['A'].isin(costanzo10_rows)) | (
~df['B'].isin(costanzo10_cols))
print(sum(needs_flipping))
# Correcly oriented and does not need duplicating
# keep = df['A'].isin(costanzo10_rows) & (~df['A'].isin(costanzo10_cols)) & \
# df['B'].isin(costanzo10_cols) & (~df['B'].isin(costanzo10_rows))
keep = ~(needs_flipping | needs_duplicates)
print(sum(keep))
dup_df = merged_df[needs_duplicates]
dup_df.rename(columns={'A': 'B', 'B': 'A'}, inplace=True)
flip_df = merged_df[needs_flipping]
flip_df.rename(columns={'A': 'B', 'B': 'A'}, inplace=True)
new_df = merged_df[needs_duplicates]
new_df = new_df.append(dup_df)
new_df = new_df.append(flip_df)
new_df = new_df.append(merged_df[keep])
print(new_df.head())
print('row genes', len(set(new_df['A'])))
print('col genes', len(set(new_df['B'])))
print(len(new_df))
new_df = new_df.pivot(index='A', columns='B')
for value_name in value_cols:
gi_mat = new_df[value_name]
print('\t-Extracting {} values'.format(value_name))
print('\t\t- GIs shape:', gi_mat.shape)
gi_mat = gi_mat.apply(pd.to_numeric)
print('\t\t- GIs shape:', gi_mat.shape)
print('\t\t- Sparsity:', sparsity(gi_mat.values))
print('\t\t- Interactions measured :',
np.sum(~np.isnan(gi_mat.values)))
print('\t\t- Interactions missing:', np.sum(np.isnan(gi_mat.values)))
gi_data = GIData(values=gi_mat.values.astype(float),
rows=gi_mat.index.values.astype(str),
cols=gi_mat.columns.values.astype(str),
check_symmetric=False)
fp = cpkl_output_fmt.format(value_name)
gi_data.save(fp)
print('\t\t- Saved values to:', fp)
def load(filename):
'''Load a coffea file from disk
'''
with lz4.frame.open(filename) as fin:
output = cloudpickle.load(fin)
return output
def main():
args = parse_args()
setup_logging(args.logfile)
log = get_logger()
assert (0 <= args.hidden_fraction <= 1)
np.random.seed(args.random_seed)
tf.set_random_seed(args.random_seed)
log.info('*' * 100)
log.info('[Starting MC experiment]')
log_dict(log.info, vars(args))
log.info('[Loading target GIs]')
with open(args.target_gis, 'rb') as f:
tgt_gis = cpkl.load(f)
log.info('[Loading source GIs]')
with open(args.source_gis, 'rb') as f:
src_gis = cpkl.load(f)
log.info('[Loading sim scores]')
with open(args.sim_scores, 'rb') as f:
sim_scores_data = cpkl.load(f)
sim_scores = sim_scores_data['values']
sim_scores = sim_scores / np.max(sim_scores) # Normalize
log.info('[Loading target PPI]')
L_tgt = get_laplacian(tgt_gis['rows'], args.target_ppi)
log.info('[Loading source PPI]')
L_src = get_laplacian(src_gis['rows'], args.source_ppi)
hp_param_space = kxsmf_param_space(args)
results, models, training_curves, trials = \
run_kxsmf_experiment(tgt_gis=tgt_gis,
src_gis=src_gis,
L_tgt=L_tgt,
L_src=L_src,
space=hp_param_space,
sim_scores=sim_scores,
val_hf=args.val_hidden_fraction,
test_hf=args.hidden_fraction,
n_repeats=args.n_repeats,
hp_iters=args.n_hyperopt_iters,
hp_seed=args.random_seed)
# Save results and other information
log_results(results['summary'])
with open(args.results_output, 'w') as f:
json.dump(results, f, indent=2)
with open(args.training_curve_output, 'wb') as f:
cpkl.dump(training_curves, f)
# TODO: save models the models cannot be pickled at the moment
# We will need to implement a from dict and a to dict method
with open(args.models_output, 'wb') as f:
cpkl.dump(trials, f)
with open(args.trials_output, 'wb') as f:
cpkl.dump(trials, f)
encoded_file = base64.b64encode(buf.read()).decode('ascii')
return (encoded_file)
# load data
engine = create_engine('sqlite:///../data/DisasterResponse.db')
df = pd.read_sql_table('Main', engine)
# load model
#random_state=1234
#np.random_state=random_state
#estimator=ExtraTreeClassifier(random_state=random_state,class_weight='balanced')
#model = build_model(estimator)
with open("../models/classifier.pkl", 'rb') as f:
model = cp.load(f)
nlp = spacy.load("en_vectors_web_lg")
seq_lens = np.unique(df.message.apply(lambda x: len(nlp(x))).values,
return_counts=True)
counts_df = pd.DataFrame(np.array(seq_lens).T)
# index webpage displays cool visuals and receives user input text for model
@app.route('/')
@app.route('/index')
def index():
ax = counts_df.plot.hist(x=1,
bins=50,
legend=False,
title="Sequence Length Distribution",
#!/usr/bin/python
import numpy as np
import cloudpickle as cp
sample = np.random.zipf(2, 10)
print(sample)
for i in range(10):
print(sample[i])
msg = b'127.0.0.1:anna:1000'
msg = cp.load(msg)
print(msg)
def load_bytes(fname):
import cloudpickle
with open(fname, "rb") as f:
obj = cloudpickle.load(f)
return obj(1, 2, 3)
def __init__(self, model_path: str):
with open(model_path, "rb") as rf:
self.model = pickle.load(rf)
def main():
args = parse_args()
setup_logging(args.logfile)
log = get_logger()
assert (0 <= args.hidden_fraction <= 1)
np.random.seed(args.random_seed)
tf.set_random_seed(args.random_seed)
log.info('*' * 100)
log.info('[Starting MC experiment]')
log_dict(log.info, vars(args))
log.info('[Loading input data]')
with open(args.input_file, 'rb') as f:
obj = cpkl.load(f)
# Set up experiments
fit_params = None
if args.mc_alg == 'PMF':
param_space = pmf_param_space(args)
run_experiment = run_pmf
elif args.mc_alg == 'PMF_b':
param_space = pmfb_param_space(args)
run_experiment = run_pmfb
elif args.mc_alg in ['KPMF', 'NGMC', 'KPMF_b']:
# Experiments that need PPI network
if args.ppi is not None:
ppi = nx.read_edgelist(args.ppi)
if args.mc_alg == 'KPMF':
L = get_ppi_data(obj['rows'], ppi, mode='laplacian')
param_space = kpmf_param_space(args)
run_experiment = run_kpmf
fit_params = dict(L=L)
elif args.mc_alg == 'KPMF_b':
L = get_ppi_data(obj['rows'], ppi, mode='laplacian')
param_space = kpmfb_param_space(args)
run_experiment = run_kpmfb
fit_params = dict(L=L)
elif args.mc_alg == 'NGMC':
fit_params = dict(P=None)
P = get_ppi_data(obj['rows'], ppi, mode='normalized_adjacency')
fit_params['P'] = P
param_space = ngmc_param_space(args)
run_experiment = run_ngmc
else:
raise (NotImplementedError(
'{} option is invalid or not implemented'.format(args.mc_alg)))
else:
raise (NotImplementedError(
'{} option is invalid or not implemented'.format(args.mc_alg)))
# Run experimental protocol
results, models, training_curves, trials = \
run_experiment(obj,
param_space = param_space,
fit_params = fit_params,
val_hidden_fraction=args.val_hidden_fraction,
hidden_fraction=args.hidden_fraction,
n_repeats=args.n_repeats,
hyperopt_iters=args.n_hyperopt_iters,
seed=args.random_seed,
logistic=args.logistic)
# Save results and other information
log_results(results['summary'])
with open(args.results_output, 'w') as f:
json.dump(results, f, indent=2)
with open(args.training_curve_output, 'wb') as f:
cpkl.dump(training_curves, f)
# TODO: save models the models cannot be pickled at the moment
# We will need to implement a from dict and a to dict method
with open(args.models_output, 'wb') as f:
cpkl.dump(trials, f)
with open(args.trials_output, 'wb') as f:
cpkl.dump(trials, f)
def _load_from_folder(model_folder: str,
install_requirements: bool = False):
code_temp_dir = os.path.join(model_folder, UnifiedModel._CODE_BASE_DIR)
data_temp_dir = os.path.join(model_folder, UnifiedModel._DATA_BASE_DIR)
model_pkl_path = os.path.join(data_temp_dir,
UnifiedModel._PICKLE_FILENAME)
if not os.path.isfile(model_pkl_path):
log.error("Model pickle file does not exist at path: " +
str(model_pkl_path))
return None
# TODO remove modules from _requirements...?
if install_requirements:
log.info("Installing model requirements")
# Execute setup.sh script if available
setup_script = os.path.join(model_folder,
UnifiedModel._SETUP_SCRIPT_FILENAME)
if os.path.isfile(setup_script):
os.chmod(setup_script, 0o777)
if subprocess.check_call([setup_script]) > 0:
log.warning("Failed to execute setup script")
else:
log.debug("Setup script does not exist.")
# Install dependencies from requirements.txt if available
requirements_file = os.path.join(
model_folder, UnifiedModel._REQUIREMENTS_FILENAME)
if os.path.isfile(requirements_file):
if subprocess.check_call([
sys.executable, '-m', 'pip', 'install', '-r',
requirements_file
]) > 0:
log.warning("Failed to install requirements")
else:
log.debug("requirements file does not exist.")
# extract code modules into temp folder
if os.path.isdir(code_temp_dir):
# append core modules to python path
sys.path.append(code_temp_dir)
else:
log.debug("Code directory does not exist.")
# add all data to stored files
stored_files = {}
if os.path.isdir(data_temp_dir):
for dir_, _, files in os.walk(data_temp_dir):
for file_name in files:
rel_dir = os.path.relpath(dir_, data_temp_dir)
if rel_dir != '.':
file_name = os.path.join(rel_dir, file_name)
if file_name != UnifiedModel._PICKLE_FILENAME:
# don't save the unified_model pickle file as stored file
stored_files[file_name] = os.path.join(dir_, file_name)
else:
log.debug("Data directory does not exist.")
try:
# , encoding='latin1', fix_imports=True
model = pickle.load(open(model_pkl_path, "rb"))
# provide information if requirements should be installed to loaded model
model._install_requirements = install_requirements
except Exception as e:
log.error("Failed to unpickle model: " + str(e), e)
if not install_requirements:
log.info("Try to set install_requirements flag to true.")
return None
if not hasattr(model, 'predict'):
log.warning("Model is not valid: predict method is missing!")
model._stored_files = stored_files
model._after_load()
return model
def _safe_deserialize(filename, mode="rb"):
with open(filename, mode) as pfile:
return cloudpickle.load(pfile)
else:
hout['sumw'][dataset] += np.sum(df['scale1fb'])
return hout
def postprocess(self, accumulator):
# set everything to 1/fb scale
lumi = 1000 # [1/pb]
scale = {}
for dataset, dataset_sumw in accumulator['sumw'].items():
scale[dataset] = lumi * self._corrections['xsections'][
dataset] / dataset_sumw.value
for h in accumulator.values():
if isinstance(h, hist.Hist):
h.scale(scale, axis="dataset")
return accumulator
if __name__ == '__main__':
with lz4f.open("corrections.cpkl.lz4", mode="rb") as fin:
corrections = cloudpickle.load(fin)
processor_instance = BoostedHbbProcessor(corrections=corrections)
with lz4f.open('boostedHbbProcessor.cpkl.lz4',
mode='wb',
compression_level=5) as fout:
cloudpickle.dump(processor_instance, fout)
def load_object(path):
import cloudpickle
with gzip.open(path, mode='rb') as file:
return cloudpickle.load(file)
def load(byte_array):
buf = io.BytesIO(byte_array.tobytes())
return cloudpickle.load(buf)
def read(name, task):
base_dir = '{}/{}'.format(DATA_BASE, name)
os.makedirs(base_dir, exist_ok=True)
with open(os.path.join(base_dir, task), 'rb') as f:
return pickle.load(f)
def get_run_command(self, filepath): """ Load the submission object and get the command to run the compiled bot """ with open (filepath, 'rb') as fo: subm = cloudpickle.load(fo) return subm.get_command(config.worker_compiled + subm.sub_id)
def __init__(self, artifacts):
with open(artifacts[key], 'rb') as f: # should not KeyError
if cloudpickle.load(f) != val:
raise ValueError # should not ValueError
from model_utils import *
from PIL import Image
import cloudpickle
import tensorflow_datasets
import tensorflow as tf
import cv2
import numpy as np
with open("weights/int2str.pkl", "rb") as f:
int2str = cloudpickle.load(f)
def load_model():
resnet50_backbone = get_backbone()
loss_fn = RetinaNetLoss(80)
model = RetinaNet(80, resnet50_backbone)
optimizer = tf.optimizers.SGD(learning_rate=learning_rate_fn, momentum=0.9)
model.compile(loss=loss_fn, optimizer=optimizer)
latest_checkpoint = tf.train.latest_checkpoint("weights")
model.load_weights(latest_checkpoint)
image = tf.keras.Input(shape=[None, None, 3], name="image")
predictions = model(image, training=False)
detections = DecodePredictions(confidence_threshold=0.5)(image,
predictions)
inference_model = tf.keras.Model(inputs=image, outputs=detections)
return inference_model
def predict(image_b, inference_model):
image_t = tf.io.decode_image(image_b)
def main():
import neptune
parser = argparse.ArgumentParser(argument_default=None)
parser.add_argument('--config', action='append', help='Gin config files.')
parser.add_argument('--debug', action='store_true', default=False)
cmd_args, unknown = parser.parse_known_args()
debug = cmd_args.debug
spec_path = cmd_args.config[0]
if not debug:
try:
with open(spec_path, 'rb') as f:
import cloudpickle
specification = cloudpickle.load(f)
except pickle.UnpicklingError:
with open(spec_path) as f:
vars_ = {'script': os.path.basename(spec_path)}
exec(f.read(), vars_) # pylint: disable=exec-used
specification = vars_['experiments_list'][0].to_dict()
print(
'NOTE: Only the first experiment from the list will be run!'
)
parameters = specification['parameters']
else:
print("debug run")
parameters = dict(env_id="toy_mr", env_size=None)
class MockArgs(object):
def add(self, key, value):
setattr(self, key, value)
args = MockArgs()
args.add('env', parameters["env_id"]) # 'chain_env' 'toy_mr'
args.add('env_size', parameters["env_size"])
args.add('seed', 0)
args.add('max_episode_steps', 300)
args.add('num_timesteps', int(1e12))
args.add('num_env', 32)
args.add('use_news', 0)
args.add('gamma', 0.99)
args.add('gamma_ext', 0.999)
args.add('lam', 0.95)
args.add('update_ob_stats_every_step', 0)
args.add('update_ob_stats_independently_per_gpu', 0)
args.add('update_ob_stats_from_random_agent', 1)
args.add('proportion_of_exp_used_for_predictor_update', 1.)
args.add('tag', '')
args.add(
'policy',
'cnn',
)
args.add('int_coeff', 1.)
args.add('ext_coeff', 2.)
args.add('dynamics_bonus', 0)
if not debug:
# TODO read more from specification
print("running with neptune")
neptune.init(
project_qualified_name="pmtest/planning-with-learned-models")
neptune.create_experiment(
name=specification['name'],
tags=specification['tags'],
params=specification['parameters'],
upload_stdout=False,
upload_stderr=False,
)
neptune.send_metric("test", 777)
baselines_format_strs = ['log', 'csv']
else:
print("running without neptune")
baselines_format_strs = ['stdout', 'log', 'csv']
logger.configure(dir="out", format_strs=baselines_format_strs)
seed = 10000 * args.seed # + MPI.COMM_WORLD.Get_rank()
set_global_seeds(seed)
hps = dict(frame_stack=4,
nminibatches=4,
nepochs=4,
lr=0.0001,
max_grad_norm=0.0,
env_size=args.env_size,
use_news=args.use_news,
gamma=args.gamma,
gamma_ext=args.gamma_ext,
max_episode_steps=args.max_episode_steps,
lam=args.lam,
update_ob_stats_every_step=args.update_ob_stats_every_step,
update_ob_stats_independently_per_gpu=args.
update_ob_stats_independently_per_gpu,
update_ob_stats_from_random_agent=args.
update_ob_stats_from_random_agent,
proportion_of_exp_used_for_predictor_update=args.
proportion_of_exp_used_for_predictor_update,
policy=args.policy,
int_coeff=args.int_coeff,
ext_coeff=args.ext_coeff,
dynamics_bonus=args.dynamics_bonus)
tf_util.make_session(make_default=True)
train(env_id=args.env,
num_env=args.num_env,
seed=seed,
num_timesteps=args.num_timesteps,
hps=hps,
use_neptune=(not debug))
solvermap = {'optunity': 'Optunity',
'tpe': 'Hyperopt',
'smac': 'SMAC',
'bayesopt': 'BayesOpt',
'random': 'random'}
all_results = {k: [] for k in datasets}
performance = {}
random90 = []
for repetition in range(start_index, stop_index + 1):
random90.append({})
for dataset in datasets:
all_results[dataset].append({})
with open('results-repeated/%s-all.pkl-%d' % (dataset, repetition), 'r') as f:
results = pickle.load(f)
performance = {}
for solver in solvers:
if solver == 'optunity':
perf = max(results[solver]['results'][:budget])
else:
perf = -min(results[solver]['results'][:budget])
all_results[dataset][-1][solver] = perf
if solver == 'random':
srtd = sorted(results[solver]['results'][:budget], reverse=True)
random90[-1][dataset] = -srtd[int(0.75 * len(srtd))]
ranks = {solver: {data: [] for data in datasets} for solver in solvers}
bests = {dataset: [] for dataset in datasets}
a, b, c = y_test.shape
y_test = np.reshape(y_test, [a, b, c, 1])
# Get the mean image files
mean_img_x = []
mean_img_y = []
if (NORMALIZED == True):
MEANS_FILE = MODEL_NAME + '_data_means.p'
MEANS_PATH = MODEL_ARCHIVE + MEANS_FILE
try:
ff = open(MEANS_PATH, 'rb')
except:
print(
'ResBrowswer: failed to open the file ' + MEANS_PATH + '.' +
' Attempting to run without the image means.', sys.stderr)
NORMALIZED = False
else:
mean_img_x, mean_img_y = pickle.load(ff)
ff.close()
rB = res_browser(loaded_model,
x_test,
y_test,
normalized=NORMALIZED,
mean_img_x=mean_img_x,
mean_img_y=mean_img_y)
rB.initialize()
plt.show()
