def test_wo_pred(model_id, model_path, data_paths_path, batch_size=16, load_epoch=None):
input_size = 256
output_classes = 7
start_time = time.clock()
model = models.get_model_from_id(model_id, input_size, output_classes)
if model is None:
return
# Load log
model_path = model_path + model_id
if not os.path.exists(model_path):
os.makedirs(model_path)
log = open(model_path + "/log_testing_" + str(load_epoch) + ".txt", "a")
log.write("\n\n\nTesting initialised: {:%Y-%m-%d %H:%M:%S}".format(datetime.datetime.now()))
if load_epoch is not None:
log.write("\nLoading past model:")
log.write("\n" + model_path + "/" + model_id + "_" + str(load_epoch) + ".h5")
model.load_weights(model_path + "/" + model_id + "_" + str(load_epoch) + ".h5")
else:
print "ERROR: Need load_epoch number to load"
return
model = models.compile_model(model_id, model)
# get data
with open(data_paths_path + "test_paths_equalised.txt") as f:
all_paths = f.readlines()
random.shuffle(all_paths) # randomise order every epoch!!
all_paths = [line.split() for line in all_paths] # split so x and y split
X_batch = None
Y_batch = []
sum_loss = 0
sum_acc = 0
past = 0
count = 0
inner_count = 0
for path in all_paths:
# print path
count += 1
x, y = models.load_input(model_id, path, input_size)
if X_batch is None:
X_batch = x
else:
if "MVK_12" in model_id:
X_batch[0] = np.append(X_batch[0], x[0], axis=0)
X_batch[1] = np.append(X_batch[1], x[1], axis=0)
else:
X_batch = np.append(X_batch, x, axis=0)
Y_batch.append(y)
if batch_size is not None:
if (count % batch_size == 0) or (count == len(all_paths)):
Y_batch = np.eye(output_classes)[Y_batch]
# train
loss, acc = model.test_on_batch(X_batch, Y_batch)
sum_loss += loss
sum_acc += acc
inner_count += 1
# clear batch
X_batch = None
Y_batch = []
if int((float(count) / len(all_paths)) * 100) > past:
tr = (len(all_paths) - count) / ((count) / (time.clock() - start_time))
print "(%d) [%.5f] Image: %d / %d; TR: %02d:%02d:%02d;" % (
past,
sum_loss / inner_count,
count,
len(all_paths),
int((tr / 60) / 60),
int((tr / 60) % 60),
int(tr % 60),
)
log.write(
"\n(%d) [%.5f] Image: %d / %d; TR: %02d:%02d:%02d;"
% (
past,
sum_loss / inner_count,
count,
len(all_paths),
int((tr / 60) / 60),
int((tr / 60) % 60),
int(tr % 60),
)
)
past += 5
if batch_size is None:
loss, acc = model.evaluate(X_batch, Y_batch, batch_size=32)
else:
loss = sum_loss / inner_count
acc = sum_acc / inner_count
log.write("\n %d samples. \t loss: %.5f \t acc: %.5f" % (count, loss, acc))
print ("\n %d samples. \t loss: %.5f \t acc: %.5f" % (count, loss, acc))
log.close()
def predict(model_id, model_path, data_paths_path, split, batch_size=None, load_epoch=None, layers=['pred'], save_path=None, equalised=False, soft=None):
start_time = time.clock()
model, output_classes = models.get_model_from_id(model_id)
if model is None:
return
# Load log
model_path = model_path + model_id
if not os.path.exists(model_path):
os.makedirs(model_path)
if load_epoch is not None:
print 'Loading model: ' + model_path + '/' + model_id + '_' + str(load_epoch) + '.h5'
model.load_weights(model_path + '/' + model_id + '_' + str(load_epoch) + '.h5')
else:
if model_id != 'MVK_17_04':
print 'ERROR: Need load_epoch number to load'
return
# model = models.compile_model(model_id, model) # dont need to compile on prediction
# get data
if equalised:
with open(data_paths_path + split + '_paths_equalised.txt') as f:
all_paths = f.readlines()
else:
with open(data_paths_path + split + '_paths.txt') as f:
all_paths = f.readlines()
# all_paths = all_paths[:500]
all_paths = [line.split() for line in all_paths] # split so x and y split
for layer_name in layers:
# model, output_classes = models.get_model_from_id(model_id)
model = Model(input=model.input, output=model.get_layer(layer_name).output)
X_batch = None
Y_batch = []
Y_gt = None
Y_pred = None
past = 0
count = 0
inner_count = 0
for path in all_paths:
# print path
count += 1
cor_path = DRIVE + path[0][path[0].find('/DATASETS/')+1:]
if path[0] != cor_path:
# print 'Paths in .txt files seem incorrect'
# print 'Changed from: '+path[0]
# print 'Changed to: '+ cor_path
path[0] = cor_path
x, y = models.load_input(model_id, path, soft)
if X_batch is None:
X_batch = x
else:
if 'MVK_12' in model_id:
X_batch = [np.append(X_batch[0], x[0], axis=0), np.append(X_batch[1], x[1], axis=0)]
else:
X_batch = np.append(X_batch, x, axis=0)
Y_batch.append(y)
if batch_size is not None:
if (count % batch_size == 0) or (count == len(all_paths)):
# Y_batch = np.eye(output_classes)[Y_batch]
# train
if Y_gt is None:
Y_pred = model.predict_on_batch(X_batch)
Y_gt = Y_batch
else:
Y_pred = np.append(Y_pred, model.predict_on_batch(X_batch),axis=0)
Y_gt = np.append(Y_gt,Y_batch,axis=0)
inner_count += 1
# clear batch
X_batch = None
Y_batch = []
else:
Y_pred = model.predict(x)
Y_gt = np.eye(output_classes)[y]
if int((float(count) / len(all_paths)) * 100) > past:
tr = (len(all_paths) - count) / ((count) / (time.clock() - start_time))
print '(%d) Image: %d / %d; TR: %02d:%02d:%02d;' % (past, count, len(all_paths), int((tr / 60) / 60),
int((tr / 60) % 60), int(tr % 60))
past += 5
# if batch_size is None:
# Y_pred = model.predict_proba(X_batch, batch_size=32)
# save predictions to file
all_array = {}
if save_path is not None:
print '\nSaving ....'
save_path += model_id + '_' + str(load_epoch)+'/'
print save_path
for p in range(len(Y_gt)):
path = all_paths[p]
# write out to npy files
image_name = path[0].split('/')[len(path[0].split('/')) - 1]
if not os.path.exists(save_path + layer_name + '/npy/ind/'):
os.makedirs(save_path + layer_name + '/npy/ind/')
np.save(save_path + layer_name + '/npy/ind/' + image_name[:-4] + '.npy', np.squeeze(Y_pred[p]))
all_array[image_name[:-4]] = [Y_gt[p], Y_pred[p]]
# np.save(save_path + layer_name + '/npy/'+split+'.npy', all_array)
return Y_gt, Y_pred
