mtcnn
mtcnn copied to clipboard
ipazc
Disable progress bars?
Apologies if this is a Keras or TF issue, but how do I disable the stdout and stderr logging with mtcnn?
Example Class:
#!/usr/bin/env python
""" Face Detection Filter for Progeny """
from __future__ import absolute_import
from filter_base import Filter
from mtcnn import MTCNN
import os
import cv2
import logging
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
class FaceDetectFilter(Filter):
pass
def __init__(self, threshold=0.70):
super().__init__()
self.threshold = threshold
self.detector = MTCNN()
return
def get_faces(self, path):
img = cv2.cvtColor(cv2.imread(path), cv2.COLOR_BGR2RGB)
faces = self.detector.detect_faces(img)
return faces
def determine_class(self, imgrow):
faces = self.get_faces(imgrow['thumb_path'])
if faces is not None and len(faces) > 0:
for face in faces:
if face['confidence'] >= self.threshold:
return 'face'
return 'noface'
Results in:
1/1 [==============================] - 0s 15ms/step
1/1 [==============================] - 0s 18ms/step
1/1 [==============================] - 0s 15ms/step
1/1 [==============================] - 0s 15ms/step
1/1 [==============================] - 0s 15ms/step
1/1 [==============================] - 0s 14ms/step
2/2 [==============================] - 0s 2ms/step
1/1 [==============================] - 0s 26ms/step
Logging is disabled but the (tqdm?) progress bars remain. This prevents my own tqdm progress bars from working.
Any ideas?
@amykweon @xeb @equ1
Those loggings are from mtcnn package.
You can suppress the logging by editting the mtcnn.py file in ~/miniconda3/envs/<env_name>/lib/<python3.x>/site-packages/mtcnn/mtcnn.py (depends on your package management) or fork this repo: https://github.com/ipazc/mtcnn
Then add verbose=0 to these linse:
Line 410: out = self._pnet.predict(img_y, verbose=0)
Line 410: out = self._rnet.predict(tempimg1, verbose=0)
Line 466: out = self._onet.predict(tempimg1, verbose=0)
I add verbose=0 in lines,but it does not work then I remove tqdm module,it is worked~
With newer versions of keras, this has worked for me:
import keras
keras.utils.disable_interactive_logging()
not sure exactly which version added this
