After yolo training tfnet give empty result.

[KumarPardeep1]

After training results = tfnet.return_predict(frame) give me empty array here is some sort of code.

import cv2 from darkflow.net.build import TFNet import zipfile import numpy as np import time

options = { 'model': 'cfg/tiny-yolo-voc-1c.cfg', 'load': 3000, 'threshold': 0.1, 'gpu': 0.8 } tfnet = TFNet(options) results = tfnet.return_predict(frame)

[S0soo]

You can try saving checkpoint as. pb and. meta files.Then modify your options options = { 'pbLoad': 'metaLoad': }

[rahuledavalath]

@KumarPardeep1 While printing results. (print(results) I am also getting empty result( [] ). Did you got any solution?

[HSHunterR]

@rahuledavalath Are you have bounding box on your test image? If not ,,,I mean the train maybe has problem.

[ludwikbukowski]

the same problem...

[hashbanger]

Guys any solutions to this? I am also getting the empty array.

[Hritik-choukikar]

Same. when I am training the model I am getting this warning "/content/drive/MyDrive/darkflow2/darkflow/darkflow/net/ops/convolution.py:79: RuntimeWarning: invalid value encountered in sqrt temp /= (np.sqrt(layer.w['moving_variance']) + 1e-5"

Does that mean anything? In my case generated .pb file is blank