Open sbharadwajj opened 5 years ago
Hi,
I fixed it by changing this in the models.py
file
with chainer.using_config('train', False):
h = F.dropout(h)
h_detection = F.relu(self.fc_detection1(h))
h_detection = F.dropout(h_detection)
h_detection = self.fc_detection2(h_detection)
h_landmark = F.relu(self.fc_landmarks1(h))
h_landmark = F.dropout(h_landmark)
h_landmark = self.fc_landmarks2(h_landmark)
h_visibility = F.relu(self.fc_visibility1(h))
h_visibility = F.dropout(h_visibility)
h_visibility = self.fc_visibility2(h_visibility)
h_pose = F.relu(self.fc_pose1(h))
h_pose = F.dropout(h_pose)
h_pose = self.fc_pose2(h_pose)
h_gender = F.relu(self.fc_gender1(h))
h_gender = F.dropout(h_gender)
h_gender = self.fc_gender2(h_gender)
But i'm unable to draw the rectangles around the predicted results for gender.
@sbharadwajj could you please share your python environment ? I cannot start training with full cuda (cupy) support
Hi, @takiyu I'm running the use_on_file.py like this
python scripts/use_on_file.py --model model_epoch_190 --img age_gender/input/jan2.jpg
The error I'm getting is this:
The model is successfully running till this step:
[2018/10/12 04:45:15] INFO (4127) __main__ : Forward the network
Is there some step I'm missing out?