Tf-pose-estimation: AttributeError: 'TfPoseEstimator' object has no attribute 'persistent_sess'

[qintaohu]

/usr/bin/python3.5 /home/hqt/tf-openpose/src/run.py
/home/hqt/.local/lib/python3.5/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type.
from ._conv import register_converters as _register_converters
Traceback (most recent call last):
File "/home/hqt/tf-openpose/src/run.py", line 32, in
e = TfPoseEstimator(get_graph_path(args.model), target_size=(w, h))
File "/home/hqt/tf-openpose/src/estimator.py", line 255, in __init__
graph_def.ParseFromString(f.read())
File "/home/hqt/.local/lib/python3.5/site-packages/tensorflow/python/lib/io/file_io.py", line 119, in read
self._preread_check()
File "/home/hqt/.local/lib/python3.5/site-packages/tensorflow/python/lib/io/file_io.py", line 79, in _preread_check
compat.as_bytes(self.__name), 1024 * 512, status)
File "/home/hqt/.local/lib/python3.5/site-packages/tensorflow/python/framework/errors_impl.py", line 473, in __exit__
c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.NotFoundError: ./models/graph/mobilenet_thin_432x368/graph_opt.pb; No such file or directory
Exception ignored in: >
Traceback (most recent call last):
File "/home/hqt/tf-openpose/src/estimator.py", line 278, in __del__
self.persistent_sess.close()
AttributeError: 'TfPoseEstimator' object has no attribute 'persistent_sess'

I run the test the inference feature with a single image. use

$ python3 run.py --model=mobilenet_thin_432x368 --image=
but has a error,Anybody meet this problem?

[elmuz]

tensorflow.python.framework.errors_impl.NotFoundError:
./models/graph/mobilenet_thin_432x368/graph_opt.pb; No such file or directory

Try laucnhing the script from the repo root (parent folder)
$python3 src/run.py --model=mobilenet_thin_432x368 --image=

[oguzhankarabulut]

run your code in tf-openpose folder not in src folder

[lesreaper]

Or you can run it as stated in the ReadMe and use a relative image path "../../Desktop"...

[pushkalkatara]

Same issue while using roslauch, please help.

[Svito-zar]

@elmuz, your solution helped me, when I used --model=mobilenet_thin_432x368, but I have got the same error again when using any other model, like cmu_640x480 or cmu_640x360.

I am getting the error about 'persistent_sess' for these models also when running src/run.py

[paperclipt]

@Svito-zar did you figure out the problem? i got the same error when i use cmu_640x480 or cmu_640x360

[elmuz]

@Svito-zar @paperclipt be sure you downloaded the trained model you need.
I pulled this morning the repository and I can confirm it works ok. 'cmu_640x480' and 'cmu_640x360' have to be taken from mediafire links inside download.sh files in their 'models/graph' folders. If you just execute the .sh file it may not work, try manually download those files through a web browser. After copying those files accordingly in their folders, I can confirm the tool works nicely.

[paperclipt]

@Svito-zar @elmuz thanks. after using the files downloaded from the web browser, i confirm that it works.