Models: Getting ImportError: No module named deployment on object_detection/train.py

[lucasharada]

I found a lot of problem about this, tried a bunch of proposed solution but didn't get it to work.
I'm following the guides found here to make a custom object detector with Tensorflow.

I already passed through the Installation guide, including the models/research and models/research/slim in the $PYTHONPATH variable. That is the most common suggestion to fix this error. But I still get the same error:

~/tensorflow/models/research$ echo $PYTHONPATH:
/home/cc/tensorflow/models/research:/home/cc/tensorflow/models/research/slim
~/tensorflow/models/research$ sudo python object_detection/train.py --logtostderr --pipeline_config_path=/home/cc/Documents/my-model-tensorflow/training/ssd_mobilenet_v1_person.config --train_dir=/home/cc/Documents/my-model-tensorflow/data
Traceback (most recent call last):
  File "object_detection/train.py", line 49, in <module>
    from object_detection import trainer
  File "build/bdist.linux-x86_64/egg/object_detection/trainer.py", line 33, in <module>
    from deployment import model_deploy
ImportError: No module named deployment

Any other solution?

[robin4690]

I'm in a similar position, It seems like you solved the issue, how did you manage that?
Cheers.

[lucasharada]

When I installed the object_detection module, I included the path in the
.bashrc file, but it didn't load automatically. All I needed to do is to
load the package manually everytime.

Em 25 de nov de 2017 1:24 AM, "robin4690" [e-mail]
escreveu:

I'm in a similar position, It seems like you solved the issue, how did you
manage that?
Cheers.

—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub
https://github.com/tensorflow/models/issues/2888#issuecomment-346917826,
or mute the thread
https://github.com/notifications/unsubscribe-auth/ARzfoRBwRqCLukDxsb7P0_AFjmDrbG8Tks5s55aNgaJpZM4Qp6Mj
.

[macro-dadt]

@lucasharada make sure that you cd .../models/research in .bashrc file

[MariamMohamedFawzy]

@lucasharadacan you explain how you load the package manually? I don't get it.
I have the same problem, but I have my code inside object detection folder. export path ... does not solve the issue.

[lucasharada]

Everytime I open a Terminal, I have to run the final step of the Installation guide. That is:

export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim

[robin4690]

I'm still having issues. Have attached a screenshot below, any ideas? Running this on Windows 10 (Macbook Pro bootcamp).

[xav12358]

I used to work on windows but now I try to learn my ssd model on windows and I also have the same problem?

robn4690 do you solve your problem?

[JackLaii]

Same problem here....

[npeirson]

Same problem here. On Linux. Anyone else tried importing it from shell? That works, which seems to imply version problems, even though I can't find any actual conflict...

[JackLaii]

If you are using windows, try to change 'pwd' to your actual path to models while setting PYTHONPATH.
i.e, 'pwd'---change to--->'\path to\tensorflow\models'

[npeirson]

It's a hacky fix, but I'm on a vm so I don't care: moved the dependencies (deployment and the others) to /usr/local/lib/python3.5/dist-packages/object_detection-0.1-py3.5.egg/
Messy, but works.
Also, don't forget that if you run into the new dct_method problem, you can follow @jqcorreia's fix here

updated 09 Apr. 18
Ran into this problem again on a VM, reset the machine to vanilla and re-installed packages fresh, and this problem disappeared.

Cool.

[fionasoon1]

I got this from another site and solved the forementioned problem.

Try the following steps and according to the sequence strictly.

1-in folder /models/research
~/.../models/research$ python setup.py build
~/.../models/research$ python setup.py install

2-
set PYTHONPATH=c:\ tensorflow1\object_detection\; c:\ tensorflow1\object_detection\models\research; c:\ tensorflow1\object_detection\models; c:\tf_tutorial\object_detection\models\research\slim
C:\tensorflow1\models\research$ echo %PYTHONPATH%

3- in folder /models/research/object_detection:
~/.../models/research/object_detection$
python train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/faster_rcnn_inception_v2_pets.config

Hope this works for you too!

[npeirson]

ok so the thread you're posting to was closed a long text ago because the problem was resolved long ago. the current error you're encountering is because train.py no longer exists in models/research/object_detection/. it was changed a few updates ago.

[TechGenius28]

Hi,
I have been stuck on this problem for a few months now! Does anyone have a solution to this frustrating error?
Thanks

[npeirson]

It hasn't existed for a few months. Try this.