Models: object_detection ImportError: No module named 'object_detection'

[jinfagang]

in research directory, Simply run:

python object_detection/builders/model_builder_test.py

as readme install suggested.
Got the error:

ImportError: No module named 'object_detection'

I know this is a low-level error, but the tutorial in readme is really misleading, follow that command just got this error message.
in object_detection many files are using from object_detection import ..., So how to solve this problem???

[jjahanip]

go to the research directory and type:

export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim

[T0MTOM1]

'export' is not recognized as an internal or external command,
operable program or batch file.

[jjahanip]

if you are in WINDOES you can use the following command (from the research directory):
SET PYTHONPATH=%cd%:%cd%\slim

if working with editors (like PyCharm) you can add path to PYTHONPATH variable from the software.

[jinfagang]

@jjahanip So many thanks, I'll check it again.

[williamsuppiger]

i am working in anaconda and i get a similar error accept it says:
File "object_detection/builders/model_builder_test.py", line 21, in
from object_detection.builders import model_builder
ModuleNotFoundError: No module named 'object_detection'

[jinfagang]

@williamsuppiger Maybe this is not supporte anymore? This is an obvious mistake.

[tombstone]

@jinfagang @williamsuppiger this error usually occurs if you fail to set the PYTHONPATH as mentioned by @jjahanip . Have you tried setting it ?

[temiklis]

if you use windows add in model_builder_test.py after import tensorflow as tf.

 import sys
   sys.path.append("{YOU_PATH}\\models-master\\research\\")
      sys.path.append("{YOU_PATH}\\models-master\\research\\object_detection\\utils")

[tombstone]

I will close this for now since things seems to work as expected. If you are still having problems please feel free to reopen this issue.

[imrandatascientist]

After I entered 'export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim' to get around the Import error for object detection, I got another import error

ImportError: No module named nets

[klauszeng]

to correct @jjahanip, on windows is SET PYTHONPATH=%cd%;%cd%\slim.
It's ';'
not ':'

[EricTheAI]

One possible reason for this might be that you are not running the Jupyter Notebook in the same terminal where you set the $PYTHONPATH value.

Or you can try to add $PYTHONPATH value into your ~/.bashrc file (or ~/.zshrc file if you are using 'oh my zsh')

[jpdoiron]

on windows the line is :

SET PYTHONPATH=%cd%;%cd%\slim

NOT
SET PYTHONPATH=%cd%:%cd%\slim

You can test with :
echo %PYTHONPATH%

it should give you somethine like
\models\research;\models\research\slim

[uthiraa]

Has anyone tried the same thing on floydhub?
I am getting the modulenotfound error there. Can anyone help me?

[vishnuvardhanreddychallapalli]

When i am running the ipython in object detection i am getting an error in line 17 saying that i do not have tensor flow how to avoid this error

[gfjiyue]

I tried add these two lines and it works for me.
import sys
sys.path.append("..")

[rbarman]

If anyone has the same issue in a Kaggle kernel, here is one fix:

import os
os.environ['PYTHONPATH'] += ':/kaggle/working/models/research/:/kaggle/working/models/research/slim/'

[ghost]

from imageai.Detection import ObjectDetection
error at this line while running as windows service.
please help me.

[ravi3222]

If anyone has the same issue in a Kaggle kernel, here is one fix:

import os
os.environ['PYTHONPATH'] += ':/kaggle/working/models/research/:/kaggle/working/models/research/slim/'

Great working on colab too!

[Sakifneoworld]

Thank you very much.....
for windows user......

import sys
sys.path.append("{YOU_PATH}\models-master\research\")
sys.path.append("{YOU_PATH}\models-master\research\object_detection\utils")

PATH_TO_LABELS = 'data/mscoco_label_map.pbtxt'

[Ashoknicky]

Thank you very much.....
for windows user......

import sys
sys.path.append("{YOU_PATH}\models-master\research\")
sys.path.append("{YOU_PATH}\models-master\research\object_detection\utils")

PATH_TO_LABELS = 'data/mscoco_label_map.pbtxt'

Thank you soo much...
its working now..

[SystemsPatrick]

pip install tensorflow-object-detection-api

[miladfa7]

If anyone has the same issue in a Kaggle kernel, here is one fix:

import os
os.environ['PYTHONPATH'] += ':/kaggle/working/models/research/:/kaggle/working/models/research/slim/'

Great working on colab too!

[poonia-amit]

if you use windows add in model_builder_test.py after import tensorflow as tf.
import sys
sys.path.append("{YOU_PATH}\models-master\research\")
sys.path.append("{YOU_PATH}\models-master\research\object_detection\utils")

This worked for me. Thanks

I also add these above three lines of code into my train.py.

[gilmotta]

I had to move my .py file to the \research folder and it worked.