Models: ImportError: No module named nets

[10183308]

when I enter : python object_detection/builders/model_builder_test.py

it shows the wrong : ImportError: No module named nets

could you help me?

thank you very much

[heroinlin]

Look at Installation！！！
When running locally, the tensorflow/models/ and slim directories should be appended to PYTHONPATH. This can be done by running the following from tensorflow/models/:

From tensorflow/models/

export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim
nets and deployment modules belong to slim

[heroinlin]

'pwd' is the path of models folder.
Try to set "pwd:pwd/slim" with "/media/media_share/linkfile/models:
/media/media_share/linkfile/models/slim".

2017-07-04 9:32 GMT+08:00 10183308 [e-mail]:

thank you .

I have added the PYTHONPATH ,

but it does not work.

hc@hc-PC:/media/media_share/linkfile/models$ echo $PYTHONPATH
/usr/local/lib/python2.7/dist-packages:/usr/lib/python2.7/
dist-packages:/home/hc/project/caffe2/caffe2/build/
caffe2:/home/hc/project/caffe2/caffe2/build/caffe2/
python:/home/hc/project/caffe_sorce/caffe/python:/usr/local:
/home/hc/project/caffe_sorce/caffe/python:/usr/local:/home/
hc/project/caffe2/caffe2/build::/home/hc:/home/hc/slim:
/home/hc:/home/hc/slim:pwd:pwd/slim

hc@hc-PC:/media/media_share/linkfile/models$ python
object_detection/builders/model_builder_test.py
Traceback (most recent call last):
File "object_detection/builders/model_builder_test.py", line 21, in
from object_detection.builders import model_builder
File "/media/media_share/linkfile/models/object_detection/builders/model_builder.py",
line 30, in
from object_detection.models import faster_rcnn_inception_resnet_v2_feature_extractor
as frcnn_inc_res
File "/media/media_share/linkfile/models/object_detection/
models/faster_rcnn_inception_resnet_v2_feature_extractor.py", line 29, in
from nets import inception_resnet_v2
ImportError: No module named nets

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/tensorflow/models/issues/1842#issuecomment-312759640,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AQrfnvlM0XfHYhlj-2gMb1e3up1qAjxoks5sKZY2gaJpZM4OL73k
.

[aweichina]

Ah, it is the problem of "path"! I have trying for several days. When I added in the program code:
_"import sys_
_print(sys.path)_
_sys.path.append("/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/slim")_
_print(sys.path)"_
It works! For the nets.py in directory"/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/slim"

[heroinlin]

Congratulate to work it well.You can find the slim folder in models folder, and nets folder in the slim folder.I work it on windows, like this:
.
so I just add this slim to the path.on the linux, you can try it similarly.

[FlashSnail]

I think it's a trick.... in the document, it said

export PYTHONPATH=$PYTHONPATH:'pwd':'pwd'/slim

pwd is your local path, however, you should keep mark like this : 'local path':'localpath'/slim
it works on my computer. 哇，喷火龙唉！

[aweichina]

Thanks a lot!

[wationgo]

Thanks a lot. It worked for me

[aweichina]

Thank you for all! I have solved this problem by run setup.py. Thank you!

[arpitsharmagit]

Thanks!!! running setup.py build and setup.py install worked

[qiaoguan]

sudo gedit ~/.bashrc #open bashrc file

then add this new line to the end of your bashrc file

export
PYTHONPATH=$PYTHONPATH=/home/hri/tensorflow/models/research:/home/hri/tensorflow/models/research/slim

then restart then bashrc to make it work

source ~/.bashrc

you can change the path to your own!

[kil0watt]

I encountered the same problem on windows. What I did was, I just copied the entire nets folder from slim directory to C:\Python\Python35\Lib\site-packages\object_detection-0.1-py3.5.egg (where I installed object_detection API). Then the error is completely gone.

[BokaiLIAN]

I have the same problem and I go to the tensorflow's github for information
You can never figure it out that I solve this by simply copy the code below from the net page
export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim

Note that it's ` instead of ' around the key pwd
That's my problem, hope it's helpful to you

[kochsnow]

I add export PYTHONPATH="$PYTHONPATH:tensorflow/models/research:/tensorflow/models/research/slim" in my .bashrc file, then source ~/.bashrc file, but I still has this problem. someone can tell me whiy

[appdotpy]

It looks like your paths are wrong (unless you have a directory called "tensorflow" at the root of your filesystem, in which case you are missing a '/' at the beginning of your first path)

try to include the absolute path. e.g.:
export PYTHONPATH="$PYTHONPATH:/home/kochsnow/tensorflow/models/research:/home/kochsnow/tensorflow/models/research/slim"

[aweichina]

Thank you! It works!

---

发件人：Jeremiah Harbach [e-mail]
发送时间：2018年8月31日(星期五) 02:15
收件人：tensorflow/models [e-mail]
抄　送：aweichina [e-mail]; Comment [e-mail]
主　题：Re: [tensorflow/models] ImportError: No module named nets (#1842)

It looks like your paths are wrong (unless you have a directory called "tensorflow" at the root of your filesystem, in which case you are missing a '/' at the beginning of your first path)
try to include the absolute path. e.g.:
export PYTHONPATH="$PYTHONPATH:/home/kochsnow/tensorflow/models/research:/home/kochsnow/tensorflow/models/research/slim"
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.

[LyonOconner]

I encountered the same problem on windows. What I did was, I just copied the entire nets folder from slim directory to C:\Python\Python35\Lib\site-packages\object_detection-0.1-py3.5.egg (where I installed object_detection API). Then the error is completely gone.

No working for me :/

[mymultiverse]

I am able to run on a local machine by adding the right slim path to PYTHONPATH. But when I am trying to run on google colab I am getting error even though I added the slim path with:

# From tensorflow/models/research/
export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim

Here is my error messages:

Traceback (most recent call last):
  File "legacy/train.py", line 51, in <module>
    from object_detection.builders import model_builder
  File "/usr/local/lib/python3.6/dist-packages/object_detection-0.1-py3.6.egg/object_detection/builders/model_builder.py", line 35, in <module>
    from object_detection.models import faster_rcnn_inception_resnet_v2_feature_extractor as frcnn_inc_res
  File "/usr/local/lib/python3.6/dist-packages/object_detection-0.1-py3.6.egg/object_detection/models/faster_rcnn_inception_resnet_v2_feature_extractor.py", line 28, in <module>
    from nets import inception_resnet_v2
ModuleNotFoundError: No module named 'nets'

[LyonOconner]

I am able to run on a local machine by adding the right slim path to PYTHONPATH. But when I am trying to run on google colab I am getting error even though I added the slim path with:

# From tensorflow/models/research/
export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim

Here is my error messages:

Traceback (most recent call last):
  File "legacy/train.py", line 51, in <module>
    from object_detection.builders import model_builder
  File "/usr/local/lib/python3.6/dist-packages/object_detection-0.1-py3.6.egg/object_detection/builders/model_builder.py", line 35, in <module>
    from object_detection.models import faster_rcnn_inception_resnet_v2_feature_extractor as frcnn_inc_res
  File "/usr/local/lib/python3.6/dist-packages/object_detection-0.1-py3.6.egg/object_detection/models/faster_rcnn_inception_resnet_v2_feature_extractor.py", line 28, in <module>
    from nets import inception_resnet_v2
ModuleNotFoundError: No module named 'nets'

set PYTHONPATH=C:\Tensorflow\models;C:\Tensorflow\models\research;C:\Tensorflow\models\research\slim
set PATH=%PATH%;%PYTHONPATH%
Set environment EVERYTIME when activate tensorflow gpu

[LyonOconner]

or copy & paste Slim folder content in
C:\ProgramData\Anaconda3\envstensorflow_cpu\Lib\site-packages\object_detection-0.1-py3.5.egg\ {here}

[mymultiverse]

set PYTHONPATH=C:\Tensorflow\models;C:\Tensorflow\models\research;C:\Tensorflow\models\research\slim
set PATH=%PATH%;%PYTHONPATH%

I think there no C in google colab. Results from above command

/bin/bash: C:Tensorflowmodelsresearch: command not found
/bin/bash: C:Tensorflowmodelsresearchslim: command not found
/bin/bash: line 0: fg: no job control

[mymultiverse]

or copy & paste Slim folder content in
C:\ProgramData\Anaconda3\envstensorflow_cpu\Lib\site-packages\object_detection-0.1-py3.5.egg\ {here}

I did that but still same import error. In case of google colab I did with following command, let me know if path is not correct.
cp -R slim/ /usr/local/lib/python3.6/dist-packages/object_detection-0.1-py3.6.egg/

Problem solved when I copy nets and deployment.

[mymultiverse]

For google colab. Inside research folder.

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

[jameswan]

To overcome this error:
ImportError: No module named 'tensorflow.contrib.slim'

All I did was:
pip install tensorflow-gpu==1.8.0

[taalaialmasova]

cd "path to the models folder"
$ protoc object_detection/protos/*.proto --python_out=.
$ export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim

[xiaJis]

Ah, it is the problem of "path"! I have trying for several days. When I added in the program code:
_"import sys_
_print(sys.path)_
_sys.path.append("/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/slim")_
_print(sys.path)"_
It works! For the nets.py in directory"/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/slim"

The path you add yourself after exiting the Python environment will automatically disappear

[codexponent]

@aweichina solution worked for me. :)

[MohammadDabbas]

Hi guys I'm having an error when I'm trying to run the training using this command:

python train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/ssd_mobilenet_v1_coco.config

Traceback (most recent call last): File "train.py", line 51, in <module> from object_detection.builders import model_builder File "C:\tensorflow1\models\research\object_detection\builders\model_builder.py", line 35, in <module> from object_detection.models import faster_rcnn_inception_resnet_v2_feature_extractor as frcnn_inc_res File "C:\tensorflow1\models\research\object_detection\models \faster_rcnn_inception_resnet_v2_feature_extractor.py", line 28, in <module> from nets import inception_resnet_v2 ModuleNotFoundError: No module named 'nets

note:
the slim folder is already added to PYTHONPATH !!!

[arascan35]

under slim folder use command 'python setup.py install' research/slim/setup.py

[cj-l]

under slim folder use command 'python setup.py install' research/slim/setup.py

This did it for me after nothing else worked, thanks!

[Tourenathan-G5organisation]

For google colab. Inside research folder.

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

thanks you very much. This solution just came al the right time. you save my day

[hoangnt2601]

I add this code sys.path.append("../../models/research/slim/") into /models/research/object_detection/models/faster_rcnn_inception_resnet_v2_feature_extrac
tor.py
It's worked for me

[devimonica]

@hoangnt2601 where exactly did you add the code ?

[wander1985]

For windows, create a text file in _your\Anaconda\environmentpath_\Lib\site-packages, rename it to tensorflow_model.pth, add the following two lines:
_your\Anaconda\environmentpath_\Lib\site-packagestensorflow\models\research
_your\Anaconda\environmentpath_\Lib\site-packagestensorflow\models\research\slim
Then save and close.
Do the test again. It works for me.

[libichong]

under slim folder use command 'python setup.py install' research/slim/setup.py
https://github.com/tensorflow/models/issues/5795#issuecomment-440562409

(python35) D:\Code\GitHub\models\research\slim>python setup.py build
running build
running build_py
creating build
error: could not create 'build': \u5f53\u6587\u4ef6\u5df2\u5b58\u5728\u65f6\uff0c\u65e0\u6cd5\u521b\u5efa\u8be5\u6587\u4ef6\u3002

delete BUILD file under the folder will solve it.

[anmol4210]

Hi,
I tried all the solutions but the one that worked for me was

export PYTHONPATH=$PYTHONPATH:pwd:pwd/research:pwd/research/slim

I ran the above command in the models folder...

Hope that helps someone......

[rikenshah2002]

For google colab. Inside research folder.

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

It worked by a small tweak: os.environ['PYTHONPATH'] += ':/:slim/'

Thanks a Lot!
It finally worked!

[luiscorujo]

copy the absolute path of your slim folder (we'll call it XYZ), which is inside models/research.
then type export PYTHONPATH=$PYTHONPATH:XYZ

[Conghung43]

under slim folder use command 'python setup.py install' research/slim/setup.py
https://github.com/tensorflow/models/issues/5795#issuecomment-440562409

(python35) D:\Code\GitHub\models\research\slim>python setup.py build
running build
running build_py
creating build
error: could not create 'build': \u5f53\u6587\u4ef6\u5df2\u5b58\u5728\u65f6\uff0c\u65e0\u6cd5\u521b\u5efa\u8be5\u6587\u4ef6\u3002

delete BUILD file under the folder will solve it.

It really work for me. Thank you!

[jsn-bln]

do i need to run this everytime ? because when i run train.py the error pops up again

[luiscorujo]

You can run from the models/research directory exactly this:
export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim
Just copy that line, just once, and it should work

[Danielson23]

The correct way, if you need only object detection module:

• copy object_detection and slim folder in a same location

• export the working path to pythonpath using this commande (this will be done in the folder where object_detection and slim folder ar located):

export PYTHONPATH=$PYTHONPATH:pwd:pwd/slim in this folder

[vatsal1301]

I slove this problem !!!!!!!!

open slim folder copy net folder and past in object _detection folder

[CHAMOD]

fo Windows users, you could solve this by setting PYTHONPATH as follows.
set PYTHONPATH=C:\Users...\models\research;C:\Users...\models\research\slim
then after run train.py inside object_detection dir.

[dhanpalrajpurohit]

Go to research/slim/ and fire this command
python setup.py build
sudo python setup.py install

[i-amgeek]

cd tensorflow/models/research . 

Check slim exists by ls | grep 'slim' 

Then run - export PYTHONPATH=$PYTHONPATH:$(pwd):$(pwd)/slim

[rsrrohit]

Go to research/slim/ and fire this command
python setup.py build
sudo python setup.py install

It worked for me on Google Colab!

[artongdou]

Go to research/slim/, delete BUILD, and run the following commands fix the 'nets' issue in windows 10 under anaconda virutal env.
python setup.py install

[yegortokmakov]

Changing PYTHONPATH manually doesn't seem like the right solution. Is there a reason why this repo is not using tf-slim package as a dependency and instead has a copy of the source code? Just trying to understand.

Update: There is a PR #8626 somehow related to this

Update2: Found the answer here https://github.com/google-research/tf-slim/issues/6

[rethamasae]

I can't run the Training, maybe someone can help me
(tensorflow1) C:tensorflow1\models\research\object_detection>python train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/faster_rcnn_inception_v2_pets.config
2020-08-10 22:08:00.754622: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudart64_101.dll
Traceback (most recent call last):
File "train.py", line 53, in
from object_detection.builders import model_builder
File "C:\Users\ASUS\anaconda3\envstensorflow1\lib\site-packages\object_detection-0.1-py3.8.egg\object_detection\builders\model_builder.py", line 39, in
from object_detection.protos import model_pb2
File "C:\Users\ASUS\anaconda3\envstensorflow1\lib\site-packages\object_detection-0.1-py3.8.egg\object_detection\protos\model_pb2.py", line 14, in
from object_detection.protos import center_net_pb2 as object__detection_dot_protos_dot_center__net__pb2
ImportError: cannot import name 'center_net_pb2' from 'object_detection.protos' (C:\Users\ASUS\anaconda3\envstensorflow1\lib\site-packages\object_detection-0.1-py3.8.egg\object_detection\protos__init__.py)

[khalidw]

On Windows 10 with Anaconda I used the following hack

import sys
import os
pwd = %pwd
sys.path.append(os.path.join(pwd,'models', 'research')) # point to your research dir
sys.path.append(os.path.join(pwd,'models', 'research', 'object_detection')) # point to your object_detection dir
sys.path.append(os.path.join(pwd,'models', 'research', 'slim')) # point ot your slim dir