Error message during Neural Network Training step

[Jaime Llodra]

Dear members of this group.

I am using the CNN for annotation of objects in a Serial Block Face dataset.

I am getting this error message after clicking the Launch button in the Train Neural Network screen

Using CPU...

loading particles...

2013 particles loaded, 2013 in training set, 0 in validation set

(2013, 128, 128)

Std of particles:  0.8236664

Setting up model...

Training...

WARNING:tensorflow:AutoGraph could not transform <function train_on_batch at 0x7f720d3f44d0> and will run it as-is.

Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.

Cause: No module named 'tensorflow_core.estimator'

Thanks in advance for any idea/suggestion.

Best,

Jaime Llodrá

[Ludtke, Steven J.]

My initial guess would be that you have an inconsistent installation of tensorflow, eg-

https://stackoverflow.com/questions/66022256/modulenotfounderror-no-module-named-tensorflow-core-estimator-for-tensorflow

Do a 'conda list | grep tensorflow' with the EMAN2 environment activated, and post it here if no obvious solution. Also the full output of e2version.py would be useful.

--------------------------------------------------------------------------------------

Steven Ludtke, Ph.D. <slu...@bcm.edu>                      Baylor College of Medicine

Charles C. Bell Jr., Professor of Structural Biology        Dept. of Biochemistry 

Deputy Director, Advanced Technical Cores                   and Molecular Biology

Academic Director, CryoEM Core

Co-Director CIBR Center

On Jun 24, 2022, at 6:25 AM, Jaime Llodra <jaime...@gmail.com> wrote:

No module named 'tensorflow_core.estimator'

[Jaime Llodra]

Thanks Steve for the reply.

I will get you the answer by Monday. I have no access to the computer at my workplace.

Have a nice weekend,

Jaime

--
--
----------------------------------------------------------------------------------------------
You received this message because you are subscribed to the Google
Groups "EMAN2" group.
To post to this group, send email to em...@googlegroups.com
To unsubscribe from this group, send email to eman2+un...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/eman2

---
You received this message because you are subscribed to the Google Groups "EMAN2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eman2+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/eman2/45A5256F-F1F2-4D54-BAEA-8342BB760AAD%40bcm.edu.

[Jaime Llodra]

Dear Steve,

Here is the output of conda list | grep tensorflow

tensorflow                           2.1.0           gpu_py37h7a4bb67_0  
tensorflow-base                2.1.0           gpu_py37h6c5654b_0  
tensorflow-estimator      2.6.0            pyh7b7c402_0  
tensorflow-gpu                  2.1.0            h0d30ee6_0 

And here is the output of e2version.py

EMAN 2.91 final ( GITHUB: 2021-03-08 11:36 - commit: 81caed2 )
Your Python version is: 3.7.9

Have a great day

Jaime Llodra

[Ludtke, Steven J.]

Hi Jaime,

Well, clearly your tensorflow-estimator is, indeed a different version. This is what mine looks like:

tensorflow                2.6.2           cuda110py39ha53fd0e_2    conda-forge

tensorflow-base           2.6.2           cuda110py39h1b3dc91_2    conda-forge

tensorflow-estimator      2.6.2           cuda110py39ha53fd0e_2    conda-forge

tensorflow-gpu            2.6.2           cuda110py39h5b0ac8e_2    conda-forge

though I am running a development build.

No way to say what caused this. As far as I know, tensorflow 2.6 should be ok with the older version of EMAN2 you're using. You could try either updating the other components to 2.6 or reverting the estimator back to 2.1.0. It is possible that something else you installed within the anaconda environment triggered the change ... no idea. eg -

conda install tensorflow-estimator=2.1.0

or

conda install tensorflow=2.6.0 tensorflow-base=2.6.0 tensorflow-gpu=2.6.0

--------------------------------------------------------------------------------------

Steven Ludtke, Ph.D. <slu...@bcm.edu>                      Baylor College of Medicine

Charles C. Bell Jr., Professor of Structural Biology        Dept. of Biochemistry 

Deputy Director, Advanced Technical Cores                   and Molecular Biology

Academic Director, CryoEM Core

Co-Director CIBR Center

To view this discussion on the web visit https://groups.google.com/d/msgid/eman2/f3eb3cee-0337-4097-9284-93ade7cd9511n%40googlegroups.com.

[Jaime Llodra]

Hello Steve,

I went with

conda install tensorflow-estimator=2.1.0

I got the downgrading message but I continued and installation was OK.

Training of the neural network proceeded without any problem.

Thanks Steve.

Have a good day,

Jaime Llodra

To view this discussion on the web visit https://groups.google.com/d/msgid/eman2/F988F377-3B12-4435-A012-CEDD28BD25DB%40bcm.edu.