Models: Failed to run optimizer ArithmeticOptimizer

[sed0724963]

I had upgrade my tensorflow object detection API, because I want to plot the validation loss on tensorboard,In the Old version is only plot the total loss for training .

and I'm use the macos system , tensorflow is cpu version.
the different on new and old version is : train.py and model_main.py.

now, I am use the model_main.py ,but it's have a error,anyone knows it?

---

System information

• What is the top-level directory of the model you are using:
  ./models/research
• Have I written custom code (as opposed to using a stock example script provided in TensorFlow):
  NO
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
  MacOS High Sierra
• TensorFlow installed from (source or binary):
  pip3
• TensorFlow version (use command below):
  1.13.0-rc2
• Bazel version (if compiling from source):
  don't know
• CUDA/cuDNN version:
  no
• GPU model and memory:
  cpu only
• Exact command to reproduce:
  don't know

[Shashi2411]

I'm experiencing the same issue.
Have you solved it?

[sed0724963]

@Shashi2411 Not yet

[v-qjqs]

I'm experiencing the same issue on win notebook with py3.7, tf-gpu1.13, cuda10

[abh2424]

I am still facing this issue. Please help me to solve the error.

[mttbx]

Me too. It seems the preprocessing problem, I can still train my network though.

[Tdoe4321]

I as well am having this issue with a faster_rcnn network, anyone have any ideas about it yet?

[mehtanaman-awl]

I have the issue, but my Memory usage and Swp are used 100%. And the system hangs(I'm guessing it may as well be due to some other reason, but not sure why the system hangs, kswapd0 consumes 50% of memory).
Ubuntu 18.04, 16GB RAM, tf-gpu 13.1.0, cuda 10, python 3.6