Models: ModuleNotFoundError: No module named 'tensorflow.compat.v1'

[hzxie]

Prerequisites

Please answer the following questions for yourself before submitting an issue.

• [ ] I am using the latest TensorFlow Model Garden release and TensorFlow 2.
• [x] I am reporting the issue to the correct repository. (Model Garden official or research directory)
• [x] I checked to make sure that this issue has not already been filed.

1. The entire URL of the file you are using

https://github.com/tensorflow/models/tree/master/research/feelvos

2. Describe the bug

I got the following error when running bash train.sh.
The README doesn't tell us which TF to use.

The backward compatibility of TF is REALLY AWFUL.

Traceback (most recent call last):
  File "/mnt/lustre/xiehaozhe/Development/feelvos/train.py", line 24, in <module>
    from feelvos import model
  File "/mnt/lustre/xiehaozhe/Development/feelvos/model.py", line 58, in <module>
    from deeplab import model
  File "/mnt/lustre/xiehaozhe/Development/feelvos/deeplab/model.py", line 58, in <module>
    from deeplab.core import feature_extractor
  File "/mnt/lustre/xiehaozhe/Development/feelvos/deeplab/core/feature_extractor.py", line 21, in <module>
    import tensorflow.compat.v1 as tf
ModuleNotFoundError: No module named 'tensorflow.compat.v1'

3. Steps to reproduce

bash train.sh

4. Expected behavior

The program runs normally without raising ModuleNotFoundError: No module named 'tensorflow.compat.v1'.

5. Additional context

None

6. System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Cent OS 7.3
• Mobile device name if the issue happens on a mobile device: N/a
• TensorFlow installed from (source or binary): Binary
• TensorFlow version (use command below): 1.12.3
• Python version: 3.6.5
• Bazel version (if compiling from source): N/a
• GCC/Compiler version (if compiling from source): 5.4.0
• CUDA/cuDNN version: 9.0
• GPU model and memory: NVIDIA TITAN Xp / 12GB

[TannerGilbert]

If you're using Tensorflow 1.x I recommend you use the r1.13.0 branch.

[hzxie]

@TannerGilbert
What if I only have CUDA 9? TensorFlow 1.13 requires CUDA 10.
DON'T TELL ME TO UPGRADE TO CUDA 10.

[hzxie]

I solved this problem by

 git reset fe748d4a4a1576b57c279014ac0ceb47344399c4 --hard

[hzxie]

Again, the backward compatibility of TF is REALLY AWFUL.