Tensorflow OD API - Tensorboard doesn't load new data

Prerequisites

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

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

1. The entire URL of the file you are using

python3 model_main_tf2.py --pipeline_config_path=/models/tf2/20200814_ssd_resnet50_v1_fpn/pipeline.config --model_dir=/model/20200818_ssd_resnet50_v1_fpn

For monitoring:
tensorboard --logdir=/model/20200818_ssd_resnet50_v1_fpn

2. Describe the bug

I am training on a remote machine. To see the tensorboard on my local computer i use local port forwarding:
ssh -L 6006:localhost:6006 myusername@remotemachine

I did this earlier with TF1.15 and everything worked fine. But now I am training with TF2 and the tensorboard isn't refreshing. That means, that i still see the old data from the TF1.15 training. I am not sure if the port is caching somehow or what i do wrong so that the tensorboard won't show my new training data.

3. Steps to reproduce

Training itself works fine. I can see the loss decreasing in the log.

Tensorboard is the problem.

4. Expected behavior

Tensorboard should display my training process.

5. Additional context

6. System information

Remote Machine:
Linux
TensorFlow installed from pip 2.3.0
Python version: 3.6.9
CUDA 10.1
GPU model: GeForce GTX 1080 Ti

Local Computer:
Windows