Which is the tensorflow and kears version?

[zk19991208]

Hello, I read your article and am very interested in your work. Can you share your keras and tensorflow versions? I would like to try running your test code.

[markveilletteLL]

Thanks for the question. We used version 2.13 of tensorflow, and version 2.13.1 of keras to train the model. However, the code and pretrained model should work in newer versions as well.

I will update the requirements file to reflect this.

[LiAlH4]

I have to comment that actually only tensorflow==2.13.1 will work for the jupyter notebook. At least the script notebooks/ModelTraining_Tensorflow.ipynb will face a compiling error for the command model.fit(...) on my local machine.

[markveilletteLL]

Thanks for letting us know -- can you please post the error? We can work on making it backwards compatible

[comlor]

I am running into errors as well during training. I am running the command in the readme.md in the repo.
python scripts/tornado_detection/train_tornado_tf.py scripts/tornado_detection/config/params.json

I am able to successfully run the Model_Train_Tensorflow notebook without issue on the same below configuration

Python environment details:
UPDATE: I scrapped by python venv and rebuilt the environment. I noticed prevously that tensorflow was not utilizing my GPU. I pulled the latest versions of tensorflow and keras and updated below details.

(venv) user@laptop:~/Documents/Projects/tornet$ pip3 show tensorflow
Name: tensorflow
Version: 2.16.1
Summary: TensorFlow is an open source machine learning framework for everyone.
Home-page: https://www.tensorflow.org/

(venv) user@laptop:~/Documents/Projects/tornet$ pip3 show keras
Name: keras
Version: 3.3.3
Summary: Deep learning for humans.
Home-page: https://keras.io/

(venv) user@laptop:~/Documents/Projects/tornet$ python3 --version
Python 3.10.12

This is the error i receive running the Train CNN Baseline Model:

2024-05-12 13:30:12.591535: W tensorflow/core/framework/local_rendezvous.cc:404] Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence
Traceback (most recent call last):
  File "~/Documents/Projects/tornet/scripts/tornado_detection/train_tornado_tf.py", line 225, in <module>
    main(config)
  File "~/Documents/Projects/tornet/scripts/tornado_detection/train_tornado_tf.py", line 135, in main
    nn = build_model(shape=in_shapes,
  File "~/venv/tornet/lib/python3.10/site-packages/tornet/models/tf/cnn_baseline.py", line 59, in build_model
    x,c = vgg_block(x,c, filters=start_filters,   ksize=3, l2_reg=l2_reg, n_convs=2, drop_rate=0.1)   # (60,120)
  File "~/venv/tornet/lib/python3.10/site-packages/tornet/models/tf/cnn_baseline.py", line 95, in vgg_block
    x,c = CoordConv2D(filters=filters,
  File "~/venv/tornet/lib/python3.10/site-packages/keras/src/utils/traceback_utils.py", line 122, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/usr/lib/python3.10/inspect.py", line 3186, in bind
    return self._bind(args, kwargs)
  File "/usr/lib/python3.10/inspect.py", line 3107, in _bind
    raise TypeError('too many positional arguments') from None
TypeError: too many positional arguments

As an aside. During the initial setup and install. After installing the requirements and running pip3 install .. The entire module did not install into my site package. I had to go through the directory structure and add "__init__.py" file in a bunch of locations for the module to fully install correctly.

[markveilletteLL]

@comlor Apologies, in a recent push I forgot to commit a file. the too many positional arguments error should now be fixed in c5fa03c