Renamed function is not updated when imported

[UriShavit]

Hello @muellerdo ,
First of all, thank you for sharing your knowledge and your experience.
I saw that this project is a modification of an older project you worked on.
In:

MIScnn/miscnn/utils/callback.py

Line 25 in e835371

from miscnn.data_io import save_evaluation

you are trying to import save_evaluation.
I think you meant to import backup_evaluation.

Hope I'm not misleading.

Cheers,
Uri

[muellerdo]

Hey @UriShavit,

thanks for pointing this out! :)

This callback is deprecated code, as you already said correctly, from an older project in which I started to create MIScnn.

I have removed this code in the dev branch, now, and will merge this update into master the next days.

I personally use and highly recommend using the Keras callback CSVLogger:
https://www.tensorflow.org/api_docs/python/tf/keras/callbacks/CSVLogger

It basically is doing the same thing as my old callback, but the metrics are not hardcoded and it is the officially supported callback from the Keras team.

Example usage:

# Import stuff...
from tensorflow.keras.callbacks import CSVLogger

# Define Callbacks
cb_csv = CSVLogger("my_logs.csv")
callback_list = [cb_csv]

# Run training
model.train(sample_list, epochs=20, callbacks=callback_list)

First of all, thank you for sharing your knowledge and your experience.

Thank you. I would be glad if you can use MIScnn in your research.

Cheers,
Dominik

[UriShavit]

Hey, Thank you for your elaborated response. I am doing a project where I'm trying to reproduce your results as a first stage. Afterwards, I'll try to add data of early stages cysts to improve the segmentation for this stages as well. I'm hoping that you might help me - I tried to download the official DB challenge, but the files seem to be corrupted.... Link to the source I used: https://gitee.com/tanjiale/kits19/tree/master/data Did you download the files from the same link? Uri Sasson Shavit אורי ששון שביט

…

On Mon, Jul 6, 2020 at 1:40 PM Dominik Müller ***@***.***> wrote: Hey @UriShavit <https://github.com/UriShavit>, thanks for pointing this out! :) This callback is deprecated code, as you already said correctly, from an older project in which I started to create MIScnn. I have removed this code in the dev branch, now, and will merge it into master the next days. I personally use and highly recommend using the Keras callback CSVLogger: https://www.tensorflow.org/api_docs/python/tf/keras/callbacks/CSVLogger It basically is doing the same thing as my old callback, but the metrics are not hardcoded and it is the officially supported callback from the Keras team. Example usage: # Import stuff...from tensorflow.keras.callbacks import CSVLogger # Define Callbackscb_csv = CSVLogger(os.path.join(fold_subdir, "logs.csv"), separator=',', append=True)callback_list = [cb_csv] # Run trainingmodel.train(sample_list, epochs=20, callbacks=callback_list) First of all, thank you for sharing your knowledge and your experience. Thank you. I would be glad if you can use MIScnn in your research. Cheers, Dominik — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#29 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AF7HVHY3BKFO5L5FIQB6NFLR2GS2HANCNFSM4OQMQ2UA> .

[muellerdo]

I am doing a project where I'm trying to reproduce your results as a first
stage.

Afterwards, I'll try to add data of early stages cysts to improve the
segmentation for this stages as well.

Good plan!

I'm hoping that you might help me - I tried to download the official DB
challenge, but the files seem to be corrupted....

Link to the source I used:

https://gitee.com/tanjiale/kits19/tree/master/data

Did you download the files from the same link?

Your provided link is not the official repository for the KiTS19 data set.

Try out this link: https://github.com/neheller/kits19
Run the following code for downloading the data:

git clone https://github.com/neheller/kits19
cd kits19
pip3 install -r requirements.txt
python3 -m starter_code.get_imaging

Cheers,
Dominik