Closed miranska closed 8 months ago
Hi @miranska !
The model file dt_output/dt_20240104144359_linear_dnn_nets/linear+dnn_nets.h5
that recorded in your log is exactly a path that relative to DeepTables's workdir
. You can check the workdir by code:
from hypernets.utils import fs
print(f"workdir: {fs.remote_root_}")
In my OS (centos 7.9), it's /tmp/workdir
, so I should check the files generated by deeptables in this directory.
For the warning below:
You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.
It does not prevent model persistence, if you want to eliminate this warning, just downgrade tensorflow<=2.11.0
For further infomation of model persistence, you can refer to this example:
@oaksharks, great thank you! So in order to persist a model in my own directory, I should first save it to fs.remote_root_
directory (which I can't change) and then move the model to my own directory using something like shutil.move()
, correct?
You can do that shutil.move()
or change the workdir
, see more https://github.com/DataCanvasIO/DeepTables/issues/85
Thank you!
System information
pip list
):Describe the current behavior
I run sample classification code from the documentation:
During execution, I see the following message in the logs:
Describe the expected behavior
I do not see any output in the
dt_output
subdirectory where my source file was located. Furthermore, I couldn't findlinear+dnn_nets.h5
anywhere on the hard drive.Standalone code to reproduce the issue