das icon indicating copy to clipboard operation
das copied to clipboard

DAS loading model error

Open emory-saba opened this issue 11 months ago • 5 comments

Hello,

Thank you all for all of the help you've provided! I'm running into an issue where when I try to load a model I've previously trained, I get the following error. Any ideas on how to resolve this?

Image

emory-saba avatar Feb 05 '25 23:02 emory-saba

Hello! Just wanted to push this issue again, if you all have any ideas as to what this might be? I spent some time in the keras source code and was not able to diagnose the issue or see any obvious issues that would have changed between the last time I ran this model and now (only a month or two)

emory-saba avatar Feb 11 '25 15:02 emory-saba

Hi, sorry about the late reply. Looks like this model was trained a while ago. The latest version of DAS uses a new keras version. Should be compatible but maybe not. I guess this does not happen with a model trained with the current version. Can you try with an older version of DAS?

postpop avatar Feb 11 '25 16:02 postpop

I tried both training a new model as well as loading older versions of DAS (up to 0.30.1) and unfortunately still encountered the same issue

emory-saba avatar Feb 17 '25 15:02 emory-saba

It just tested the latest version of DAS with a zebra finch dataset - training and predicting work as expected.

Could you share your training config with me? You can save all network settings from the dialog through which you start the training to a file via the "Save config" button of that dialog. This will save a yaml file which I can then use to reproduce your error.

And just to be safe, could you also share the output of das version?

Thanks!

postpop avatar Feb 18 '25 12:02 postpop

Hello! I apologize for the delay, I was traveling!

The das version I have been working with is 0.32.5

I have attached the yaml file for the model I most recently trained here - please let me know if the file is not attached for some reason or if there is anything else I can provide to help solve this.

Thank you!

On Tue, Feb 18, 2025 at 7:59 AM Jan Clemens @.***> wrote:

It just tested the latest version of DAS with a zebra finch dataset - training and predicting work as expected.

Could you share your training config with me? You can save all network settings from the dialog through which you start the training to a file via the "Save config" button of that dialog. This will save a yaml file which I can then use to reproduce your error.

And just to be safe, could you also share the output of das version?

Thanks!

— Reply to this email directly, view it on GitHub https://github.com/janclemenslab/das/issues/82#issuecomment-2665647530, or unsubscribe https://github.com/notifications/unsubscribe-auth/BL2UWAQLEQLEU7ZZ2QOM2U32QMVENAVCNFSM6AAAAABWSITJB2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNRVGY2DONJTGA . You are receiving this because you authored the thread.Message ID: @.***> [image: postpop]postpop left a comment (janclemenslab/das#82) https://github.com/janclemenslab/das/issues/82#issuecomment-2665647530

It just tested the latest version of DAS with a zebra finch dataset - training and predicting work as expected.

Could you share your training config with me? You can save all network settings from the dialog through which you start the training to a file via the "Save config" button of that dialog. This will save a yaml file which I can then use to reproduce your error.

And just to be safe, could you also share the output of das version?

Thanks!

— Reply to this email directly, view it on GitHub https://github.com/janclemenslab/das/issues/82#issuecomment-2665647530, or unsubscribe https://github.com/notifications/unsubscribe-auth/BL2UWAQLEQLEU7ZZ2QOM2U32QMVENAVCNFSM6AAAAABWSITJB2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNRVGY2DONJTGA . You are receiving this because you authored the thread.Message ID: @.***>

emory-saba avatar Mar 10 '25 17:03 emory-saba