ML skill Error: The model is getting successfully trained but not deploying

Hello,

Despite of training my model following the right steps its failing on deployment even though I have retrained the model if my package has changed.

This is the error log I’m getting

Hi @Nidhi_Baheti ,

Could you let us know if you are using the Trained Model version while deploying or creating the ML Skill ?

I am using the ‘Light text Classification model’ and the training and validation has run succesfully

@Nidhi_Baheti ,

I believe there must be an additional version created after training pipeline was successful, so perhaps maybe checking if minor version 1 is present and using it to create the ML Skill ?

Yes it worked for me, thank you!
what is the reason we do this though?

Yes it worked for me, thank you!
what is the reason we do this though?

@Nidhi_Baheti ,

The version 1 is the one where the model is trained with the data that you have provided and version 0 is the initial base version, if it is not a pre-trained model, it would need to be trained before using the model and hence the training generates the version 1. Some models need not be trained and can be used without training that is we can select the version 0 for deploying the ML Skill.

Oh, Got it now… Thanks alot

1 Like