ML Skill - Pipeline Run

[rnahasnahasuddin (RN)]

Hi All,

I have created a project in AI Center and started data labelling a set of invoices and exported that into dataset. Done the pipeline train run for this batch and updated the ML Skill.

Now I have created another batch in the same project and exported that into same dataset.

My question is : If I train run pipeline for this particular batch and update the ML Skill version. Will it affect the set of invoices trained in pipeline earlier?
Do the recent pipeline trained batch will only be active?

[Anil_G (Anil Gorthi)]

@rnahasnahasuddin

Ideally as per recommendation you need to train the base version that is minor 0 with full both datasets together again

If you train the minor 1 with new batch then there might be differences in the batches which can effect the accuracy

Why To Choose Minor Version As 0 Or Base Model Of The Version?

Why should the minor version be chosen as 0 or base model of the version every time the ML Model is re-trained? As per the recommendation and the documentation that whenever new documents are added to increase the dataset size, it is recommended to "Always add Validation Station data to same dataset and train on ML Package minor version 0 (zero)". Read more on the Training High Performing Models The reason behind this is fallacy of Deep Learning pre trained models, where iterative trainin…

Choosing Correct Minor Version For Pipelines And Skill Deployments

Which is the correct Minor version of the ML package to choose when running pipelines and deploying ML Skill? When a package is newly deployed there is only minor version '0' present since there haven't been any pipelines executed yet. Training Pipeline: It is strongly recommended to always use minor version 0 (zero).Full Pipeline: As above, always choose minor version '0' while running a Full Pipeline. Evaluation Pipeline: Evaluation Pipeline is used to evaluate a trained machine learning …

Hope this helps

Cheers