What is the difference between running the full pipeline and running evaluation and training pipeline separately in UiPath ai center?
What is the best practise?
Generally it is good to run fullpipeline…for debugging if issue in training the we go with seaprate
also to check model accuracy we need evaluation pipeline…to check how better model is…
in cases like retraining autoamtically by uploading new data after validation and all we use train pipeline alone…
similarly for random analysis we use evaluation pipeline
cheers
This will save your time to run Training & Evaluation Pipelines one after another.
Will give you more control when to train model with more data and when to evaluate.
This will depend on your model purpose and architecture.
If your model is mission critical and should be trained on each and every exception then better to use Full Pipeline.
If your model have a threshold like if confidence score dropped below 60-70 % then start training with the provided data and then run evaluation pipeline.
Thanks,
Ashok 
This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.