Hi , I have developed a process to train document using action Center.
It’s working in Studio Debug Mode using following steps:
Target item available in uipath Q
Once I started Debug in Studio
Target item will be picked by the bot and turned status as inprogress
Data moving into Storage bucket
Documents are appears in Action Center in Unassigned
I manullay pulled to Pening and validated and submitted
Studio Debug status was turned into Resumed and again I clicked on Resumed , it was processed all other steps (training data sets and marking queue items ) successfully.
The Problem I am facing , when I start running in robot tray or unattended mode it was showing resumed in Jobs and robot running icon disappeared but nothing happened after submitting documents. Any idea?
When Training/Retraining the models, it is required to choose the export folder, as the fine-tune folder only contains only the data submitted through the ML Trainer activity and not the whole of the Dataset.
The Image doesn’t show Auto Training being enabled using the environment variable, Do Check the post below. It contains most of the clarifications to your questions.
The Automatic Update of ML Package retrained version in ML Skills can be applied by Managing the ML Skill by Clicking on Modify Current Deployment on the ML Skill and Enabling Auto Update.
If the concern is Training on which minor version, then whatever is the selected version it will retrain on it. But there is an explanation provided in the documentation as the version to be chosen is always recommended to be 0 (Base version).
This case is also often the real case scenario, hence this should be fine. But understanding the documentation will help you realise that it will not only be these documents, it will be these new Data + Old Data that the Training Pipeline will be executed on.