Hi,
I am currently working on a new Use-case within AI Center that revolves around training and test data as Excel files. My data has been already loaded. For the next step for data labeling. is this required? what does this step entail if I want to use text classification model for out-of-box ML Packages.
Hey @ashton.lalchan
Data labeling is a crucial step for training ML models, especially for tasks like text classification. The purpose of data labeling is to provide a dataset with pre-defined labels that the machine learning model can learn from.
Hi @ashton.lalchan ,
Could you let us know the Name of the OOB Model that you are trying to use ?
For the English Text Classification, we could go ahead with the CSV or Json file format for training the Models/Packages directly and then deploying the ML Skill after the training pipeline is complete. There is no need for labelling the data as we do for Invoices or custom document types since we have the training data already in csv format.
For more on the English Text Classification model, you could check the docs below :
hi @pikorpa,
Thanks for this info, however I am training with excel files and not pdfs,text files etc…would I have to “choose from OOB labeling template” ? if so, how would that be done since I already created datasets for training and test data.
hi @supermanPunch,
I an using English text classification model from OOB.
Is the Training data in CSV format and is as per the documentation provided ? If so, you could go ahead with creating a training pipeline for the package.
No, the files are saved as .xlsx files to .csv.
Will this work or no?
Do save the files as CSV and then use them to perform the Training for the Package. As it does mention that A CSV or Json format is the required data set format.
Hi @supermanPunch
So I have changed the training files to csv however whenever I run the pipeline, they fail for the ML model english text classification. is there a reason why?