Train Particular files in ML skills - Pipeline

Hello Community

Could you please let me know if a particular type of files can be trained and get scores during evaluation.
For example, Initialy we provide 10 files for training ML skills in Document understanding. Out of 10, 5 are PDFs and 5 files are JPEGs. I need to train only those JPEG files alone and get prediction score…

Thank you in advance for fast reply…

do you even want to process those PDFs?

I assume if they don’t contribute to your training then you cna filter well before they get loaded? Why to confuse rhe robot in his learning phase?

still thinking, why the PDFs will be needed to pass?

Thanks for the reply @rahulsharma

Am thinking of how ML model predicts when trained with 2 different file formats…

2 different pipelines for both…

i still have the same question. Why the training is provided via different pipeline for a format that won’t be even used for processing?

May be you csn give more idea about what you are trying to achieve using this usecase that might clear things out.

Always happy to connect and discuss if you wish to.