Document Understanding model with 1000 vendors

Help Document Understanding
, ,
1

I have a question regarding model training when we have huge number of different vendors (~1000).

It seems unrealistic to label 10-15 documents for each vendor in this case - dataset will be huge and it will require a lot of effort to label, also training will take very long time.

What would be the best approach in this situation?
Anyone had this issue and was able to find solution that worked?

2

@Karolina_Debowska

Now you have a new model docpath which is for a situation like this

Cheers

3

Thanks, but I think it won’t help in this situation.

I look for a solution in already existing project (we already have trained model on 800+ documents) that we want to expand with new vendors, and at the end the number of vendors will be ~1000

(sorry didn’t mentioned that in original post :stuck_out_tongue:)