Is this model trained on US dataset?
#2
by
ycffm
- opened
I am using this model to standardize the address text in my own project for addresses in Europe, Middle East and LATAM region. It seems that the model is prone to return an address with "US" in the end. Therefore I am wondering if it is mostly trained on US dataset.
Yes, the model was primarily trained on datasets with a US.
I m not sure where to find good data for the other regions