For the latter, unfortunately, we do not have any data for India. The model we’ve prepared for field delineation (see also accompanying blog post) was trained on high quality reference data (GSAA data provided by national paying agency, containing vector data of declared agricultural fields).
Unfortunately, such data is rarely open, and often does not exist. If that is the case for you, you have two possible paths forward (possibly more, but these two come to my mind):
create the dataset yourself by manually drawing polygons (beware: this is costly and quite difficult to do, but done properly you can get very high quality data)
train the model someplace else - where you can get reference data - apply it to your AOI, and then fine tune the model by adding manually drawn polygons (only) where the predictions are most wrong
I hope this helps. Kind regards and keep us posted.