Spaces:

nielsr

/

donut-docvqa

like 77

Running

Hi,

This model is fine-tuned on DocVQA, a dataset of document-question pairs, so it's not really possible to ask follow-up questions in a conversational setup.

However if you mean sending a batch of images + questions through the model, that's possible. You can just provide a batch of pixel_values + decoder_input_ids to the generate method, and use the batch_decode method of the tokenizer to turn the generated ID's into text.

Could you provide an example on this??

However if you mean sending a batch of images + questions through the model, that's possible. You can just provide a batch of pixel_values + decoder_input_ids to the generate method, and use the batch_decode method of the tokenizer to turn the generated ID's into text.

Does anyone have an example on how to do this? Thanks.