Skip to content

How to perform text infilling #438

@oierIM

Description

@oierIM

I'm attempting to perform text infilling, but it doesn't seem to work. I'm not sure if the input that I am providing is correct or not, or how to change it so that it works propertly. Here is the code that I'm using:

from transformers import OFAModel, OFATokenizer

model = OFAModel.from_pretrained('path', use_cache=True)
tokenizer = OFATokenizer.from_pretrained('path')

txt = ' what is the complete text of "A <mask> woman"?'

inputs = tokenizer([txt], return_tensors="pt").input_ids
gen = model.generate(inputs)

If I print 'gen', I get the following: ['\x1a "what's the woman"?"? "what's the woman"?"? "what's']

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions