How to read a score a batch of input sentences?

[oersoy1]

Hi, this is great and really interesting. Thanks for the great blog post. I have a large amount of misspelled input sentences. Is there a way of feeding them into a scoring function in batches and get the corrected output using the fine-tuned model instead of using fairseq-interactive from the command line? Since the training code is not available, I am not sure how to stitch it all together. Thanks.