OpenSeq2Seq icon indicating copy to clipboard operation
OpenSeq2Seq copied to clipboard

Published 20 hours ago • verified publisher icon NVIDIA

Phrase input in seed tokens

Open sruteesh-pivot opened this issue 7 years ago • 2 comments

Hi, How to input phrase in seed tokens eg: ['this is test for', this is another'] to generate new sentences. Also how do i get the softmax probabilities of new generated word or sentence.

Thanks

sruteesh-pivot avatar Nov 27 '18 10:11 sruteesh-pivot

Is this a question for our lm models? Perhaps @chiphuyen or @RaymondNie would be better suited to answer.

blisc avatar Jan 10 '19 02:01 blisc

@chiphuyen and @RaymondNie Any updates? I couldn't find any built-in function to calculate sentence probabilities. I am trying to built this functionality from the output_values in https://github.com/NVIDIA/OpenSeq2Seq/blob/master/open_seq2seq/models/lstm_lm.py#L190. Is there any function already available?

sruteesh-pivot avatar Jan 25 '19 13:01 sruteesh-pivot