Saeed Najafi

Beam search decoding is correct?

3

Hello, I am just curious how you implemented beam search decoding. Did you use the output logits of the default Seq2Seq model? I am working around these dialogue papers and...

Bug in Logit Processor?

1

Don't you have a bug in the logit processor? It seems you don't even use the processed scores in the wrapper. https://github.com/microsoft/LMOps/blob/96e7d9d086cba6cd878c98ac810902eae1323d8b/minillm/transformers/src/transformers/generation/utils.py#L2942