axlearn icon indicating copy to clipboard operation
axlearn copied to clipboard
verified publisher icon apple

expose max_decode_len and eos_token_id in decoding

Open gyin94 opened this issue 1 year ago • 1 comments

Expose

  • max_decode_len for beam_search to causal_lm module
  • eos_token_id for sample_decode to causal_lm module

It looks like both have been added unit test in decoding_test.py. This change is to expose the parameter to causal_lm and decoder module.

gyin94 avatar Feb 17 '24 09:02 gyin94

Hi Guoli, what's the use case? Should we first discuss in an internal PR?

sg. let's discuss in an internal PR firstly.

gyin94 avatar Feb 26 '24 05:02 gyin94