feifeibear / LLMSpeculativeSampling

Fast inference from large lauguage models via speculative decoding
Apache License 2.0
530 stars 51 forks source link

a change in the output of the model #22

Closed wenxin-zhu closed 1 year ago

wenxin-zhu commented 1 year ago

Hello. I noticed a change in the output of the model when using specific sampling. However, the original paper stated that the reasoning results of the model will not be changed.