Hi @sgrvinod, thanks for your code.
When using beam search, how de we perform ensemble testing (testing multiple models and average predictions across models). Should we add all the log softmax output in the beam search fuction and then divide them by the number of models?
Hi @sgrvinod, thanks for your code. When using beam search, how de we perform ensemble testing (testing multiple models and average predictions across models). Should we add all the log softmax output in the beam search fuction and then divide them by the number of models?