Minibatch calls into GPT2 #3
Description
Right now each phrase is run through the GPU one at a time. This is not efficient. Combine them into minibatches
{{ message }}
Right now each phrase is run through the GPU one at a time. This is not efficient. Combine them into minibatches