megatron.model.language_model.parallel_lm_logits#

megatron.model.language_model.parallel_lm_logits(input_, word_embeddings_weight, parallel_output, bias=None)#: LM logits using word embedding weights.