Skip to content
CICD NeMo

Add attention_bias argument in transformer block and transformer layer modules, addressing change in MCore #6344

Sign in to view logs

Add attention_bias argument in transformer block and transformer layer modules, addressing change in MCore

Add attention_bias argument in transformer block and transformer layer modules, addressing change in MCore #6344

L2_Megatron_Change_Partitions_Increase_TP_Num_Partitions_-2_to_4-_and_PP_Num_Partitions_-1_to_2  /  main

L2_Megatron_Change_Partitions_Increase_TP_Num_Partitions_-2_to_4-_and_PP_Num_Partitions_-1_to_2 / main
succeeded Nov 15, 2024 in 1m 58s