Skip to content

Update llama32 vision (mllama) use attention bias #6560

Update llama32 vision (mllama) use attention bias

Update llama32 vision (mllama) use attention bias #6560

L2_Megatron_Change_Partitions_Reduce_TP_Num_Partitions_-2_to_1-_and_PP_Num_Partitions_-1_to_2  /  main

succeeded Nov 23, 2024 in 2m 1s