[TorchToTMTenor] enable broadcasting query- and key_seq_len of attention mask#4464

ziereis wants to merge 2 commits intollvm:mainllvm/torch-mlir:mainfrom

ziereis:fix_broadcast_attention_maskziereis/torch-mlir:fix_broadcast_attention_maskCopy head branch name to clipboard