DeepSpeed

bd71eed6 - fix gemma4 num attention head bugs (from #7975) (#7990)

fix gemma4 num attention head bugs (from #7975) (#7990) This PR is based on #7975 and fix CI errors. Thanks for @mingxiang1006 for providing the fix. --------- Signed-off-by: Guokai Ma <guokai.ma@intel.com>