eager mode flakiness in rocm inductor

eager.md

eager mode flakiness in MI300

	model	inductor_config	suite	mode	dtype	pass_rate	speedup	absolute_execution_time	comp_time	TS
561	alexnet	cpp_wrapper	torchbench	training	amp	eager_two_runs_differ	1.16	3.88	0	42006231922
563	alexnet	cudagraphs	torchbench	training	amp	eager_two_runs_differ	1.13	3.91	0	42006231922
565	alexnet	cudagraphs_dynamic	torchbench	training	amp	eager_two_runs_differ	1.1	4.0	0	42006231922
568	alexnet	default	torchbench	training	amp	eager_two_runs_differ	1.13	3.95	0	42006231922
1934	vgg16	aot_inductor	torchbench	inference	bfloat16	eager_two_runs_differ	1.13	1.28	0	42006231926
1935	vgg16	cpp_wrapper	torchbench	inference	bfloat16	eager_two_runs_differ	1.09	1.31	0	42006231926
1936	vgg16	cpp_wrapper	torchbench	training	amp	eager_two_runs_differ	1.27	24.35	0	42006231926
1937	vgg16	cudagraphs	torchbench	inference	bfloat16	eager_two_runs_differ	1.05	1.32	0	42006231926
1938	vgg16	cudagraphs	torchbench	training	amp	eager_two_runs_differ	1.26	24.54	0	42006231926
1939	vgg16	cudagraphs_dynamic	torchbench	inference	bfloat16	eager_two_runs_differ	1.07	1.3	0	42006231926
1940	vgg16	cudagraphs_dynamic	torchbench	training	amp	eager_two_runs_differ	1.24	24.96	0	42006231926
1941	vgg16	cudagraphs_freezing	torchbench	inference	bfloat16	eager_two_runs_differ	1.06	1.32	0	42006231926
1942	vgg16	default	torchbench	inference	bfloat16	eager_two_runs_differ	1.1	1.27	0	42006231926
1943	vgg16	default	torchbench	training	amp	eager_two_runs_differ	1.26	24.6	0	42006231926
1944	vgg16	export	torchbench	inference	bfloat16	eager_two_runs_differ	?	?	0	42006231926