[QNN] Fix LiftConstantScalarOperands to handle aten.pow.Scalar (#18994) #3082

Job	Run time
test-xnnpack / test-backend-linux (xnnpack_static_int8_per_tensor, models) / linux-job	25m 55s
test-xnnpack / test-backend-linux (xnnpack_dynamic_int8_per_channel, models) / linux-job	18m 33s
test-xnnpack / test-backend-linux (xnnpack, operators) / linux-job	14m 25s
test-xnnpack / test-backend-linux (xnnpack_static_int8_per_channel, models) / linux-job	26m 43s
test-xnnpack / test-backend-linux (xnnpack_dynamic_int8_per_channel, operators) / linux-job	18m 31s
test-xnnpack / test-backend-linux (xnnpack_static_int8_per_tensor, operators) / linux-job	27m 16s
test-xnnpack / test-backend-linux (xnnpack, models) / linux-job	52m 24s
test-xnnpack / test-backend-linux (xnnpack_static_int8_per_channel, operators) / linux-job	26m 16s
test-xnnpack / test-backend-macos	-1s
test-xnnpack / package-golden-artifacts	5m 25s
	3h 35m 27s

Provide feedback