Met's biggest impact is reducing liver insulin resistance, which means that your liver gets better at knowing when to stop producing new glucose ("gluconeogenesis") because it's better at knowing how much is already running around in your blood.
If you lose weight and reduce visceral fat, hopefully your liver's insulin resistance reduces and the benefits of metformin dwindle away.
The easy rough'n'ready test for this is to check your waking/fasting BG level. If it's "normal" (so <= 5.6 mmol/l) then it probably means your liver hasn't been chugging out glucose all night & so it's probably not insulin resistant.
Metformin can also improve peripheral insulin resistance - ie muscle etc insulin resistance - and this has an impact on post-eating BG levels independent of liver insulin resistance. But the impact is usually pretty minor.
Also, for many people, it seems metformin can actually attenuate the benefits you can get to peripheral insulin resistance from exercise - less benefit with exercise+metformin than from exercise alone (due to ummm some complicated biochemistry stuff possibly involving mitochondria as well).
Anyway, after I zapped my T2D via weight loss I thought I might as well keep taking Met, until I found the exercise studies, and experiment showed that I did do slightly better without it after exercise. Dropping it "cold turkey" was just fine.