Something is opening up and letting excess air into the engine, it could be the vacuum line to the headlights or the carb mounts or some other tapping into the inlet manifold.
If you can get the car to do it when stationary then squirt some (unlit!) butane/propane gas from a blowtorch around each joint and see if the idle speed changes. If it does then you’ve found the leak (there may be more than one)
I suspect you have a vacuum leak as others have said, but dont be too surprised if there is still a difference in idle between cold and hot.
I replaced the mounting O-rings, balanced the carbs carefully at idle (otherwise whichever is more closed will hold the other butterflies open through the linkage between the carbs); changed the throttle cable, and fitted an extra return spring. I am now happy there is no leak. But it still runs faster when hot. In my case, 600 rpm when cold, 900 when warm.
one other thing. depending on your distributor, if it starts to advance at idle, the adavnce will make it idle faster still, and exaggerate the effect. Use a timing light and check where your distributor starts to advance. Ideally set your idle (and timing if it’s wrong) so it doesnt start to advance on idle, but you’ll have to find any vacuum leaks first.
I believe it is quite normal for a warmed up engine to idle faster than when cold. After all, oil will thin out, so there is less bearing drag, running clearances will be re-established, and combustion mixture may be better, especially if the overall idle mixture is a little on the lean side to begin with. This warming up process may increase idle rpm 100-150 rpm. If at the slightly higher idle speed, ingnition timing advance begins, due to weak distributor springs, then engine idle speed could increase another 100 rpm or so.