The idea is that the (entirely hypothetical) terraforming technology would accomplish this in something like centuries or tens of centuries while the processes

The idea is that the (entirely hypothetical) terraforming technology would accomplish this in something like centuries or tens of centuries while the processes that led to the loss of the atmosphere took millions or tens of millions of years - similar to how the Earth's atmosphere doesn't just whoosh away during geomagnetic reversal. Another thing to keep in mind is that what atmosphere Mars had, when, how it lost it is something actively studied so understanding of it might still change. There's a probe (Maven) launching late this year that's going to measure the rate of loss due to solar wind, among other things - it will have an elliptical orbit taking several 'low' (150km or so) passes over Mars.