The most common method used is to add a non solvent to the QD dispersion and precipitate the QDs to extract them (As Alexander suggested). The downside of this is that you may also remove some surfactants from the QDs in the process resulting in non-passivated surfaces which acts as traps, there by reducing the emission quantum yield.
DMF is a Miscible solvent. To precipitate the QDs, you need to add an immiscible solvent such as cyclohexane, heptane, hexane, pentane. Then, to collect the QDs do centrifugation. Please do this by taking small portions of QDs-DMF solution.
Add a solvent (50 % by volume) which is miscible with DMF, but can not disperse the QDs (or non solvent for QDs) and then centrifuge the entire solution.
Sorry about the late response, and I see that your question is already answered Jeotikanta. Indeed what I meant is that you add a solvent in which the QDs are not not soluble. Hexane should do the trick. To test whether it works, first take 1 mL of your solution and add 1- 2mL of Hexane to it. You will see the solution turning turbid, if everything works as expected. Centrifuge the solution and throw out the supernatant. The remaining precipitate is your QD. You might also want to play around with other nonsolvents and also with the amount of the nonsolvent to optimize your yield.