How to get the result from torch.multiprocessing

It’s suggested using pytorch multiprocessing insted of multiprocessing from python ,However in python3 I can use something like

def toy_fun(boxes):
    boxes += 20
    return boxes

with concurrent.futures.ProcessPoolExecutor as executor : 
    results = [ executor.submit(toy_fun,boxes) for _ in range(10) ] 
    
    for f in concurrent.futures.as_completed(results):
        #do something with f.result()

However I do not know how can I process or get the result get from method using in MULTIPROCESSING BEST PRACTICES

How I can I do this ? Thanks!