Nipype and fmri: single task paradigm

Hi, I am trying to get a first level analisys for my fMRI using nipype and fsl.
I have one run per subject (120 volumes, TR 3) with a single task paradigm (in feat I would select ArAr) with 30s Task and 30s Rest.

for the subject_info input of the SpecifyModel node i used this:

evs_run = Bunch(
                onsets=[list(range(0, int(TR*120), 60))],

and for contrasts input of the Level1Design node i used this:
cont1 = ['Task>Rest', 'T', ['Task'], [1]]

Can someone help me and check if my nipype setup is correct? I’m getting an error in the FILMGLS step (Warning: Number of voxels = 0. Spatial smoothing of autocorrelation estimates is not carried out) and I wonder if the problem is in the model design or in the preprocessing.