aharley / pips

Hi,

Section 3.7 of the paper says "To lock the model into tracking the “original” target, we simply re-use the original F0 across all re-initializations."

But in chain_demo.py, the query point's feature representation seems not locked?

line 54  outs = model(traj_e[:,cur_frame].reshape(1, -1, 2), rgb_seq, iters=6, feat_init=feat_init, return_feat=True)
line 57  feat_init = outs[3]

Best

Hi, it will always be fixed on the initial ffeat, you may confuse it with ffeats

pips/nets/pips.py

Line 463 in bde038a

    
           ffeat = utils.samp.bilinear_sample2d(fmaps[:,0], coords[:,0,:,0], coords[:,0,:,1]).permute(0, 2, 1) # B, N, C

pips/nets/pips.py

Line 609 in bde038a

return coord_predictions, coord_predictions2, vis_e, ffeat, losses

The return term in L609 is ffeat, not ffeats.

Thanks, you are right. I did confuse it with feats.

Feature representation not locked in test-time trajectory linking?