NickleDave / visual-search-nets

neural network models of visual search behavior
BSD 3-Clause "New" or "Revised" License
6 stars 2 forks source link

run VSD the same way as searchstims: not multi-label #41

Closed NickleDave closed 3 years ago

NickleDave commented 4 years ago

training should "prompt" network with class and output should be present / absent

NickleDave commented 4 years ago

documenting alternatives in #52 but I think using just softmax actually could be okay based on this paper: https://arxiv.org/pdf/1612.03663.pdf

they specifically use Pascal VOC (so I can cite it)

NickleDave commented 3 years ago

done, closing