no code implementations • 17 Mar 2022 • Lina Achaji, Thierno Barry, Thibault Fouqueray, Julien Moreau, Francois Aioun, Francois Charpillet
Finally, we justify our usage of the parallel decoding technique by showing that the trajectory prediction task can be better solved as a non-autoregressive task.
no code implementations • 16 Jul 2021 • Lina Achaji, Julien Moreau, Thibault Fouqueray, Francois Aioun, Francois Charpillet
Interestingly, we showed that pre-training our Transformer model on the CP2A dataset and then fine-tuning it on the PIE dataset is beneficial for the action prediction task.