Zobrazeno 1 - 3
of 3
pro vyhledávání: '"Ruiz, Carlos Riquelme"'
Autor:
Chen, Xi, Djolonga, Josip, Padlewski, Piotr, Mustafa, Basil, Changpinyo, Soravit, Wu, Jialin, Ruiz, Carlos Riquelme, Goodman, Sebastian, Wang, Xiao, Tay, Yi, Shakeri, Siamak, Dehghani, Mostafa, Salz, Daniel, Lucic, Mario, Tschannen, Michael, Nagrani, Arsha, Hu, Hexiang, Joshi, Mandar, Pang, Bo, Montgomery, Ceslee, Pietrzyk, Paulina, Ritter, Marvin, Piergiovanni, AJ, Minderer, Matthias, Pavetic, Filip, Waters, Austin, Li, Gang, Alabdulmohsin, Ibrahim, Beyer, Lucas, Amelot, Julien, Lee, Kenton, Steiner, Andreas Peter, Li, Yang, Keysers, Daniel, Arnab, Anurag, Xu, Yuanzhong, Rong, Keran, Kolesnikov, Alexander, Seyedhosseini, Mojtaba, Angelova, Anelia, Zhai, Xiaohua, Houlsby, Neil, Soricut, Radu
We present the training recipe and results of scaling up PaLI-X, a multilingual vision and language model, both in terms of size of the components and the breadth of its training task mixture. Our model achieves new levels of performance on a wide-ra
Externí odkaz:
http://arxiv.org/abs/2305.18565
Autor:
Komatsuzaki, Aran, Puigcerver, Joan, Lee-Thorp, James, Ruiz, Carlos Riquelme, Mustafa, Basil, Ainslie, Joshua, Tay, Yi, Dehghani, Mostafa, Houlsby, Neil
Training large, deep neural networks to convergence can be prohibitively expensive. As a result, often only a small selection of popular, dense models are reused across different contexts and tasks. Increasingly, sparsely activated models, which seek
Externí odkaz:
http://arxiv.org/abs/2212.05055
Autor:
Allingham, James Urquhart, Wenzel, Florian, Mariet, Zelda E, Mustafa, Basil, Puigcerver, Joan, Houlsby, Neil, Jerfel, Ghassen, Fortuin, Vincent, Lakshminarayanan, Balaji, Snoek, Jasper, Tran, Dustin, Ruiz, Carlos Riquelme, Jenatton, Rodolphe
Machine learning models based on the aggregated outputs of submodels, either at the activation or prediction levels, often exhibit strong performance compared to individual models. We study the interplay of two popular classes of such models: ensembl
Externí odkaz:
http://arxiv.org/abs/2110.03360