Zobrazeno 1 - 10
of 15 312
pro vyhledávání: '"Pagnoni, A."'
RLHF-aligned LMs have shown unprecedented ability on both benchmarks and long-form text generation, yet they struggle with one foundational task: next-token prediction. As RLHF models become agent models aimed at interacting with humans, they seem to
Externí odkaz:
http://arxiv.org/abs/2407.02446
Publikováno v:
Il Foro Italiano, 2002 May 01. 125(5), 1479/1480-1481/1482.
Externí odkaz:
https://www.jstor.org/stable/23198362
We present QLoRA, an efficient finetuning approach that reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU while preserving full 16-bit finetuning task performance. QLoRA backpropagates gradients through a frozen, 4-bi
Externí odkaz:
http://arxiv.org/abs/2305.14314
Autor:
Francesca Porto, Sara Cisternino, Emiliano Cazzola, Giorgia Speltri, Liliana Mou, Alessandra Boschi, Lorenza Marvelli, Giovanni Di Domenico, Antonella Pagnoni, Lucia De Dominicis, Irene Calliari, Claudio Gennari, Licia Uccelli, Gaia Pupillo, Giancarlo Gorgoni, Juan Esposito, Petra Martini
Publikováno v:
EJNMMI Radiopharmacy and Chemistry, Vol 9, Iss 1, Pp 1-19 (2024)
Abstract Background The integration of positron emission tomography (PET) and magnetic resonance imaging (MRI) holds promise for advancing diagnostic imaging capabilities. The METRICS project aims to develop cyclotron-driven production of 52Mn for PE
Externí odkaz:
https://doaj.org/article/b83fd672a65c4c9eba796db6c1f9dc0d
Autor:
Renzo Guarnieri, Rodolfo Reda, Dario Di Nardo, Gabriele Miccoli, Francesco Pagnoni, Alessio Zanza, Luca Testarelli
Publikováno v:
Journal of Dental Research, Dental Clinics, Dental Prospects, Vol 18, Iss 2, Pp 135-142 (2024)
Background. This study evaluated the gingival crevicular fluid (GCF) and Peri- implant crevicular fluid (PICF) concentrations of interleukin-1 beta (IL-1β), interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α) and active metalloproteinase-8 (a
Externí odkaz:
https://doaj.org/article/efc48577c9e543baa26192d009e6380a
In long document controllable summarization, where labeled data is scarce, pretrained models struggle to adapt to the task and effectively respond to user queries. In this paper, we introduce Socratic pretraining, a question-driven, unsupervised pret
Externí odkaz:
http://arxiv.org/abs/2212.10449
Autor:
Sowmya SV, Dominic Augustine, Jagdish Hosmani, Francesco Pagnoni, Rodolfo Reda, Luca Testarelli, Shankargouda Patil
Publikováno v:
Frontiers in Dental Medicine, Vol 5 (2024)
IntroductionNanoparticles have orchestrated a paradigm shift in the landscape of cancer diagnosis and therapy, presenting a multifaceted approach to tackle the intricacies of malignancies. This comprehensive exposition delves deep into the forefront
Externí odkaz:
https://doaj.org/article/dca2cb8f5c004e028117c7b212f13fca
Autor:
Marco de Pinto, Francesca Coppi, Amelia Spinella, Gianluca Pagnoni, Vernizia Morgante, Pierluca Macripò, Matteo Boschini, Anna Francesca Guerra, Francesca Tampieri, Ottavio Secchi, Martina Orlandi, Gabriele Amati, Federica Lumetti, Gilda Sandri, Rosario Rossi, Giuseppe Boriani, Anna Vittoria Mattioli, Clodoveo Ferri, Dilia Giuggioli
Publikováno v:
Frontiers in Cardiovascular Medicine, Vol 11 (2024)
IntroductionReduced TAPSE/sPAP ratio has recently emerged as a predictive parameter risk factor for PH, however its role in SSc has been poorly investigated. The aim of the study was to investigate the prognostic value of the TAPSE/sPAP ratio for the
Externí odkaz:
https://doaj.org/article/57c52c2ede4d498997e93639fe755fa5
This paper investigates models of event implications. Specifically, how well models predict entity state-changes, by targeting their understanding of physical attributes. Nominally, Large Language models (LLM) have been exposed to procedural knowledg
Externí odkaz:
http://arxiv.org/abs/2211.05392