Zobrazeno 1 - 10
of 277
pro vyhledávání: '"Siriwardhana P"'
Merging in a Bottle: Differentiable Adaptive Merging (DAM) and the Path from Averaging to Automation
Autor:
Gauthier-Caron, Thomas, Siriwardhana, Shamane, Stein, Elliot, Ehghaghi, Malikeh, Goddard, Charles, McQuade, Mark, Solawetz, Jacob, Labonne, Maxime
By merging models, AI systems can combine the distinct strengths of separate language models, achieving a balance between multiple capabilities without requiring substantial retraining. However, the integration process can be intricate due to differe
Externí odkaz:
http://arxiv.org/abs/2410.08371
Autor:
Siriwardhana, Shamane, McQuade, Mark, Gauthier, Thomas, Atkins, Lucas, Neto, Fernando Fernandes, Meyers, Luke, Vij, Anneketh, Odenthal, Tyler, Goddard, Charles, MacCarthy, Mary, Solawetz, Jacob
We conducted extensive experiments on domain adaptation of the Meta-Llama-3-70B-Instruct model on SEC data, exploring its performance on both general and domain-specific benchmarks. Our focus included continual pre-training (CPT) and model merging, a
Externí odkaz:
http://arxiv.org/abs/2406.14971
Autor:
Goddard, Charles, Siriwardhana, Shamane, Ehghaghi, Malikeh, Meyers, Luke, Karpukhin, Vlad, Benedict, Brian, McQuade, Mark, Solawetz, Jacob
The rapid expansion of the open-source language model landscape presents an opportunity to merge the competencies of these model checkpoints by combining their parameters. Advances in transfer learning, the process of fine-tuning pretrained models fo
Externí odkaz:
http://arxiv.org/abs/2403.13257
Autor:
Jithmi Siriwardhana, D.M.D. Rasika, Dinusha Yapa, W.A.D.V. Weerathilake, Hasitha Priyashantha
Publikováno v:
Food Chemistry Advances, Vol 5, Iss , Pp 100792- (2024)
This study aimed to assess the impact of bael fruit pulp on the viability of probiotic Lacticaseibacillus rhamnosus GG (LGG) and some physicochemical properties of bael-goat milk-based beverages during 21 days of refrigerated storage. Bael fruit pulp
Externí odkaz:
https://doaj.org/article/c9e948b2985946be82d6f8a04a5a0e6e
Autor:
Samarakoon, Sehan, Siriwardhana, Yushan, Porambage, Pawani, Liyanage, Madhusanka, Chang, Sang-Yoon, Kim, Jinoh, Kim, Jonghyun, Ylianttila, Mika
With a plethora of new connections, features, and services introduced, the 5th generation (5G) wireless technology reflects the development of mobile communication networks and is here to stay for the next decade. The multitude of services and techno
Externí odkaz:
http://arxiv.org/abs/2212.01298
Autor:
Siriwardhana, Shamane, Weerasekera, Rivindu, Wen, Elliott, Kaluarachchi, Tharindu, Rana, Rajib, Nanayakkara, Suranga
Retrieval Augment Generation (RAG) is a recent advancement in Open-Domain Question Answering (ODQA). RAG has only been trained and explored with a Wikipedia-based external knowledge base and is not optimized for use in other specialized domains such
Externí odkaz:
http://arxiv.org/abs/2210.02627
We propose a novel personalized concept for the optimal treatment selection for a situation where the response is a multivariate vector, that could contain right-censored variables such as survival time. The proposed method can be applied with any nu
Externí odkaz:
http://arxiv.org/abs/2209.15068
Autor:
Shashanka Rajapakse, Tharusha Chamanthi Siriwardhana, Vimansha Sumanapala, Thiweda Subhanee, Savithri Sulakkhana, Periyasami Sivabalan Sridharan, Sajeewa Thennakoon
Publikováno v:
BMJ Paediatrics Open, Vol 8, Iss 1 (2024)
Background Asthma is the most common chronic disease affecting children. However, the epidemiology of asthma in adolescents from rural geographies is lacking.Methods An analytical cross-sectional study was conducted in secondary schools located in th
Externí odkaz:
https://doaj.org/article/4925978e639c419f8d98b48f8d60b8cf
In this paper, we illustrate how to fine-tune the entire Retrieval Augment Generation (RAG) architecture in an end-to-end manner. We highlighted the main engineering challenges that needed to be addressed to achieve this objective. We also compare ho
Externí odkaz:
http://arxiv.org/abs/2106.11517
Multimodal emotion recognition from speech is an important area in affective computing. Fusing multiple data modalities and learning representations with limited amounts of labeled data is a challenging task. In this paper, we explore the use of moda
Externí odkaz:
http://arxiv.org/abs/2008.06682