Search results

Filters

  • Journals
  • Authors
  • Keywords
  • Date
  • Type

Search results

Number of results: 2
items per page: 25 50 75
Sort by:
Download PDF Download RIS Download Bibtex

Abstract

This paper explores cost-effective alternatives for resource-constrained environments in the context of language models by investigating methods such as quantization and CPUbased model implementations. The study addresses the computational efficiency of language models during inference and the development of infrastructure for text document processing. The paper discusses related technologies, the CLARIN-PL infrastructure architecture, and implementations of small and large language models. The emphasis is on model formats, data precision, and runtime environments (GPU and CPU). It identifies optimal solutions through extensive experimentation. In addition, the paper advocates for a more comprehensive performance evaluation approach. Instead of reporting only average token throughput, it suggests considering the curve’s shape, which can vary from constant to monotonically increasing or decreasing functions. Evaluating token throughput at various curve points, especially for different output token counts, provides a more informative perspective.
Go to article

Authors and Affiliations

Bartosz Walkowiak
1
Tomasz Walkowiak
1

  1. Faculty of Information and Communication Technology, Wroclaw University of Science and Technology, Wroclaw, Poland
Download PDF Download RIS Download Bibtex

Abstract

Southern elephant male seals (Mirounga leonina Linnaeus, 1758) were studied at King George Island (62°14´S, 58°40´W) from September to December 1999. The first males came ashore at the beginning of September. Twenty-five adults were immobilized, hot iron branded, and measured. Thirteen out of the 25 marked males spent an average of 66 (+/-8) days on land. Early arrival was positively correlated with the time spent ashore (r = 0.88, P < 0.05). Nine harems were formed in the study area. At the maximum haul-out of females (28 October) mean harem size was 32ą42 females (range 3–107). During the course of harem development, 10 changes in male harem dominance were observed. These changes were more frequent during the early (1–20 October, n = 6) than during the mid (21 October – 10 November, n = 2) and late (11–29 November, n = 2) periods of harem development. Overall, there were 14 dominant males; five of these in two different harems and nine in one harem. Of the 25 marked males, 44% were resighted in the following breeding or moulting season, and 16% seemed to improve their potential breeding success.

Go to article

Authors and Affiliations

Alejandro R. Carlini
Sebastián Poljak
Gustavo A. Daneri
Maria E.I. Márquez
Joachim Plötz

This page uses 'cookies'. Learn more