Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 3

Liczba wyników na stronie
first rewind previous Strona / 1 next fast forward last
Wyniki wyszukiwania
Wyszukiwano:
w słowach kluczowych:  CLARIN-PL
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 1 next fast forward last
EN
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.
EN
The paper presents dependability analysis of CLARIN-PL Centre of Language Technology (CLT). It describes infrastructure, high availability aspects and micro-service architecture used in CLARIN-PL applications. Microservices architecture improves dependability in respect to availability and reliability and to some extent safety. It is comprised of the mechanisms of reliable communication of applications, replication, recovery, and transaction processing. CLT has also a set of components for failure detection, monitoring and autonomic management, and distributed security policy enforcement.
EN
The paper presents reliability analysis of CLARIN-PL Centre of Language Technology (CLT). The CLT is a Polish part of the language technology infrastructure developed by CLARIN project. The main goal of which is to support researchers in humanities and social sciences. The infrastructure is a complex computer system that enables combining language tools with language resources into processing chains. Authors present the system structure, analyse types of faults and define the CLT reliability model. The model takes into account the fact the time gap, between the failure and the repair of the system is not exponential since repair actions are taken only when administrators are at work (assuming that administrators are not working 24/7). The model is used to estimate reliability metrics (mean time and 90th percentile of relative down time and relative partial operational time) by a use of Monte-Carlo simulation. Moreover, analysis of possible improvement in the CLT organisation and its influence on the estimated metrics is given.
first rewind previous Strona / 1 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.