Distribuovaný systém pro akcelerovaný vývoj AI, lokální okrajové výpočty a robotické systémy

TENDER DESCRIPTION

The subject of this public contract is the supply of a comprehensive computing infrastructure for the development, training and deployment of artificial intelligence (AI) models, which consists of three successive hardware levels. The core of the system is a pair of high-performance central workstations designed for AI computing. These stations will provide massive computing performance of accelerators for local training, optimization and tuning of complex neural networks under laboratory conditions. The aim is to ensure fast iterations when developing algorithms before their final deployment. This central part is followed by a set of compact accelerated computing nodes with performance for real-time inference. These units will be physically deployed directly in the operations and local sites of the industrial partners. Their hardware architecture must guarantee high data throughput for massive and immediate evaluation of industrial data in the target operation, in particular for complex tasks for which standard low-voltage edge platforms are not sufficient. The third infrastructure level consists of a specialized hardware platform intended for the development of advanced…

TENDER BRIEF

TENDER DOCUMENTS