Training of neural networks using CUDA technology

iv vPod¥kování Cht¥l bych touto cestou vyjád°it osobní pod¥kování panu Ing. Zde¬ku Bukovi za jeho trp¥livé vedení, v¥cné p°ipomínky a osobní p°ístup p°i tvorb ¥ této práce. Nemalé díky pat°í také mým rodi£·m za jejich morální podporu p°i studiu. vi vii Declaration I hereby declare that I have completed this thesis independently and that I have listed all the literature and publications used. I have no objection to usage of this work in compliance with the act 60 Zákon £. 121/2000Sb. (copyright law), and with the rights connected with the copyright act including the changes in the act. In Prague on May 27, 2010............................................................. viii Purpose of this document is to familiarize reader with parallel technologies used in graphic processing units and provide general overview over nowadays trends in general purpose par-allel computing on graphic processing units. Reader will be told about advantages of massive parallel computing on graphic processing units and about their applications in scientific cal-culations, especially in artificial recurrent neural network technology. It is not purpose of this document to describe every aspect of recurrent neural networks, but to demonstrate means of graphic processing units in this matter. Abstrak