深度学习入门经典：《The Little Book of Deep Learning》简介

深度学习

需积分: 0 134 浏览量更新于2024-06-25 1 收藏 3.74MB PDF 举报

身份认证购VIP最低享 7 折!

30元优惠券

资源详情

资源推荐

Both regression and classication are generally

referred to as

sususususususususususuperperperperperperperperperperpervisedvisedvisedvisedvisedvisedvisedvisedvisedvisedvised learnlearnlearnlearnlearnlearnlearnlearnlearnlearnlearninginginginginginginginginginging

since the value

to predict, which is required as a target during

training, has to be produced, for instance by hu-

man experts. On the contrary density modeling

is usually seen as

unununununununununununsususususususususususuperperperperperperperperperperpervisedvisedvisedvisedvisedvisedvisedvisedvisedvisedvised learnlearnlearnlearnlearnlearnlearnlearnlearnlearnlearninginginginginginginginginginging

since

it is sucient to take existing data, without the

need for producing an associated ground-truth.

These three categories are not disjoint, for in-

stance classication can be cast as class score

regression, or discrete sequence density model-

ing as iterated classication. Also, they do not

cover all cases. One may want to predict com-

pounded quantities, or multiple classes, or model

a density conditional to a signal.

16 140

2.1 GPUs, TPUs, and batches

Graphical Processing Units were originally de-

signed for real-time image synthesis, which re-

quires highly parallel architectures that happen

to be tting to deep models. As their usage

for AI has increased, GPUs got equipped with

dedicated sub-components referred to as

tententententententententententen

sorsorsorsorsorsorsorsorsorsorsor corescorescorescorescorescorescorescorescorescorescores

, and deep-learning specialized chips

such as Google’s

TenTenTenTenTenTenTenTenTenTenTensorsorsorsorsorsorsorsorsorsorsor ProProProProProProProProProProProcesscesscesscesscesscesscesscesscesscesscessinginginginginginginginginginging UnitsUnitsUnitsUnitsUnitsUnitsUnitsUnitsUnitsUnitsUnits

(

TPUsTPUsTPUsTPUsTPUsTPUsTPUsTPUsTPUsTPUsTPUs

)

have been produced.

A GPU possesses several thousands of parallel

units, and its own fast memory. The limiting fac-

tor is usually not the number of computing units

but the

readreadreadreadreadreadreadreadreadreadread

writewritewritewritewritewritewritewritewritewritewrite opopopopopopopopopopoperererererererererereraaaaaaaaaaationstionstionstionstionstionstionstionstionstionstions tototototototototototo memmemmemmemmemmemmemmemmemmemmemoryoryoryoryoryoryoryoryoryoryory

. The

slowest link is between the CPU memory and

the GPU memory and consequently one should

avoid copying data across devices. Moreover

the structure of the GPU itself involves multiple

levels of

cachecachecachecachecachecachecachecachecachecachecache memmemmemmemmemmemmemmemmemmemmemoryoryoryoryoryoryoryoryoryoryory

, which are smaller but

faster, and computation should be organized to

avoid copies between these dierent caches.

This is achieved in particular by organizing the

computation in

batchesbatchesbatchesbatchesbatchesbatchesbatchesbatchesbatchesbatchesbatches ofofofofofofofofofofof samsamsamsamsamsamsamsamsamsamsamplesplesplesplesplesplesplesplesplesplesples

that are pro-

cessed in parallel. When an operator combines

a sample and model parameters, both have to be

moved to the cache memory near the actual com-

puting units. Proceeding by batches allows for

20 140

剩余139页未读，继续阅读

死磕代码程序媛

粉丝: 110
资源: 315

深度学习入门经典：《The Little Book of Deep Learning》简介

深度学习算法教程(Deeplearning Algorithms Tutorial) 完整版PDF

深度学习 最新中文版 pdf

ucdos98袖珍怎么运行

python实现袖珍计算器

袖珍示波器开源 csdn

使用switch语句实现模拟袖珍计算器

Python实现袖珍计算器可重复计算

用python实现袖珍计算机

Python实现袖珍计算器

鑫恒新牌hx-6b型袖珍收音机组装

C＃中使用switch语句实现袖珍模拟器，要求输入两个操作数字和一个操作符，根据操作符输出运算结果

python编程实现袖珍计算器

使用switch语句实现袖珍模拟器，要求输入两个操作数字和一个操作符，根据操作符输出运算结果

c语言编一个袖珍计算器的完整程序

编程实现袖珍计算机，要求输入两个操作数和一个运算符（+，-，*，/，%），要求输出运算结果，注意零除异常问题

Python实现袖珍计算器，并且利用循环操作，可以重复计算，直到按0结束运算

postgresql WITH RECURSIVE

g1sniper b5 nvme

java 魔方解算算法

0805封装的PCB图

最新资源

深度学习最新中文版 pdf