Next Page: 10000

          Google 为树莓派 Zero W 发布了基于TensorFlow 的视觉识别套件      Cache   Translate Page   Web Page Cache   

Google 发布了一个 45 美元的 “AIY Vision Kit”,它是运行在树莓派 Zero W 上的基于 TensorFlow 的视觉识别开发套件,它使用了一个带 Movidius 芯片的 “VisionBonnet” 板。

为加速该设备上的神经网络,Google 的 AIY 视频套件继承了早期树莓派上运行的 AIY 项目 的语音/AI 套件,这个型号的树莓派随五月份的 MagPi 杂志一起赠送。与语音套件和老的 Google 硬纸板 VR 查看器一样,这个新的 AIY 视觉套件也使用一个硬纸板包装。这个套件和 Cloud Vision API 是不一样的,它使用了一个在 2015 年演示过的基于树莓派的 GoPiGo 机器人,它完全在本地的处理能力上运行,而不需要使用一个云端连接。这个 AIY 视觉套件现在可以 45 美元的价格去预订,将在 12 月份发货。

 

AIY 视觉套件,完整包装(左)和树莓派 Zero W

这个套件的主要处理部分除了所需要的 树莓派 Zero W 单片机之外 —— 一个基于 ARM11 的 1 GHz 的 Broadcom BCM2836 片上系统,另外的就是 Google 最新的 VisionBonnet RPi 附件板。这个 VisionBonnet pHAT 附件板使用了一个 Movidius MA2450,它是 Movidius Myriad 2 VPU 版的处理器。在 VisionBonnet 上,处理器为神经网络运行了 Google 的开源机器学习库 TensorFlow。因为这个芯片,使得视觉处理的速度最高达每秒 30 帧。

这个 AIY 视觉套件要求用户提供一个树莓派 Zero W、一个 树莓派摄像机 v2、以及一个 16GB 的 micro SD 卡,它用来下载基于 Linux 的 OS 镜像。这个套件包含了 VisionBonnet、一个 RGB 街机风格的按钮、一个压电扬声器、一个广角镜头套件、以及一个包裹它们的硬纸板。还有一些就是线缆、支架、安装螺母,以及连接部件。

 

AIY 视觉套件组件(左)和 VisonBonnet 附件板

有三个可用的神经网络模型。一个是通用的模型,它可以识别常见的 1000 个东西,一个是面部检测模型,它可以对 “快乐程度” 进行评分,从 “悲伤” 到 “大笑”,还有一个模型可以用来辨别图像内容是狗、猫、还是人。这个 1000 个图片模型源自 Google 的开源 MobileNets,它是基于 TensorFlow 家族的计算机视觉模型,它设计用于资源受限的移动或者嵌入式设备。

MobileNet 模型是低延时、低功耗,和参数化的,以满足资源受限的不同使用情景。Google 说,这个模型可以用于构建分类、检测、嵌入、以及分隔。在本月的早些时候,Google 发布了一个开发者预览版,它是一个对 Android 和 iOS 移动设备友好的 TensorFlow Lite 库,它与 MobileNets 和 Android 神经网络 API 是兼容的。

AIY 视觉套件包装图

除了提供这三个模型之外,AIY 视觉套件还提供了基本的 TensorFlow 代码和一个编译器,因此用户可以去开发自己的模型。另外,Python 开发者可以写一些新软件去定制 RGB 按钮颜色、压电元素声音、以及在 VisionBonnet 上的 4x GPIO 针脚,它可以添加另外的指示灯、按钮、或者伺服机构。Potential 模型包括识别食物、基于可视化输入来打开一个狗门、当你的汽车偏离车道时发出文本信息、或者根据识别到的人的面部表情来播放特定的音乐。

 

Myriad 2 VPU 结构图(左)和参考板

Movidius Myriad 2 处理器在一个标称 1W 的功耗下提供每秒万亿次浮点运算的性能。在被 Intel 收购之前,这个芯片最早出现在 Tango 项目的参考平台上,并内置在 2016 年 5 月由 Movidius 首次亮相的、Ubuntu 驱动的 USB 的 Fathom 神经网络处理棒中。根据 Movidius 的说法,Myriad 2 目前已经在 “市场上数百万的设备上使用”。

更多信息

AIY 视觉套件可以在 Micro Center 上预订,价格为 $44.99,预计在(2017 年) 12 月初发货。更多信息请参考 AIY 视觉套件的 公告Google 博客、以及 Micro Center 购物页面


via: http://linuxgizmos.com/google-launches-tensorflow-based-vision-recognition-kit-for-rpi-zero-w/

作者:Eric Brown 译者:qhwdw 校对:wxy

本文由 LCTT 原创编译,Linux中国 荣誉推出


          El Snapdragon 670 ya está aquí: 10 nanómetros, 8 núcleos y mucha IA para la súper gama media      Cache   Translate Page   Web Page Cache   

Snapdragon 670

El pasado año, Qualcomm lanzó un procesador que no estaba destinado a recibir el nombre que finalmente tuvo. El Snapdragon 660 estaba destinado a cotas mayores, pero una reducción, o aumento, de sus nanómetros hizo que acabase enclavado en la línea 600 del fabricante, y su renovación se atisbó hace unos meses en una filtración.

Ahora parece que la renovación ha dado la cara, y no es finalmente la del Snapdragon 710 que vimos llegar hace ya unos meses. El Snapdragon 670 está entre nosotros y apuesta por una línea continuista frente a otros modelos ya presentados este año, pero dando un salto adelante en la línea 600, que recibe un procesador para gamas medias premium en 10 nanómetros.

Así es el nuevo Snapdragon 670

Snapdragon 670

Como hemos comentado, Qualcomm ha apostado al fin por los 10 nanómetros para el sucesor de su Snapdragon 660. Este Snapdragon 670 llega en 10 nanómetros LPP y apuesta por una configuración de ocho núcleos semipersonalizada. Eso es que tenemos dos núcleos de alta potencia Kryo 360, basados en los Cortex A75, y otros seis núcleos para rendimiento moderado Cortex A55.

Los dos núcleos de alto rendimiento corren a un máximo de 2GHz mientras que los otros seis llegan a 1,7GHz. También tenemos una caché L1 de 64KB y 32KB, cada una para un clúster, y una caché L2 de 256KB y 128KB, con la misma configuración. La caché L3 es compartida y cuenta con una capacidad de 1MB, equiparándose así a los más potentes del catálogo.

El Snapdragon 670 apuesta por 10 nanómetros, ocho núcleos en dos clústeres y por la inteligencia artificial

Si hablamos de la GPU, nos encontramos con una vieja conocida, la Adreno 615, pero hay otros chips insertados en este SoC que nos suenan de otras ediciones. Como el Hexagon 685 DSP para el procesamiento de todos, o casi todos, los códigos de inteligencia artificial. Además, es compatible con el SDK de procesamiento neural de Snapdragon, así como con la Hexagon NN y la Android NN, además de otros frameworks como Caffe2 o TensorFlow.

En cuanto al ISP, tenemos un Spectra 250 que nos permitirá operar con una única cámara de 25 megapíxeles de capacidad, o con una configuración dual de 16 y 16 megapíxeles. En ambos se soportará la estabilización óptica y electrónica, además de contar con desenfoque de profundidad nativo, o bokeh, y composiciones multifotograma con reducción de ruido. En el apartado de vídeo tendremos un máximo de 4K a 30 fotogramas por segundo.

Este Snapdragon 670 será capaz de soportar hasta 8GB de RAM LPDDR4x y contará con la posibilidad de navegar en 4G a 600Mbps gracias a su modem Snapdragon X12, pudiendo subir datos a 150Mbps. Por cierto, también contará con audio Aqstic y aptX, además de con la capacidad de carga rápida Quick Charge 4+.

Vía | XDA


          Podcast FS Hebdo - 9 au 15 novembre      Cache   Translate Page   Web Page Cache   
De Google qui libère le code source de TensorFlow, en passant par un simple collyre contre la cataracte, Jupiter qui aurait expulsé une planète géante, ou encore les gaz à effet de serre qui dépassent de nouveau leurs records, découvrez l’actualité scientifique pour la semaine du 9 au 15 novembre.
          Principal Data Scientist - Rio Tinto - Montréal, QC      Cache   Translate Page   Web Page Cache   
Python (Sci-kit Learn, numpy, pandas, Tensorflow, Keras), R, Matlab, SQL. Scientifique principal des données — Montréal....
From Rio Tinto - Wed, 01 Aug 2018 11:02:26 GMT - View all Montréal, QC jobs
          Ingénieur en apprentissage automatique - Groom & Associates - Montréal, QC      Cache   Translate Page   Web Page Cache   
Expérience avec tensorflow ou d'autres backends, keras ou autres frameworks, scikit-learn, OpenCV, Pandas. Experience with tensorflow or other backends, keras...
From Groom & Associates - Thu, 07 Jun 2018 17:10:46 GMT - View all Montréal, QC jobs
          Machine Learning/AI Engineer - Groom & Associates - Montréal, QC      Cache   Translate Page   Web Page Cache   
Expérience avec tensorflow ou d'autres backends, keras ou autres frameworks, scikit-learn, OpenCV, Pandas. Experience with tensorflow or other backends, keras...
From Groom & Associates - Thu, 07 Jun 2018 14:58:16 GMT - View all Montréal, QC jobs
          (USA-TX-Plano) Advertising & Analytics - Principal Data Scientist (AdCo)      Cache   Translate Page   Web Page Cache   
The Data Scientist will be responsible for designing and implementing processes and layouts for complex, large- scale data sets used for modeling, data mining, and research purposes. The purpose of this role is to conceptualize, prototype, design, develop and implement large scale big data science solutions in the cloud and on premises, in close collaboration with product development teams, data engineers and cloud enterprise teams. Competencies in implementing common and new machine learning, text mining and other data science driven solutions on cloud based technologies such as AWS are required. The data scientist will be knowledgeable and skilled in the emerging data science trends and must be able to provide technical guidance to the other data scientists in implementing emerging and advanced techniques. The data scientist must also be able to work closely with the product and business teams to conceptualize appropriate data science models and methods that meet the requirements. Key Roles and Responsibilities + Uses known and emerging techniques and methods in data science (including statistical, machine learning, deep learning, text and language analytics and visualization) in big data and cloud based technologies to conceptualize, prototype, design, code, test, validate and tune data science centric solutions to address business and product requirements + Conceptualizes data science enablers required for supporting future product features based on business and product roadmaps, and guides cross functional teams in prototyping and validating these enablers + Mentors and guides other data scientists + Uses a wide range of existing and new data science and machine learning tools and methods as required to solve the problem on hand. Skilled in frameworks and libraries using but not limited to R, python, spark, scala, pig, hive, mllib, mxnet, tensorflow, keras, theanos etc. + Is aware of industry trends an collaborates with the platform and engineering teams to update the data science development stack for competitive advantage + Collaborates with third party data science capability vendors and provides appropriate recommendations to the product development teams + Works in a highly agile environment **Experience** Typically requires 10 or more years experience or PhD in an approved field with a minimum of 6 years of relevant experience. **Education** Preferred Masters of Science in Computer Science, Math or Scientific Computing; Data Analytics, Machine Learning or Business Analyst nanodegree; or equivalent experience.
          Generalist Software Developer - AIRY:3D - Montréal, QC      Cache   Translate Page   Web Page Cache   
Computer vision, image processing, computational imaging, 2D/3D computer graphics, or machine learning algorithms and toolkits (OpenCV, OpenGL, CUDA, TensorFlow...
From Indeed - Wed, 01 Aug 2018 17:08:19 GMT - View all Montréal, QC jobs
          (USA-MA-Boston) Software Engineer (Node.js/Python)      Cache   Translate Page   Web Page Cache   
Software engineers on our team work closely with Machine Learning engineers to create a smarter, personalized learning journey for our users. You will be working on a cross-functional team with a Product Manager, UX Designer, DevOps Engineer, Machine Learning Engineers, and Software Engineers. You’ll be part of a team that is user focused, has a mentality for experimentation, and iterates quickly. *Ways we work:* * Autonomous & responsible teams - making their own product & dev choices (https://www.pluralsight.com/tech-blog/team-responsibilty) * Data first - we work with large volumes of data to build scalable solutions for our products * Software Craftsmanship - we want to be proud of our work o Pair programming - we value collaborative development o Test-driven development - we take responsibility for our code without QA engineers o Continuous delivery - teams independently ship code to production every day o Kanban & Lean - no more backlog grooming, no more T-shirt size estimating o Continual improvement - we hold weekly retrospectives in order to assess and improve system processes *What we create with:* * Backend - Node.js/Python * Testing - Mocha/Pytest * Declarative UI - React * Styling - CSS Modules <3 * Messaging - RabbitMQ * Database - PostgreSQL * Source Control - Github * Frameworks - Airflow/TensorFlow EOE Statement Be Yourself. Pluralsight is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, gender, gender identity or expression, sexual orientation, national origin, genetics, disability, age, or veteran status.
          tf-nightly-gpu 1.11.0.dev20180810      Cache   Translate Page   Web Page Cache   
TensorFlow is an open source machine learning framework for everyone.
          tf-nightly 1.11.0.dev20180810      Cache   Translate Page   Web Page Cache   
TensorFlow is an open source machine learning framework for everyone.
          A little bit of Machine Learning: Playing with Google's Prediction API      Cache   Translate Page   Web Page Cache   
Before we get started, let’s begin by making clear that this isn’t going to be a deep dive on TensorFlow, neural networks, inductive logic, Bayesian networks, genetic algorithms or any other sub-heading from the Machine Learning Wikipedia article. Nor is this really a Go-heavy article, but rather an introduction to machine learning via a simple consumption of the Google Prediction API. How the Google Prediction API works The Google Prediction API attempts to guess answers to questions by either predicting a numeric value between 0 and 1 for that item based on similar valued examples in its training data (“regression”), or choosing a category that describes it given a set of similar categorized items in its training data (“categorical”).
          Principal Data Scientist - Rio Tinto - Montréal, QC      Cache   Translate Page   Web Page Cache   
Python (Sci-kit Learn, numpy, pandas, Tensorflow, Keras), R, Matlab, SQL. Scientifique principal des données — Montréal....
From Rio Tinto - Wed, 01 Aug 2018 11:02:26 GMT - View all Montréal, QC jobs
          Ingénieur en apprentissage automatique - Groom & Associates - Montréal, QC      Cache   Translate Page   Web Page Cache   
Expérience avec tensorflow ou d'autres backends, keras ou autres frameworks, scikit-learn, OpenCV, Pandas. Experience with tensorflow or other backends, keras...
From Groom & Associates - Thu, 07 Jun 2018 17:10:46 GMT - View all Montréal, QC jobs
          Machine Learning/AI Engineer - Groom & Associates - Montréal, QC      Cache   Translate Page   Web Page Cache   
Expérience avec tensorflow ou d'autres backends, keras ou autres frameworks, scikit-learn, OpenCV, Pandas. Experience with tensorflow or other backends, keras...
From Groom & Associates - Thu, 07 Jun 2018 14:58:16 GMT - View all Montréal, QC jobs
          Platform Developer, Machine Learning - Kinaxis - Ottawa, ON      Cache   Translate Page   Web Page Cache   
Experience with Machine Learning projects, familiarity with platforms or languages such as scikit-learn, Pandas, NumPy, SciPy, R, TensorFlow....
From Kinaxis - Wed, 08 Aug 2018 20:38:15 GMT - View all Ottawa, ON jobs
          Machine Learning avec Scikit-Learn. Mise en oeuvre et cas concrets, un livre de Aurélien Géron, critique par Thibaut Cuvelier      Cache   Translate Page   Web Page Cache   
Cet ouvrage, conçu pour tous ceux qui souhaitent s'initier au Machine Learning (apprentissage automatique) est la traduction de la première partie du best-seller américain Hands-On Machine Learning with Scikit-Learn & TensorFlow.

Il ne requiert que peu de connaissances en mathématiques et présente les fondamentaux du Machine Learning d'une façon très pratique à l'aide de Scikit-Learn qui est l'un des frameworks de ML les plus utilisés actuellement.

Des exercices corrigés...
          TensorFlow 501 Week 6      Cache   Translate Page   Web Page Cache   

Hi,

 

I am working my way through the TensorFlow 501 course and have got slightly stuck on Week 6. I'm slightly unsure what the question "If I tell you that the inner-most values come as `W,b` pairs, can you get the *shapes* of the coefficients?" is asking of me. What does it mean by "inner-most"?  Does it mean 'fc6'?

Apart from that I am struggling with how you would add these to the convolutional layer slightly further on in the notebook.

 

I have completed the Machine Learning 501 course and the first 5 weeks of TensorFlow 501 and throughout both of them it references a set of answers. I would like to check what I have done against these but I cannot find them anywhere! Could anyone help me locate them? It would help a lot.

 

Thanks,

Andrew


          Create text classification with TensorFlow      Cache   Translate Page   Web Page Cache   
Looking for someone to build NLP models using Keras and Tensorflow (CNN/RNNs) There will be many tasks and putting one task price here, which we can discuss. (Budget: €250 - €750 EUR, Jobs: Machine Learning, Python, Tensorflow)
          Vue CLI 3.0, Babel 7 RC, and a playground for Electron coding      Cache   Translate Page   Web Page Cache   

#398 — August 10, 2018

Read on the Web

JavaScript Weekly

Electron Fiddle: A Playground for Electron Experiments — Want to quickly experiment with Electron (a cross-platform JS desktop app environment) development? Electron Fiddle is a downloadable cross-platform app that makes it easy. Downloads here.

Felix Rieseberg

Pts: A Library for Visualization and Creative Coding — An impressive, lightweight and modular library, written in TypeScript and compiled to ES6. Plenty of demos to play around with.

William Ngan

🔨 JavaScript: The 'New' Hard Parts — In this new course, you'll develop an intuitive understanding of the new features of JavaScript in ES6+: iterators, generators, promises, and async/await. Go under the hood of async JS and learn to solve complex async challenges with ease.

Frontend Masters sponsor

How to Deal with Dirty Side Effects in Your Pure Functional JavaScript“A pure function is a function that has no side effects. But if you know anything about programming, you know that side effects are the whole point.”

James Sinclair

Getting Alexa to Respond to Sign Language Using a Webcam and TensorFlow.js — A look at the underlying tech behind a proof of concept to get an Amazon Echo to respond to sign language, plus how the system was built using TensorFlow.js. Live demo here.

Abhishek Singh

First Babel 7.0 Release Candidates AvailablePlanning for 7.0 and Nearing the 7.0 Release will bring you up to date with what’s new in this long awaited released.

Babel

Vue CLI 3.0 Released — Vue CLI provides the standard build toolchain for Vue applications and this new version is a rewrite that uses pre-configured webpack features, ES2017 transpilation, and PostCSS out of the box.

Evan You

V8 Version 6.9: An Efficiency and Performance Release — Features reduced memory usage through embedded built-ins, faster WebAssembly startup through the new Liftoff compiler, better DataView and WeakMap performance, and more.

Mathias Bynens

💻 Jobs

Senior Software Engineer (NYC) — Learn from the best and lead by example. Do your finest work, with purpose, freedom, and a great community.

Reaktor

Frontend Engineer in Fast-Growing FinTech Start-up in Munich — Help building, running & deploying our SSR React, GraphQL, D3 frontend stack.

Scalable Capital

Find A JavaScript Job Through Vettery — Create a profile to connect with 4,000+ companies seeking top tech talent.

Vettery

📘 Tutorials and Opinions

Hello WebAssembly — An introductory yet low-level introduction to WebAssembly.

Casper Beyer

Building a Long Press Directive in Vue

Obaseki Nosa

Full Stack Essentials with React: Bootcamp — It’s immersive. It’s Intense. It’s Oct 15-19. Few seats remain.

Big Nerd Ranch sponsor

Turning an Angular 6 App Into a Progressive Web App — How to add PWA capabilities to your Angular 6 app.

Yassine Benabbas

Variable Scope in Modern JavaScript — Examples to show how global, local and block scopes work.

Andy Carter

Tips for Creating a 13KB JavaScript Game in 30 Days — We linked to the contest last week, but GitHub has shared some tips and techniques to make the most of the 13KB limitation. It starts on Monday.

Lee Reilly (GitHub)

A Quick Guide to Mocha Testing — Want to write and run tests for your JavaScript code with Mocha? Start here.

Glad Chinda

The PRPL Pattern for Progressive Web Apps using Angular 6+ — A primer on PRPL (which stands for Push, Render, Pre-cache, and Lazy-load) — this pattern is a relatively new technique used for building PWAs.

Vlado Tesanovic

Write Complex Languages. Online & Evening M.S. Software Engineering Program

Regis University sponsor

▶  Five Ways to Be A Happier JavaScript Developer — How to deal with the language and its ecosystem now without repeating the mistakes of the past.

Christian Heilmann

Why Ember? — A quick look at where Ember can provide value to your workflow.

Melanie Sumner

🔧 Code and Tools

Angular Console: A UI for the Angular CLI — There’s also a blog post with the rationale.

Narwhal Technologies

Angular 7 Beta 0 Released: What's To Come for Angular 7?

Valentyn Yakymenko

Catch JavaScript Errors Before Users Do — Detect JavaScript errors in real-time, then debug them in minutes. Works with Node, React, Angular, and more.

ROLLBAR sponsor

Azure CosmosDB JavaScript SDK 2.0 Now in Public Preview — Features a new object model and introduces support for promises for working with Azure’s multi-model database service. Here’s the repo.

Deborah Chen (Microsoft)

Dumper.js: A Pretty Variable Inspector for Node — Ideal for ‘print-style’ debugging.

Zeeshan Ahmed

Franc: Detect a Text's Natural Language from JavaScript — Even in the browser. Demo here (keep refreshing the page). Supports up to 402 languages.

Titus Wormer

SuperSlide.js: A Smooth, GPU Accelerated Sliding Menu for Your Next PWA

OSREC Technologies

Xterm.js: A Terminal Component for the Front End — Now used in lots of projects like VS Code, cPanel and Azure Cloud Shell.

SourceLair


          DeepBrain Chain Goes Online: DLT Could Power AI Development      Cache   Translate Page   Web Page Cache   
DeepBrain Chain's platform offers optimized operations on top of CUDA GPUs to support popular deep-learning frameworks, like TensorFlow, CNTK ...
          Analista Machine Learning-Scikitlearn - Gfi Informática - Madrid, España      Cache   Translate Page   Web Page Cache   
En Gfi Informática queremos incorporar un Analista Machine Learning para importante entidad bancaria. Requisitos: Experiencia demostrable de al menos 5 años como desarrollador machine learning con Scikitlearn. Recomendable Catboost o Tensorflow. Experiencia en despliegue de modelos para su consumo en producción. Experiencia en captura de requisitos con áreas de negocio. Funciones: Tratamiento de datos internos que apoyen la estrategia de las áreas auxiliares de la Compañía. ...


Next Page: 10000

Site Map 2018_01_14
Site Map 2018_01_15
Site Map 2018_01_16
Site Map 2018_01_17
Site Map 2018_01_18
Site Map 2018_01_19
Site Map 2018_01_20
Site Map 2018_01_21
Site Map 2018_01_22
Site Map 2018_01_23
Site Map 2018_01_24
Site Map 2018_01_25
Site Map 2018_01_26
Site Map 2018_01_27
Site Map 2018_01_28
Site Map 2018_01_29
Site Map 2018_01_30
Site Map 2018_01_31
Site Map 2018_02_01
Site Map 2018_02_02
Site Map 2018_02_03
Site Map 2018_02_04
Site Map 2018_02_05
Site Map 2018_02_06
Site Map 2018_02_07
Site Map 2018_02_08
Site Map 2018_02_09
Site Map 2018_02_10
Site Map 2018_02_11
Site Map 2018_02_12
Site Map 2018_02_13
Site Map 2018_02_14
Site Map 2018_02_15
Site Map 2018_02_15
Site Map 2018_02_16
Site Map 2018_02_17
Site Map 2018_02_18
Site Map 2018_02_19
Site Map 2018_02_20
Site Map 2018_02_21
Site Map 2018_02_22
Site Map 2018_02_23
Site Map 2018_02_24
Site Map 2018_02_25
Site Map 2018_02_26
Site Map 2018_02_27
Site Map 2018_02_28
Site Map 2018_03_01
Site Map 2018_03_02
Site Map 2018_03_03
Site Map 2018_03_04
Site Map 2018_03_05
Site Map 2018_03_06
Site Map 2018_03_07
Site Map 2018_03_08
Site Map 2018_03_09
Site Map 2018_03_10
Site Map 2018_03_11
Site Map 2018_03_12
Site Map 2018_03_13
Site Map 2018_03_14
Site Map 2018_03_15
Site Map 2018_03_16
Site Map 2018_03_17
Site Map 2018_03_18
Site Map 2018_03_19
Site Map 2018_03_20
Site Map 2018_03_21
Site Map 2018_03_22
Site Map 2018_03_23
Site Map 2018_03_24
Site Map 2018_03_25
Site Map 2018_03_26
Site Map 2018_03_27
Site Map 2018_03_28
Site Map 2018_03_29
Site Map 2018_03_30
Site Map 2018_03_31
Site Map 2018_04_01
Site Map 2018_04_02
Site Map 2018_04_03
Site Map 2018_04_04
Site Map 2018_04_05
Site Map 2018_04_06
Site Map 2018_04_07
Site Map 2018_04_08
Site Map 2018_04_09
Site Map 2018_04_10
Site Map 2018_04_11
Site Map 2018_04_12
Site Map 2018_04_13
Site Map 2018_04_14
Site Map 2018_04_15
Site Map 2018_04_16
Site Map 2018_04_17
Site Map 2018_04_18
Site Map 2018_04_19
Site Map 2018_04_20
Site Map 2018_04_21
Site Map 2018_04_22
Site Map 2018_04_23
Site Map 2018_04_24
Site Map 2018_04_25
Site Map 2018_04_26
Site Map 2018_04_27
Site Map 2018_04_28
Site Map 2018_04_29
Site Map 2018_04_30
Site Map 2018_05_01
Site Map 2018_05_02
Site Map 2018_05_03
Site Map 2018_05_04
Site Map 2018_05_05
Site Map 2018_05_06
Site Map 2018_05_07
Site Map 2018_05_08
Site Map 2018_05_09
Site Map 2018_05_15
Site Map 2018_05_16
Site Map 2018_05_17
Site Map 2018_05_18
Site Map 2018_05_19
Site Map 2018_05_20
Site Map 2018_05_21
Site Map 2018_05_22
Site Map 2018_05_23
Site Map 2018_05_24
Site Map 2018_05_25
Site Map 2018_05_26
Site Map 2018_05_27
Site Map 2018_05_28
Site Map 2018_05_29
Site Map 2018_05_30
Site Map 2018_05_31
Site Map 2018_06_01
Site Map 2018_06_02
Site Map 2018_06_03
Site Map 2018_06_04
Site Map 2018_06_05
Site Map 2018_06_06
Site Map 2018_06_07
Site Map 2018_06_08
Site Map 2018_06_09
Site Map 2018_06_10
Site Map 2018_06_11
Site Map 2018_06_12
Site Map 2018_06_13
Site Map 2018_06_14
Site Map 2018_06_15
Site Map 2018_06_16
Site Map 2018_06_17
Site Map 2018_06_18
Site Map 2018_06_19
Site Map 2018_06_20
Site Map 2018_06_21
Site Map 2018_06_22
Site Map 2018_06_23
Site Map 2018_06_24
Site Map 2018_06_25
Site Map 2018_06_26
Site Map 2018_06_27
Site Map 2018_06_28
Site Map 2018_06_29
Site Map 2018_06_30
Site Map 2018_07_01
Site Map 2018_07_02
Site Map 2018_07_03
Site Map 2018_07_04
Site Map 2018_07_05
Site Map 2018_07_06
Site Map 2018_07_07
Site Map 2018_07_08
Site Map 2018_07_09
Site Map 2018_07_10
Site Map 2018_07_11
Site Map 2018_07_12
Site Map 2018_07_13
Site Map 2018_07_14
Site Map 2018_07_15
Site Map 2018_07_16
Site Map 2018_07_17
Site Map 2018_07_18
Site Map 2018_07_19
Site Map 2018_07_20
Site Map 2018_07_21
Site Map 2018_07_22
Site Map 2018_07_23
Site Map 2018_07_24
Site Map 2018_07_25
Site Map 2018_07_26
Site Map 2018_07_27
Site Map 2018_07_28
Site Map 2018_07_29
Site Map 2018_07_30
Site Map 2018_07_31
Site Map 2018_08_01
Site Map 2018_08_02
Site Map 2018_08_03
Site Map 2018_08_04
Site Map 2018_08_05
Site Map 2018_08_06
Site Map 2018_08_07
Site Map 2018_08_08
Site Map 2018_08_09
Site Map 2018_08_10