<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[RSS Feed]]></title><description><![CDATA[RSS Feed]]></description><link>https://ecency.com</link><image><url>https://ecency.com/logo512.png</url><title>RSS Feed</title><link>https://ecency.com</link></image><generator>RSS for Node</generator><lastBuildDate>Wed, 01 Jul 2026 12:32:33 GMT</lastBuildDate><atom:link href="https://ecency.com/created/keras/rss.xml" rel="self" type="application/rss+xml"/><item><title><![CDATA[Build a CNN Model for Retinal Image Diagnosis ]]></title><description><![CDATA[👁️ CNN Image Classification for Retinal Health Diagnosis with TensorFlow and Keras! 👁️ How to gather and preprocess a dataset of over 80,000 retinal images, design a CNN deep learning model , and train]]></description><link>https://ecency.com/@feitgemel/build-a-cnn-model-for-retinal-image-diagnosis</link><guid isPermaLink="true">https://ecency.com/@feitgemel/build-a-cnn-model-for-retinal-image-diagnosis</guid><category><![CDATA[python]]></category><dc:creator><![CDATA[feitgemel]]></dc:creator><pubDate>Sat, 07 Dec 2024 17:10:54 GMT</pubDate><enclosure url="https://i.ecency.com/p/NTy4GV6ooFRmaCXZ8UYgPhoud1kjiNX8QokLEZtbBKLuLWQ9yt7K3o4GiYnsWVBEhyAUiS34bcQKWMXBFYdXNWtbodEkwwb2ecCwbMNw343VcfGVo9f51rT8TnhwwgnaKWju1cpq9CirB3LVnh5Xrq5ncxSzW7St2Z14ihta?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[学习keras第二天]]></title><description><![CDATA[今天继续学习keras，今天看的是 Introduction to Keras for Engineers 与 Text classification from scratch 两部分。 我发现从吴恩达的教程开始学一些基本理论，无聊学不下去的时候，就直接开始看框架的教程，不懂的名词直接去查，配合起来更加的适合我。 我喜欢实战，喜欢丢到水里去蹦跶，实践对我而言更合适一些。]]></description><link>https://ecency.com/@jywahaha/keras</link><guid isPermaLink="true">https://ecency.com/@jywahaha/keras</guid><category><![CDATA[hive-105017]]></category><dc:creator><![CDATA[jywahaha]]></dc:creator><pubDate>Thu, 10 Dec 2020 14:53:18 GMT</pubDate></item><item><title><![CDATA[Tensorflow 2.0 —— RNN实战路透社新闻分类]]></title><description><![CDATA[Image from unsplash.com by @Jason_xj 之前的文章我们介绍了 RNN 循环网络，并用循环网络成功地预测了牛奶产量。这篇文章我们继续使用 RNN 以及 LSTM 和 GRU 处理分类问题。我们使用的是 Keras 自带的数据集——路透社新闻分类问题。 关注微信公众号获取源代码（二维码见文末） 1. RNN 回顾 与卷积神经网络处理空间局部相关性数据不同，循环网络主要用于处理时间序列]]></description><link>https://ecency.com/@hongtao/3m5a3e-tensorflow-2-0-rnn</link><guid isPermaLink="true">https://ecency.com/@hongtao/3m5a3e-tensorflow-2-0-rnn</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Mon, 20 Jan 2020 13:02:57 GMT</pubDate><enclosure url="https://i.ecency.com/p/QVdSJhgNm7s1isdYsJFBDW9ZzoDuLBosbeESi96iEy3ZtCTNyfcfNR7xGPFqD1oRXg6ayDjWmz3kNocq2vqVtPjmGoHjLqXQTq4Y31yaHDB5BcUAaAnaNjLChpXjzRy2B6TuQuKdYU8yCRdCmYdJbtbmqWoMPMktoGrGpY1SJTwPDtyQdSD1PQ3?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow2.0--可视化工具Tensorboard]]></title><description><![CDATA[image from unsplash.com by Tara Hegerty 之前的文章介绍过如何在 Keras 中快速调用 Tensorboard 这个可视化工具，这篇文章我们再深入探讨一下 Tensorboard 在T Tensorflow 2.0 下如何使用。 本文源代码请关注微信公众号(二维码见文末)获取 1. Tensorboard 工作原理 简单来说，tensorboard 就是通过监听定目录下的]]></description><link>https://ecency.com/@hongtao/tensorflow2-0-tensorboard</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow2-0-tensorboard</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Sat, 14 Dec 2019 21:58:27 GMT</pubDate><enclosure url="https://i.ecency.com/p/QVdSJhgNm7s1isdYsJFBDW9ZzoDuLBosbeESi96iEy3ZtCTNyfcaoBBkhzZpTabzZLZetqA2kgSkmSVnRPas85vnTXcqBuTL4dDeGjYrVkd8pMYnjEgDu2oYWSimfVK6FtXWmpBYb1hv2xuD6GDDvm9QFfFq3W3PvRnihYkiZgksctzPcZSHitF?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow2.0-数据加载和预处理]]></title><description><![CDATA[Image from unsplash.com by Adrian Cuj 前面的文章我们讨论了机器学习的各种问题，但是还没有认真讨论一下数据加载的问题。作为机器学习的第一步，如何正确和快速地加载数据以及预处理对于机器学习的项目成败是至关重要的。 我们已经很熟悉用 pandas.read_csv 来读取csv文件数据，然后用其内建函数或借助 numpy 和 scikit-learn]]></description><link>https://ecency.com/@hongtao/tensorflow2-0</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow2-0</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Mon, 09 Dec 2019 15:03:06 GMT</pubDate><enclosure url="https://i.ecency.com/p/QVdSJhgNm7s1isdYsJFBDW9ZzoDuLBosbeESi96iEy3ZtCTNyfcZK3zBHhyosTJ8io9s7d3HHXoXD2ycCjcPfPAyPTWmUqMdHq6qjRmH2eBDhX4UQvmypgXmiXHUN9auyrqzM1knMkecY4q2eSGmDsEsm2uiu7QvHhEm5VrryoxwepES5jEYZwD?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow 2.0 —— 与 Keras 的深度融合]]></title><description><![CDATA[image from unsplash by Zhang Kenny 前面的几篇文章从线性回归，到手写数字识别再到预测牛奶产量，我们用 Tensorflow 2.0 与 Keras 结合完成了全链接神经网络，卷积神经网络以及循环神经网络的搭建和训练。那 Tensorflow 2.0 和 Keras 到底是什么关系，我们应该如何选择和搭配二者来完成自己的项目呢？这篇文章就来探讨一下这个问题。 1.]]></description><link>https://ecency.com/@hongtao/69fpgj-tensorflow-2-0-keras</link><guid isPermaLink="true">https://ecency.com/@hongtao/69fpgj-tensorflow-2-0-keras</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Wed, 20 Nov 2019 16:03:51 GMT</pubDate><enclosure url="https://i.ecency.com/p/QVdSJhgNm7s1isdYsJFBDW9ZzoDuLBosbeESi96iEy3ZtCTNyfVsaMYed4arLt2uFcBmVK66rgGbojaCDooihGVZXqAYzu7AnrCKPoK1Yyv738MwfH63mBy1NJPuXZyhGRzZXd4zpMyXHJCrbdmBmRhZn5RkFM3pHhdYqpn1L4uN3Q8rbWG6CkF?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Bitcoin Price Up / Bakkt Launches in Sept / Keras LSTM Deep Learning Modules.]]></title><link>https://ecency.com/@britvr/bitcoin-price-up-bakkt-launches-in-sept-keras-lstm-deep-learning-modules</link><guid isPermaLink="true">https://ecency.com/@britvr/bitcoin-price-up-bakkt-launches-in-sept-keras-lstm-deep-learning-modules</guid><category><![CDATA[bitcoin]]></category><dc:creator><![CDATA[britvr]]></dc:creator><pubDate>Fri, 06 Sep 2019 04:44:06 GMT</pubDate><enclosure url="https://i.ecency.com/p/S5Eokt4BcQdk7EHeT1aYjzebg2hC7hkthT45eKYCFEpX1S7QH4wtA4FweN8juToZVrHFjMg?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Deep Learning Frameworks: A Comparison]]></title><description><![CDATA[Introduction: I started working on a new classification problem and the first question in my mind: which framework to use? So decided to explore the pros and cons with the available frameworks before to]]></description><link>https://ecency.com/@utpalkant/deep-learning-frameworks-a-comparison</link><guid isPermaLink="true">https://ecency.com/@utpalkant/deep-learning-frameworks-a-comparison</guid><category><![CDATA[deeplearning]]></category><dc:creator><![CDATA[utpalkant]]></dc:creator><pubDate>Tue, 11 Jun 2019 23:01:54 GMT</pubDate><enclosure url="https://i.ecency.com/p/D5zH9SyxCKd9GJ4T6rkBdeqZw1coQAaQyCUzUF4FozBvW73duqf6zjADYiPithe7J7hMKxxQATsoi6bbgqhphobqjKJRW4qf66gZ6jD51qkeezVnhMoBi3QroLWdWfNRPsCEeS?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow入门——改进RNN预测牛奶产量]]></title><description><![CDATA[image source from unsplash by Mehrshad Rajabi 上一篇文章我们用Keras搭建GRU神经网络，通过对前13年牛奶产量的学习，成功预测了地最后1年牛奶的产量。 该模型是多对一的输入/输出结构，也就意味着12个月的数据输入，只能输出1个月的数据。有没有可能改进模型，让输出输入的数量一致，以提高预测效率呢？这篇文章我们就来改进GRU模型，实现多对多的结构。]]></description><link>https://ecency.com/@hongtao/addlb-tensorflow-rnn</link><guid isPermaLink="true">https://ecency.com/@hongtao/addlb-tensorflow-rnn</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Tue, 16 Apr 2019 11:51:00 GMT</pubDate><enclosure url="https://i.ecency.com/p/ADdPNihJzmPcacC5s8k339azk2Jje3XyKPD2u8HLwQv3qaXFX2YWabTw2cDtGhQixBCU5tt8kyMuNJri1yhP8usvE?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow入门——RNN预测牛奶产量]]></title><description><![CDATA[早前，我介绍了循环神经网络(RNN)的基本概念，这篇文章我们就用Keras搭建一个RNN网络。该RNN能够通过对历史数据的学习，预测未来数据。 同样的，为了方便与读者交流，所有的代码都放在了这里： Repository: 1. GRU和LSTM简介 GRU(Gated Recurrent Unit)和LSTM(Long Short Term]]></description><link>https://ecency.com/@hongtao/tensorflow-rnn</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow-rnn</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Mon, 15 Apr 2019 23:22:12 GMT</pubDate><enclosure url="https://i.ecency.com/p/ADdPNihJzmPcacCSWGVBrMPbSTN5r1n3bkWuQnrxw73YCMwBDrnkmVSTVqeCuaf5FYjD9RYfEWwmTZBiZcqs1nn6i?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Deep learning in Python-Tensorflow and Keras| learn in lovescience.online]]></title><description><![CDATA[The TensorFlow applications are high level, as well as highly advanced technology use, and large-scale AI undertakings in the realms of the machine learning as well as the deep learning. It empowered the]]></description><link>https://ecency.com/@laithsharba/deep-learning-in-python-tensorflow-and-keras-or-learn-in-lovescience-online</link><guid isPermaLink="true">https://ecency.com/@laithsharba/deep-learning-in-python-tensorflow-and-keras-or-learn-in-lovescience-online</guid><category><![CDATA[keras]]></category><dc:creator><![CDATA[laithsharba]]></dc:creator><pubDate>Tue, 02 Apr 2019 11:52:06 GMT</pubDate><enclosure url="https://i.ecency.com/p/RtfsyWaiw3sRGUqTfQrCUwG3tq2nXYYbdZTkvV7eGosCUFQfawPnycE23ZePDYoeHGSUWdj3yZ5t9PfztQVbKvGEfQBRUDr28mUd86oFX19hNJFzBsq5MG3nJSWVwYfo3iojb54fE5y4o5rdv9L5iEEQ5cXRTCyZPSzA3DaPLNBPJvYFL1VQ8yVKaNgcXodVpqmWyTAUMyHB6FGGZB4K9MdrCepJ7JFZbar4eAfkt5F2MXrHuhGwx8MFLa5rUt?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow入门——单层神经网络识别MNIST手写数字]]></title><description><![CDATA[Image source wikipedia MNIST database 手写数字识别是一个非常经典的机器学习项目，这篇文章，我们就通过Kaggle上这个经典项目，学习如何用Tensorflow和Keras搭建最简单的单层神经网络，来识别手写数字。 同样的，为了方便与读者交流，所有的代码都放在了这里： Repository: 1. 数据下载和预处理]]></description><link>https://ecency.com/@hongtao/tensorflow-mnist</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow-mnist</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Fri, 29 Mar 2019 17:16:39 GMT</pubDate><enclosure url="https://i.ecency.com/p/23KQwnti57su7HZA5xrZrYQY4ZmwmqQtUPDSTh1dFN8Zf9wUbHnoQxPwdXSP96NwWEhaE6smv6jcXkT42mYGWGTJUqHMgxN?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Bakau]]></title><description><![CDATA[Hutan bakau menjadi suatu alasan untuk dijadikan sebagai tameng untuk menahan abrasi dipinggir sungai dan laut. Huta bakau terkenal cukup keras untuk membentengi perairan dari abrasi. Kemudian, bibit bakau]]></description><link>https://ecency.com/@preliminary18/bakau</link><guid isPermaLink="true">https://ecency.com/@preliminary18/bakau</guid><category><![CDATA[bakau]]></category><dc:creator><![CDATA[preliminary18]]></dc:creator><pubDate>Thu, 28 Mar 2019 17:56:39 GMT</pubDate><enclosure url="https://i.ecency.com/p/JvFFVmatwWHRfvmtd53nmEJ94xpKydwmbSC5H5svBACH81Lvvmk9bfBgUygs1DJR91d412wBxozc8aZQ3vjc9MmANMNPV1pJEdoz3gbRJFQmnZiX1xHK73y1jVV8uQCw6A7o2ERhTt?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow入门——处理overfitting的问题]]></title><description><![CDATA[在之前的文章中，我们发现训练组(篮)和验证组(红)的损失函数在20个Epoch之后，向着相反方向变化。训练组损失函数继续下降，验证组损失函数反而在上升，这就是典型的Overfitting(过拟合)现象。 过拟合就是模型过度地学习了训练集的特征，反而没法处理测试集中更一般化的问题。处理过拟合最根本的解决方法当然是获得更多的训练样本。]]></description><link>https://ecency.com/@hongtao/tensorflow-overfitting</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow-overfitting</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Tue, 19 Mar 2019 17:14:24 GMT</pubDate><enclosure url="https://i.ecency.com/p/ADdPNihJzmPcacBjDzztDwnQ3bFPS5Hu31uAPThiwinZUnufLzAPjz7uusT6YXMHAXGJMsfUP63Dub6mRWVC7CWJ2?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow入门——Keras处理分类问题，Classification with Keras]]></title><description><![CDATA[Tensorflow 和 Keras 除了能处理前一篇文章提到的回归(Regression，拟合&预测)的问题之外，还可以处理分类(Classfication)的问题。 这篇文章我们就介绍一下如何用Keras快速搭建一个线性分类器或神经网络，通过分析病人的生理数据来判断这个人是否患有糖尿病。 同样的，为了方便与读者交流，所有的源代码都放在了这里： 1. 数据的导入]]></description><link>https://ecency.com/@hongtao/tensorflow-keras-classification-with-keras</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow-keras-classification-with-keras</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Sat, 16 Mar 2019 23:55:54 GMT</pubDate><enclosure url="https://i.ecency.com/p/ADdPNihJzmPcacC5s8k339azk2Jje3XyKPD2u8HLwQv3qaKb3pQdvt6SP6pyeGgVSHurBMDdEX2kk4RYbrbSrjWPQ?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Tensorflow入门——Keras简介和上手]]></title><description><![CDATA[前面两篇文章介绍了Tensorflow，以及用Tensorflow快速上手一个线性回归的项目。 实际上Tensorflow对于机器学习新手来说并不是很友好，Tensorflow自己也有高层API，比如Estimator和tf.data就相对来说更容易上手。 Keras本来是独立的机器学习框架，但已经逐渐被整合进了Tensorflow, 今天我们就来简单上手一下吧。 同样的，为了方便与读者交流，所有源代码都放在了]]></description><link>https://ecency.com/@hongtao/tensorflow-keras</link><guid isPermaLink="true">https://ecency.com/@hongtao/tensorflow-keras</guid><category><![CDATA[cn-stem]]></category><dc:creator><![CDATA[hongtao]]></dc:creator><pubDate>Tue, 12 Mar 2019 17:32:18 GMT</pubDate><enclosure url="https://i.ecency.com/p/3RTd4iuWD6NV7fkvydeHL8YGLc5rJssTjG45wdmKLsGKeQGsLrnaVELoSTkDq4hnfG8CJbtGUa63o9vnQ12RMWgMCVAiPnWQ1y2bGyzaXZcELfVSpR28vryQgxxV2hB8zhC9Jv1jeEmKfiPzeN9r5LZqf8g93u5gYzPBRMiQWakWTH?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[16 Hot Essential Cheat Sheets for Machine Learning and Deep Learning]]></title><description><![CDATA[If a worker wants to do something good, he must first sharpen his tools. In machine learning and deep learning research, excellent reference materials and manuals can often help us get twice the result]]></description><link>https://ecency.com/@evanli/16-hot-essential-cheat-sheets-for-machine-learning-and-deep-learning</link><guid isPermaLink="true">https://ecency.com/@evanli/16-hot-essential-cheat-sheets-for-machine-learning-and-deep-learning</guid><category><![CDATA[ai]]></category><dc:creator><![CDATA[evanli]]></dc:creator><pubDate>Fri, 25 Jan 2019 12:55:30 GMT</pubDate><enclosure url="https://i.ecency.com/p/7ohP4GDMGPrUMp8dW6yuJTR9MKNu8P8DCXDU9qmmhwUpE8CoNErTqgd6mGpt8i5UT5CQsH4yPTdHGr5u98M9jhaDnwGqh42J5QzR?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[Perjuangan Menuju Kemenangan]]></title><description><![CDATA[Kemenangan akan dapat diperoleh melalui kerja keras. Kerja keras itu sendiri akan dapat dilihat melalui usaha yang dilakukan sungguh-sungguh. Tanpa melakukan usaha secara sunguh-sungguh, maka belum dapat]]></description><link>https://ecency.com/@banglie-1/perjuangan-menuju-kemenangan</link><guid isPermaLink="true">https://ecency.com/@banglie-1/perjuangan-menuju-kemenangan</guid><category><![CDATA[kerja]]></category><dc:creator><![CDATA[banglie-1]]></dc:creator><pubDate>Sun, 20 Jan 2019 13:34:24 GMT</pubDate><enclosure url="https://i.ecency.com/p/qjrE4yyfw5pEPvDbJDzhdNXM7mjt1tbr2kM3X28F6SraZePAxH3qJpP5KrhqvCsVz2Rsdqjx7r6NHu69N4qXxq1NVcEv8PoD1VusggjJS6C9m6mmU19RJNSa?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[How to build a cat detector with a Raspberry Pi and a Pi Noir camera using Deep Learning - Part IV]]></title><description><![CDATA[Part I is here: Part II: Part III: Introduction In this part I will describe how we can deploy the model we trained in part III on the raspberry pi. You can then choose what you want to do when a cat is]]></description><link>https://ecency.com/@mragic/how-to-build-a-cat-detector-with-a-raspberry-pi-and-a-pi-noir-camera-using-deep-learning-part-iv</link><guid isPermaLink="true">https://ecency.com/@mragic/how-to-build-a-cat-detector-with-a-raspberry-pi-and-a-pi-noir-camera-using-deep-learning-part-iv</guid><category><![CDATA[raspberrypi]]></category><dc:creator><![CDATA[mragic]]></dc:creator><pubDate>Tue, 01 Jan 2019 17:26:03 GMT</pubDate><enclosure url="https://i.ecency.com/p/5ShzsKnKF7vppGeV6VN3m3GSDcLoRruAhMmifZtFSDkYScsLm1eH5mHCrhJPdcXedAjhfQ7aEjofJKg7TgjfXRqDNajzAZRB9VtZKdMhfdUGkKMKDbbSCDZKSmDCooEo7u6sYn4A2CkZqBmjBfHT4x5k?format=match&amp;mode=fit" length="0" type="false"/></item><item><title><![CDATA[How to build a cat detector with a Raspberry Pi and a Pi Noir camera using Deep Learning - Part III]]></title><description><![CDATA[Part I is here: Part II: Introduction In this part III I will describe how the the images taken in can be used to train a deep neural network to distinguish images with cats against images without cats.]]></description><link>https://ecency.com/@mragic/how-to-build-a-cat-detector-with-a-raspberry-pi-and-a-pi-noir-camera-using-deep-learning-part-iii</link><guid isPermaLink="true">https://ecency.com/@mragic/how-to-build-a-cat-detector-with-a-raspberry-pi-and-a-pi-noir-camera-using-deep-learning-part-iii</guid><category><![CDATA[raspberrypi]]></category><dc:creator><![CDATA[mragic]]></dc:creator><pubDate>Tue, 01 Jan 2019 13:32:54 GMT</pubDate><enclosure url="https://i.ecency.com/p/21PRtjKRXPQybj4WUXScWv5QPLninWRxfbcWNsx7SenD7FpVNYXGdMz1wteQvuTM2CtSQzYHEH3SBJuQTVm65LXNLgpu9CJAr8YsLEpmMETFcuPwLE5J23s9wX7132fXDcHZx5BRoxWttKv9c72R62i?format=match&amp;mode=fit" length="0" type="false"/></item></channel></rss>