<?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" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Sudharshan Chakravarthy on Medium]]></title>
        <description><![CDATA[Stories by Sudharshan Chakravarthy on Medium]]></description>
        <link>https://medium.com/@cs_bot?source=rss-34b69a74ecdb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*3mThcys9Cdl6l7pqjxNLHw.jpeg</url>
            <title>Stories by Sudharshan Chakravarthy on Medium</title>
            <link>https://medium.com/@cs_bot?source=rss-34b69a74ecdb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 06:26:16 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@cs_bot/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Semantic Segmentation in Self-driving Cars]]></title>
            <link>https://medium.com/@cs_bot/semantic-segmentation-in-self-driving-cars-3cb89aa08e9b?source=rss-34b69a74ecdb------2</link>
            <guid isPermaLink="false">https://medium.com/p/3cb89aa08e9b</guid>
            <category><![CDATA[pytorch]]></category>
            <category><![CDATA[semantic-segmentation]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[autonomous-cars]]></category>
            <category><![CDATA[deep-learning]]></category>
            <dc:creator><![CDATA[Sudharshan Chakravarthy]]></dc:creator>
            <pubDate>Wed, 01 Jul 2020 21:07:25 GMT</pubDate>
            <atom:updated>2020-07-03T13:03:06.106Z</atom:updated>
            <content:encoded><![CDATA[<h4>A bare-bones intro to distinguish objects in a road scene with Autonomous Vehicles using PyTorch 🔥</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*szjFpqqvOdOsQepq" /><figcaption>Photo by <a href="https://unsplash.com/@ort?utm_source=medium&amp;utm_medium=referral">Bram Van Oost</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Driving a car is often taken for granted. Little do we know that this effortless task is a result of millions of years of evolution. We’re able to identify the objects on the road, estimate their speeds and positions while having an intuitive, futuristic sense of what these objects are going to do. All these happen within a split-second.</p><blockquote>Can we give computers the same ability?</blockquote><p>Video inputs are an array of images, displayed at a large pace. Images themselves, are matrices with integer values representing different intensities of colors. The advent of Deep Learning algorithms and compute power has finally enabled us to take a shot at these previously untouchable problems.</p><h3>Introduction</h3><p>In <strong>Semantic segmentation</strong>, the model is tasked with partitioning the image into several components, each representing a different entity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/858/1*WHxTMCP9veUim40UyEXlTA.jpeg" /><figcaption>We input this image (left), the model identifies the cat, grass, the mountain, and the sky in the output (right)</figcaption></figure><p>We are about to attempt the same problem but in a complex road scenario. We use a subset of the <a href="https://www.cityscapes-dataset.com/">Cityscapes Dataset</a>, which has image pairs of a road scene and its segmented counterpart. The dataset can be downloaded <a href="https://www.kaggle.com/dansbecker/cityscapes-image-pairs">here</a>. The dataset contains images taken from vehicles in Germany.</p><p>This dataset has <strong>2975 training images</strong> files and <strong>500 validation image</strong> files. Each image file is <strong>256 x 512 pixels</strong>, and each file is a composite with the original photo on the left half of the image, alongside the labeled image (output of semantic segmentation) on the right half.</p><h3>Step 1. Building the Model</h3><h4>Setup</h4><p>Before we proceed, we have to import the necessary libraries.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D7&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D7&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/9ff7e6a8c91da0b656487d9770eac9a3/href">https://medium.com/media/9ff7e6a8c91da0b656487d9770eac9a3/href</a></iframe><h4>Configure GPU</h4><p>Secondly, let’s configure our GPU options. If there is a GPU available, this snippet detects it and sets the <strong>device</strong> to GPU.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D8&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D8&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/decc9fdadac3f5ebc4025259ebc0c6d3/href">https://medium.com/media/decc9fdadac3f5ebc4025259ebc0c6d3/href</a></iframe><h4>File System</h4><p>Now, let’s proceed to tidy up our file system locations.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D4%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D4%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/d83987d10d52cf226ab7f6ea7b002b0b/href">https://medium.com/media/d83987d10d52cf226ab7f6ea7b002b0b/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D10&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D10&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/80e39c7d0ec6f844c15384346f94d44b/href">https://medium.com/media/80e39c7d0ec6f844c15384346f94d44b/href</a></iframe><h4>Sample Image Retrieval</h4><p>Now that we’ve established a file path, let’s fetch a sample image and see if the retrieval process works fine.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D12&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D12&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/488c051c8dbc509d0a3aa927c66d9d59/href">https://medium.com/media/488c051c8dbc509d0a3aa927c66d9d59/href</a></iframe><h4>Define Output Labels</h4><p>We now define the labels in the output image. Ideally, every pixel of the input image is annotated and assigned labels by hand. Since we already have the color-coded output image, we just need to segregate these colors as different labels.</p><p>I have used a simple K-means algorithm to give these colored labels into 10 different classes. This will be the <strong>ground truth by proxy</strong> for our model.</p><p>To train our clustering model, we initialize color_array, which is a randomly initialized matrix.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D17&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D17&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/412c8fe8fe4dce8f7ea3f8a6d6bc0b31/href">https://medium.com/media/412c8fe8fe4dce8f7ea3f8a6d6bc0b31/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D18&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D18&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/f0d1a2ee5aaecf7739f4a4a84c97ea84/href">https://medium.com/media/f0d1a2ee5aaecf7739f4a4a84c97ea84/href</a></iframe><p>Here, we take the output image (which is color-coded for different objects) and apply our K-means clustering model on it. Let’s see how it turns out.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D14%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D14%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/5aeab0ba098a6e2d12bc5a18a840c987/href">https://medium.com/media/5aeab0ba098a6e2d12bc5a18a840c987/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D19&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D19&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/7c51cf6c3810e62b0f15c29f4a908a3c/href">https://medium.com/media/7c51cf6c3810e62b0f15c29f4a908a3c/href</a></iframe><h4>Define the Dataset</h4><p>We define the class CityscapeDataset, which acts as an iterator that returns a single input road scene image X and its corresponding label image Y. This vital function is carried out by the function__getitem__(). Furthermore, the image is <a href="https://machinelearningmastery.com/how-to-manually-scale-image-pixel-data-for-deep-learning/">normalized</a> using the transform() function. This reduces the impact that any single channel (R, G, B) can have on the activations and eventually outputs.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D22%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D22%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/783ef194a630e878121bb690b761f5a4/href">https://medium.com/media/783ef194a630e878121bb690b761f5a4/href</a></iframe><p>Let’s instantiate this class and verify the function.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D23&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D23&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/6baa997b2c3c9621cb2d9b2c1e8a2cbc/href">https://medium.com/media/6baa997b2c3c9621cb2d9b2c1e8a2cbc/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D24&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D24&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/218caf0ae74eb9eae51e46a31c5880b9/href">https://medium.com/media/218caf0ae74eb9eae51e46a31c5880b9/href</a></iframe><h4>The UNet Model</h4><p>The end-end model which we are going to use is the <a href="https://towardsdatascience.com/u-net-b229b32b4a71">UNet</a> model. In a naive image classification task, the primary objective is to convert the input feature map into a vector. However, in semantic segmentation, we have to reconstruct an image from this vector as well. That’s where UNet comes into the picture. We use the same feature mapping that helped us convert the feature map into the vector, to perform the reverse function. Thus, we have the <strong>contracting</strong> <strong>block </strong>and the <strong>expansive block</strong>.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D26%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D26%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/5e1c10ba35dd374143484d73b848ce70/href">https://medium.com/media/5e1c10ba35dd374143484d73b848ce70/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D27%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D27%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/3fe1dd8b49ee2584ef927172c6203223/href">https://medium.com/media/3fe1dd8b49ee2584ef927172c6203223/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D28&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D28&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/4ab663845fe0971c31277ae9c8c2e346/href">https://medium.com/media/4ab663845fe0971c31277ae9c8c2e346/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D29&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D29&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/2cdfccdd7f383124dc4a7e58388d82f4/href">https://medium.com/media/2cdfccdd7f383124dc4a7e58388d82f4/href</a></iframe><p>Voila! The model has been created. Let’s proceed to train the model. To get a more comprehensive view of the model, check out <a href="https://towardsdatascience.com/u-net-b229b32b4a71">this article</a>.</p><h3>Step 2. Training the model</h3><p>We train the model for 10 epochs using a learning_rate of 0.01. Further, we use the CrossEntropyLoss and Adamoptimizer to evaluate the model and guide the gradients through the state space respectively.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D31%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D31%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/70a468e7e335077ff62fee5ac33a9d56/href">https://medium.com/media/70a468e7e335077ff62fee5ac33a9d56/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D32%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D32%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/6afd89658eb1693331ffae1b17cc21de/href">https://medium.com/media/6afd89658eb1693331ffae1b17cc21de/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D33%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D33%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/0771be8910443ca46b894b2d06aef31d/href">https://medium.com/media/0771be8910443ca46b894b2d06aef31d/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D34%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D34%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/80884b8276d5724bb9d05640bc3a2ce8/href">https://medium.com/media/80884b8276d5724bb9d05640bc3a2ce8/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D35&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D35&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/6eb63f5fc5aae569e91980966c63431e/href">https://medium.com/media/6eb63f5fc5aae569e91980966c63431e/href</a></iframe><p>Let’s plot the training losses to observe the trend.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D36&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D36&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/ec059c17e82b79355252fe81d2d2381c/href">https://medium.com/media/ec059c17e82b79355252fe81d2d2381c/href</a></iframe><p>Let’s save the model and move on to the model predictions section.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D37%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D37%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/51766fa8fa1eef2738c7e667c3388e8a/href">https://medium.com/media/51766fa8fa1eef2738c7e667c3388e8a/href</a></iframe><h3>Step 3. Check model predictions</h3><p>We load the saved model and fetch the test set. We then iterate through the set and predict the new, unseen images.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D39%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D39%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/1bbaa76a2f1687e6fd78a28219db3f53/href">https://medium.com/media/1bbaa76a2f1687e6fd78a28219db3f53/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D40%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D40%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/edf895cf8a5194d4ba6b0b9f2bf3b360/href">https://medium.com/media/edf895cf8a5194d4ba6b0b9f2bf3b360/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D41&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D41&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/4dcdbfe77f04f482b5c2e9d35b7aa6fc/href">https://medium.com/media/4dcdbfe77f04f482b5c2e9d35b7aa6fc/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D42%26hideOutput%3Dtrue&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D42%26hideOutput%3Dtrue&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/75fd674740a0a21ba8c2d0794e981d5f/href">https://medium.com/media/75fd674740a0a21ba8c2d0794e981d5f/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D43&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D43&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/3ba854d3fe76c657d3e5573bf0e4d066/href">https://medium.com/media/3ba854d3fe76c657d3e5573bf0e4d066/href</a></iframe><h3>Step 4. IOU Score</h3><p>Although we can visually observe that the model has done alright, we need to quantify this by providing a metric score. We have <a href="https://github.com/BardOfCodes/seg_metrics_pytorch">these metrics</a> to do just that. I have chosen the mean overall IOU. Finally, we calculate this value. Read more on IOU <a href="https://www.pyimagesearch.com/2016/11/07/intersection-over-union-iou-for-object-detection/">here</a>.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fjovian.ml%2Fembed%3Furl%3Dhttps%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D44&amp;dntp=1&amp;display_name=Jovian&amp;url=https%3A%2F%2Fjovian.ml%2Fchakravarthy-sudharshan%2Fsemantic-segmentation-cityscapes%2Fv%2F1%26cellId%3D44&amp;image=https%3A%2F%2Fstorage.googleapis.com%2Fjvn%2Fassets%2Fchakravarthy-sudharshan%2Fprofile_images%2Fa2d2c9ca6ce54b6b92d9841bb01ded56%2Fmedium.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;scroll=auto&amp;schema=jovian" width="800" height="200" frameborder="0" scrolling="no"><a href="https://medium.com/media/860d8387d913018c6a6986f78f818036/href">https://medium.com/media/860d8387d913018c6a6986f78f818036/href</a></iframe><p>This shows that the overlap between the actual labels and the predicted outputs. In our case, we get around <strong>97%</strong>. Not bad for a first try!</p><h3>Final thoughts</h3><p>The sensitivity of the application domain means that we require much greater accuracy than this. To improve the model further (or to get the whole source code given above), you can have this <a href="https://jovian.ml/chakravarthy-sudharshan/semantic-segmentation-cityscapes">starter notebook</a> and build upon it.</p><p>Do check out <a href="https://medium.com/u/fcbd78e83361">Gokul Karthik</a>’s implementation <a href="https://github.com/GokulKarthik/Deep-Learning-Projects.pytorch/blob/master/6-Image-Segmentation-with-UNet.ipynb">here</a>. Give his <a href="https://github.com/GokulKarthik">repo</a> a visit for some cool Deep Learning projects. This <a href="https://github.com/meetshah1995/pytorch-semseg">project</a> implements an array of models that tackles semantic segmentation.</p><p>A big shoutout to <a href="https://jovian.ml/"><strong>Jovian.ml</strong></a><strong> </strong>for the platform to save and embed the notebooks.</p><p><em>Show some love by clapping off folks!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3cb89aa08e9b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Key to Happiness and How I may have Accidentally Found It]]></title>
            <link>https://medium.com/@cs_bot/key-to-happiness-and-how-i-may-have-accidentally-found-it-8b3406642e45?source=rss-34b69a74ecdb------2</link>
            <guid isPermaLink="false">https://medium.com/p/8b3406642e45</guid>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[travel]]></category>
            <category><![CDATA[personal-development]]></category>
            <category><![CDATA[happiness]]></category>
            <dc:creator><![CDATA[Sudharshan Chakravarthy]]></dc:creator>
            <pubDate>Mon, 10 Sep 2018 13:00:01 GMT</pubDate>
            <atom:updated>2025-09-13T21:51:08.919Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Frl-tVoeCZACFJlq0sJFuA.jpeg" /></figure><p>It has almost been a year since I embarked on a road-trip &amp; trek to <a href="https://www.google.co.in/maps/place/Ubbalamadugu+Waterfalls/@13.6042026,79.8416523,17z/data=!4m7!3m6!1s0x3a4d6ea4fdda9ce1:0xb64f97329717627e!8m2!3d13.6041974!4d79.843841!9m1!1b1">Tada</a>. I was accompanied by three awesome people: Kaushik, Praneetha, and <a href="https://medium.com/u/6ea77984c2b2">Godwyn EdinBaro</a> .</p><p>We convened at a pre-set spot (Guindy, central Chennai, India) started at around 7 AM in our motor-bikes. After a 3.5-hour <em>ass-burning</em> journey, we clearly started regretting the fact that we ditched the idea of taking an Uber.</p><blockquote>If you think adventure is dangerous, try routine. It’s lethal -Paulo Coelho</blockquote><p>Admittedly, upon reaching Tada (Ubbalamadugu Falls) we were already half drained. To make proceedings worse, there was a 2 Km walk from our parking spot to the base location. And ugh! the weather. It was around 40 degree Celsius but felt like ’twas 43252003274489856000. Moreover, we all had our mobiles switched to airplane mode because of no connectivity whatsoever.</p><p>Thirty minutes into the trek and it starts pouring like crazy. This forced us out of our shoes. But none of these caveats stifled us since we’d already set our eyes on the summit.</p><p>It was an upward ratchet of a trek wherein you follow a “river” (of sorts, it had a minimal flow) upstream. The idea was to reach the origin or the river so that we could enjoy some of the fresh-water, only now devoid of dirt and animal feces.</p><p>Imagine yourself standing near a cliff sans its fences. Your senses and eyes go into a hyper-focus mode. You start simulating different ways of falling and tripping to your own tragic death. With a stiff back, shaking legs, blinding fear and abnormally increased heart-rate (now the brain needs more oxygen and blood for those simulations, right?) the fear starts to blind you. Furthermore, at times this leads to physical catatonic immobilization and a mentally induced trance-like state.</p><p>[<strong>cue</strong>: pun-exclusive] Even though you’re tired, you can’t <em>Everest</em>. You start drawing a <em>Blanc </em>after a while, but I suppose a protein bar <em>Alps</em>. [ Intended af ]</p><p>On top of all this, imagine this happening to you every few minutes. Crazy rains made the rocks slippery. One fall and a fracture can prove to be futile because of the lack of medical facilities nearby.</p><p>Your pals will have to literally carry you back down to the base, like all the way down, then to the parking lot and hope there’s some help in there. Now that’s a big ask.</p><p>Mathematically speaking, there’s more probability now for you to be penciled in for a sudden visit by the angel of Death than me <em>not</em> moving out of my mother’s basement.</p><p>This is not the Himalayas. Agreed. But apart from slippery rocks, Tada also had its own ‘make it’ or ‘break your skull’ moments. Godwyn almost busted his balls in one such instance. One slip and you go into oblivion, forever.</p><p>This alarming proximity to death somehow brought a <em>sense of comfort </em>and <em>total acceptance.</em></p><p>Frivolous entities like taste, feelings, hunger, etc don’t register. Significantly more remarkable factors such as <em>people</em> close to you, your <em>legacy </em>and what you <em>really </em>wanted to do with your life — flash in front of your eyes.</p><p>If your (inner-)listening skills are well-honed, this encounter can be instrumental in finding what you’d always <strong>truly</strong> wanted to become. Such an occurrence can help you find and connect with your inner-self.</p><p>It also paves way for revelations such as this:</p><blockquote>You are your ultimate soul-mate.</blockquote><p>After almost an hour of relentless squabbling, we reached a mini-lake of sorts. This, in my perception of the overall anatomy of the mountain range, was check-point One. The difficulty exponentially increased w.r.t the increase in altitude. Passageways became steeper. We had to find the optimal path to a higher point by trial and error. After two such similar check-points, we finally made it to the summit.</p><p>Absorbed in unfathomable euphoria, it took me twenty odd minutes to come in terms with whatever had happened in the past three-four hours. After this mind vortex, I was distracted by something splendid.</p><p>Oh, the view from up top!</p><p>It was <em>pretty ordinary</em> (shoot! anti-cliché, much?) as the trees around were still engulfing our view.</p><p>The pièce de résistance however, was the waterfall and its source, where we spent a good hour and a half enjoying in the fresh-water outlet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*6C1i07M3-N0gLuR_GtWezg.jpeg" /><figcaption>Notice that lingering nimbus around our heads? :P</figcaption></figure><p>We had to resist the illusive pull of that place which would have otherwise kept us there throughout the eerie night; amongst the backdrop of crickets in the long grass.</p><p>The passage downstream (although not as extreme as the upward trail) had no mercy either. There were one too many déjà vu feelings’ of the path that we took up during the ascent which sent chills through the spine. The ensuing darkness made it all the more spooky.</p><p>It took us three hours to reach the base parking lot. Soaking wet, head-splitting migraines, voraciousness best described the state of each and every one of us. On top of this, we still had a four-hour bike ride to reach home (Okay, just kiddin’. I’m a hobo).</p><p>I’ve been to quite a few hill stations — Ooty, Kodaikanal, Ponmudi, Coorg, Kurungudi, Munnar, and a minor part of The Himalayas (Nepal). Each and every one of these places had magnificently breathtaking landscapes spanning tens of miles at the least. And yet . . .</p><p>Through the test of time after all the retrospection, I can aver that none of the above places have quite been as reminiscent-worthy as Tada.</p><p>In all of these places, there was a bus/car to take me through the onerous path; and the view from the summit was a lil’ stroll away. And boy that was ridiculously easy.</p><p>Watching the peaks unravel right in front of you is magical, for sure. However, it almost always left behind a choking feel that somehow made me think that I hadn’t worked hard enough to enjoy that view. And now that is blatantly analogous to the concept of <em>participation prize </em>in competitions. It makes you feel <strong>entitled </strong>to win everythin’ there is to, without any effort whatsoever (cue: generations — Y and Z).</p><p>Shallow. Superficial. And a vicious shortcut. This has always been a constricting factor for me in truly experiencing what was in store.</p><p>A random muse-session over a coffee much later was serendipitous enough to hit me:</p><blockquote>What is happiness, if not the hustle?</blockquote><p><em>Show some love by clapping off, folks!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8b3406642e45" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[To be or not to be — A moral dilemma]]></title>
            <link>https://medium.com/@cs_bot/to-be-or-not-to-be-a-case-study-51cd39c07655?source=rss-34b69a74ecdb------2</link>
            <guid isPermaLink="false">https://medium.com/p/51cd39c07655</guid>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[philanthropy]]></category>
            <category><![CDATA[transportation]]></category>
            <category><![CDATA[morality]]></category>
            <category><![CDATA[life-lessons]]></category>
            <dc:creator><![CDATA[Sudharshan Chakravarthy]]></dc:creator>
            <pubDate>Tue, 21 Aug 2018 07:08:45 GMT</pubDate>
            <atom:updated>2018-09-08T14:19:54.005Z</atom:updated>
            <content:encoded><![CDATA[<p>Let’s face it, problems are inevitable. Especially problems that you foresee, and have a (generic) solution road-map for, is what I call a continually design-patterened set of problems. Let me delineate. Take an assignment at school or a project at work, for instance. Without any loss of generality, it can be assumed that at least the concept of it is not truly novel. This requires you to follow an already “documented” elucidation and extend it for your solution, thereby marking your trail.</p><p>But are all problems continually design-patterned? Nope.</p><p>At times, certain situations have the potential to subject you to intense contemplative moments. During times like these, even clogged views become glaringly transparent. It just feels as if people, who see you, can actually see through you. Facades get blown away, and you become so utterly and pathetically readable. Psychologically speaking, there’s sense to avoid this social awkwardness. After all, unpredictability is a major valuation in coming off as interesting. No experience is more naked than that of being figured out.</p><p>So what situation am I talking about, specifically?</p><p>Riding the metro*. Yup, Local trains. That was what this whole set-up was about.</p><p>*(MRTS, Metro, Subway — but since metro is concise, sticking with it)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/660/1*BSjlv9xFbpvdgyCafzA1vg.jpeg" /><figcaption>Image courtesy: The Hindu</figcaption></figure><p>Okay, let me elaborate this a bit further. What do you expect in local trains? Music. Meaningless staring. Books. Social media. And alas, the kinetic panhandlers.</p><p>One of the key takeaways while riding the metro, is that there’ll almost always be no interaction whatsoever, among the riders. People are alone, together. It is a short span (10–20 minutes end-end) to effectively make contact. Agreed. But think about this. It is THE place to make in-person social associations. People who belong to the same vicinity, socio-economic class, and even of similar industries travel together (for work). And yet, we all choose to fade into darkness to never show up during transit.</p><p>Yet, there is an exceptional anomaly to account for.</p><p>The panhandlers. Urban beggars who bring with them exceptional power and talent in speech and songs. Situations were exacerbated for them by the wealth difference created by our economy. They create intense, revealing moments where we, the spectators become as clear as a still lake-water. This is because the unmentioned rule of zero interaction is shattered. This violation gives us a sense of unease.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*M7Ayuu-KBerO30R4fMPO3A.jpeg" /></figure><p>But there is a moral problem at hand here. Should we give them money? We, as riders are confronted by desperation. Some donate. Some shake heads. Most choose to stare aggressively into the ground, mobiles and their books. And many pretend-sleep. Refusing and ignoring the shameless person’s existence altogether. This is just one instance of powerful, yet subconsciously disregarded set of situations which certainly spans its own cosmos.</p><p>People who feel the whole thing is forced, start justifying their intent not to give, with one of the following reasons — what if they use the money for all the wrong reasons? Can’t really afford to donate every day, to everyone. Ten bucks is a bit too much, innit? And finally, the cliched one — They need to get a job, not beg.</p><p>In conclusion . . .</p><p>Not every problem you face on a day-to-day basis has definitive, extensible solutions. Although having no right or wrong answers, these problems tend to blow away your emotional shield, which is unfazed otherwise. One such situation I observed is when you face panhandlers while riding the subway.</p><p>If your inclination is to not donate, then the case is fairly restrictive. Done and dusted.</p><p>But if you intend to give, there’s a humongous spectrum of moral web you’ll have to clear off before you proceed. How much do you give? Every time? Who among all gets it? Should I carry spare change just for this?</p><p>Most of us don’t donate. Many of us decide to do it, but don’t, too. This tension ignites a pain in our heart and you end up becoming a victim of a monster within you.</p><p>Much of the content is a coalescence of my own experience riding the MRTS (metro / subway trains) and <a href="https://www.youtube.com/watch?v=T21v0zQDtUw&amp;t=79s">this vlog</a> in the YouTube channel <a href="https://www.youtube.com/user/Nerdwriter1">Nerdwriter1 </a>featuring Evan.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=51cd39c07655" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Validation of the Buddha’s diagnosis on suffering]]></title>
            <link>https://medium.com/@cs_bot/validation-of-the-buddhas-diagnosis-on-suffering-1bdd19ede31d?source=rss-34b69a74ecdb------2</link>
            <guid isPermaLink="false">https://medium.com/p/1bdd19ede31d</guid>
            <category><![CDATA[mind]]></category>
            <category><![CDATA[minimalism]]></category>
            <category><![CDATA[mindfulness]]></category>
            <category><![CDATA[buddhism]]></category>
            <category><![CDATA[psychology]]></category>
            <dc:creator><![CDATA[Sudharshan Chakravarthy]]></dc:creator>
            <pubDate>Tue, 06 Feb 2018 16:39:39 GMT</pubDate>
            <atom:updated>2018-02-12T09:41:20.970Z</atom:updated>
            <content:encoded><![CDATA[<p>The diagnosis provided by the Buddha on the topic ‘suffering’ or ‘dhukka’ (in Pali) pertains to the fact that we all constantly suffer and must strive to attain ‘nirvana’/ ’nibbana’ (the sole truth). I believe that although his discourse is concise, it has embraced all the vital aspects there is to it.</p><p>It is based on the assumption that all humans seek the state of happiness and are in a perennial struggle to attain it. And yet, some of us fall into the trap of temporary happiness (more suffering, when it ends). The Buddha offers an incisive insight into the issue. The prime idea is to provide a minimalist view of a convoluted and a grey area (which we hardly contemplate upon). Thus, the diagnosis served via ‘The four Noble Truths’, in my view point is valid.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*CBcjEdw7j5iiNZrzkPJNcA.jpeg" /></figure><p>The Buddha takes his teachings to the level of what a great scientist (in his or her field) would do now. The four vital steps in this avenue:</p><blockquote>(i) to identify the problem (dhukka)</blockquote><blockquote>(ii) deduce the cause of the problem (craving)</blockquote><blockquote>(iii) realize that a solution is possible (nirvana)</blockquote><blockquote>(iv) tread the path to the solution (towards nirvana).</blockquote><p>One has to take extra care and effort in discerning the meta-references made in the discourses. The Bhikku Bodhi in one of his lectures emphasizes this issue. Though many interpretations refer nirvana as the ‘one path to righteousness’, there can be many pathways (actions) which lead to this ‘path’ (figuratively). At a meditation retreat last year, all of us were asked to ‘observe’ the chain of thoughts that occur during the session. One key aspect is not the observation, but the realization that you are observing it. This seemed like a step above the usual thinking level, the one which is maintained by us in our perfunctory lives. Moreover, it was very much like playing a game ‘Temple Run’. Only that ‘falling’ into the traps meant start rambling along with the mundane thoughts.</p><p>The common frustration among many people is that the Buddha’s teachings seem to be incomplete. Many argue that,</p><blockquote>“Though the concepts are fine for studying and preaching, it is insufficient for ‘understanding’, since a disciple learns by examples. There is not one monk who has accounted for his configuration (the analysis of his state of ‘mind’) upon his own path to the nirvana (end of suffering)”.</blockquote><p>The above misconception can be refuted by many important inferences. First, the mind does not exist. It is the element of control that does. Contrary to the writings of Ramanuja (Vishitadwaita) and Adi-Shankaracharya (advaita), the concept of anatman (not-soul) profoundly exists in Buddhism. In the lecture, “Does Your-Self Exist?” by Dr. Wright, he elucidates the fact that we are made of 5 impermanent components. The ideal ‘configuration’ (immunity to suffering) is attained, not by some definitive path of actions, but by ultimate control (no materialistic craving). Second, the generality of the sermons indicate that there is no single ‘way’ to attain it. Each one has their own set of strengths and weaknesses, and only by exploring it further will the subsequent steps toward the end of ‘dhukka’ will burgeon.</p><blockquote>“Has the Buddha ignored some vital aspects of human life?”</blockquote><p>He says that we must be free from craving, as it brings ‘dhukka’. On the contrary, for you to attain ‘nirvana’, you have to first take care of your health! (For good health, one needs food; this craving toward food, is it a necessary evil?). For many of us, this seems almost paradoxical. But, as the Bhikku Bodhi elegantly handles this issue in his sermon, one must view this as an autonomous entity in itself. An ailing man, even with his physical difficulties, can attain ‘nirvana’. His path to nirvana is not affected by either his physical or mental shortcomings, as they are impermanent in nature. Thus, the ideal state does not necessarily mean a sound body or a sound mind; it is just immune to worry and suffering. Moreover, unlike in other religions, the Buddha’s teachings, there is no classification of a human based on birth, demographic or geographic identities. It is just the stage of a person’s configuration towards ‘saccha’ (the noble truth). This is because of his realization that these classifications in themselves were not permanent. Thus, the decision to not to divulge into impermanent specifics is justified.</p><p><strong>In conclusion</strong>, whether you are gloomy or elated, keep calm and say:</p><blockquote>“This too shall pass”.</blockquote><p>Sudharshan Chakravarthy</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1bdd19ede31d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>