<?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:media="http://search.yahoo.com/mrss/"><channel><title><![CDATA[Ashutosh Sanzgiri]]></title><description><![CDATA[Thoughts, stories and ideas.]]></description><link>https://www.ashutoshsanzgiri.com/</link><image><url>https://www.ashutoshsanzgiri.com/favicon.png</url><title>Ashutosh Sanzgiri</title><link>https://www.ashutoshsanzgiri.com/</link></image><generator>Ghost 4.48</generator><lastBuildDate>Thu, 19 Mar 2026 15:51:14 GMT</lastBuildDate><atom:link href="https://www.ashutoshsanzgiri.com/rss/" rel="self" type="application/rss+xml"/><ttl>60</ttl><item><title><![CDATA[Getting started with SageMaker Studio Lab]]></title><description><![CDATA[<p>I have just started playing with SageMaker Studio Lab and think it is a worthy contender to Google Colab. &#xA0;</p><p>Registration Link: <a href="https://studiolab.sagemaker.aws/">https://studiolab.sagemaker.aws/</a></p><p>FAQ: <a href="https://studiolab.sagemaker.aws/faq">https://studiolab.sagemaker.aws/faq</a></p><p>Like Google Colab, it is a free service, and is separate from your AWS subscription. </p><p>Some of the</p>]]></description><link>https://www.ashutoshsanzgiri.com/sagemaker-studio-lab/</link><guid isPermaLink="false">61bd3975b534fe000165ab4b</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Sun, 26 Dec 2021 23:58:40 GMT</pubDate><content:encoded><![CDATA[<p>I have just started playing with SageMaker Studio Lab and think it is a worthy contender to Google Colab. &#xA0;</p><p>Registration Link: <a href="https://studiolab.sagemaker.aws/">https://studiolab.sagemaker.aws/</a></p><p>FAQ: <a href="https://studiolab.sagemaker.aws/faq">https://studiolab.sagemaker.aws/faq</a></p><p>Like Google Colab, it is a free service, and is separate from your AWS subscription. </p><p>Some of the pros I see are:</p><ul><li>You get 15 GB persistent storage, so although you have to install all the python packages you need, they persist when you stop/restart your server.</li><li>You can use a Git workflow with your notebooks (clone git repo, push to github etc.). Note that for github, you can save your personal access token using the command <code>git config --global credential.helper store</code></li><li>You get a JupyterLab IDE and can install extensions.</li><li>You get a newer version of Python 3.9.7 (with Colab you get 3.7.12).</li></ul><p>Unlike Google Colab, there are no packages installed by default. So you do have to install them (albeit just once).</p><p>Like Google Colab, you can add an &quot;Open in Studio Lab&quot; &#xA0;badge to your github repo. </p>]]></content:encoded></item><item><title><![CDATA[Emotion Classifiers]]></title><description><![CDATA[<p>Just a collection of links at this point...</p><ul><li><a href="https://github.com/EleutherAI/gpt-neo/">GPT-Neo</a>: open-source implementation of GPT-2 / GPT-3 models (in progress)</li><li><a href="https://github.com/google-research/google-research/tree/master/goemotions">GoEmotions</a>: dataset from Google from Reddit comments</li><li>Emotion Classification for chatbots:</li><li>	<a href="https://github.com/google-research/google-research/tree/master/goemotions">ConveRT models</a></li><li>	<a href="https://arxiv.org/abs/2003.04807">Intent Detection paper</a></li><li>	ScenarioSA: Conversational <a href="https://github.com/anonymityanonymity/ScenarioSA">Dataset</a> and <a href="https://arxiv.org/abs/1907.05562">Paper</a></li></ul>]]></description><link>https://www.ashutoshsanzgiri.com/emotion-classifiers/</link><guid isPermaLink="false">6131acdea5bdf700012a71e9</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Mon, 25 Jan 2021 06:49:39 GMT</pubDate><content:encoded><![CDATA[<p>Just a collection of links at this point...</p><ul><li><a href="https://github.com/EleutherAI/gpt-neo/">GPT-Neo</a>: open-source implementation of GPT-2 / GPT-3 models (in progress)</li><li><a href="https://github.com/google-research/google-research/tree/master/goemotions">GoEmotions</a>: dataset from Google from Reddit comments</li><li>Emotion Classification for chatbots:</li><li>	<a href="https://github.com/google-research/google-research/tree/master/goemotions">ConveRT models</a></li><li>	<a href="https://arxiv.org/abs/2003.04807">Intent Detection paper</a></li><li>	ScenarioSA: Conversational <a href="https://github.com/anonymityanonymity/ScenarioSA">Dataset</a> and <a href="https://arxiv.org/abs/1907.05562">Paper</a></li></ul>]]></content:encoded></item><item><title><![CDATA[Getting Started with TradingView & PineScript]]></title><description><![CDATA[<p>The popular social trading site <a href="https://www.tradingview.com/">TradingView</a> uses a language called PineScript for creating indicators (that you can plot very easily) and strategies (that can be backtested directly on the platform).</p><p>Here are links to some resources for getting started with TradingView &amp; PineScript (mostly for my edification when I get</p>]]></description><link>https://www.ashutoshsanzgiri.com/pinescript/</link><guid isPermaLink="false">6131acdea5bdf700012a71e5</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Mon, 25 Jan 2021 06:02:51 GMT</pubDate><content:encoded><![CDATA[<p>The popular social trading site <a href="https://www.tradingview.com/">TradingView</a> uses a language called PineScript for creating indicators (that you can plot very easily) and strategies (that can be backtested directly on the platform).</p><p>Here are links to some resources for getting started with TradingView &amp; PineScript (mostly for my edification when I get back to this):</p><ul><li><a href="https://www.youtube.com/watch?v=qYMpPokST-k&amp;list=PLSP_1DBafH-Ew_z-tCVGav1XUnYqKIq_2">PineScript Tutorials on YouTube from the Art of Trading</a>: Also <a href="https://zenandtheartoftrading.com/pine-script-video-lessons/">here</a>.</li><li>Udemy Course: <a href="https://www.udemy.com/course/learn-tradingview-pine-script-programming-from-scratch">Learn TradingView PineScript Programming from Scratch</a></li><li>Another set of <a href="https://www.youtube.com/watch?v=YxDSFdvAei4&amp;list=PLM0BBafRCnRNhvMxyiqScZWV_h41raut6&amp;index=10">PineScript Tutorials on Youtube</a></li><li>Great Intro to the <a href="https://www.youtube.com/watch?v=iENn5rBHIAo">TradingView UI</a></li></ul><p>TradingView itself has some great documentation. &#xA0;And another great resource is <a href="https://kodify.net/tradingview-programming-articles/">Kodify</a>.</p><p></p>]]></content:encoded></item><item><title><![CDATA[How this website was created]]></title><description><![CDATA[<p>This website as well as <a href="https://www.hoopedondata.com/">HoopedOnData</a> has been created using Ghost. The site is hosted on an AWS t2.micro EC2 instance. More details are below.</p><hr><h2 id="aws-configuration-">AWS Configuration:</h2><ul><li>Use <a href="https://docs.bitnami.com/aws/apps/ghost/">Bitnami</a> Ghost AMI (free) from AWS Marketplace while creating your EC2 instance. </li><li>Set up Elastic IP to get a public IP</li></ul>]]></description><link>https://www.ashutoshsanzgiri.com/how-this-website-was-created/</link><guid isPermaLink="false">6131acdea5bdf700012a71e4</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Mon, 25 Jan 2021 05:49:58 GMT</pubDate><content:encoded><![CDATA[<p>This website as well as <a href="https://www.hoopedondata.com/">HoopedOnData</a> has been created using Ghost. The site is hosted on an AWS t2.micro EC2 instance. More details are below.</p><hr><h2 id="aws-configuration-">AWS Configuration:</h2><ul><li>Use <a href="https://docs.bitnami.com/aws/apps/ghost/">Bitnami</a> Ghost AMI (free) from AWS Marketplace while creating your EC2 instance. </li><li>Set up Elastic IP to get a public IP and point your IP to your domain using Route 53 as described <a href="https://medium.com/@sanzgiri/creating-ml-apps-the-easy-way-using-streamlit-ed480e369db2">here</a>. (the domain registrar for HoopedOnData is GoDaddy)</li><li>Login to your EC2 instance via ssh and obtain initial account password for Ghost admin user (user@example.com)</li><li>Next update the site for https access as described <a href="https://docs.bitnami.com/aws/faq/administration/generate-configure-certificate-letsencrypt/">here</a>:</li></ul><!--kg-card-begin: markdown--><pre><code>sudo cat /home/bitnami/bitnami_credentials
</code></pre>
<!--kg-card-end: markdown--><!--kg-card-begin: markdown--><pre><code>sudo /opt/bitnami/configure_app_domain --domain hoopedondata.com
sudo /opt/bitnami/bncert-tool
# At the prompt, enter: hoopedondata.com www.hoopedondata.com 
</code></pre>
<!--kg-card-end: markdown--><h2 id="ghost-configuration-">Ghost Configuration:</h2><ul><li>Refer to the latest <a href="https://docs.bitnami.com/aws/apps/ghost/">Ghost docs</a>.</li><li>Login to <a href="https://www.hoopedondata.com/ghost/#/staff">https://www.hoopedondata.com/ghost</a>/ using the initial set of credentials</li><li>Configure Ghost settings (General) and replace the default admin user with your profile.</li><li>Remove profile for &quot;Ghost&quot; user. This also removes the default posts.</li><li>Remove the banner by running the following command from the shell:</li></ul><!--kg-card-begin: markdown--><pre><code>sudo /opt/bitnami/apps/ghost/bnconfig.disabled --disable_banner 1
</code></pre>
<!--kg-card-end: markdown--><p>Finally, thanks to <a href="http://the-odd-dataguy.com/why-do-i-switch-from-github-pages-to-ghost/">this blog</a> for the inspiration to set up a blog using Ghost.</p>]]></content:encoded></item><item><title><![CDATA[Streamlit Jeopardy]]></title><description><![CDATA[<p>Sometime back, I created a <a href="https://www.streamlit.io/">Streamlit</a> App to practice Jeopardy questions. Initially deployed on heroku, it is now deployed on <a href="https://docs.streamlit.io/en/stable/deploy_streamlit_app.html">Streamlit&apos;s sharing service</a> and you can access it <a href="https://share.streamlit.io/sanzgiri/st_labeler/jeopardy.py">here</a>.</p><p>Here&apos;s a screen shot (the score tracking is still wonky!)</p><figure class="kg-card kg-image-card"><img src="/content/images/2021/01/image-1.png" class="kg-image" alt loading="lazy" width="1560" height="988" srcset="/content/images/size/w600/2021/01/image-1.png 600w, /content/images/size/w1000/2021/01/image-1.png 1000w, /content/images/2021/01/image-1.png 1560w" sizes="(min-width: 720px) 720px"></figure><p>The question bank for the app comes</p>]]></description><link>https://www.ashutoshsanzgiri.com/streamlit-deployments/</link><guid isPermaLink="false">6131acdea5bdf700012a71df</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Wed, 20 Jan 2021 07:34:07 GMT</pubDate><content:encoded><![CDATA[<p>Sometime back, I created a <a href="https://www.streamlit.io/">Streamlit</a> App to practice Jeopardy questions. Initially deployed on heroku, it is now deployed on <a href="https://docs.streamlit.io/en/stable/deploy_streamlit_app.html">Streamlit&apos;s sharing service</a> and you can access it <a href="https://share.streamlit.io/sanzgiri/st_labeler/jeopardy.py">here</a>.</p><p>Here&apos;s a screen shot (the score tracking is still wonky!)</p><figure class="kg-card kg-image-card"><img src="/content/images/2021/01/image-1.png" class="kg-image" alt loading="lazy" width="1560" height="988" srcset="/content/images/size/w600/2021/01/image-1.png 600w, /content/images/size/w1000/2021/01/image-1.png 1000w, /content/images/2021/01/image-1.png 1560w" sizes="(min-width: 720px) 720px"></figure><p>The question bank for the app comes from <a href="https://j-archive.com/">JArchive</a>. I have github action running in <a href="https://github.com/sanzgiri/jarchive">this repo</a> that updates the question bank on a weekly basis. </p><p>The code for the app is inspired by the <a href="https://github.com/gesteves/trebekbot">Jeopardy slack bot</a> and can be accessed from this <a href="https://github.com/sanzgiri/st_labeler/blob/master/jeopardy.py">github repo</a>.</p>]]></content:encoded></item><item><title><![CDATA[Non-Transitive Dice]]></title><description><![CDATA[<p>I came across non-transitive dice recently while reading <a href="https://www.amazon.com/dp/B07R8QSXYB/">Humble Pi</a> by Matt Parker.</p><figure class="kg-card kg-image-card"><img src="/content/images/2021/01/image.png" class="kg-image" alt loading="lazy" width="446" height="113"></figure><p>This triggered a bunch of research, wherein I found this <a href="https://www.mathartfun.com/dSpecial.html">Dice Shop</a>, this gem of a book by Bogomolny (<a href="https://www.cut-the-knot.org/index.shtml">Cut the Knot</a>) and I have ordered these <a href="https://www.amazon.com/gp/product/B07SJS3HMR/">blank dice</a> from Amazon to try some of these ideas</p>]]></description><link>https://www.ashutoshsanzgiri.com/non-transitive-dice/</link><guid isPermaLink="false">6131acdea5bdf700012a71e2</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Wed, 20 Jan 2021 07:15:11 GMT</pubDate><content:encoded><![CDATA[<p>I came across non-transitive dice recently while reading <a href="https://www.amazon.com/dp/B07R8QSXYB/">Humble Pi</a> by Matt Parker.</p><figure class="kg-card kg-image-card"><img src="/content/images/2021/01/image.png" class="kg-image" alt loading="lazy" width="446" height="113"></figure><p>This triggered a bunch of research, wherein I found this <a href="https://www.mathartfun.com/dSpecial.html">Dice Shop</a>, this gem of a book by Bogomolny (<a href="https://www.cut-the-knot.org/index.shtml">Cut the Knot</a>) and I have ordered these <a href="https://www.amazon.com/gp/product/B07SJS3HMR/">blank dice</a> from Amazon to try some of these ideas out with real dice.</p><p>In the interim, I created this <a href="https://colab.research.google.com/drive/1M5_0LpkxG5wQi5DOokUhYzlyfZZW08gm#scrollTo=I74mZeQMVqAO">Google Colab</a> to check on the probabilities of various events with the Grime Dice.</p><p>Also check this <a href="https://singingbanana.com/dice/article.htm">Singing Banana</a> site!</p><p><strong>To be continued...</strong></p>]]></content:encoded></item><item><title><![CDATA[SNS Scrape]]></title><description><![CDATA[<p><a href="https://github.com/JustAnotherArchivist/snscrape">SNS Scrape</a> is a scraping library in Python for scraping social media sites.</p><p>I tried this out in this <a href="https://colab.research.google.com/drive/1kvyIUKy2rr2OB8J9jOWaA7q2RwdgGR-_#scrollTo=m__hVpV9PHN5">Google Colab</a>. Does not work with Instagram or Facebook (at least the accounts I tried), but can be used for Twitter.</p><p>The first time I used it, I tried it with</p>]]></description><link>https://www.ashutoshsanzgiri.com/sns-scrape/</link><guid isPermaLink="false">6131acdea5bdf700012a71e0</guid><dc:creator><![CDATA[Ashutosh Sanzgiri]]></dc:creator><pubDate>Wed, 20 Jan 2021 06:56:13 GMT</pubDate><content:encoded><![CDATA[<p><a href="https://github.com/JustAnotherArchivist/snscrape">SNS Scrape</a> is a scraping library in Python for scraping social media sites.</p><p>I tried this out in this <a href="https://colab.research.google.com/drive/1kvyIUKy2rr2OB8J9jOWaA7q2RwdgGR-_#scrollTo=m__hVpV9PHN5">Google Colab</a>. Does not work with Instagram or Facebook (at least the accounts I tried), but can be used for Twitter.</p><p>The first time I used it, I tried it with <code>@realDonaldTrump</code> but that account is now permanently suspended and tweets are no longer available...</p><p>So instead, here are the tweets since the start of 2021 from Andrew Ng :-)</p><pre><code>2021-01-19 06:55:35+00:00: Today is Martin Luther King Jr. day. After the violence in the U.S. Capitol on Jan 6th, I am struck by how far MLK  advanced civil rights through nonviolent activism. Let us build on his legacy and keep finding peaceful ways to advance civil rights. https://t.co/nnCgFRi5oF

2021-01-14 00:57:40+00:00: Democracy has taken a lot of damage recently. Last Wednesday was a nadir for the U.S., and the path back will be long and hard. But I believe the country has reached a turning point. Here&apos;re my thoughts on the role AI can play to support democracy. https://t.co/rjaaetR5kW

2021-01-06 22:31:31+00:00: U.S. democracy is going through a dark day today with the Capitol building breached. But we will get through this, democracy will prevail, and we will learn from this  and strengthen our nation against future assaults. &#x1F1FA;&#x1F1F8;

2021-01-05 20:38:51+00:00: Congrats @OpenAI on really cool image generation result from a text+image prompt! Here&apos;s my favorite set of generated examples: https://t.co/sGXFNOc93n

2021-01-04 18:53:47+00:00: @JCastroS @RyanCampagna84 @andresbarreto @antoniogm @rabois @skillshare @AI_Fund @QuinteroCalle @landingAI Thanks @JCastroS. Yes I&apos;m really proud of our Colombia team, and am very glad we have an office there!

2021-01-01 20:06:43+00:00: Happy New Year, friends! &#x1F389; 2020 was a tough year for many, so lets make sure to build, support each other, and lift everyone up in 2021.
</code></pre><p></p><p></p>]]></content:encoded></item></channel></rss>