Hacker News Top Stories with Summaries (November 20, 2023)
<style>
p {
font-size: 16px;
line-height: 1.6;
margin: 0;
padding: 10px;
}
h1 {
font-size: 24px;
font-weight: bold;
margin-top: 10px;
margin-bottom: 20px;
}
h2 {
font-size: 18px;
font-weight: bold;
margin-top: 10px;
margin-bottom: 5px;
}
ul {
padding-left: 20px;
}
li {
margin-bottom: 10px;
}
.summary {
margin-left: 20px;
margin-bottom: 20px;
}
</style>
<h1> Hacker News Top Stories</h1>
<p>Here are the top stories from Hacker News with summaries for November 20, 2023 :</p>
<div style="margin-bottom: 20px;">
<table cellpadding="0" cellspacing="0" border="0">
<tr>
<td style="padding-right: 10px;">
<div style="width: 200px; height: 100px; border-radius: 10px; overflow: hidden; background-image: url('https://opengraph.githubassets.com/20a05aa82dffb376d6810e6a14e80c532dde4c061c53cc1d0516d03e58717bdb/cxli233/FriendsDontLetFriends'); background-size: cover; background-position: center;">
Friends don't let friends make bad graphs
Summary: GitHub user cxli233's repository "FriendsDontLetFriends" provides a guide on good and bad practices in data visualization. The guide covers various topics, including avoiding bar plots for means separation, not using violin plots for small sample sizes, and being cautious with bidirectional color scales for unidirectional data. It also discusses the importance of reordering rows and columns in heatmaps, checking data range at each factor level, and using appropriate color scales.
<div style="margin-bottom: 20px;">
<table cellpadding="0" cellspacing="0" border="0">
<tr>
<td style="padding-right: 10px;">
<div style="width: 200px; height: 100px; border-radius: 10px; overflow: hidden; background-image: url('https://substackcdn.com/image/fetch/w_1200,h_600,c_fill,f_jpg,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5dfbd169-eb7e-41e1-a050-556ccd6fb679_1600x672.png'); background-size: cover; background-position: center;">
Practical Tips for Finetuning LLMs Using LoRA (Low-Rank Adaptation)
Summary: Sebastian Raschka shares practical tips for finetuning large language models (LLMs) using Low-Rank Adaptation (LoRA). Key takeaways include: consistency in outcomes across multiple runs, QLoRA offering 33% memory savings at the cost of a 39% increase in runtime, optimizer choice being less critical, and the importance of adjusting LoRA rank and alpha value. LoRA enables efficient finetuning of 7B parameter models on a single GPU. The article also addresses common questions related to LoRA, such as dataset importance, domain adaptation, and selecting the best rank.