add README and add image example

blog/README.md

@@ -0,0 +1,36 @@
+# GitHub actions to a Blog from Markdown
+
+This is a simple Blog model, generated from posts in the `posts` folder.
+
+Create a new article in posts/blog-post-slug.md:
+
+    ---
+    title: Inference for fun and profit
+    excerpt: Generative AI is all the rage, and for good reason, but actually implementing an inference pipeline has many challenges, not the least of which is getting and serving a model to our peers.
+    categories:
+    - hands-on
+    - blog
+    layout: default.liquid
+    is_draft: false
+    ---
+    Generative AI is all the rage, and for good reason, but actually implementing an 
+    inference pipeline has many challenges, not the least of which is getting and serving
+    a model to our peers.
+
+
+    We want to build something like: ![Transformer Image](../assets/transformers-diffusion-image.png)
+
+    Let's build a simple inference API in a Jupyter notebook to help get us started.
+
+    ```python
+    import transformers
+    import fastapi
+
+    @app
+    def route("/"):
+        print("Hello World")
+    ```
+
+currently categories are not used, but could be exposed (or may be) via an updated index.
+
+create a PR against origin/gh-pages to post your article.

blog/_cobalt.yml

@@ -1,7 +1,7 @@
-
 site:
   title: Artificial Wisdom
   description: Our musings and thoughts on artificial intelligence, machine learning and all things wise
   base_url: https://blog.artificialwisdom.cloud
 posts:
   rss: rss.xml
+  jsonfeed: atom.json

blog/_site/README.html

@@ -0,0 +1,50 @@
+<h1>GitHub actions to a Blog from Markdown</h1>
+<p>This is a simple Blog model, generated from posts in the <code>posts</code> folder.</p>
+<p>Create a new article in posts/blog-post-slug.md:</p>
+<pre style="background-color:#2b303b;">
+<code><span style="color:#c0c5ce;">---
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">title: Inference for fun and profit
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">excerpt: Generative AI is all the rage, and for good reason, but actually implementing an inference pipeline has many challenges, not the least of which is getting and serving a model to our peers.
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">categories:
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">- hands-on
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">- blog
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">layout: default.liquid
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">is_draft: false
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">---
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">Generative AI is all the rage, and for good reason, but actually implementing an 
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">inference pipeline has many challenges, not the least of which is getting and serving
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">a model to our peers.
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">Let&#39;s build a simple inference API in a Jupyter notebook to help get us started.
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">```python
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">import transformers
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">import fastapi
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">@app
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">def route(&quot;/&quot;):
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">    print(&quot;Hello World&quot;)
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">```
+</span></code></pre>

blog/_site/index.html

@@ -0,0 +1,27 @@
+<!DOCTYPE html>
+<html>
+<head>
+  <meta charset="utf-8">
+  <title>Artificial Wisdom</title>
+  <description></description>
+  <link href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css" rel="stylesheet" integrity="sha384-eOJMYsd53ii+scO/bJGFsiCZc+5NDVN2yr8+0RDqr0Ql0h+rP48ckxlpbzKgwra6" crossorigin="anonymous">
+</head>
+<body>
+  <div class="container">
+    <h2 class="text-center">Artificial Wisdom</h2>
+    <div class="row">
+      <div class="col-lg-8 offset-lg-2">
+        <p>We hope to enlighten you on your journey in the world of Machine Learning and Artificial Intelligence,
+and in so doing, guide you along a path to Artificial Wisdom.</p>
+<h4><a href="posts/getting-started-with-retrieval.html">Getting Started with Retrieval</a></h4>
+<p>Getting started with retrieval, build your own RALM with RALMify</p>
+<h4><a href="posts/inference-for-fun-and-profit.html">Inference for fun and profit</a></h4>
+<p>Generative AI is all the rage, and for good reason, but actually implementing an inference pipeline has many challenges, not the least of which is getting and serving a model to our peers.</p>
+<h4><a href="posts/retro-and-RAG.html">RETRO and Retrieval Augmented Generation</a></h4>
+<p>A few thoughts on retrieval augmentation.</p>
+
+      </div>
+    </div>
+  </div>
+</body>
+</html>

blog/_site/posts/getting-started-with-retrieval.html

@@ -0,0 +1,26 @@
+<!DOCTYPE html>
+<html>
+<head>
+  <meta charset="utf-8">
+  <title>Getting Started with Retrieval</title>
+  <description></description>
+  <link href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css" rel="stylesheet" integrity="sha384-eOJMYsd53ii+scO/bJGFsiCZc+5NDVN2yr8+0RDqr0Ql0h+rP48ckxlpbzKgwra6" crossorigin="anonymous">
+</head>
+<body>
+  <div class="container">
+    <h2 class="text-center">Getting Started with Retrieval</h2>
+    <div class="row">
+      <div class="col-lg-8 offset-lg-2">
+        <p>We're going to cover RALMifying your data, which encompases a few steps:</p>
+<ol>
+<li>collect your datasets</li>
+<li>encode the dataset into a vector database</li>
+<li>incorporate the dataset into a RALM base model</li>
+<li>serve for inference and generative requests</li>
+</ol>
+
+      </div>
+    </div>
+  </div>
+</body>
+</html>

blog/_site/posts/inference-for-fun-and-profit.html

@@ -0,0 +1,32 @@
+<!DOCTYPE html>
+<html>
+<head>
+  <meta charset="utf-8">
+  <title>Inference for fun and profit</title>
+  <description></description>
+  <link href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css" rel="stylesheet" integrity="sha384-eOJMYsd53ii+scO/bJGFsiCZc+5NDVN2yr8+0RDqr0Ql0h+rP48ckxlpbzKgwra6" crossorigin="anonymous">
+</head>
+<body>
+  <div class="container">
+    <h2 class="text-center">Inference for fun and profit</h2>
+    <div class="row">
+      <div class="col-lg-8 offset-lg-2">
+        <p>Generative AI is all the rage, and for good reason, but actually implementing an 
+inference pipeline has many challenges, not the least of which is getting and serving
+a model to our peers.</p>
+<p>We want to build something like: <img src="transformers-diffusion-image.png" alt="Transfomer Image" /></p>
+<p>Let's build a simple inference API in a Jupyter notebook to help get us started.</p>
+<pre style="background-color:#2b303b;">
+<code><span style="color:#b48ead;">import </span><span style="color:#c0c5ce;">transformers
+</span><span style="color:#b48ead;">import </span><span style="color:#c0c5ce;">fastapi
+</span><span style="color:#c0c5ce;">
+</span><span style="color:#c0c5ce;">@</span><span style="color:#bf616a;">app
+</span><span style="color:#b48ead;">def </span><span style="color:#8fa1b3;">route</span><span style="color:#c0c5ce;">(&quot;/&quot;):
+</span><span style="color:#c0c5ce;">    </span><span style="color:#96b5b4;">print</span><span style="color:#c0c5ce;">(&quot;</span><span style="color:#a3be8c;">Hello World</span><span style="color:#c0c5ce;">&quot;)
+</span></code></pre>
+
+      </div>
+    </div>
+  </div>
+</body>
+</html>

blog/_site/posts/retro-and-RAG.html

@@ -0,0 +1,25 @@
+<!DOCTYPE html>
+<html>
+<head>
+  <meta charset="utf-8">
+  <title>RETRO and Retrieval Augmented Generation</title>
+  <description></description>
+  <link href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css" rel="stylesheet" integrity="sha384-eOJMYsd53ii+scO/bJGFsiCZc+5NDVN2yr8+0RDqr0Ql0h+rP48ckxlpbzKgwra6" crossorigin="anonymous">
+</head>
+<body>
+  <div class="container">
+    <h2 class="text-center">RETRO and Retrieval Augmented Generation</h2>
+    <div class="row">
+      <div class="col-lg-8 offset-lg-2">
+        <p>A few thoughts on retrieval augmentation.</p>
+<ol>
+<li>You should do it</li>
+<li>Building embeddings is the source of truth and the source of pain</li>
+<li>Managing datasets will lead you to retrieval salvation</li>
+</ol>
+
+      </div>
+    </div>
+  </div>
+</body>
+</html>

blog/_site/rss.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="utf-8"?><rss version="2.0"><channel><title>Artificial Wisdom</title><link>https://blog.artificialwisdom.cloud</link><description>Our musings and thoughts on artificial intelligence, machine learning and all things wise</description><item><title>Getting Started with Retrieval</title><link>https://blog.artificialwisdom.cloud/posts/getting-started-with-retrieval.html</link><description><![CDATA[<p>Getting started with retrieval, build your own RALM with RALMify</p>
+]]></description><guid>https://blog.artificialwisdom.cloud/posts/getting-started-with-retrieval.html</guid></item><item><title>Inference for fun and profit</title><link>https://blog.artificialwisdom.cloud/posts/inference-for-fun-and-profit.html</link><description><![CDATA[<p>Generative AI is all the rage, and for good reason, but actually implementing an inference pipeline has many challenges, not the least of which is getting and serving a model to our peers.</p>
+]]></description><guid>https://blog.artificialwisdom.cloud/posts/inference-for-fun-and-profit.html</guid></item><item><title>RETRO and Retrieval Augmented Generation</title><link>https://blog.artificialwisdom.cloud/posts/retro-and-RAG.html</link><description><![CDATA[<p>A few thoughts on retrieval augmentation.</p>
+]]></description><guid>https://blog.artificialwisdom.cloud/posts/retro-and-RAG.html</guid></item></channel></rss>

blog/posts/inference-for-fun-and-profit.md

@@ -4,14 +4,15 @@ excerpt: Generative AI is all the rage, and for good reason, but actually implem
 categories:
 - hands-on
 - blog
-published_date: 2023-12-14 21:44:45.167379 +0000
 layout: default.liquid
 is_draft: false
 ---
 Generative AI is all the rage, and for good reason, but actually implementing an 
 inference pipeline has many challenges, not the least of which is getting and serving
 a model to our peers.
 
+We want to build something like: ![Transfomer Image](../assets/transformers-diffusion-image.png)
+
 Let's build a simple inference API in a Jupyter notebook to help get us started.
 
 ```python