chore: add llm friendly resources.

rag-advanced/resources/Chapter 0 - extras/Chapter00.md

@@ -0,0 +1,65 @@
+## Chapter 0: Setup
+
+<a target="_blank" href="https://colab.research.google.com/github/wandb/edu/blob/main/rag-advanced/notebooks/Chapter00.ipynb">
+  <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
+</a>
+
+<!--- @wandbcode{rag-course-00} -->
+
+Let's install the required packages and check our setup for this course.
+
+### 🎉 Free Cohere API key
+
+Before you run this colab notebook, head over to this [link to redeem a free Cohere API key](https://docs.google.com/forms/d/e/1FAIpQLSc9x4nV8_nSQvJnaINO1j9NIa2IUbAJqrKeSllNNCCbMFmCxw/viewform?usp=sf_link).
+
+Alternatively if you have a Cohere API key feel free to proceed. :)
+
+
+```
+!pip install -qq weave cohere
+```
+
+## 1. Setup Weave
+
+
+The code cell below will prompt you to put in a W&B API key. You can get your API key by heading over to https://wandb.ai/authorize.
+
+
+```
+# import weave
+import weave
+
+# initialize weave client
+weave_client = weave.init("rag-course")
+```
+
+## 2. Setup Cohere
+
+The code cell below will prompt you to put in a Cohere API key.
+
+
+```
+import getpass
+
+import cohere
+
+cohere_client = cohere.ClientV2(
+    api_key=getpass.getpass("Please enter your COHERE_API_KEY")
+)
+```
+
+## A simple-turn chat with Cohere's command-r-plus
+
+
+```
+response = cohere_client.chat(
+    messages=[
+        {"role": "user", "content": "What is retrieval augmented generation (RAG)?"}
+    ],
+    model="command-r-plus",
+    temperature=0.1,
+    max_tokens=2000,
+)
+```
+
+Let's head over to the weave URL to check out the generated response.

rag-advanced/resources/Chapter 0 - extras/RAG-0.1-course intro.srt

@@ -0,0 +1,100 @@
+1
+00:00:00,000 --> 00:00:03,000
+Hi, we are happy to welcome you to take our RAG++ course.
+
+2
+00:00:03,000 --> 00:00:08,000
+If you have built a RAG-powered system or have developed a proof of concept
+
+3
+00:00:08,000 --> 00:00:10,000
+but lack the confidence to deploy it into production,
+
+4
+00:00:10,000 --> 00:00:12,000
+this course will help bridge that gap.
+
+5
+00:00:13,000 --> 00:00:16,000
+Over the last 21 months, we have been running wandbot,
+
+6
+00:00:16,000 --> 00:00:19,000
+a live customer support bot for weights and biases,
+
+7
+00:00:19,000 --> 00:00:23,000
+and we're excited to share what we have learned from the experience.
+
+8
+00:00:23,000 --> 00:00:27,000
+We will highlight what separates a good RAG system from a great one
+
+9
+00:00:27,000 --> 00:00:29,000
+by focusing on evaluation-driven development,
+
+10
+00:00:29,000 --> 00:00:32,000
+emphasizing the importance of an effective evaluation pipeline.
+
+11
+00:00:33,000 --> 00:00:36,000
+We will explore different metrics for evaluation,
+
+12
+00:00:36,000 --> 00:00:39,000
+like MRR, NDCG, and LLM evaluators.
+
+13
+00:00:39,000 --> 00:00:44,000
+We will also see how we can align an LLM evaluator with human feedback.
+
+14
+00:00:44,000 --> 00:00:50,000
+Next, we will dive into advanced RAG components for sophisticated data ingestion strategies,
+
+15
+00:00:50,000 --> 00:00:55,000
+effective metadata utilization, and query enhancement techniques like query decomposition.
+
+16
+00:00:55,000 --> 00:00:59,000
+We will improve retrieval quality using methods like metadata filtering,
+
+17
+00:00:59,000 --> 00:01:00,000
+routing, and re-ranking.
+
+18
+00:01:01,000 --> 00:01:03,000
+We also have two guest lecturers.
+
+19
+00:01:03,000 --> 00:01:07,000
+Charles from Weaviate will cover production-grade hybrid retrieval systems,
+
+20
+00:01:07,000 --> 00:01:13,000
+and Meor from Cohere will show how to incorporate tool-use capabilities into your RAG system.
+
+21
+00:01:13,000 --> 00:01:18,000
+The course is packed with practical insights to enhance response synthesis
+
+22
+00:01:18,000 --> 00:01:21,000
+and methods to optimize your RAG pipeline's latency.
+
+23
+00:01:22,000 --> 00:01:26,000
+By the end of this course, you will have the tools and confidence to take your RAG systems
+
+24
+00:01:26,000 --> 00:01:29,000
+from POC to production-led deployments.
+
+25
+00:01:29,000 --> 00:01:34,000
+We hope you enjoyed the course as much as we enjoyed creating it.
+

rag-advanced/resources/Chapter 0 - extras/RAG-0.2-setup.srt

@@ -0,0 +1,84 @@
+1
+00:00:00,000 --> 00:00:04,000
+Before you get started, let us set you up to get the most out of this course.
+
+2
+00:00:04,000 --> 00:00:07,000
+Throughout the course we will be recommending useful tools.
+
+3
+00:00:07,000 --> 00:00:12,000
+However we will mainly be using Weave, Cohere and Weaviate.
+
+4
+00:00:12,000 --> 00:00:17,000
+Weave is a lightweight toolkit for tracking and evaluating your LLM applications.
+
+5
+00:00:17,000 --> 00:00:24,000
+We will be using Cohere family of models for text generation, re-ranking and embedding.
+
+6
+00:00:24,000 --> 00:00:27,000
+Cohere is also providing with free credits for this course.
+
+7
+00:00:27,000 --> 00:00:32,000
+Check out the instructions under this lesson to redeem your free Cohere API key for this
+
+8
+00:00:32,000 --> 00:00:33,000
+course.
+
+9
+00:00:33,000 --> 00:00:37,000
+Once you have the API keys, head over to the Colab notebook for this lesson.
+
+10
+00:00:37,000 --> 00:00:40,000
+We first install the required packages.
+
+11
+00:00:40,000 --> 00:00:42,000
+We then initialise our Weave client.
+
+12
+00:00:42,000 --> 00:00:45,000
+This is where you will put in your wandb API key.
+
+13
+00:00:45,000 --> 00:00:48,000
+We then set up our Cohere client.
+
+14
+00:00:48,000 --> 00:00:52,000
+We will then use the Cohere model to ask about retrieval augmented generation.
+
+15
+00:00:52,000 --> 00:00:56,000
+If everything works fine, you will have a Weave URL.
+
+16
+00:00:56,000 --> 00:00:59,000
+Let's head over to it.
+
+17
+00:00:59,000 --> 00:01:03,000
+This is a single trace of the Cohere client chat method.
+
+18
+00:01:03,000 --> 00:01:08,000
+We will automatically capture the inputs to the client as well as the generated response.
+
+19
+00:01:08,000 --> 00:01:15,000
+We also keep track of the used tokens and the time taken to complete the generation.
+
+20
+00:01:15,000 --> 00:01:17,000
+Now let's get started with the course.
+
+21
+00:01:17,000 --> 00:01:18,000
+Enjoy!
+

rag-advanced/resources/Chapter 0 - extras/RAG-0.3-course outro.srt

@@ -0,0 +1,52 @@
+1
+00:00:00,000 --> 00:00:02,000
+That concludes our RAG++ course.
+
+2
+00:00:02,000 --> 00:00:04,000
+Thank you for joining us and taking the time
+
+3
+00:00:04,000 --> 00:00:05,000
+to enhance your skills
+
+4
+00:00:05,000 --> 00:00:08,000
+in building production-ready RAG systems.
+
+5
+00:00:08,000 --> 00:00:09,000
+We appreciate your commitment
+
+6
+00:00:09,000 --> 00:00:12,000
+and hope you found the course valuable.
+
+7
+00:00:12,000 --> 00:00:14,000
+Your feedback is important to us.
+
+8
+00:00:14,000 --> 00:00:16,000
+Please take a moment to leave a review
+
+9
+00:00:16,000 --> 00:00:18,000
+and share your thoughts.
+
+10
+00:00:18,000 --> 00:00:19,000
+Thank you once again,
+
+11
+00:00:19,000 --> 00:00:21,000
+and we look forward to seeing the innovative solutions
+
+12
+00:00:21,000 --> 00:00:23,000
+you will be creating.
+
+13
+00:00:23,000 --> 00:00:25,000
+Best of luck and happy building.
+