- 12
- 12 376
Maximilian Jesch
Приєднався 13 бер 2023
Everything you need to know about AI-Agents! From ReAct to Bee-Framework
AI Agents are the next big thing! And they are really easy to understand!
In those 30 Minutes I will explain to you everything you need to know! From the basics of "Reasoning and Acting" (ReAct) to implementing them using the Bee Framework, this video covers everything you need to know. Learn how to create AI Agents locally on your machine and explore advanced concepts like ReAct prompting, system dissection, and more.
Timestamps:
00:00 - Intro and Motivation
00:28 - What actually are AI-Agents?
01:52 - the building blocks: The Framework
03:13 - the building blocks: ReAct Prompting
04:20 - Chain of Thought and ReAct in Action
10:15 - Tools in React
12:09 - The Bee Framework
14:00 - getting started
17:25 - looking under the hood! Disecting the system prompt
20:22 - looking under the hood! Following the whole interaction
22:21 - why you need small models!
23:09 - creating agents in the bee UI
26:19 - AI for Enterprise vs. AI for Consumer
27:05 - Outro
In those 30 Minutes I will explain to you everything you need to know! From the basics of "Reasoning and Acting" (ReAct) to implementing them using the Bee Framework, this video covers everything you need to know. Learn how to create AI Agents locally on your machine and explore advanced concepts like ReAct prompting, system dissection, and more.
Timestamps:
00:00 - Intro and Motivation
00:28 - What actually are AI-Agents?
01:52 - the building blocks: The Framework
03:13 - the building blocks: ReAct Prompting
04:20 - Chain of Thought and ReAct in Action
10:15 - Tools in React
12:09 - The Bee Framework
14:00 - getting started
17:25 - looking under the hood! Disecting the system prompt
20:22 - looking under the hood! Following the whole interaction
22:21 - why you need small models!
23:09 - creating agents in the bee UI
26:19 - AI for Enterprise vs. AI for Consumer
27:05 - Outro
Переглядів: 1 233
Відео
RAG in IBM watsonx! Get your data into your LLM system! Easy, intuitive, scalable
Переглядів 1 тис.3 місяці тому
RAG is the workhorse of the ongoing Gen-AI Revolution! And watsonx is one of the best places to make it happen. The streamlit app github.com/aseelert/chat-with-documents Timestamps: 00:00 - Intro 00:34 - The Scenario 00:50 - What is RAG? 04:53 - Like and Subscribe! 05:01 - Set up a simple RAG system 07:16 - Deploy the RAG system 10:10 - Consume the RAG system via a Streamlit app 11:04 - Make th...
Instructlab in 20 minutes! (Or: How to train your LLMs in style?)
Переглядів 4,1 тис.6 місяців тому
Looking for a way to finetune your Large Language Models in an efficient, reproducible and scalable way? Want to use Llama or Mixtral to make your small models better? You have come to the right place! Meet Instructlab! github.com/instructlab the original github repo github.com/Max-Jesch/instructlab_demo the code that I used Timestamps: 00:00 - Intro 01:17 - What is Instructlab? 02:42 - Concept...
watsonx Code Assistant for Ansible in under 10 minutes
Переглядів 56610 місяців тому
Dive into the world of Ansible automation with Watsonx Code Assistant, and find out why it's the go-to for big companies managing critical systems. Forget about using ChatGPT for this stuff; when you're in the big leagues, Watsonx brings the A-game with privacy, transparency, and top-notch data. This walkthrough under 10 minutes not only gets you into what Ansible's all about but also rolls out...
watsonx.data in 10 minutes!
Переглядів 3,5 тис.11 місяців тому
This video provides an in-depth look at the concept of data lakehouses, highlighting their role in bridging the gap between data lakes and data warehouses. It includes a hands-on demonstration of IBM's watsonx.data, showcasing how it implements the data lakehouse concept through its architecture and querying capabilities. Timestamps: 00:00 - Intro 00:25 - Datalake vs Datawareouse 02:17 - Datala...
Build better applications with LLMs leveraging AI governance!
Переглядів 168Рік тому
In this video I want to show you how you can use watsonx.governance to build better applications that use LLMs. Timestamps: 00:00 - Intro 00:36 - check out watsonx 01:04 - IDE for prompt engineering 02:20 - getting some examples 03:40 - the old way to deploy 04:30 - evaluate your prompt! 06:35 - AI Factsheets 07:10 - deploy your prompt model 08:30 - use deployed model 09:40 - Outro
Analyze instagram comments using Large Language Models
Переглядів 700Рік тому
In this video I will show you how you can use LLama2 hosted in IBM watsonx.ai to analyse instagram comments. Find all the code here: github.com/Max-Jesch/watsonx_for_social_media If you have any questions feel free to contact me directly. www.linkedin.com/in/maximilian-jesch/ Timestamps: 00:00 - Intro 00:57 - getting the data 02:28 - watsonx.ai and prompt engineering 07:08 - granite model 07:33...
Install IBM Cloud Paks the easy way
Переглядів 635Рік тому
In this video I will show you how to install Cloud Pak for Data (you can also install other Cloud Paks) using an awesome Tool called the "Cloud Pak deployer". We will be deploying to IBM Cloud, but AWS and Azure are supported as well. link to the repository containing the configuration: github.com/Max-Jesch/cpd-config-for-yt link to the cloud pak deployer: github.com/IBM/cloud-pak-deployer Time...
Machine Learning with IBM Watson Studio
Переглядів 268Рік тому
In this video I will show you how to upload your data to a DB2 database and use it to train a neural network in watson studio and deploy it in watson machine learning Code used in Video: github.com/Max-Jesch/mlops_demo Timecodes 0:00 - Intro 0:12 - Setup DB2 Database 1:04 - Setup Watson Studio 1:45 - Setup Watson Machine Learning 2:05 - Setup Cloud Object Storage 2:45 - First Look at Cloud Pak ...
Introduction to AI 2023 The Low Math way to understand Neural Networks
Переглядів 89Рік тому
The recording of a talk I gave at the "Data, Cloud and AI in Switzerland" meetup group: www.meetup.com/big-data-developers-switzerland/ The code and the slides can be found here: github.com/Max-Jesch/Introduction-to-AI Timestamps: 00:00 Introduction 02:15 Agenda 03:35 Historical overview 10:04 Overview of machine learning approaches 13:20 Simple real world example 15:49 Hands-on Neural Network ...
Doing AI governance and ML ops right! part 1 of 2
Переглядів 107Рік тому
In this video I will show you how to use IBM Watson AutoAI to create a really good ML model and how to use Watson Studio with AIfactsheets and Watson Knowledge Catalog to keep track of your models so you avoid trouble with the regulators. Additionaly we will use and Watson Machine Learning to deploy our model. This is the link to the repository: github.com/Max-Jesch/aigovernance_video Timestamp...
That is exactly I was looking for. Precisely, on the point and enjoyable to watch. Keep going and I'm glad to see the next vid - great work Max. Thank you.😀
Great stuff Maxi. Really looking forward to give it a shot to ilab :)
Not sure why would anybody do a video using an open-source project, specifically made for people who want to train models locally and use cloud.
Because this is the first stage of what will be part of the IBM portfolio in the future. Right now it purely open source though
Awesome explanation. Very good Max. Hope this helps other audiences.
is this whole process free?
Consuming the LLMs through watsonx does produce some cost. It is probably fairly little and you should be fine with the credits you get in the free trial
Hi Max! Great video. Very informative. I had a question, do all foundation models allow for chain of thought and ReAct prompting? Is this something new that developed in the granite-3-8b instruct model?
Thanks :-) This works with all models, even fairly old ones. Actually newer models (like granite-3-8b) have a lot chain-of-thought examples in their training data so they tend to do that automatically without you even having to tell them
I don't agree, we can easily build simple agents with chatpt. Don't need yet another costly tool to create simple agents.
Agents are basicly Prompt Chaining, where each prompt has access to real world connectivity and include it in the prompt. It can target different llm to generate its basic needs based on the capabilities given to that prompt, pack it into a box, and call it an agent.
Pretty good summary. Now just add some basic logic to the framework to handle all the errors that the LLM produces when trying to connect to the real world and you have created a descently reliable system
That was a great video. I also really like instructlabs approach to training.
Thanks a Lot 😁
I have a requirement to upload Excel files and query to the same , can we do that using the same
Great video! Thank you! : )
I feel the most useful way to take advantage of instruct lab is to fine tune it with the functionality AND the structured outputs you require for your specific agent workflow. What do you think?
That makes a lot of sense! Fine-tuning InstructLab to match the specific functionality and structured outputs your workflow needs is a smart way to get the most out of it and ensure it works exactly how you want.
Thanks , can you add more details about the cloud setup?
Do you mean the cloudsetup that I used? It is super straight forward. Setup the VM and then do a pip install. Or do you mean the SaaS offering of instructlab?
@@maxjesch thanks
There isn't much about this type of analysis anywhere 😅
Very informative video!
Thank you, Please never stop posting .....
Thanks a lot. I will try my best :-)
Like always great content keep it up brother
Thanks a lot! Will do :-)
Thanks this is really helpful!
Awesome content! Thanks a lot for sharing! Great how far watsonx has come....
Great intro 👏👏. Thanks a lot.
Thank you, and interesting video which covers a lot of questions very quickly!
This is a great video!
Can you share some more insights from your experience on what is additionally required in step one which may be added to Instructlab by the community?
Hey and thanks for your comment :-) What do you mean with step 1?
Hallo Max, tolle Darstellung - und zeigt was Opensource in LLMs wirklich bedeuten kann und das smaller besser sein kann.
Is the Infrastructure manager available in the watsonx trial?
nice!
Awesome content 👍
Gooo Max! great one!
Excellent video, really informative - please make more!
Great work!👏
Cool video. Please more of it 👍🏻
awesome video! Thanks a lot!
Amazing! Thanks for sharing 😊
My pleasure 😊
Thanks Max, brilliant video and content, just a question if we need some credits $ on IBM cloud trail account how we would get it? Could you please help on this?
Thanks for the feedback :-) The trial account usually comes with some free quota I think. You can just register and start working. If you have any further question shoot me a message either here or on linkedin
Awesome! Truely easy! Thanks Max for sharing.
Thank you Max great content. 🚀
Thx Max for sharing, I like 😀