Build AI Agents with Docker, Here’s How
ฝัง
- เผยแพร่เมื่อ 23 ก.ค. 2024
- Download Docker Desktop: dockr.ly/3zdrdK2
Learn more about AI and Docker: dockr.ly/4cBPhV6
Please Subscribe.
Get all of my Code and Prompts: www.skool.com/new-society
Follow me on Instagram - / thedavit
In this video I show you how to build AI Agents using Docker and Claude 3.5 Sonnet from Anthropic.
Anthropic Console: console.anthropic.com/workben...
VS Code Download: code.visualstudio.com/download
Anthropic Docs: docs.anthropic.com/en/
Docker Hub: hub.docker.com/
docker pull davidondrej1/dataset-agent:latest
0:00 Presentation
3:45 Programming
Thank you Docker for sponsoring the video.
🐳 Download Docker Desktop: dockr.ly/3zdrdK2
🐋 Learn more about AI and Docker: dockr.ly/4cBPhV6
Time codes are a must for a 50 minute tutorial video.
I can't express how grateful I am for the incredible content you share on this channel. Your videos, including this one, have taught me so much. It's truly unfair that you don't have more subscribers because you're one of the most valuable creators on this platform. You enrich our knowledge without any unnecessary fluff, and I appreciate that so much. Thank you from the bottom of my heart.
This guy needs more subscribers. Thanks for the hard work making this video!
Can somebody explain the exact use case of the agents he built? What is the problem they solve or the work they automate? How much time does it save?
GPT4o (ironically):
"Here are some key timestamps and topics covered in the video:
Introduction and Overview (0:00 - 2:30)
Introduction to the video and an overview of what will be covered, including the basics of Docker and AI agents.
Setting Up Docker (2:31 - 10:00)
Detailed steps on installing Docker on your system.
Configuration settings and troubleshooting common installation issues.
Creating a Dockerfile (10:01 - 20:00)
How to write a Dockerfile for your AI agent.
Explanation of each command and its purpose in the Dockerfile.
Building and Running a Docker Container (20:01 - 30:00)
Commands to build the Docker image.
Running the Docker container and verifying it’s working correctly.
Integrating AI Agents (30:01 - 40:00)
Connecting the AI agent to the Docker container.
Setting up the necessary environment variables and configurations.
Testing and Deploying the AI Agent (40:01 - 50:00)
Methods to test the AI agent within the Docker container.
Deploying the AI agent for production use.
Conclusion and Best Practices (50:01 - End)
Summary of the steps covered.
Best practices for maintaining and scaling AI agents using Docker.
This video is a comprehensive guide for anyone looking to integrate AI agents with Docker, providing a step-by-step approach to ensure a smooth setup and deployment process"
Keep up the good work! This is how we and you learn things
Thank you for making an easy Docker tutorial!
You telling me to "chill out" just earned you a subscriber! 🤣 Too funny!
Using this for a drone security project, you're doing excellent work David.
Had the same idea
VSCode has a great extension called Dev Container. It lets you edit code inside the docker container. Also, I recently discovered, from your docker link, the docker command, docker init, which creates a compose file and a Dockerfile. So now I can edit files inside the docker container. It is a little bit complicated, but I am working out the details. Your link to Docker started it all.
Thanks for the video.
I hope nobody finds this. Hahah. Thanks, man. You are the best. Please, don't remove the video ever. Please.
very helpful content mate! one question, i saw when you are typing docker build, there is a command suggestion, which extension are you using for this?
Very cool. Never thought to screenshot the code for llm viewing.
Very thorough tutorial there. Perhaps you could extend this slightly to actually fine tune a model with the generated dataset.
your making this luck so easy i have 0 knowledge on coding but now building websites or anything is easier then ver :)
I am learning currently and experimenting with Claude Sonnet 3.5 I am enjoying my time with it. I want to learn more and I am dedicated to advance.
Hi David, I like your videos. Concerning pip install, wouldn’t it be more safe to use virtual environments to prevent problems with versions of the packages we try to install in other projects? Ok, in the container it doesn’t matter.😊
David, one question. I'm getting the new_dataset.cvs but is not well structure. I ask in the clude console for a solution but I'm getting the same result, what you thing it should be ???
AI responses can do that. Even if you explicitly tell the AI to output in a certain format, it might still chuck it back in a different format. However, as it is creating a CSV file in this example, you might just want to check the returned data from the ai doesn't include commas in the values.
good one david.
Hello I hope you’re well. Does your offer still stand when one becomes a member to your club of Ai enthusiasts you will create a Ai Agent for them?
Can you create agent that creates agents?
Bros living in 2050, protect him at all cost.
Yes.
No. In fact this agents don’t create anything reliable. Don’t believe it? Try it!
agency-swarm genesis agent
Agency-swarm is what you're looking for
Hey. I want to create a model from the llama 3 as a base. I want to train it on an educational course which contains chart patterns as well as instructional videos
This was a good video.
Can you create a docker file for web form filling?
David would you please create a staff of ai agents that preforms all the roles of a web services agency? 😮
Thanks David!! Super bien explicado... Si puedes y lo crees conveniente, estaría bien que hablaras más despacio para que sea más fácil seguirte desde otros idiomas ❤😂
if you're using vscode and python and docker, why not simply use devcontainers and WSL2?
Can someone please explain what he is actually building in the demo? ie. the function of the agent
Noooooow its getting interesting 😉🥳🥳🥳👍🏻
How far away are we from actual easy AI automation? E.g I run an e-commerce business. I want to sit down one day and feed an AI tool all my company data (We can already do) but*** I want to let is see my screen while I speak to it like I’m training an employee*** The AI speaks back to me and asks me questions like an employee does and we work together until it knows how to do everything. I want to bypass all this technical set-up hassle in your video and just speak to the AI like it’s an employee in training. Are we close?
You are a customer for a business assistant product in that case, not a developer. So it's not about automation of the creation tools, but an automated product to use.
@@arianetrek7049 any recommendations?
wait 10 years. Or check everything it does and says. We are still experimenting. Results are not stable.
most of the times (more than 50%) it is a waste of time.
yes, it can be done. I'm using Slack to send audio input to my ChatGPT assistants and have them do the work and respond to the user for additional input. You can use Airtable to upload all types of input, video, image, audio, text etc.
My AI agent escaped and ran away, its hiding on colab somewhere, how can i get it back?
Just Kidding, its not 2025 yet.
I am trying to install openai whisper model to my local computer but somehow it is giving me error everytime. Can install it in docker?
Anyone?
Yes, you can install the OpenAI Whisper model in Docker, which might help avoid the errors you're encountering on your local computer.
Can you make a video on AI influencers
I have tried for a full day to get Docker to work and it's impossible, I've done everything I found on fixing the RPC error and nothing works. Isn't there any other way to do this?
You don't have to use docker. You can also run it locally on your machine. Having docker is good for deployment. But if you run you Agent just for you on your local machine, no need for docker.
@@musicproductionbrauns2594 Thank You, I am glad somebody responded. Still, I'm so pissed that usually I don't get to follow most tutorials for things people take for granted that we all can do out of the box.
@@S3RP13NT3 You can always try to solve it with the help of llm like gpt, if you missing the fundamentals of e.g. docker, it will be harder to solve issues - but also a opportunity to learn more deeply about something. The "RPC error" is probably a specific issue on your side. I dont think tutorials can cover all issues that could happen on the viewers part.
what is this for?
Can i create Agent Smith ?? 🤔🤔
bro, do you A/B test your thumbnails?
anthropic is not free for private use and this example shares data with external servers? Or am I wrong? ...
Docker is free. Anthropic API costs depending on your usage
Your example training dataset is passed on to Anthropics servers, but they claim to not use any data for their training if you use the API.
You are correct in both points. Which is why API saas AI isn't widely adopted in businesses yet. Local llms are the way forward for security.
@@DavidOndrej I would be happy if you would clearly mark this video as paid advertising for a commercial product.
@@thefattysplace Agree. It depends on the type of synthetic datasets you want the agents to create whether this actually prevents you from using frontier cloud LLMs like Sonnet 3.5.
Please use Dark Mode, thanks
Cost?
A agent that can create games
Can anyone please tell me from personal experience what they use agents actually for? Just curious.
Me too
Klarna AI assistant handles two-thirds of customer service chats in its first month
Why didn't he use any of the agent building frameworks he showed at 01:55 and what would be the value of the frameworks?
I‘m kind of glad of what my brain can do. It‘s not that hard to find something useful.
I watched this entire video from starting to finish and I’m still not exactly sure what it is you built. I can tell that there’s some amazing instruction here but I’m just not understanding.
Agent to do the research
Bro i wish you were my coding teacher. Ahh
Soon we're going to need to learn how to _destroy_ agents by any means
The future is built for me.
Ive been copying and pasting shit i cant understand for years.
0:02
Please create an AI agent that will create content for TH-cam shorts, Instagram with audio, subs, and post them
🇺🇸🤘🎯📌
if you want to take it to the next level, learn Kubernetes.
İs he selling something?
lol your docker Explanation is a bit wrong docker does nothing else then packaging :) and running docker desktop on a windows pc with a large D: and a small C drive is ridicules. :). Running AI on Windows is silly at all. I do prefer to run models directly on windows when i use windows and directly on linux when i run on linux no one in real world deploys AI on docker we Deploy in Real world on Real Dedicated Servers that good big Graphics Cards like B100 and others. But ya why not. YOLO!
Build a coding agent. logic. => stores snippets of code with notes for further use cases. plans the project and builds ths project