- 33
- 123 248
Adam Lucek
Приєднався 14 гру 2014
teach them to long for the endless immensity of the sea
Documenting my learning journey with the world
For inquiries, refer to the email in my links section.
Documenting my learning journey with the world
For inquiries, refer to the email in my links section.
Unleash a SWARM of AI Agents: Reliable Multi-Agent Orchestration
Should we be worried about the hive mind?
Resources:
github.com/openai/swarm/tree/main
github.com/ALucek/swarm-meal-planner
cookbook.openai.com/examples/orchestrating_agents
Chapters:
00:00 - Swarm Introduction
01:16 - What Problem Does Swarm Solve?
03:51 - The Swarm Method
06:22 - Routines
08:28 - Handoffs
10:40 - Building: Agent Architecture Planning
11:35 - Building: Defining the Tools
13:47 - Building: Creating the Handoffs
14:41 - Building: System Prompts
16:54 - Building: Putting Together the Agents
17:12 - Building: Context Variables
18:11 - Demonstration!
21:22 - Post Demo Observations
23:26 - Looking Forward
#ai #programming #openai
Resources:
github.com/openai/swarm/tree/main
github.com/ALucek/swarm-meal-planner
cookbook.openai.com/examples/orchestrating_agents
Chapters:
00:00 - Swarm Introduction
01:16 - What Problem Does Swarm Solve?
03:51 - The Swarm Method
06:22 - Routines
08:28 - Handoffs
10:40 - Building: Agent Architecture Planning
11:35 - Building: Defining the Tools
13:47 - Building: Creating the Handoffs
14:41 - Building: System Prompts
16:54 - Building: Putting Together the Agents
17:12 - Building: Context Variables
18:11 - Demonstration!
21:22 - Post Demo Observations
23:26 - Looking Forward
#ai #programming #openai
Переглядів: 383
Відео
Make An AI Agent with OpenAI’s Advanced Voice Mode
Переглядів 576День тому
To try everything Brilliant has to offer-free-for a full 30 days, visit brilliant.org/AdamLucek/ You’ll also get 20% off an annual premium subscription! Resources: github.com/ALucek/openai-realtime-rag github.com/openai/openai-realtime-console openai.com/index/introducing-the-realtime-api/ platform.openai.com/docs/guides/realtime/quickstart platform.openai.com/docs/api-reference/realtime open.u...
Master Reinforcement Learning With These 3 Projects
Переглядів 68614 днів тому
Too locked in to realize my hair was sticking up most the time Resources: github.com/ALucek/three-RL-projects gymnasium.farama.org/ huggingface.co/learn/deep-rl-course/ Chapters: 00:00 - Intro 00:52 - What Is Reinforcement Learning? 03:48 - Q-Learning: Introduction 05:33 - Q-Learning: Environment Setup 07:46 - Q-Learning: Hyperparameters Explained 11:09 - Q-Learning: Defining Rewards 13:18 - Q-...
How I Made A Deep Learning Robot
Переглядів 2,5 тис.Місяць тому
I made an AI powered robot! If you want to see more videos like this and support the channel, you can subscribe and join my UA-cam membership for $0.99 a month, or donate a custom amount through Super Thanks! Resources: LeRobot Repo: github.com/huggingface/lerobot Koch v1.1 Repo: github.com/jess-moss/koch-v1-1 Full Guide: github.com/huggingface/lerobot/blob/main/examples/7_get_started_with_real...
Does Fine Tuning Embedding Models Improve RAG?
Переглядів 1,4 тис.Місяць тому
Can fine tuning embedding models improve your RAG application? Yes! And it doesn’t even have to be that complicated. In this video we show how to train a query only linear adapter on your own RAG data to improve your document retrieval accuracy- a lightweight approach that can be applied to any embedding model without needing to fully fine tune the model itself, OR re-embed your knowledgebase. ...
Create AI Images of YOU with FLUX (Training and Generating Tutorial)
Переглядів 6 тис.Місяць тому
One of the major use cases of AI image generators is being able to generate consistent photos of yourself in any situation or scene. With the recent release of the FLUX.1 models from Black Forest Labs, being able to do this in an open source setting is becoming easier and higher quality. In this video we cover training FLUX.1 Dev on a few pictures of yourself to reliably generate AI images of y...
E-Commerce Will Never Be the Same With AI
Переглядів 5152 місяці тому
Many articles are out there claiming AI will revolutionize the world of e-commerce, but how specifically? In this video we cover 5 direct examples and applications of modern AI models into e-commerce scenarios, covering: Virtual Try On Diffusion Models Personalized Product Photos via Diffusion Model Training Embedding model based product recommendations Language Model Shopping Copilots Language...
Wait... What REALLY Is A Vector Database?
Переглядів 3,1 тис.2 місяці тому
Vector databases and similarity based search/retrieval have had a massive increase in popularity with the rise of language models and retrieval augmented generation (RAG) pipelines. Many people are using VDBs like pinecone, chroma, and mongodb, but how many of us actually know how they works? In this video we cover how vector databases operate, diving into what an embedding is, how similarity i...
Language Model Merging - Techniques, Tools, and Implementations
Переглядів 1,2 тис.2 місяці тому
Model merging is an innovative approach in the field of language modeling that allows researchers and practitioners to combine multiple models into a single, more capable model without the need for additional training. This technique addresses the challenges of building high-performance models, which typically require significant time, resources, and computational power. Resources: Code: github...
Model Distillation: Same LLM Power but 3240x Smaller
Переглядів 8 тис.2 місяці тому
Foundation model performance at a fraction of the cost- model distillation is a powerful technique to leverage the advanced generation capabilities of foundation models like Llama 3.1 405B, GPT-4, or Claude Opus as teachers, distilling their knowledge and performance on a given task to a student model. The result is a task-specific lightweight language model that provides the same performance, ...
Is Synthetic Data The Future of AI? (And How To Make Your Own)
Переглядів 1,6 тис.3 місяці тому
Language model’s text generating skills have hit a point where the quality is good enough to start using as synthetic data for other LLMs- Literally LLMs teaching LLMs! In this video we go over why synthetic data is useful for modern day AI, how language models are being optimized through generated data, and how to make your own using GPT-4o and LangChain. Resources: Code: github.com/ALucek/syn...
LLM Function Calling - AI Tools Deep Dive
Переглядів 8 тис.3 місяці тому
Tool and Function calling with LLMs is becoming one of the most crucial to understand capabilities, and can elevate the way you interact with and build AI applications to the next level. I’ve put together this video to give a comprehensive overview of what tool calling is, how it works, and how you can make your own. Cheers! I put these videos together on my own time with my own funding, if you...
LangGraph Cloud: Build & Deploy an AI Agent
Переглядів 1,4 тис.3 місяці тому
Recently announced LangGraph Cloud offers the ability to package up LangGraph based agents into APIs, allowing easy and rapid sharing and deployment of advanced LLM based applications! Resources: Code: github.com/ALucek/byo-chatgpt/tree/main LangGraph Cloud Docs: langchain-ai.github.io/langgraph/cloud/ LangGraph: langchain-ai.github.io/langgraph/ Other videos I have on Graph Agents: ua-cam.com/...
True Multimodal RAG - Audio/Image/Video/Text
Переглядів 3,2 тис.3 місяці тому
Everyone knows general text based vector databases, and text based RAG for LLM applications, but as it turns out thats just the beginning! Taking advantage of CLIP & CLAP models along with some fancy tricks, we embed 25,000 text entries, 1999 pictures, 2000 audio files, and 99 videos into a single vector database, allowing us to run direct text to text/audio/image/video retrieval! Resources: Mu...
Building a Thinking Machine: My Gemma 2 9B Reflection Agent
Переглядів 1,7 тис.3 місяці тому
Is Gemma 2 - 9B the new best local model? Let’s put that to the test! In this video we create a reflection agent that runs Gemma 2 through a series of complex tasks, analyzing UA-cam data and synthesizing findings into a report, applying what’s been analyzed to new title generations, and self reflecting at each step to ensure quality responses and additional instructions if necessary! This give...
Make YOUR OWN Images With Stable Diffusion - Finetuning Walkthrough
Переглядів 6 тис.4 місяці тому
Make YOUR OWN Images With Stable Diffusion - Finetuning Walkthrough
How AI Creates Images/Videos/Audio - Diffusion Models Explained
Переглядів 9484 місяці тому
How AI Creates Images/Videos/Audio - Diffusion Models Explained
Multimodal RAG!? - Pushing the Boundaries of AI
Переглядів 11 тис.4 місяці тому
Multimodal RAG!? - Pushing the Boundaries of AI
Build Your Own Finance LLM for FREE with SEC Data
Переглядів 5 тис.5 місяців тому
Build Your Own Finance LLM for FREE with SEC Data
LLM Agents Team Up: The Future of Automated Collaboration
Переглядів 2,1 тис.5 місяців тому
LLM Agents Team Up: The Future of Automated Collaboration
LLMs & AI Benchmarks! - GenAI Eval Deep Dive
Переглядів 1,5 тис.5 місяців тому
LLMs & AI Benchmarks! - GenAI Eval Deep Dive
Function Calling Local LLMs!? LLaMa 3 Web Search Agent Breakdown (With Code!)
Переглядів 10 тис.5 місяців тому
Function Calling Local LLMs!? LLaMa 3 Web Search Agent Breakdown (With Code!)
Breaking Down & Testing FIVE LLM Agent Architectures - (Reflexion, LATs, P&E, ReWOO, LLMCompiler)
Переглядів 17 тис.6 місяців тому
Breaking Down & Testing FIVE LLM Agent Architectures - (Reflexion, LATs, P&E, ReWOO, LLMCompiler)
Speak Any Language With AI - Realtime Speech-to-Speech Translation & Voice Synthesis (w/Code)
Переглядів 7 тис.6 місяців тому
Speak Any Language With AI - Realtime Speech-to-Speech Translation & Voice Synthesis (w/Code)
AI Agents! Giving Reasoning and Tools to LLMs - Context & Code Examples
Переглядів 7 тис.6 місяців тому
AI Agents! Giving Reasoning and Tools to LLMs - Context & Code Examples
Fine Tuning OpenAI Models Walkthrough - How and Why
Переглядів 9677 місяців тому
Fine Tuning OpenAI Models Walkthrough - How and Why
Turn Videos Into Blog Posts With AI! - GPT-4, Whisper-1, and Embedding Model Approach
Переглядів 12 тис.7 місяців тому
Turn Videos Into Blog Posts With AI! - GPT-4, Whisper-1, and Embedding Model Approach
LLM Model Output Comparison - Mistral Large/Gemma 7b/Llama 70b/Claude 3 Opus & More Pt.2
Переглядів 1,3 тис.7 місяців тому
LLM Model Output Comparison - Mistral Large/Gemma 7b/Llama 70b/Claude 3 Opus & More Pt.2
How To Easily Run & Use LLMs Locally - Ollama & LangChain Integration
Переглядів 6748 місяців тому
How To Easily Run & Use LLMs Locally - Ollama & LangChain Integration
How To Summarize & Chunk Technical Documents with AI Language Models
Переглядів 4668 місяців тому
How To Summarize & Chunk Technical Documents with AI Language Models
The content you provide is truly very good. I wonder if it is possible to make a swarm of swarms. For example, there would be a main swarm (we may call it a router or triage agent, whatever) that is responsible for choosing the suitable swarm (not an agent but a swarm of agents that are responsible for specific domain actions). For example: web research swarm python coding swarm math and calculations swarm accounting swarm translations swarm ... etc. And when the domain includes complex sub-domains, we might make the domain swarm connect to sub-swarms (a sub-swarm for each sub-domain) I am not sure, but it seems that the agents that belong to a single swarm share the history, so if we are using a weak LLM, it might not be able to consume all of that history. If that is true, then I think the above suggestion would solve this problem while keeping the power of the SWARM idea. I am a beginner in Python, so if the above suggestion can be done, I hope to see you doing it. Thanks for the amazing useful content. 🌹🌹🌹
But can’t Langgraph already do this?
Do you know how langraph works? This doesnt involve having to defining any edges or a pre-designed graphs. LLM is the router of agents.
isn't the new lerobot using sts3215 servo motor?
Can't we get the better interface and deploy on render or replit with credentials and custom domain to access anywhere?
Can this be done on CPU? or non NVIDIA GPU?
Thanks for creating this video. It’s another example and use case (and in video format!) of what the hugging face blog posts explained, but using different models and tools.
😍
Thank you so much for your channel! I am learning so much! I have subscribed. I have a new Windows 11 machine with the following specs: System Manufacturer ASUS System Type x64-based PC Processor AMD Ryzen 7 7800X3D 8-Core Processor, 4201 Mhz, 8 Core(s), 16 Logical Processor(s) Installed Physical Memory (RAM) 32.0 GB Total Physical Memory 31.2 GB Available Physical Memory 1.46 GB I do not have any graphics card installed, using onboard graphics. My plan was to add a graphics card, but have not done so yet. The file nvidia-smi.exe is not on my PC anywhere, looked where various online resources said it would be and performed a search. I am showing hidden files if that makes a difference. How do I proceed from here? Thank again!
very great video thanks for clearly explaning it
Of course!
Great video, thanks
I appreciate it!
nice! where did you learn all this?
📚To try everything Brilliant has to offer-free-for a full 30 days, visit brilliant.org/AdamLucek/ You’ll also get 20% off an annual premium subscription! 💡
You really presented your video in a wonderful and beautiful way. I love this stuff and I love you too. You are so creative.
Thanks for this great work! Very useful fora beginner. Could you please elaborate on integrating this with langsmith? I've created there an API Key and thought it would be all but it doesn't see my logs and docs are not very helpful, is there anything else in your project to be done for it work?
Great content! By the way, do you have any videos on deploying a LangGraph app on the Cloud (Google, AWS, Azure)?
Wait so does Langchain create the whole Json for you using the Tool decorator?
i tried to do this and getting a "OutputParserException: This output parser can only be used with a chat generation" can anyone help
Such a great video! Thank you for the docs as well.
thanks for sharing this mate.
I love this part of youtube!!! Time to learn more, thanks for posting this!
Of course! More to come :)
@AdamLucek So, by your opinion, do you recommend FLUX is better than SDXL or SD3 medium that its(Flux's) trained Lora model can generate more consistent portrait of myself? (with the same training images)
Flux for sure! But requires more compute
@@AdamLucek how if buy one more gpu card, my current one is 3090 ti
For your own machine, its usually recommended to have at least 24gb of VRAM for Flux training
thanks for the video .. i trry to undesrtand something the dataset is made from clean image and noisy image but the model itself its composed by what ??
Thank you, excellent tutorial. I have 2 questions 1. Where is the name that will be used in the prompt indicated? 2. It is possible the same metho with flux?
Hi! Sure- 1. For the style transfer part, that will not require any special keywords/prompting like the dreambooth method, and will affect every image output 2. Yes! Flux works better actually, check out my channel, I recently posted a similar video about how to do this with Flux
I am AI developer and music producer. I built a similar mobile app and web app to talk with my wife. I speak Czech and she speak Vietnamese and if we argue, I can say it is a mistake of "Google Translate", haha. She cannot speak Czech yet. We have video call, the audio is muted and processed using serverless container doing the AI. What's the funny thing is that I am going to train a model of my voice and then apply the timbre on generated speech, so my wife will actually hear my real voice speak vietnamese!!! kakakaka
why is there no ReAct mentioned?
Wonderfully done
how will the robot work when the envirenment is complex?
Hi, I need to translate audios (wav or mp3 files) from English to Brazilian Portuguese in the same format. That is, dubbing (speech-to-speech). Is it possible? Thanks.
Great video. Subscribed. Great topics
The Hugging Face video series is helpful for those interested in learning more.
Thanks a lot for this video.
AI neural networks are ideal for robotics because so many degrees of freedom are complex. You could model the same robot in 3D instead of building a real one too.
Thank you.
Very cool. Thanks!
You bet!
Cool nails
Love it
Simple and effective, great demonstration! Love the subscription, thank you so much! Just a side note, I got the same nails as you
Why is ChatGPT the first thing people picture after hearing the word "AI"?
Because it was the first major MLM dumbass
I think people picture ChatGPT as the first thing "AI" because it's the most human acting machine (aka a parrot). I heard from somewhere the phrase "If it looks like us, we'll think it is like us" and I think it would apply to ChatGPT. *if the question was rhetorical then I have no clue what to say.
@@cromputer_stanley Yes, I agree. ChatGPT is pretty good at imitating humans. Especially the new OpenAI o1 language model. But some people could surely argue that there are many types of AI's.
What a great video! Thanks for sharing the complete process, including the mistakes you made! First video I've seen and I'm already convinced to subscribe!
Glad you enjoyed!
underrated super underrated
How is that only 300 views? Wow
Good question! :)
Why do my images get horrendously stretched when using my newly trained Lora?
Best video today!
Glad you think so!
The googly eyes added +10 intelligence 🧠
👁️👁️
That is a very nice home project!
It definitely was!
dang thats cool
Can i use flux on my rtx3060 12gb vram?
Yes! The setup I have for generating images is on my RTX 3060 with 12gb vram, the training however requires atleast 24gb with this method
Is there a training you can do on mac? Or is it purely nvidia?