Want to learn how to build a Notion x Arxiv Agent in 1 minute 23 seconds? Watch this. In all honesty though, it's 11:29pm right now and I got the MCP server working ...so i couldn't be more stoked 😅 In this you'll learn how to: 1. Build an agent in Langflow 2. Connect it to a...
Want to learn how to build a Notion x Arxiv Agent in 1 minute 23 seconds?
Watch this.
In all honesty though, it's 11:29pm right now and I got the MCP server working
...so i couldn't be more stoked 😅
In this you'll learn how to:
1. Build an agent in Langflow
2. Connect it to a MCP server
3. Add tools, agents, llms (Claude 3.5 Haiku in this case)
Bonus: you'll also see my confusion trying to understand how quantum applies to option pricing!
This week I learned how hard it is coding in VS Code in 1080x1920 format. But...I also learned how to use Docling for getting your docs ready for LLMs. In this I show how to do it in a single command...RUNNING ON A MACBOOK YES! No hardcore GPUs and its insanely fast. I did a...
This week I learned how hard it is coding in VS Code in 1080x1920 format.
But...I also learned how to use Docling for getting your docs ready for LLMs.
In this I show how to do it in a single command...RUNNING ON A MACBOOK
YES! No hardcore GPUs and its insanely fast.
I did a bunch more tests including extracting Lilian Weng's Reward Hacking article and it was actually pretty good. Setup is literally
"""
# create a uv environment
uv init
# install deps
uv add docling mlx-vlm transformers
# convert away
uv run docling --to md --pipeline vlm --vlm-model granite_docling "yourfilehere.whatever"
"""
My default mode when I'm building stuff at the moment is how do I make this easy for anyone to replicate it. This is an example of that. How to quickly add some extra safety measures when building agents. We're finally getting to the point where people are asking how to...
My default mode when I'm building stuff at the moment is how do I make this easy for anyone to replicate it.
This is an example of that.
How to quickly add some extra safety measures when building agents.
We're finally getting to the point where people are asking how to harden agents and whilst there isn't a perfect answer for every solution. There's DEFINITELY some steps you can take to prevent catastrophic failures.
This is one of those, adding in a simple guardian layer at the start and end of your agent to block any harmful prompts an outputs.
Delivered my biggest AI keynote ever this week. Maybe not the biggest in terms of audience size but definitely in terms of effort. It’s been an intense two weeks getting it done. I was flying back and forth around the world while trying to prep. Coding in airports, in the...
Delivered my biggest AI keynote ever this week.
Maybe not the biggest in terms of audience size but definitely in terms of effort.
It’s been an intense two weeks getting it done. I was flying back and forth around the world while trying to prep. Coding in airports, in the back of ubers and hotel rooms.
Then coming back and getting it all to stick together. It was all worth it though.
Thank you to everyone who came and checked out a talk, said hi, asked questions, shared photos and gave out amazing energy. Trust me it all helped!! 😅
In this, you're going to learn (through some of my painful mistakes): 1. How to hook up a Stable Baselines 3 Agent to GoDotRL 2. Best practices when it comes to reward shaping for Reinforcement Learning 3. Building optimal gaming environments in GoDot for RL training (def...
In this, you're going to learn (through some of my painful mistakes):
1. How to hook up a Stable Baselines 3 Agent to GoDotRL
2. Best practices when it comes to reward shaping for Reinforcement Learning
3. Building optimal gaming environments in GoDot for RL training (def some gems in here)
All the code, logs and stuff is available via GitHub as always:
Code: https://github.com/nicknochnack/ReinforcementLearningBrackeysGame
Brackeys Original Tutorial: https://youtu.be/LOhfqjmasi0?si=3ib-ZIREVJq4P_GY
P.s. I'm still talking crap to Claude.
Although we're on much better terms now. Learning and recording this was an absolute blast, it's 8:29pm on a Sunday right now as I'm typing this but I had a ton of fun building this out. I feel like it's laid the groundwork to do a lot more with Reinforcement Learning and custom environments, especially those built with GoDot.
I was presenting at this think tank and saw a gun engineer demo using Reinforcement Learning for mining optimization. Got me thinking about how we could start applying this to simulated environments...what better way to start than by making a damn hard game 😅.
Hope you like it, I'm grateful for each and every one of you. I'm flying out to present in the states tomorrow so if you're around Florida, hit me up.
Code is in the GitHub repo there's like 14 final versions right now. 👀 I need to add the environments to Git LFS, if you want them let me know they're like 167MB for each of the dmg files but I can push them up if you want them.
Also a HUGEEE shout out to Ed Beeching the creator of GoDotRL and of course @Brackeys (you are hilarious and an absolute legend!)
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
Well, it's been a while since we revisited this hey? I was watching the comments on the other vids and it was eating me up. "NICK FIX THESE DAMN ERRORS". So...i took it one step further and just built my own Sign Language Detection model from scratch, literally the whole...
Well, it's been a while since we revisited this hey? I was watching the comments on the other vids and it was eating me up. "NICK FIX THESE DAMN ERRORS". So...i took it one step further and just built my own Sign Language Detection model from scratch, literally the whole library doesn't rely on any other deep learning libraries apart from PyTorch.
This was a labour of love but you guys gave me so much of your trust that I want to repay the favour. I wouldn't be here without you guys and I truly, deeply appreciate it. So....I give you NODDY v1.
It's a full end to end DETR object detection pipeline, in this video its fine tuned for detecting signs but it could definitely be repurposed for other object detection tasks. It covers:
⚡️ Detections using my pre-trained transformer
💾 Collecting custom data for you to train your own
🏋🏽♀️ Fine tuning, with or without pre-trained weights
📹 ANNNNNNND Real time detections...ay ay ayyyy
Okay I've been finishing the editing on this video for what feels like six hours now and I'm a little delirious so I'll leave it there. Enjoy guys, hope you like it, I'm going to McDonalds.
Love y'all. Happy coding!
Code: https://github.com/nicknochnack/SignDETR
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
Been doing a ton more with Langgraph lately…maybe I have seen the light. Figured I’d whip something up for my weekly video to share with y’all. This pretty much covers all the fundamentals you need to get up and running with langgraph. Importing a LLM, setting up the state,...
Been doing a ton more with Langgraph lately…maybe I have seen the light. Figured I’d whip something up for my weekly video to share with y’all.
This pretty much covers all the fundamentals you need to get up and running with langgraph. Importing a LLM, setting up the state, constructing the graph itself and then calling it. Also noticed a little bug with the yfinance documentation but I covered it in the video, just make sure you use size=5 instead of count=5 and you should be good to go.
Anyway, love ya all. Happy coding!
Code: https://github.com/nicknochnack/LanggraphCrashCourse
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
So this one’s been a while in the making. Honestly I’ve been pretty keen to truly get my head around vision transformers and figured this project could be the one. It took a month and a half to get it to this point and ton of banging my head against the keyboard to finally...
So this one’s been a while in the making. Honestly I’ve been pretty keen to truly get my head around vision transformers and figured this project could be the one. It took a month and a half to get it to this point and ton of banging my head against the keyboard to finally get it working (I’m looking at you ColorJitter 😒).
I want to keep this one going as a live project so i’ll be monitoring any issues you guys have a bit more closely. If you get stuck let me know and I’ll make as many updates as I can. Anyway, hope you enjoy it guys!
Code: https://github.com/nicknochnack/ViTCandlesticks
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
Stuff I found useful:
Albumentations: https://albumentations.ai/docs
CutMix/Mixup: https://docs.pytorch.org/vision/main/auto_examples/transforms/plot_cutmix_mixup.html
Lucid Rains ViT for PyTorch: https://github.com/lucidrains/vit-pytorch/tree/main/vit_pytorch
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
Bit of a dream come true with this one 🎉! We just released a couse on Deeplearning.AI 🤖 🦾 🚀 around all the amazing stuff you can do with ACP! In it you'll learn: 1. The differences between MCP vs ACP vs A2A 2. How to build ACP compliant agents 3. How to run pre-setup ACP...
Bit of a dream come true with this one 🎉!
We just released a couse on Deeplearning.AI 🤖 🦾 🚀 around all the amazing stuff you can do with ACP! In it you'll learn:
1. The differences between MCP vs ACP vs A2A
2. How to build ACP compliant agents
3. How to run pre-setup ACP agents like gpt-researcher, aider and more
4. Getting it all to work together as an agent system
Thanks to Hawraa, Andrew and the team for the backing on this was an amazing project to get together. And a huge shoutout to my amazing co-lecturer Sandi, you smashed it.
This was months of planning, coding, editing and recording to get this out but I'm super proud of the final product. Go check it out, link in the comments!
P.s. I still have the photo in my office 😉
That gameboy blender animation...took 6 hours to render 😅. Anyway, had a ton of fun coding this up and finally getting back to some proper ML. I've been thinking about just swapping out the pdf with some specific libraries to get an ultimate Blender bot or Manim bot (also...
That gameboy blender animation...took 6 hours to render 😅.
Anyway, had a ton of fun coding this up and finally getting back to some proper ML. I've been thinking about just swapping out the pdf with some specific libraries to get an ultimate Blender bot or Manim bot (also learned it for the video...we're boosting animation quality this year!!).
Would love to see what you end up fine tuning and deploying. If you need a hand just ping me I'm flying out tonight but will answer as much as i can. Love ya, thanks for watching!
Code: https://github.com/nicknochnack/EndToEndLoRA
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
All the code and stuff is available via GitHub as always: https://github.com/nicknochnack/ACPWalkthrough Documentation: https://agentcommunicationprotocol.dev/ GitHub Repo for the Library: https://github.com/i-am-bee/acp So I think getting LLMs to work together is going to be...
All the code and stuff is available via GitHub as always: https://github.com/nicknochnack/ACPWalkthrough
Documentation: https://agentcommunicationprotocol.dev/
GitHub Repo for the Library: https://github.com/i-am-bee/acp
So I think getting LLMs to work together is going to be critical to getting huge amounts of value from them. Right now most deployments that I've seen have been single use case, single agent type stuff.
But imagine...if we could get them to work together?
This is where Agent Communication Protocol comes in, it provides a vendor, framework and language agnostic way to deploy agent servers. The biggie is that this is going to make it easier for agents to talk to each other across the net and between organizations. I show a pretty simple example here but collaboration is the big play, I think there's interesting things to come.
Happy coding ✌️
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
Disclaimer: This has been developed for academic purposes. Nothing herein is financial advice, and NOT a recommendation to trade real money. Please use common sense and always first consult a professional before trading or investing.
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
Ooooft, this one has been in the pipe for a while now but finally got around to finishing the edit. Seriously interesting experiment, I'm wondering if I should do some livestreams and we can tweak and backtest params live to see what works? 🤷🏽♂️ Probably will get back to...
Ooooft, this one has been in the pipe for a while now but finally got around to finishing the edit. Seriously interesting experiment, I'm wondering if I should do some livestreams and we can tweak and backtest params live to see what works? 🤷🏽♂️
Probably will get back to reinforcement learning for ML and some deep visualisation stuff I've been wanting to get my hands on, alas I digress. In this, you're going to learn;
1. How to hook up a bot to Kraken
2. How to setup a LLM to help make trading decisions (see performance for yourself don't at me with "you should've used {insert fav llm here} llm" 😅
3. How to refactor all the code you just wrote and use finbert.
All the code, logs and stuff is available via GitHub as always: https://github.com/nicknochnack/LLMAgentCrypto
Happy coding ✌️
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
Disclaimer: This has been developed for academic purposes. Nothing herein is financial advice, and NOT a recommendation to trade real money. Please use common sense and always first consult a professional before trading or investing.
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick
Sup...good to be back. :) Wanted to give MCP a crack so here ya go. Bit of slightly different style of vid but I'm experimenting in the new YouTube studio. Hope y'all enjoy it, if you get stuck, drop me a comment and I'll be checking it out. I've included pretty detailed...
Sup...good to be back. :)
Wanted to give MCP a crack so here ya go. Bit of slightly different style of vid but I'm experimenting in the new YouTube studio. Hope y'all enjoy it, if you get stuck, drop me a comment and I'll be checking it out. I've included pretty detailed instructions in the GitHub repo as well and allll the code is there.
Happy coding ✌️
👨💻 Sign up for the Full Stack course and use YOUTUBE50 to get 50% off:
https://www.coursesfromnick.com/bundles/fullstackml
🤖 Get the Code: https://github.com/nicknochnack/MCPin10
Disclaimer: This has been developed for academic purposes. Nothing herein is financial advice, and NOT a recommendation to trade real money. Please use common sense and always first consult a professional before trading or investing.
Oh, and don't forget to connect with me!
LinkedIn: https://bit.ly/324Epgo
Facebook: https://bit.ly/3mB1sZD
GitHub: https://bit.ly/3mDJllD
Patreon: https://bit.ly/2OCn3UW
Join the Discussion on Discord: https://bit.ly/3dQiZsV
Happy coding!
Nick