![Unclecode](/img/default-banner.jpg)
- 74
- 21 577
Unclecode
Malaysia
Registrace 7. 06. 2021
Unclecode is a passionate computer scientist, musician, lecturer, techpreneur, and astronomy fan. He composed various music and also became a CEO of an EdTech company called Kidocode.
He created KidoCode, which is a platform that challenges the conventional education system by providing personalized and adaptive learning for kids & teens. It brings a change in the future of children's education all around the world through online learning using technology like AI, machine learning, and computational thinking.
During his break from making these big steps in education, Unclecode either plays with his cats or figures out the mathematical formula and scientific terms in Big Bang Theory... by heart.
Unclecode has settled in Malaysia (although there's a plan of expanding Kidocode to other SEA countries soon). Oh, and not to forget, Unclecode truly believes that he came all the way from Andromeda.
He created KidoCode, which is a platform that challenges the conventional education system by providing personalized and adaptive learning for kids & teens. It brings a change in the future of children's education all around the world through online learning using technology like AI, machine learning, and computational thinking.
During his break from making these big steps in education, Unclecode either plays with his cats or figures out the mathematical formula and scientific terms in Big Bang Theory... by heart.
Unclecode has settled in Malaysia (although there's a plan of expanding Kidocode to other SEA countries soon). Oh, and not to forget, Unclecode truly believes that he came all the way from Andromeda.
Set Up Your Own Super-Fast Whisper Server on Colab! 🚀
Set Up Your Own Super-Fast Whisper Server on Colab! 🚀
Working on function calling projects, I needed a speech-to-text solution without third-party APIs. Found an open-source model that's super fast-converted 15 mins of audio to text in just 20 seconds!
In this video, I show how to:
1️⃣ Launch a Colab server
2️⃣ Use a script to record & transcribe instantly
3️⃣ Extract audio from CZcams videos
Thought it might be useful for everyone. Check it out!
🔗 Repo: github.com/unclecode/fast-whisper
Please subscribe to my channel and Follow me on X x.com/unclecode
Working on function calling projects, I needed a speech-to-text solution without third-party APIs. Found an open-source model that's super fast-converted 15 mins of audio to text in just 20 seconds!
In this video, I show how to:
1️⃣ Launch a Colab server
2️⃣ Use a script to record & transcribe instantly
3️⃣ Extract audio from CZcams videos
Thought it might be useful for everyone. Check it out!
🔗 Repo: github.com/unclecode/fast-whisper
Please subscribe to my channel and Follow me on X x.com/unclecode
zhlédnutí: 487
Video
Garvis: Almost Real-time Voice-Assistant Using Groq 🚀
zhlédnutí 2,7KPřed 3 měsíci
Welcome to Garvis! In this video, I showcase Garvis, a project that leverages the speech-to-text endpoint from Groq, utilizing their cutting-edge LPU (Language Processing Unit) chips. These chips offer unprecedented speed, faster than traditional GPUs, allowing for near real-time user experiences. Garvis chunks audio based on silence and translates instantly, thanks to the high-speed capabiliti...
Mistral v0.3: Multi-Function Calling & Dependencies! 🚀
zhlédnutí 303Před 3 měsíci
Mistral just dropped version 0.3, and it natively supports function calling! 🛠️ I tested multi-functions & dependencies. It didn't work initially, but I found a solution that works like a charm. Check out my video and try it on Colab. Enjoy!
Crawl4AI - Crawl the web in an LLM-friendly Style
zhlédnutí 7KPřed 3 měsíci
Welcome to the detailed walkthrough of Crawl4AI v0.2.0! 🚀 In this video, I'll dive deep into the code base of Crawl4AI, our powerful web crawling tool designed for AI enthusiasts and developers. We'll explore all the new and exciting features that make this release a game-changer: - 🕷️ Efficient web crawling to extract valuable data from websites - LLM-friendly output formats (JSON, cleaned HT...
LLaMA3 & TinyLlama Fine-Tuned for Function Calls
zhlédnutí 2,4KPřed 3 měsíci
In this video, I discuss my latest project on fine-tuning LLaMA3 and TinyLlama to natively support function calls, which is crucial for the development of AI agents in the open-source community. I cover the following topics: - Colab notebooks demonstrating how to run the models using helper - classes and GGFU versions - Examples of using the models locally and with the Ollama server - Prompt te...
Ep.1 Upskilling Employees with AI | Fireside Chats with Unclecode | Dr. Ram Raj Gopal
zhlédnutí 29Před 4 měsíci
Welcome to the first episode of Fireside Chats which I'm joined by my first guest and a dear friend, Dr. Ram Gopal Raj, the Director of Strategy and Innovation at Air Asia Academy. Together, we delve into the transformative role of AI in the workplace and its crucial part in upskilling employees. It's a journey into understanding how AI isn't just a tool for enhancing productivity but a powerfu...
Ai Assistant using Groq: Mix of GroqCall (Groq-Function-Call proxy) with PhiData
zhlédnutí 995Před 5 měsíci
Hello beautiful people. This is Unclecode and in this video, I demonstrate how to use my Groq-function-call-proxy to build an AI assistant, using a cinema assistant as an example. - The cinema assistant can suggest movies based on user preferences, find showtimes, and book tickets. - I integrate GroqCall.ai (previously FunckyCall.ai) with the PhiData library to enhance existing AI/ML workflows....
Technology Degree | Mastering Game Development
zhlédnutí 25Před 5 měsíci
Welcome to the thrilling world of Game Development! Dive into the Phaser framework, the premier HTML5 game development platform used across browsers, mobile, and desktops. In this journey, we'll explore 2D and 3D graphics, animation, sound, and collision detection using powerful frameworks like Unity and LÖVE 2D, and languages like Lua and C#. Discover the art of pixel creation through various ...
Technology Degree | Building and Publishing Mobile Apps
zhlédnutí 15Před 5 měsíci
Technology Degree | Building and Publishing Mobile Apps
Behind Podcast | AI4ALL | Upskilling Towards Artificial Intelligence (AI)
zhlédnutí 13Před 5 měsíci
Starting this podcast journey feels surreal! The first episode will be out very soon and it means a lot to me because it's all about democratizing Al and upskilling employees - something l'm deeply passionate about. Want to take a guess who l'm bringing on as my very first guest? #podcast #democratizingai #ai #edutech #upskilling #podcasting #aipodcast
GroqCall: Function call for Groq & beyond
zhlédnutí 3,3KPřed 5 měsíci
In this video, I dive into the technical details of GroqCall.ai (previously called FunckyCall.ai), a proxy server that enables function calls for Groq's high-performance Language Processing Unit (LPU) and other AI providers. I discuss my motivation for creating this project, stemming from the desire to leverage Groq's impressive speed and the need for accessible function call capabilities to bu...
Debunking Nvidia CEO, Jensen Huang’s Stance: The Crucial Role Of Programming in Children’s Education
zhlédnutí 178Před 5 měsíci
Nvidia CEO, Jensen Huang's views on children not learning computer science. Here, I contrast it with Code.org's mission, backed by tech leaders and celebrities, emphasizing early computational thinking. Watch the full video on why I believe Huang's take is misleading and why programming is essential for our kids' future.
From Seed to Series A: Navigating the Startup Funding Journey
zhlédnutí 32Před 5 měsíci
In this video, you'll gain invaluable insights into the intricacies of startup fundraising, exploring the distinct stages of seed, pre-seed, and Series A financing. By understanding the unique characteristics and requirements of each stage, you'll learn how to identify where your company stands in the fundraising journey and tailor your approach accordingly. From seed rounds that focus on valid...
To Fund or Not To Fund: Unveiling the Real Necessity for Business Investment
zhlédnutí 28Před 6 měsíci
To Fund or Not To Fund: Unveiling the Real Necessity for Business Investment
Technology Degree | Discover the World of Web Development
zhlédnutí 20Před 6 měsíci
Technology Degree | Discover the World of Web Development
Technology Degree | Navigating the Creative Process with Python Coding
zhlédnutí 58Před 6 měsíci
Technology Degree | Navigating the Creative Process with Python Coding
Ep. 2 | Demystifying Technology | Sleeper Agent in Large Language Models (LLMs)
zhlédnutí 28Před 6 měsíci
Ep. 2 | Demystifying Technology | Sleeper Agent in Large Language Models (LLMs)
AccessibleAl | Al for each and every student
zhlédnutí 52Před 7 měsíci
AccessibleAl | Al for each and every student
Ep. 2 | Introduction to AI Course : C.H.A.R.G.E Method : Humanity in AI Course
zhlédnutí 59Před 7 měsíci
Ep. 2 | Introduction to AI Course : C.H.A.R.G.E Method : Humanity in AI Course
Prompt Engineering' | Mastering The Dialogue with AI | AI Course
zhlédnutí 164Před 7 měsíci
Prompt Engineering' | Mastering The Dialogue with AI | AI Course
Ep. 1 | Introduction to AI Course : C.H.A.R.G.E Method : Communication in AI Course
zhlédnutí 38Před 7 měsíci
Ep. 1 | Introduction to AI Course : C.H.A.R.G.E Method : Communication in AI Course
Ep.1 | Demystifying Technology | Attention in Large Language Models (LLMs)
zhlédnutí 49Před 7 měsíci
Ep.1 | Demystifying Technology | Attention in Large Language Models (LLMs)
This is really cool dude!
Thanks for the cool video! Were you able to actually use this model and call functions? especially interested in calling the chain of functions that were in your video
Very useful Project, I must admit! Is it a recursive crawler, when I say recursive, I mean it, (not restricted to depth threshold). Also How differet is this from FireCrawl, in terms of functionality and other stuffs. I can't wait to get started on using this project, and give it a shot! Thanks!
Hey man. I'm going to be honest but i'm new to data scraping and wanted to ask if crawl4ai can be used to scrape data from tiktok. They have implemented some harsh measures with request rate limits and login requirements. From what i saw crawl4ai has some login feature but just wanted to ask you if i'm going in the right direction. Otherwise looks great
This video looks super interesting! I'm really impressed with your work. That's a huge step for open-source AI agent development. I'm curious about how large language models decide which API to call. It seems like it's all happening internally, like a "black box," and developers aren't using tools to analyze the output in real-time or directly control the process with functions. Is my understanding correct?
How is it used for local computing?
cool
Looks exciting. Have you considered a nix script?
awesome transition at 2:30 🎶😄
So we can use Faster Whisper, Groq, and Crewai using Agents and build a bot/personal Assitant kind
Absolutely! That's one of the immediate applications of such a stack. Eventually, you can achieve this with a true multimodal LLM that understands audio, image, and text natively. Right now, it makes sense to use Whisper to transcribe audio to text and then use an LLM to generate a response from the transcribed text. But imagine a multimodal model where you pass audio and text together and directly get text in return. Currently, models like GPT-4 or Gemini Pro support image and text, but soon they'll release models that support audio as well.
possible to put up a prebuilt docker image, including the 'models'? I had problem downloading the models during build docker. Thanks!
I will work on that. Trying to have a version without model dependency as well
Really cool man! Can I crawl all accessible subpages from a main page? So I crawl 2 levels in total?
You can send multiple links, so first crawl the main page, then get links and send them again. However soon I will release the ability to se the depth and get a cool result for that
you can use whisper small for fast spech recognicion when it doesnt use it and the phi-3 models
Agree! I actually made another version using an open-source model for Fast Whisper, running a server on a Colab L4, and it is surprisingly fast. I will make a video about it soon.
thanks
I have not yet an idea as if this video is valuable but I saw the flute intro and directly subscribed
😄😄 Perhaps, I should make some content for mix of flute and AI! You gave me an idea, I try to come up with something
Amazing, but is not available in LM Studio 😥😥 anyway, liked and subscribed. I should learn to use ollama.
You can load the GGFU version over there. Anyway, I'm working on a research project, and we're close to releasing what might be the smallest ever language model solely for function calls. Stay tuned!
YESSSSS 🔥🔥🔥
This fix definitely going to help me.
Got a new subscriber! Thanks.
WHAT HAPPENED TO THE FLUTE UNCLE CODE
Hahahaha!! Ok, ok, message received
Love how you so excited of your project! Keep it up man! Great project
Thanks! Will do!
i got a result object. how to parse it
Result is an object like this: class CrawlResult(BaseModel): url: str html: str success: bool cleaned_html: str = None markdown: str = None extracted_content: str = None metadata: dict = None error_message: str = None So you can access using this property (cleaned_html, markdown, extracted_content), or dump the model into a python dictionary using "result.model_dump()`
Excelente Trabajo..!!
can it run on free google colab ram?
Yes, L4 would be good enough
Muito legal, ficou simples de compreender
de nada.
Very nice work but please make your code fullscreen, the video like this with dark screen and darkened symbosl is super hard to follow.
Haha you right, I will consider that for the next video, stay tuned
Hi, The GGUF colab notebook cannot be accesed
It should be by now. Please check again
Can you use tool arguments? [if tool_calls, for tool_call in tool_calls] (for functions with params that need to be parsed from the chat or parse info that is stored in a database?)
Yes, it should support that. I'm compiling everything into a library to provide a seamless function call experience across any LLM. I'll be releasing it soon.
@@unclecode788 what’s your discord? Working on a few major project that would be perfect for this
@@unclecode788 I am working on some large projects that will be perfect for this. Please let me know when you release this.
Hi UncleCode I would love to see this series but can’t find the other videos. Please can you advise where I can find them?
Thanks for your interest, working on it, meanwhile you can follow me on X where I shared more of these contents @unclecode
Promo-SM
Awesome
Very Cool project❤, i wish it was written in javascript or typescript. I am not a python guy.😢
You are very right, I have to do that. Btw I'm developing a new library focused solely on tool (function) calls. I've expanded it to include any LLM, like Ollama, and it's working like a charm. I'm also incorporating "tools binding." I'll be releasing it soon and will record a video to showcase it. I definitely make it available in JS as well.
@@unclecode788 I have one question, Is openAI, groq and other providers using the same approach for function calling by using phidata/webLLM or they are doing something different like their own in-house library under the hood.
@@unclecode788 Curiously waiting for that? Do OpenAI and others use WebLLM and PhiData for function calling, or are they using their own secret library behind the wall?
Grok-1, Claude Opus, ChatGPT 4, Gemini Ultra, and Poe: Which one is the best compared to Copilot?
There are two ways to use Copilot: the code completion feature, which I love, and Copilot Chat, which I rarely use because ChatGPT works better for me, though Copilot Chat does have extra tools for interacting with your code. Honestly, for the last 40 days, I've switched to Claude Opus, and it's been the best experience so far. Regarding Gemini model series, I'm not sure why, but I couldn't get the feelings I am looking for. And In real work, I use many models at the same time, including Mistral chat and a few open sources with my Ollama client. Sometimes, one answers a specific question much better, and I've gotten used to their personalities and know which model to use for what task.
Great work
Awesome work solving a real issue (gap) in the community
The JSON advice was golden 🤌
25:30 I emphatically agree. I believe this is also related to the issue of how adding to many rules in system prompt can confuse the model etc. I find better results just letting the model do it's thing rather than writing highly detailed prompts. I think that's why chains are also really good.
Exactly! Expecting a model trained to generate tokens to deliver the perfect answer on the first try is a misunderstanding of how transformers work. Allowing the model to "think" by generating more content can refine the attention matrix, preparing it to statistically converge on the desired conclusion. This approach prevents forcing an output prematurely when the model isn't ready, leading to more accurate and reliable responses. Generating tokens is essentially a sampling problem from the word distribution. If you need to adjust the distribution, you have two main options: fine-tuning the model or letting the model navigate towards your desired output while maintaining the integrity of the original sampling distribution. This approach ensures that modifications are more natural and aligned with the model's learned patterns.
Fnx calling with groq speed is kinda wild! Mixtral good enough model to handle fxn calling reliably?
Absolutely, while speed doesn't inherently affect the quality of Mixtral, it does open up numerous possibilities with prompt engineering. For instance, self-ranking involves generating multiple responses to a single query and then using an LLM to select the top K answers to combine. Here, speed allows you to leverage a model like Mixtral-or any model-to its fullest potential. You can experiment with techniques like the tree of thought among others. In my new library, I'm compiling these "techniques of thought" to maximize the advantages offered by Groq's speed. This is where speed truly makes a difference.
Bro you really did it with this groq function calling! Just curious can we extend it with other provider like together and other local beside ollama like using lm studio?
Yes, I've primarily used Groq, the Mistral model, but I'm developing a new library focused solely on tool (function) calls. I've expanded it to include any LLM, like Ollama, and it's working like a charm. I'm also incorporating "tools binding." I'll be releasing it soon and will record a video to showcase it.
How the hell does this have 665 views??
Great work! Will def use your repo. The combo Groq, Mixtral and GroqCall is so cheated 😁
I'm developing a new library tailored specifically for tool (function) calls. This focus stems from my belief that function calls are a crucial component of AI as agents, providing a dynamic framework where LLMs can not only execute but also generate their own functions dynamically. This capability is essential for adaptive, intelligent systems. I've expanded this functionality to work seamlessly with any LLM, including testing with Ollama, and the results are promising. I'm introducing "tools binding" to enhance this integration. I'll be releasing this library soon and will create a detailed video to explain its features and the concept behind focusing on tool calls.
Hey Uncleode , really nice video ! I was wondering if I could help you with Highly Engaging Thumbnail + Video Editing which will surely increase your CTR & AVD which will help your videos to reach to a wider audience ! Please Lmk what do you think ?
He is referring to the fact that your children should not exclusively major in computer science in university or prepare themselves with only that path in mind. AI is reducing the resource burden to generate applications thus reducing the demand for people with only a computer science background over time. Employers will be seeking people with other specializations like mathematics/physics/chemistry who also have some fundamentals of computer science in their background to complement their skillset. The job market will change. It is not to say programming is not useful for children's development but to warn people that the world in the future will be a different ball game for these kids.
Hey Uncleode , really nice video ! I was wondering if I could help you with Quality Editing in your videos and also make a highly engaging Thumbnail which will help your videos to reach to a wider audience ? Please Lmk what do you think ?
That was insightful, thank you for providing these valuable information
Glad you enjoyed it!
Really informative video! You are absolutely right, the testing stage is so important. It is important to let that ego go away so we can effectively solve the problem.
this is a great show! the kids are brilliant! 😍