Loading video...

Video Failed to Load

Go Home

Hack of the day: Llama on a microcontroller 🦙🔬 Details and source code at Thanks to Andrej Karpathy, whose llama2.c inspired this!

maxbraun's profile picture

Max Braun

6,992 subscribers

463,611 views • 2 years ago •via X (Twitter)

Science & TechnologyEducation
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Andrej Karpathy llama2.c running on galaxy watch 4
0:18

Andrej Karpathy llama2.c running on galaxy watch 4

Joey (e/λ)

224,298 views • 2 years ago

Here are ten 42M TinyStories models running simultaneously in the browser. Andrej Karpathy's Llama2.c code has been ported to Rust using the Candle framework and compiled to Wasm.
0:14

Here are ten 42M TinyStories models running simultaneously in the browser. Andrej Karpathy's Llama2.c code has been ported to Rust using the Candle framework and compiled to Wasm.

Radamés Ajna

45,424 views • 2 years ago

Inspired by Andrej Karpathy and the idea of using games to compare LLMs, I've built a version of the game Codenames where different models are paired in teams to play the game with each other. Fun to see o3-mini team with R1 against Grok and Gemini! Link and repo below.
0:15

Inspired by Andrej Karpathy and the idea of using games to compare LLMs, I've built a version of the game Codenames where different models are paired in teams to play the game with each other. Fun to see o3-mini team with R1 against Grok and Gemini! Link and repo below.

Ilya Abyzov

378,698 views • 1 year ago

Andrej Karpathy: Pick almost anything. 10k hours of focused practice and you're an expert. What's your 10k on?
0:16

Andrej Karpathy: Pick almost anything. 10k hours of focused practice and you're an expert. What's your 10k on?

tetsuo

41,416 views • 1 month ago

Meta just announced that Code Llama was now free for both research and commercial. This might the strongest competitor to ChatGPT: ▸ Can generate, explain, and debug your code ▸ Handles input 100,000 tokens ▸ Free for research + commercial use ▸ Outperforms most open models ▸ Comes in 7B, 13B, and 34B ▸ Supports Python, C++, Java, PHP, Typescript (Javascript), C#, and Bash Available in: ▸ Foundation base models (Code Llama) ▸ Python specializations (Code Llama - Python), ▸ Instruction-following models (Code Llama - Instruct)
0:11

Meta just announced that Code Llama was now free for both research and commercial. This might the strongest competitor to ChatGPT: ▸ Can generate, explain, and debug your code ▸ Handles input 100,000 tokens ▸ Free for research + commercial use ▸ Outperforms most open models ▸ Comes in 7B, 13B, and 34B ▸ Supports Python, C++, Java, PHP, Typescript (Javascript), C#, and Bash Available in: ▸ Foundation base models (Code Llama) ▸ Python specializations (Code Llama - Python), ▸ Instruction-following models (Code Llama - Instruct)

Lior Alexander

196,384 views • 2 years ago

Code LLaMA is now on Perplexity’s LLaMa Chat! Try asking it to write a function for you, or explain a code snippet: 🔗 This is the fastest way to try AI at Meta’s latest code-specialized LLM. With our model deployment expertise, we are able to provide you with this model less than 24 hours of it’s release. What’s next? We’ll integrate code LLaMA into Perplexity, all in service of providing you with the best answers to your most technical questions!
0:12

Code LLaMA is now on Perplexity’s LLaMa Chat! Try asking it to write a function for you, or explain a code snippet: 🔗 This is the fastest way to try AI at Meta’s latest code-specialized LLM. With our model deployment expertise, we are able to provide you with this model less than 24 hours of it’s release. What’s next? We’ll integrate code LLaMA into Perplexity, all in service of providing you with the best answers to your most technical questions!

Perplexity

274,523 views • 2 years ago

Andrej Karpathy In the next version of Cursor, stuffing context will get much easier :D
0:12

Andrej Karpathy In the next version of Cursor, stuffing context will get much easier :D

Michael Truell

170,318 views • 1 year ago

Day 6 of Stealing Designs and converting them into Code I have made the code for the Dashboard open source Check this out in replies
0:14

Day 6 of Stealing Designs and converting them into Code I have made the code for the Dashboard open source Check this out in replies

Karan Kendre

21,958 views • 6 months ago

day 23 of forking Paperclip today i added knowledge bases for projects and you company. based on Andrej Karpathy idea below. added a 'system chores' and 'agent chores' to manage memory. after each heartbeat run memories get added ('agent chores') and every 4 hours memeries get managed. This is in a db with an API for all agents to use and is seperated form the agents memeroes which they get at instantiation time.
0:17

day 23 of forking Paperclip today i added knowledge bases for projects and you company. based on Andrej Karpathy idea below. added a 'system chores' and 'agent chores' to manage memory. after each heartbeat run memories get added ('agent chores') and every 4 hours memeries get managed. This is in a db with an API for all agents to use and is seperated form the agents memeroes which they get at instantiation time.

Michael Houghton 🐺

18,188 views • 2 months ago

Tired of manual prompt tweaking? Watch the latest Llama tutorial on how to optimize your existing GPT or other LLM prompts for Llama with `llama-prompt-ops`, the open-source Python library! In this video, Partner Engineer Justin Lee demonstrates installation, project setup, migrating your first prompt, and analyzing performance gains. Watch now to discover: ✨ Why systematic prompt optimization is crucial for migrating from GPT to Llama. 💻 A live code walkthrough of `llama-prompt-ops` in action for a customer service classification task. Watch the full tutorial here:
0:10

Tired of manual prompt tweaking? Watch the latest Llama tutorial on how to optimize your existing GPT or other LLM prompts for Llama with `llama-prompt-ops`, the open-source Python library! In this video, Partner Engineer Justin Lee demonstrates installation, project setup, migrating your first prompt, and analyzing performance gains. Watch now to discover: ✨ Why systematic prompt optimization is crucial for migrating from GPT to Llama. 💻 A live code walkthrough of `llama-prompt-ops` in action for a customer service classification task. Watch the full tutorial here:

Meta for Developers

38,047 views • 11 months ago

Finally Neuraxon is out!!! Our new paper & code & demo by Jose Sánchez & David Vivancos - e/acc expanding on the Qubic #Aigarth #intelligenttissue by Come-from-Beyond the path towards #TrueAI "A New Neural Growth & Computation Blueprint" one step forward to #AGI & #ASI Read the preprint Paper: (All the details of our ongoing research) Explore the Code: (Please give us a⭐fork and build!) Play with the Interactive Demo: (Build your own Neuraxon NetWork) Andrej Karpathy do you want to give it a try? Beff (e/acc) can we try it on Extropic hw? And more to come stay tuned! #EAGI #Neuraxon #Trinarystates #Continuousprocessing #Synapticdynamics #Neuralplasticity #Spontaneousactivity #Temporalsynchronisation #Bioinspiredcomputation #Artificiology
0:19

Finally Neuraxon is out!!! Our new paper & code & demo by Jose Sánchez & David Vivancos - e/acc expanding on the Qubic #Aigarth #intelligenttissue by Come-from-Beyond the path towards #TrueAI "A New Neural Growth & Computation Blueprint" one step forward to #AGI & #ASI Read the preprint Paper: (All the details of our ongoing research) Explore the Code: (Please give us a⭐fork and build!) Play with the Interactive Demo: (Build your own Neuraxon NetWork) Andrej Karpathy do you want to give it a try? Beff (e/acc) can we try it on Extropic hw? And more to come stay tuned! #EAGI #Neuraxon #Trinarystates #Continuousprocessing #Synapticdynamics #Neuralplasticity #Spontaneousactivity #Temporalsynchronisation #Bioinspiredcomputation #Artificiology

David Vivancos - e/acc

110,097 views • 7 months ago

felt cute, did some Andrej Karpathy style cozy coding. now i can run GPT 2 124M in pure TypeScript at 7 tps. played with implementing the GEMV via C/WASM, but that only got me a 1.7x speed up.
0:17

felt cute, did some Andrej Karpathy style cozy coding. now i can run GPT 2 124M in pure TypeScript at 7 tps. played with implementing the GEMV via C/WASM, but that only got me a 1.7x speed up.

Mario Zechner

29,165 views • 1 month ago

☃️On the 19th day of #Kwikmas, Kwik Fit gifted me a an iPad thanks to our friends at Pirelli️ ☃️ Like this post and tag a friend to enter!!! T&C's
0:15

☃️On the 19th day of #Kwikmas, Kwik Fit gifted me a an iPad thanks to our friends at Pirelli️ ☃️ Like this post and tag a friend to enter!!! T&C's

Kwik-Fit

39,156 views • 2 years ago

Here’s the #Flutter implementation of this cool Airbnb interaction 🤩 The recipe is just the good old built-in #Hero widget with some 3D transform tricks 🪄! 🔗 Source code: Read on for the code details👇🏼 1/5
0:12

Here’s the #Flutter implementation of this cool Airbnb interaction 🤩 The recipe is just the good old built-in #Hero widget with some 3D transform tricks 🪄! 🔗 Source code: Read on for the code details👇🏼 1/5

Roaa 🦄💙

429,035 views • 3 years ago

I just created my own LaTeX-OCR app using Llama 3.2 Vision! Upload the LaTeX code as an image, and it gives you the corresponding LaTeX code using Llama 3.2 multimodal! Here's what I used: - Ollama for serving Llama 3.2 vision locally - Streamlit for the UI Everything is just 50 lines of code! Find the code in the next tweet. -- Every day, I share tutorials and insights on DS, ML, LLMs, and RAGs. Find me → Avi Chawla
0:15

I just created my own LaTeX-OCR app using Llama 3.2 Vision! Upload the LaTeX code as an image, and it gives you the corresponding LaTeX code using Llama 3.2 multimodal! Here's what I used: - Ollama for serving Llama 3.2 vision locally - Streamlit for the UI Everything is just 50 lines of code! Find the code in the next tweet. -- Every day, I share tutorials and insights on DS, ML, LLMs, and RAGs. Find me → Avi Chawla

Avi Chawla

15,608 views • 1 year ago

This is WILD. Andrej Karpathy ranked every job in america by AI exposure > scraped 342 occupations from the bureau of labor statistics > used an LLM to score each job > built a treemap showing job size and AI risk main signal: fully digital work from home jobs are most exposed average score: 5.3/10 The whole thing is open source. Yeah, we’re cooked.
0:13

This is WILD. Andrej Karpathy ranked every job in america by AI exposure > scraped 342 occupations from the bureau of labor statistics > used an LLM to score each job > built a treemap showing job size and AI risk main signal: fully digital work from home jobs are most exposed average score: 5.3/10 The whole thing is open source. Yeah, we’re cooked.

CG

21,403 views • 3 months ago

While at #OpenSourceAIWeek, Andrej Karpathy received one of our first NVIDIA DGX Sparks, personally delivered by Nader Khalil🍊 and @Baxate_carter. 🙌 Excellent work by Andrej on introducing Nanochat -- an #opensource project that makes it easy for developers to explore and train their own LLMs from scratch. We can’t wait to see what you’ll do with your Spark! #SparkSomethingBig ✨
0:14

While at #OpenSourceAIWeek, Andrej Karpathy received one of our first NVIDIA DGX Sparks, personally delivered by Nader Khalil🍊 and @Baxate_carter. 🙌 Excellent work by Andrej on introducing Nanochat -- an #opensource project that makes it easy for developers to explore and train their own LLMs from scratch. We can’t wait to see what you’ll do with your Spark! #SparkSomethingBig ✨

NVIDIA AI Developer

48,804 views • 7 months ago

late to the party but built a UI layer on top of the wiki my OpenClaw🦞 agent writes and maintains, based on exactly what Andrej Karpathy describes here. obsidian vault underneath, react frontend on top: search, an interactive knowledge graph, topic browsing, article pages with tables of contents + neighborhood mini-graph and many more.
0:19

late to the party but built a UI layer on top of the wiki my OpenClaw🦞 agent writes and maintains, based on exactly what Andrej Karpathy describes here. obsidian vault underneath, react frontend on top: search, an interactive knowledge graph, topic browsing, article pages with tables of contents + neighborhood mini-graph and many more.

Ansub

110,250 views • 2 months ago

They’re letting anyone in these days 🙂‍↔️ Huge congratulations to Roger Federer on his Hall of Fame induction. A true legend of the sport whose sportsmanship, brilliance, and humility have inspired millions on and off the court.
0:15

They’re letting anyone in these days 🙂‍↔️ Huge congratulations to Roger Federer on his Hall of Fame induction. A true legend of the sport whose sportsmanship, brilliance, and humility have inspired millions on and off the court.

Served with Andy Roddick

15,063 views • 7 months ago

🔍 Code Decompiler arrives on #SuiVision! Our commitment to transparency for developers and users just got stronger. It brings Sui Code transparent by converting bytecode into readable source code. Special thanks to Sam Blackshear for the insights and step-by-step guides. 🌟 Watch the video for the full scoop!
0:18

🔍 Code Decompiler arrives on #SuiVision! Our commitment to transparency for developers and users just got stronger. It brings Sui Code transparent by converting bytecode into readable source code. Special thanks to Sam Blackshear for the insights and step-by-step guides. 🌟 Watch the video for the full scoop!

BlockVision

44,726 views • 2 years ago