jeffheaton

YouTuber (90K+ subs), phd computer science, data scientist, and adj faculty at @WUSTL. Any opinions expressed are my own. #InsurTech #FinTech

[jeffheaton @jeffheaton]

Run 70B-parameter local LLMs at home with an NVIDIA DGX Spark (1 PFLOP, 128 GB Unified Memory). Access from any laptop via WebUI or Python API for full control. #AI #LLM #NVIDIA #DGXSpark #GPU @NVIDIAAIDev
https://t.co/EaJzmd0Zgt

[jeffheaton @jeffheaton]

NVIDIA DGX Spark, 1 PFLOP super computer that fits in your hands. It has 128 GB of unified RAM, this has allowed me to generate 1024 square stable diffusion images and 70b param LLM's. This is an amazing machine, everything worked great out of the box. More videos to come. #gpu #llm #GenAI @NVIDIAAIDev
https://t.co/y0DpwBvl1t

[jeffheaton @jeffheaton]

Similarities and differences between human brains and GPUs. Both are about the same size/weight. So the internal processing is considerably different. GPU's draw much more power under load, brains no so much. Brains last much longer and require vastly much less infrastructure to produce. #genai #gpu

[jeffheaton @jeffheaton]

Quick review of the upcoming NVIDIA DGX Spark, a small desktop supercomputer with 128GB unified GPU RAM on Grace-Blackwell @NVIDIAAIDev #gpu https://t.co/Jea157TGcN

[jeffheaton @jeffheaton]

What is "Unified Memory"? Basically the CPU and GPU can share memory. We seem to be seeing this more and more, from Macs to game consoles to high end server computers. The NVIDIA DGX Spark will soon add this in a "Mac Mini" sized 128GB desktop super computer. @NVIDIAAIDev #gpu https://t.co/EKEgFz0eYB

[jeffheaton @jeffheaton]

Want a 50X GPU speed up from sklearn with no code changes? Have a look at the latest version of CUML from @NVIDIAAIDev. #gpu #sklearn
https://t.co/Lpe8ofHXoM

[jeffheaton @jeffheaton]

Don’t ask me how long I’ve been working on this! #2048

[jeffheaton @jeffheaton]

Polars was already amazingly fast, but now Polars GPU Engine Powered by RAPIDS cuDF is Available in Open Beta! @RAPIDSai @NVIDIAAIDev https://t.co/DnpdrNkJQ0

[jeffheaton @jeffheaton]

@LastPassHelp Please fix this soon, it keeps popping in front of what I am working on. Do you have an ETA on the fix, I might need to migrate password vaults if this persists too long.

[jeffheaton @jeffheaton]

@LastPass Is there a way that I can keep LastPass from popping up this tab all the time? Sometimes it even pops up while I am using a website, interrupting. As you can see here, it just keeps opening a new one every few minutes.

[jeffheaton @jeffheaton]

@LastPassHelp Mac, Chrome Version 129.0.6668.103 (Official Build) (arm64)

[jeffheaton @jeffheaton]

@Sentdex The better question…. What is AGI’s next moves?

[jeffheaton @jeffheaton]

#GTC24 is in full swing and Jensen Huang's keynote introduced the Blackwell platform, which includes the newest generation of NVIDIA GPUs. In this video I give some of my main reactions. @NVIDIAAI https://t.co/VeGCAkvdz4

[jeffheaton @jeffheaton]

I am giving a way an #NVIDIA RTX 4080 SUPER, which was kindly provided by NVIDIA AI, GTC 2024 starts next week on March 18th! See this video for my top picks and details on this giveaway! #GTC24
https://t.co/i9VvG1iLlu

[jeffheaton @jeffheaton]

NVIDIA #gtc24 is just around the corner, here are my picks of the sessions I will attend. This is the event for AI! You can register here, https://t.co/o0XbIqyNYQ Also, thank to NVIDIAI will have a Geforce GTX 4080 SUPER to give away, details soon! https://t.co/7Xb2z6uwHp

[jeffheaton @jeffheaton]

#GTC24 is just around the corner. Thank you to @NVIDIAAI for providing this 4080 Super for a giveaway. More to come! Register here. https://t.co/o0XbIqyNYQ My video intro! https://t.co/0sfu5XR8Sn

[jeffheaton @jeffheaton]

@EwinRacing is running a nice holiday sale! This is the chair you see behind me on my newer videos, I really like it! https://t.co/6Ibce8aeZl

[jeffheaton @jeffheaton]

68 YouTube videos, 65 Jupyter notebooks, and 10 programming assignments! Last night I completed the refresh of the now PyTorch-based, Generative AI infused, course. #python #PyTorch #deeplearning GitHub: https://t.co/XauAYZZZDz YouTube: https://t.co/kPhv9xCZC0

[jeffheaton @jeffheaton]

Would you like to get up to 150x speed improvements with PANDA using a GPU without the major code changes usually needed? In this video I introduce the latest version of cudf, from @NVIDIAAI that can automatically fall back to CPU Pandas for the portions of your pandas code that cannot be automatically moved to the #gpu. #MachineLearning #pandas https://t.co/FHoEQrZTsg

[jeffheaton @jeffheaton]

@chimeldfr @BertChakovsky That is on the list, I want to give Python on IOS a try. Right now #flet looks like the best option, but I want to see it first hand.

[jeffheaton @jeffheaton]

My first Mac App Store app! Mostly just to prove that I can release something non-trivial, written in Python, to the MacOS App Store. My program is open source, I will provide a link to the code. I will do a much more in-depth write up of the somewhat convoluted process soon, if there is interest. The actual application is a cellular automata rule evolution that system that I published in a journal. Think, evolve your own Conway's Game of Life. #machinelearning #python #osx https://t.co/VMOu4RXntC