You can read a bit more in the "expert Parallelism" section in the Ultra-Scale Playbook: https://
huggingface.co/spaces/nanotro
n/ultrascale-playbook
…
@thom_wolf
-
Ultra-Scale Parallelism Guide for Advanced AI Training
By
–
-
First Open-Source Software Conference for Fusion Energy 2025
By
–
Don't know who need to hear it but in 2025 there is a first-of-its-kind conference on *Open-Source* Software for *Fusion* Energy We're living in a (cool) future
-
Ultra-Scale Playbook mobile optimization needs community contribution
By
–
Making the open-source “Ultra-Scale Playbook” mobile/phone friendly would be quite amazing if someone feels like contributing and wants to give it a try
-
Ultrascale Playbook Podcast: 1h30 AI Learning Summary
By
–
if you didn't have time yet to read the 100-page "Ultrascale Playbook", I gathered all the NotebookLM audios spread along the main sub-sections of the book in a single 1h30 podcast below
— Thomas Wolf (@Thom_Wolf) 20 février 2025
factuality is pretty good – great listen for the car https://t.co/vjgtUAaZW2 pic.twitter.com/LwmpE6iBaQif you didn't have time yet to read the 100-page "Ultrascale Playbook", I gathered all the NotebookLM audios spread along the main sub-sections of the book in a single 1h30 podcast below factuality is pretty good – great listen for the car
-
Ultra-Scale Playbook: 5D Parallelism and CUDA Optimization Guide
By
–
After 6+ months in the making and burning over a year of GPU compute time, we're super excited to finally release the "Ultra-Scale Playbook" Check it out here: http://
hf.co/spaces/nanotro
n/ultrascale-playbook
… A free, open-source, book to learn everything about 5D parallelism, ZeRO, fast CUDA kernels, -
Platform hosts 3,500 terabytes of GGUF model files
By
–
we currently host 3,500 terabytes of .gguf files – just saying
-
8GB High Quality Math Reasoning Model Released
By
–
holy shit 8GB of high quality reasoning math just dropped
-
Macron and Vance Champion AI Infrastructure Investment and Collaboration
By
–
Three highlight from the AI summit:
– Emmanuel Macron ‘s « Plug, baby, plug » and $109B investment plan. Let’s scale AI infrastructures now!
– JD Vance’s call to « embark on the AI revolution before us with the spirit of openness and collaboration » pushing the frontier of what -
60k Students Join New AI Agent Course Tomorrow
By
–
we’re now way past 60k+ student joining the new AI agent course! Totally mind-blown It's starting tomorrow, you can still join at hf[dot]co/agents-course Syllabus:
-
DeepSeek-R1 RL Training Simplicity Versus Process Reward Models
By
–
From an open-research point of view, probably the greatest thing about DeepSeek–R1 is how its RL training technique appears so straightforward and simple in comparison to the cumbersome approaches people were starting to think necessary for learning reasoning like Process Reward