Mastodawn

Create a one element tensor in PyTorch with torch.tensor today.

Run a 1T parameter model on a 32gb Mac by streaming tensors from NVMe

#HackerNews #Run #a #1T #parameter #model #on #a #32gb #Mac #by #streaming #tensors #from #NVMe #https://github.com/t8/hypura #MachineLearning #Tensors #NVMe #Mac #Optimization

GitHub - t8/hypura: Run models too big for your Mac's memory

Run models too big for your Mac's memory. Contribute to t8/hypura development by creating an account on GitHub.

GitHub

Stephen Blum Feb 7

Arrays store data; matrices are 2D; tensors generalize across dimensions.

#math #ai #tensors

Hackaday Jan 31

Writing an Optimizing Tensor Compiler from Scratch

https://fed.brid.gy/r/https://hackaday.com/2026/01/30/writing-an-optimizing-tensor-compiler-from-scratch/

13 Jan 31

Writing an Optimizing Tensor Compiler from Scratch https://hackaday.com/2026/01/30/writing-an-optimizing-tensor-compiler-from-scratch/
#SoftwareDevelopment #Tensors

Writing An Optimizing Tensor Compiler From Scratch

Not everyone will write their own optimizing compiler from scratch, but those who do sometimes roll into it during the course of ever-growing project scope creep. People like [Michael Moroz], who w…

Hackaday

3rdEye Dec 25

1st Open-Sourced Neural Network Tensor Activation Visualization for GGUF Models??

https://github.com/3rdEyeVisuals/Spectra-Vis

Heavily stripped down version of my private flow. Open to questions, comments, and suggestions. This repo will be maintained and occasionally enhanced.

#MachineLearning #LLMs #Tensors #InterpretabilityTools

GitHub - 3rdEyeVisuals/Spectra-Vis

Contribute to 3rdEyeVisuals/Spectra-Vis development by creating an account on GitHub.

GitHub

Hacker News Oct 14

All in on MatMul? Don’t Put All Your Tensors in One Basket!

https://www.sigarch.org/dont-put-all-your-tensors-in-one-basket-hardware-lottery/

#HackerNews #MatMul #Tensors #HardwareLottery #AIResearch #DataScience

All in on MatMul? Don’t Put All Your Tensors in One Basket!

Matrix multiplication dominates AI hardware and research. Betting everything on MatMul risks an innovation monoculture — it’s time to diversify our compute bets.

SIGARCH