Using uv with PyTorch (via) PyTorch is a notoriously tricky piece of Python software to install, due to the need to provide separate wheels for different combinations of Python version and GPU accelerator (e.g. different CUDA versions).
In the meantime, you can make sense of tinygrad’s architecture on google colab.
uv now has dedicated documentation for PyTorch which I’m finding really useful - it clearly explains the challenge and then shows exactly how to configure a pyproject.toml such that uv knows which version of each package it should install from where.
You don’t have to manually install packages anymore! Instead, you write a program that automates that process.
quartz config
display source property
Rust has made it easier to build command line tools which made it easier to build package managers like uv and pixi. We are about to topple the local first domino!
There is an ongoing shift from data silos, SaaS to markdown files, dynamic documents as personal software.