馃殌 The feature, motivation and pitch
What would it take to continue support for intel macOS? is this a technical problem, or an issue of human-power?
Projects (e.g. github.com/orbital-materials/orb-models) make use of features in later versions of pytorch (they require 2.6.0 and there are errors when I manually bypass this requirement), but it appears pytorch is capped for macOS at python3.12 and the latest wheels are 2.2.2. this means some code is simply non-runnable, on computers that are a mere 5 years old. I wrote an "issue" on their github too, but this feels like less of a "them problem" and more of a pytorch problem.
Alternatives
No response
Additional context
No response
cc @seemethere @malfet @atalman @tinglvv @nWEIdia @frank-wei @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10
馃殌 The feature, motivation and pitch
What would it take to continue support for intel macOS? is this a technical problem, or an issue of human-power?
Projects (e.g. github.com/orbital-materials/orb-models) make use of features in later versions of pytorch (they require 2.6.0 and there are errors when I manually bypass this requirement), but it appears pytorch is capped for macOS at python3.12 and the latest wheels are 2.2.2. this means some code is simply non-runnable, on computers that are a mere 5 years old. I wrote an "issue" on their github too, but this feels like less of a "them problem" and more of a pytorch problem.
Alternatives
No response
Additional context
No response
cc @seemethere @malfet @atalman @tinglvv @nWEIdia @frank-wei @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10