I saw that blog post about running local models on a Xeon based server so I benchmarked a bunch of local models on it. The server is dual Xeons with 256 GB of DDR4 and I found it capable of running some things locally. It's slow but still faster than coding things manually. Also that was just an example. Even my 6 year old 6900XT is faster.
alt Hacker News
I saw that blog post about running local models on a Xeon based server so I benchmarked a bunch of local models on it. The server is dual Xeons with 256 GB of DDR4 and I found it capable of running some things locally. It's slow but still faster than coding things manually. Also that was just an example. Even my 6 year old 6900XT is faster.