Fabric looks like a great tool for a Linux user familiar with pipes. I’ve had a bit of a play with it but only touched the surface.
I may be wrong but I have conceptualised fabric as letting you take some data, pipe it into one or more “verbs” and get a hopefully desired output. By “verbs” I mean analyse, convert, create, extract, improve, rate, etc. (See patterns).
I recent came across Alex Jones’ llmfit, a “terminal tool that right-sizes LLM models to your system’s RAM, CPU, and GPU” which should work well with fabric.
It lists DeepSeek‑R1‑Distill‑Qwen‑14B as the first choice for my hardware although I have yet to give that a go.
I’m looking forward to your talk on Thursday night about fabric. I’ve titled it, “Fabric - the AI pipeline”