small speech stack for local use: setup advice

[cobaltworks25]

I think the bottleneck is somewhere else. I've seen people blame speculative decoding when the real issue was much more mundane. Could be a workload difference, but I wouldn't generalize too fast from one setup.