Show HN: FusionHarness – An Open-source Mixture-of-Agents compound-model servergithub.com5 points·by jackxlau·24 giorni fa·1 comments
jackxlau·28 giorni fa·discussIn my own testing I have seen peak performance happen usually within 15-20% of the intended context limit, albeit there are a few optimizations depending on the task quality.
jackxlau·mese scorso·discussthe handshake process to unlock the A100 cores is pretty interesting it reminds me of driving CPU cores like a GPU command queue.
jackxlau·mese scorso·discussI came across the conclusion here since a change sometimes spans several repos, per-repo history optimizes the wrong target.