You're right. I'm getting ~33tok/sec w/ dflash on it, even bursts up to 60tok/sec, using my personal home-built-for-Spark inference engine (not vLLM or llama.cp

You're right. I'm getting ~33tok/sec w/ dflash on it, even bursts up to 60tok/sec, using my personal home-built-for-Spark inference engine (not vLLM or llama.cpp based) That's pretty respectable. Still working on optimizing and cleaning up before I push it.