strata serve: prompt 94 tokens = 0 reused + 94 read in 3132 ms (30.0 tok/s), 256 generated in 7696 ms (33.3 tok/s), drafts accepted 162 of 214, 1 checkpoints
strata serve: decode expert cache hit rate: 78.3% (99566 hits / 127102 lookups)
strata serve: KV streaming: 99.50% of 209376 block reads hit VRAM, 4.3 MiB read from RAM
