strata serve: prompt 4022 tokens = 0 reused + 4022 read in 12752 ms (315.4 tok/s), 256 generated in 7120 ms (36.0 tok/s), drafts accepted 164 of 213, 1 checkpoints
strata serve: decode expert cache hit rate: 81.4% (104553 hits / 128423 lookups)
strata serve: KV streaming: 99.34% of 1922532 block reads hit VRAM, 51.3 MiB read from RAM
strata serve: suffix drafts: 2 windows, 5 of 6 drafts accepted
