 Alexander Korolev
|
06590549ff
mimo-vllm: drop sparkring 20GiB KV pin — GB10 UMA auto-sizing keeps ~54GiB pool; pin would strand 34GiB
|
пре 3 недеља |
|
Alexander Korolev
|
1f59f0586e
mimo-vllm: adopt sparkring tp4 serve contract (spec5, 16x8192, 20GiB KV pin, bounce loader, BLHNC, A16 lm-head, extended-GID prefill); ring-only transport NOT ported (we're switched)
|
пре 3 недеља |
|
Alexander Korolev
|
9b9109112f
mimo: vLLM lane serving — 168/184 parity + slope sweep flattens the 256K cliff (0.30→0.74 of d0) at −34% d0; shift log 09-28
|
пре 3 недеља |
|
Alexander Korolev
|
1440b6e841
mimo26-vllm: quality bench vllm-kk-nightly-dflash — 168/184 (91/100), parity with sglang anchor
|
пре 3 недеља |
|
Alexander Korolev
|
4a33f594df
mimo-vllm: garble root cause = instanttensor mis-maps MiMo split-GDN weights; b12x loader contract (mm-encoder-tp-mode weights, full Omni, bf16 KV, VLLM_PLUGINS=b12x_loader); dflash re-enabled per Pro launcher; quality bench started
|
пре 3 недеља |
|
Alexander Korolev
|
787ea12644
mimo-vllm: spec RESOLVED note (nightly fixes all four walls); lane live, quality bench running
|
пре 3 недеља |
|
Alexander Korolev
|
c4668aa86f
mimo-vllm: reopen lane on eugr/spark-vllm-b12x nightly (karmic-kraken: MiMo aux, BLHNC layouts, b12x noncausal, eagle3 unwrap) + b12x linear/moe, CUTE_DSL_ARCH=sm_121a
|
пре 3 недеља |
|
Alexander Korolev
|
02846405b1
mimo: vLLM+DFlash experiment PARKED (r39 KV-layout wall); findings spec + recipe banner, patches kept for reopen
|
пре 3 недеља |
|
Alexander Korolev
|
1d0bb43ece
mimo26: vLLM B12X A/B lane recipe + sglang 4x fp8-KV trial rollback
|
пре 3 недеља |
|
Alexander Korolev
|
4ace7f4d06
bench roce-quality: 168/184 (91/100), resp 79, decode +10-15% — RoCEnante is quality-neutral
|
пре 3 недеља |
|
Alexander Korolev
|
615e526ecd
mimo26-flash-mopd-4x: RoCEnante live on TP4 (a15), tg +20-36%
|
пре 3 недеља |
|
Alexander Korolev
|
7a55726869
mimo26-spark: a15 — relax sglang RoCEnante gate from world==8 to b12x's 2..16
|
пре 3 недеља |
|
Alexander Korolev
|
7eba2f933f
report: bench completion is self-describing — no pool-size hardcode
|
пре 3 недеља |
|
Alexander Korolev
|
050f870355
mimo: 4x lane in production (GMU 0.94, bench 91/100, resp 76), JIT cache mounts
|
пре 3 недеља |
|
Alexander Korolev
|
3f71f5b346
mimo26-flash-mopd-4x: new 4-node contingency lane, GMU 0.94 prod (pool 4.14M)
|
пре 3 недеља |
|
Alexander Korolev
|
f39449c4ca
bench: --parallel passthrough + config C bench 166/184 (90/100, two-stream)
|
пре 3 недеља |
|
Alexander Korolev
|
ae420833fc
history: Pro-MOPD download was complete — GB/GiB misread (535 GiB == 573.5 GB), snapshot adea8e2c verified 130/130
|
пре 3 недеља |
|
Alexander Korolev
|
7ec09146a3
mimo26-flash-mopd-2x: config C verified — 512K context live (pool 570,752)
|
пре 3 недеља |
|
Alexander Korolev
|
848f902f3e
orchestrate: re-probe once before failing 'not enough free nodes'
|
пре 3 недеља |
|
Alexander Korolev
|
2f6c39e270
bench(mopd-kv088): MOPD config B — 168/184, pool 499K; shift notes
|
пре 3 недеља |
|
Alexander Korolev
|
c8676d703b
recipes(mimo-mopd-2x): mem-fraction 0.85→0.88 — KV pool 306,752→498,752 tokens
|
пре 3 недеља |
|
Alexander Korolev
|
6bc35921d8
bench(mopd-base): MiMo-V2.6-Flash-MOPD 2x — 170/184 (92/100, ★★★★★), +9 over Flash-RL
|
пре 3 недеља |
|
Alexander Korolev
|
ba3f771fc2
recipes: mimo26-flash-mopd-2x — MOPD flash lane (same arch, NAS download started)
|
пре 3 недеља |
|
Alexander Korolev
|
ba2d898771
docs(AGENTS): mikrotik power-off relay semantics
|
пре 3 недеља |
|
Alexander Korolev
|
5c82201e40
power: mikrotik off must cut the outlet relay too
|
пре 3 недеља |
|
Alexander Korolev
|
c01f43cdc8
bench(engram-local): aiden16 localized-Engram verification — 91/100, pp 10x, penalty gone
|
пре 3 недеља |
|
Alexander Korolev
|
b07f853b50
docs: aiden16 localized-engram cutover + liveness fix shift note
|
пре 3 недеља |
|
Alexander Korolev
|
6ae3c5159c
readiness: watch() must not call an API that never opened 'dead'
|
пре 3 недеља |
|
Alexander Korolev
|
789983fc62
staging(sparse-copy): aiden base+rem row tiling + aiden16 localized engram staging
|
пре 3 недеља |
|
Alexander Korolev
|
bc269457c2
bench(exl3-staged): 92/100 declarative-staging verification + bench --tokenizer passthrough
|
пре 3 недеља |
