Alexander Korolev
|
5e0d7c2306
specs: DeepInfer assessment — offline-only + SM100-only, not runnable on GB10; the transferable ideas (SWA bounded replay, confidence-scheduled verification)
|
2 天之前 |
Alexander Korolev
|
7a8c147767
night shift 2026-10-20: FLARE 1017 tested and rejected (2x slower decode, crashes on agent prompts); Beam-1.0 NVFP4 lane added at 90/100 (fleet best); four-lane comparison recorded
|
2 天之前 |
Alexander Korolev
|
5a4eec608c
spec: 1017 TensorFold lane is now an sm121 recipe; record the TF+vLLM blocker
|
4 天之前 |
Alexander Korolev
|
8f734c7926
spec: TensorFold 2x 1017 lane measured — 89/100 with half the nodes, decode 50.4 vs 39-43, but turn latency 6.7s vs 1.4s
|
4 天之前 |
Alexander Korolev
|
b17dc421d2
spec: 1017 reasoning-effort A/B — 75 buys nothing over 100 on the new model (86 vs 87, same 1.4s turn); depth sweep closed
|
4 天之前 |
Alexander Korolev
|
0deb95f896
spec: 1017 depth sweep (16K-256K) on the EP2/1M config — single-stream and suffix decode hold, concurrent deep prefills starve decode
|
4 天之前 |
Alexander Korolev
|
000f022481
spec: dsv41-1017 tuning A/B (EP2 decode win, fp4-indexer negative, engram MAX_IDS vs chunked-prefill coupling, slow-NAS boot desync)
|
4 天之前 |