 Alexander Korolev
|
8e49f12e5a
vision 2x: parked note — B12X sparse o_proj DeepGEMM fp8_einsum assert at TP2 (his verified values); 4x DCP2 is the profile lane
|
1 miesiąc temu |
|
Alexander Korolev
|
33cde9c3ef
vision 2x: his verified TP2 values (GMU 0.85, KV pin off) — 0.80+pin hit a DeepGEMM layout assert
|
1 miesiąc temu |
|
Alexander Korolev
|
7bb4c8a7b1
lab-dflash2: spec swap applied (mtp-3 -> dflash + lab drafter), name updated
|
1 miesiąc temu |
|
Alexander Korolev
|
200aae23f0
new lanes: aiden 4x + lab DFlash2 drafter (incoai swap); lab image + lab DFlash2 (MTP swap) — competitive A/Bs on the lab's speculative model
|
1 miesiąc temu |
|
Alexander Korolev
|
923e6b2bfc
vision recipes: args block was absorbed into [mods] table (TOML section scoping) — moved top-level
|
1 miesiąc temu |
|
Alexander Korolev
|
2b52f97ac3
aiden lanes: drop manual HF mounts (launcher auto-mounts ~/.cache/huggingface = shared NAS on all 4 nodes); GLM 4x -> NVFP4 (in shared cache, +23% vs FP8)
|
1 miesiąc temu |
|
Alexander Korolev
|
01c4ccad57
vision aiden lanes: flag route with --entrypoint sh override (baked-in env-script doubles the tail); 4x gets same fix + shared NAS HF mount
|
1 miesiąc temu |
|
Alexander Korolev
|
7497729957
vision 2x: MODEL_PATH -> cached snapshot (his script's unconditional --revision '' breaks offline resolution)
|
1 miesiąc temu |
|
Alexander Korolev
|
f8ab43bcd5
vision 2x: MODEL_PATH -> cached snapshot (his script's unconditional --revision '' breaks offline resolution)
|
1 miesiąc temu |
|
Alexander Korolev
|
8c292e8a5d
vision 2x: point SPARK_HF_HOME at the shared NAS HF cache (DFlash2/DSpark drafts live there; the NVMe copy lacked the DSpark draft)
|
1 miesiąc temu |
|
Alexander Korolev
|
b2ec9376cc
vision 2x: drop the revision pin (offline refs/main resolves; pinned hash broke the draft lookup)
|
1 miesiąc temu |
|
Alexander Korolev
|
dc13bd0afe
vision 2x: aiden image via HIS env-script entrypoint — mods derive NODE_RANK/MASTER_ADDR from VLLM_HOST_IP + sed off the $@ tail (env-driven contract, launcher-tail dropped)
|
1 miesiąc temu |
|
Alexander Korolev
|
f2ba69ee5d
vision 2x: use his /opt/vllm-src/.venv python (dist-packages wheel lacks the multimodal wrap) + his env block
|
1 miesiąc temu |
|
Alexander Korolev
|
41f2808d7a
vision 2x: cudagraph cap from sizes list (pydantic consistency); MAX env kept
|
1 miesiąc temu |
|
Alexander Korolev
|
a0c57b9e2f
aiden lanes: entrypoint=[] for built-in-entrypoint images (launcher doubles CMD otherwise); vision 2x uses proven overlay img + aiden DSpark-greedy additions (his env-script needs a dedicated pass)
|
1 miesiąc temu |
|
Alexander Korolev
|
ce4727d2a3
aiden lanes: 2x+4x recipes from aidenle.com guides (glm prod-1.1 DFlash2-7 + dsv4 vision prod-3.72 DSpark-6); 2x to verify, 4x deferred (FP8 ckpt downloading)
|
1 miesiąc temu |
|
Alexander Korolev
|
fd4c74dcc1
checkpoint A/B on ds4: NVFP4 lab ckpt = same 88 quality on r17 stack but +4-7 tok/s; lab day-0 stack retains 93; r17abi recipe added
|
1 miesiąc temu |
|
Alexander Korolev
|
9079638ee4
lab recipes: instanttensor baked into lab image + enabled (verified: 631s vs 700s, two-pass loader, all boot markers + smoke 3901)
|
1 miesiąc temu |
|
Alexander Korolev
|
04984a169d
ledger: one-backend bisect solved (linear b12x = the toxin); ship config locked
|
1 miesiąc temu |
|
Alexander Korolev
|
2d379f9752
final bisect: moe-b12x kept (88, d8192+2.8), linear-b12x is the toxin (44), align perf-negative — recipe locked
|
1 miesiąc temu |
|
Alexander Korolev
|
582343c9c6
r17 composition ship anchor: backend-flags are the quality toxin (bisect 46/47/74 vs 87); knobs env clean; graphs+MTP fixed on public r17 refs; ledger
|
1 miesiąc temu |
 kilork
|
08a930e477
r12 image: adopt eugr lab build patches (12x patch_vllm_*.py) + pre-staged build sub-deps (SRC_DIR escapes) for flaky node github; ledger: lil render reference + perf push state
|
1 miesiąc temu |
|
kilork
|
cc41c48e03
glm53 spark recipe: full lil policy alignment (b12x linear/moe backends, k=5 MTP + moe_backend, instanttensor load, mamba-cache-mode align, INSTANTTENSOR env) + mounted pre-baked JIT caches (~/sm121-cache/glm53-jit-121f, dsv4 pattern); boot 35m -> 7m, exact-391 verified
|
1 miesiąc temu |
|
kilork
|
5283d11ce1
clean r12 image shipped: 90/100 c1 (was 88 day-0-base), entrypoint+MAX_JOBS boot fixes, ledger
|
1 miesiąc temu |
|
kilork
|
236ce5a7f3
spark recipe: entrypoint (clean image is ENTRYPOINT []) + MAX_JOBS=2 (JIT nvcc OOM at boot); clean r12 image serving, exact-391 verified
|
1 miesiąc temu |
|
kilork
|
3d084c54e4
cli: proper --no-follow flag (SetTrue opt-out; the SetFalse+default combo inverted semantics and lacked the negation)
|
1 miesiąc temu |
|
kilork
|
2d17cd0356
r12 clean build: SHA-pinned PR heads (550 closed), deepgemm pre-clone, lil launch-policy diff recorded
|
1 miesiąc temu |
|
kilork
|
9395d6d955
r12 clean build: flashinfer-python 0.6.17 from PyPI (whl-index path dead); recipe MTP:3 comment clarified (built-in MTP head, not dspark)
|
1 miesiąc temu |
|
kilork
|
8de6340763
clean r12 rebuild: raw-cuda base per solar pattern (r12-specific refs, torch 2.13, flashinfer 0.6.17, b12x PR-head composition); docker_image_dir fixes; bench 88/100 (69-suite)
|
1 miesiąc temu |
|
kilork
|
ac5c6c9108
NVFP4-Spark serving on our r12 fork build: spec-config attention_backend=B12X fixes MTP draft selection; intel recipe retargeted; ledger
|
1 miesiąc temu |
