 Alexander Korolev
|
51912523c5
feat(registry): publish --registry upload + sm121 registry list/show/install
|
1 miesiąc temu |
|
Alexander Korolev
|
79d1a76c74
feat(cli): sm121 publish — recipe -> standalone reproducible bundle (and gist)
|
1 miesiąc temu |
|
Alexander Korolev
|
ed4e0f99ea
build(images): digest-pin glm53 base image — tag glm53-flash-arm64-cu130 is mutable; pin sha256:905c0293 (the base our verified images were built on). Checksum changes → one no-op rebuild on next serve.
|
1 miesiąc temu |
|
Alexander Korolev
|
96c6bfe777
feat(recipes): lab defaults to --enforce-eager — MTP+CUDA-graphs quality bug confirmed by A/B bench
|
1 miesiąc temu |
|
Alexander Korolev
|
28e08afa73
feat(recipes): redhat @ 524288 ctx — pinned 10 GiB KV, autotune off, 4096 chunks, lab image for CC-12.0 topk guards
|
1 miesiąc temu |
|
Alexander Korolev
|
3d30d85655
docs(specs): lab /174 explained — TC-53 excluded on 120s infra timeout (score unchanged at 89 either way)
|
1 miesiąc temu |
|
Alexander Korolev
|
a0383302e3
docs(specs): redhat MTP reality check — RedHatAI ships mtp as separate -mtp-FP8 variant repos (GLM-5.2 pattern); none for 5.3-Flash yet
|
1 miesiąc temu |
|
Alexander Korolev
|
5730abbfee
docs(specs): GLM-5.3-Flash three-way quant benchmark on ds4 (measured)
|
1 miesiąc temu |
|
Alexander Korolev
|
9d72c8b804
feat(recipes): lab GLM-5.3 recipe = FujitsuPolycom proven stack, verified on ds4
|
1 miesiąc temu |
|
Alexander Korolev
|
b710a09bcb
feat(recipes): GLM-5.3-Flash NVFP4 variants — local-inference-lab + RedHatAI
|
1 miesiąc temu |
|
Alexander Korolev
|
01fd6bd817
chore(deps): drop unused russh-keys dependency
|
1 miesiąc temu |
|
Alexander Korolev
|
25c53b5f50
refactor(recipes): drop redundant entrypoint override from glm53-flash-nvfp4
|
1 miesiąc temu |
|
Alexander Korolev
|
095dec1d73
refactor(recipes): quant-agnostic served-model-name + rename glm53 to -nvfp4
|
1 miesiąc temu |
|
Alexander Korolev
|
995e45c6c6
fix(recipes): GLM-5.3-Flash -> vLLM NVFP4 on 2x Spark with NoPE-MLA mod
|
1 miesiąc temu |
|
Alexander Korolev
|
cdedc90e5b
feat(recipes): Qwen3.8-Flash-Next FP8 on vLLM via expert parallelism
|
1 miesiąc temu |
|
Alexander Korolev
|
5ee2326c40
fix(recipes): switch Qwen3.8-Flash-Next to NVFP4 for 2x Spark
|
1 miesiąc temu |
|
Alexander Korolev
|
08d005ffb5
feat(config): global [before_serve."*"] hooks + optional master_ib_ip
|
1 miesiąc temu |
|
Alexander Korolev
|
e9212feb72
feat(recipes): add Qwen3.8-Flash-Next-FP8 and GLM-5.3-Flash SGLang day-0 recipes
|
1 miesiąc temu |
|
Alexander Korolev
|
7742fbf290
feat(deepseek-v4-flash): fix current parameters
|
1 miesiąc temu |
|
Alexander Korolev
|
0949d63f47
web: fix log window dying after close (onClosed compared win objects)
|
2 miesięcy temu |
|
Alexander Korolev
|
b845cd33e1
web: DMMono Nerd Font — the actual Alacritty font
|
2 miesięcy temu |
|
Alexander Korolev
|
32e5bab765
web: monospace content in Menlo (the Alacritty font)
|
2 miesięcy temu |
|
Alexander Korolev
|
b1854d10ab
web: fix recipes-empty-after-load + dead log streams
|
2 miesięcy temu |
|
Alexander Korolev
|
aca3310f31
docs: record the UI↔API parity audit (4f1837d)
|
2 miesięcy temu |
|
Alexander Korolev
|
4f1837d856
web: close the UI<->API gap — every server capability reachable
|
2 miesięcy temu |
|
Alexander Korolev
|
606e0b65cf
docs: mark web P4 done — all dashboard phases shipped
|
2 miesięcy temu |
|
Alexander Korolev
|
ec9b0d2e4f
clippy: drop imports left dangling by the test split
|
2 miesięcy temu |
|
Alexander Korolev
|
0950d9dc26
fmt
|
2 miesięcy temu |
|
Alexander Korolev
|
887bcac7ba
web tests: split API round-trips into api_tests.rs (500-line cap)
|
2 miesięcy temu |
|
Alexander Korolev
|
3ca1d59e4e
web P4: raw content windows + animation polish
|
2 miesięcy temu |
