 Alexander Korolev
|
bbe9605380
Plan: sm121 web dashboard (design doc)
|
il y a 2 mois |
|
Alexander Korolev
|
4275e1cd7c
config: global-config mgmt commands + ssh creds from cluster sources
|
il y a 2 mois |
|
Alexander Korolev
|
8446509baf
Global config: recipe_sources + cluster_sources
|
il y a 2 mois |
|
Alexander Korolev
|
48bbec1f66
serve: --clear flag to wipe host-side extra_mount folders (sudo)
|
il y a 2 mois |
|
Alexander Korolev
|
65f366427a
Add Qwen3.8-27B recipes + measured state doc
|
il y a 2 mois |
|
Alexander Korolev
|
08871e3dfa
Add --fit flag to serve: scale recipe to N nodes or whole cluster
|
il y a 2 mois |
|
Alexander Korolev
|
551ea7d8d7
feat(nccl): auto-detect NCCL_IB_GID_INDEX per node, opt-out via config pin
|
il y a 2 mois |
|
Alexander Korolev
|
08d1e9fe03
feat(serve): add --mirror — serve from a per-node local HF mirror
|
il y a 2 mois |
|
Alexander Korolev
|
3aa9d380b2
feat(recipes): agent-serving profile + GMU/env fixes from benchmark findings
|
il y a 2 mois |
|
Alexander Korolev
|
afa6999213
feat(tmux): support inline venv_setup + venv_setup_script
|
il y a 2 mois |
|
Alexander Korolev
|
5f646966b8
feat(tmux): add venv_setup field for declarative venv creation
|
il y a 2 mois |
|
Alexander Korolev
|
d9b922e01b
feat(recipe): add venv field for declarative tmux venv activation
|
il y a 2 mois |
|
Alexander Korolev
|
c4a74cc152
feat(recipe): add deploy field for docker/tmux variant selection
|
il y a 2 mois |
|
Alexander Korolev
|
240002c125
fix(glm52): patch vLLM structured-output reasoning-boundary bug under MTP
|
il y a 2 mois |
|
Alexander Korolev
|
dbbdc3efbf
fix(composed_model): add exclude field to skip overlay files
|
il y a 2 mois |
|
Alexander Korolev
|
f9abbeb509
feat(compose): recursive whole-hierarchy symlink when no glob is set
|
il y a 2 mois |
|
Alexander Korolev
|
0208fc12f1
feat(compose): support subdirectory globs in composed_model sources
|
il y a 2 mois |
|
Alexander Korolev
|
892c28d1a4
feat(serve): add --force flag to ignore occupied nodes
|
il y a 2 mois |
|
Alexander Korolev
|
de18759712
feat(serve): add composed_model for multi-repo model directories
|
il y a 2 mois |
|
Alexander Korolev
|
febdc3b9d7
feat(build): add Solar Open2 NVFP4 recipe with recipe-pinned versions
|
il y a 2 mois |
|
Alexander Korolev
|
23ae7a6484
feat(serve): add solo mode and fix single-node launch bug
|
il y a 2 mois |
|
Alexander Korolev
|
65198e60fc
feat(recipe): add built-in script aliases and after_stop hook
|
il y a 2 mois |
|
Alexander Korolev
|
1ba1d99e05
fix: resolve review findings across modules
|
il y a 3 mois |
|
Alexander Korolev
|
4767f6fe4d
docs(agents): warn that cluster may be serving live before stop/power off
|
il y a 3 mois |
|
Alexander Korolev
|
b5905ac702
docs(agents): clarify path-ref resolves to container-side HF cache path
|
il y a 3 mois |
|
Alexander Korolev
|
a7cd759af1
feat(recipe): model-subfolder path refs via {{name}} substitution
|
il y a 3 mois |
|
Alexander Korolev
|
6cd245f9a8
feat(recipe): model-subfolder mounts resolved per-node from HF cache
|
il y a 3 mois |
|
Alexander Korolev
|
2b0deeaffa
feat(recipe): extra_mounts with config alias resolution + tilde + env injection
|
il y a 3 mois |
|
Alexander Korolev
|
12f1fa44f9
feat!: rename binary from sm121-launcher to sm121
|
il y a 3 mois |
|
Alexander Korolev
|
acadda1259
feat: auto-sync Docker image (ANY source node) + --sync-image / --sync-image-source flags
|
il y a 3 mois |
