 Alexander Korolev
|
551ea7d8d7
feat(nccl): auto-detect NCCL_IB_GID_INDEX per node, opt-out via config pin
|
2 ay önce |
|
Alexander Korolev
|
df4149920b
fix: remediate F1 (image-sync host-key verification) and F2 (docker run quoting)
|
2 ay önce |
|
Alexander Korolev
|
0181645b9e
fix: make IB config optional for single-node clusters
|
2 ay önce |
|
Alexander Korolev
|
1b4d07309a
fix: make NCCL config optional for single-node, capture stderr, support --on for tmux
|
2 ay önce |
|
Alexander Korolev
|
afa6999213
feat(tmux): support inline venv_setup + venv_setup_script
|
2 ay önce |
|
Alexander Korolev
|
5f646966b8
feat(tmux): add venv_setup field for declarative venv creation
|
2 ay önce |
|
Alexander Korolev
|
d9b922e01b
feat(recipe): add venv field for declarative tmux venv activation
|
2 ay önce |
|
Alexander Korolev
|
c4a74cc152
feat(recipe): add deploy field for docker/tmux variant selection
|
2 ay önce |
|
Alexander Korolev
|
de18759712
feat(serve): add composed_model for multi-repo model directories
|
2 ay önce |
|
Alexander Korolev
|
febdc3b9d7
feat(build): add Solar Open2 NVFP4 recipe with recipe-pinned versions
|
2 ay önce |
|
Alexander Korolev
|
23ae7a6484
feat(serve): add solo mode and fix single-node launch bug
|
3 ay önce |
|
Alexander Korolev
|
65198e60fc
feat(recipe): add built-in script aliases and after_stop hook
|
3 ay önce |
|
Alexander Korolev
|
1ba1d99e05
fix: resolve review findings across modules
|
3 ay önce |
|
Alexander Korolev
|
2b0deeaffa
feat(recipe): extra_mounts with config alias resolution + tilde + env injection
|
3 ay önce |
|
Alexander Korolev
|
9e28b49887
feat(recipes): add volumes support and GLM-5.2 Hybrid FP8-MXFP4 recipe
|
3 ay önce |
|
Alexander Korolev
|
731937a625
feat: --build-node, docker_image_dir recipe field
|
3 ay önce |
|
Alexander Korolev
|
acadda1259
feat: auto-sync Docker image (ANY source node) + --sync-image / --sync-image-source flags
|
3 ay önce |
|
Alexander Korolev
|
ff23a80a81
feat: extend NcclConfig with per-cluster NCCL overrides
|
3 ay önce |
|
Alexander Korolev
|
0ae454c259
feat: add --node parameter to logs command
|
4 ay önce |
|
Alexander Korolev
|
4936530710
feat: RecipeArgs enum — inline string or list for recipe args
|
4 ay önce |
|
Alexander Korolev
|
5931869d69
refactor: DistributedBackend as Option — remove None_ variant
|
4 ay önce |
|
Alexander Korolev
|
dbdbc7ed62
refactor: replace String distributed_backend with DistributedBackend enum
|
4 ay önce |
|
Alexander Korolev
|
6081ccf443
feat: mods system — entrypoint wrapping + global/recipe mods with volume mount
|
4 ay önce |
|
Alexander Korolev
|
f55cd70082
refactor: replace #[allow] with structs for start_cluster and build_docker_command
|
4 ay önce |
|
Alexander Korolev
|
04413a8991
feat: auto-scale TP when recipe runs on larger cluster
|
4 ay önce |
|
Alexander Korolev
|
9d54d3f94c
gemma4-12b: working recipe with distributed_backend=none
|
4 ay önce |
|
Alexander Korolev
|
7b01e19ec2
fix: HF_HOME -> /root/.cache/huggingface (fixes stale cache)
|
4 ay önce |
|
Alexander Korolev
|
9241bdc682
fix: don't pass --distributed-executor-backend mp (auto-detected)
|
4 ay önce |
|
Alexander Korolev
|
6f2c33a82f
fix: only add --shm-size when recipe explicitly sets it
|
4 ay önce |
|
Alexander Korolev
|
d9c3f845c2
fix(deepseek): align parameters with working system
|
4 ay önce |
