LoRA adapters trained on the nodes: llmgw ctl train

A job names the catalog model the adapter is for, the safetensors weights to
train on, the node and the MiB of a card it needs, and question/answer pairs.
The control plane sends it when the node's agent is there and nothing else
trains on it; the agent trains with its installed vLLM's torch and a pinned
overlay of peft and accelerate (no second environment), as the engines'
user, on the card with that much free — beside an engine, if it leaves that
much — converts the adapter to GGUF with the installed llama.cpp's own
conversion code when the model is served by llama.cpp, and sends what it
made back in pieces, kept in the store to download. Progress, loss and peak
card memory are reported as it goes; a node that goes fails its job.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Ilker ManapandClaude Opus 5.5 committed 2026-10-08 20:25:57 +02:00
1 parent ae95ac8261
commit 79b3e9f887
24 files changed
+3185 -72

No files matched your search

+4 -1
View File
@@ -70,7 +70,9 @@ type API struct {
// guard is the guardrail policies' library and attachments.
guard *GuardAdmin
// kb is knowledge bases; nil without a store.
kb *KBAdmin
kb *KBAdmin
// train is adapter training; nil without a store and agents.
train *TrainAdmin
settings *Settings
// restartAgents tells every connected agent to start again.
restartAgents func(reason string) []string
@@ -117,6 +119,7 @@ func (a *API) Register(mux Mux) {
mux.HandleFunc("POST /api/v1/models/{name}/resources", a.setResources)
a.registerQuotas(mux)
a.registerKB(mux)
a.registerTrain(mux)
a.registerGuard(mux)
a.registerSettings(mux)
mux.HandleFunc("POST /api/v1/agents/restart", a.restartAgentsHandler)