akka optimize training endpoints info
Show what an endpoint here can serve
Synopsis
Show what an endpoint of this deployment can serve and what one claims.
"Adapter ranks" is the largest LoRA rank an endpoint loads. An adapter of a lower rank loads too, and a run that trains above it is refused at submission. "Adapters loaded" is how many adapters one endpoint has loaded at once beside its base model. More can attach, and one is loaded when a request first names it. "GPUs per endpoint" and the device type are what one endpoint claims for as long as it answers. The deployment sets each of these, and "training endpoints create" takes none of them. The JSON output also lists the rank limits the serving engine accepts, which an operator chooses the deployment’s limit from.
On the vLLM engine, each architecture of a selected base model is listed with what an endpoint returns for a model of it beside its text: tool calls, where the architecture has a tool-call parser, and reasoning apart from the answer, where it has a reasoning parser.
An endpoint serves a base model and adapters over it. A model registered as whole weights is not served.
akka optimize training endpoints info [flags]
Examples
# Show what an endpoint here can serve
akka optimize training endpoints info
# Read the largest adapter rank using jq
akka optimize training endpoints info --jq .maxAdapterRank
Options inherited from parent commands
--cache-file string location of cache file (default "~/.akka/cache.yaml")
--config string location of config file (default "~/.akka/config.yaml")
--context string configuration context to use
--disable-prompt Disable all interactive prompts when running akka commands. If input is required, defaults will be used, or an error will be raised.
This is equivalent to setting the environment variable AKKA_DISABLE_PROMPTS to true.
--force Skip the confirmation prompt
--jq string Filter the JSON output with a jq expression
--json string Output JSON with only these fields; --json= lists the fields a command offers
--no-color Print without color, as NO_COLOR does
-o, --output string set output format to one of [text,json,json-compact,go-template=] (default "text")
--page-mode string the mode for paging, either paged, buffered or auto. (default "auto")
-q, --quiet set quiet output (helpful when used as part of a script)
--region string Region to use if the project has more than one region
--timeout duration client command timeout (default 10s)
--trainer-url string Trainer service URL (default TRAINER_URL, then AKKA_OPTIMIZE_URL, then http://localhost:9001)
--url string Optimize service URL (default OPTIMIZE_URL, then AKKA_OPTIMIZE_URL, then http://localhost:9010)
--use-grpc-web use grpc-web when talking to Akka APIs. This is useful when behind corporate firewalls that decrypt traffic but don't support HTTP/2.
--verbose set verbose output
SEE ALSO
-
akka optimize training endpoints - Serve a base model or a candidate to try it by hand