Skip to main content

Settings of models.

GET 

/model-settings

Only models available to the client will be listed.

Responses

OK

Schema
  • Array [
  • namestring
    statusstring

    The current availability status of the model. Currently supported states are:

    • available: The model is configured and a matching worker is connected to serve it.
    • unavailable: The model is configured but no worker has shown recent activity to serve it.

    Possible values: [available, unavailable]

    worker_typestring

    The worker type that is used to serve the configured model. The following fields are supported:

    • luminous: The model is served by a Luminous worker.
    • vllm: The model is served by a vLLM worker.
    • translation: Worker type to serve translation requests.
    • transcription: Worker type to serve transcription requests.

    Possible values: [luminous, vllm, transcription, translation]

    descriptionstring
    multimodalboolean

    Feature flag for whether or not multimodal prompts are available to users.

    max_context_sizenumber

    The maximum context size of this model.

    semantic_embeddingboolean

    True if this model supports semantic embeddings.

    completion_typestring

    The completion type supported by the model.

    • none: The model has not been trained to support completions. Trying to trigger a completion request will lead to a validation error.
    • full: The model has been trained to support completions.

    Possible values: [none, full]

    embedding_typestring

    The embedding type supported by the model.

    • none: The model cannot be used for embeddings. The scheduler will reject requests for embeddings to this model.
    • raw: The model has not explicitly been trained to support embeddings. However, it is possible to retrieve the embedding details technically. This option maps to the /embed endpoint.
    • semantic: The model has been trained with a switchable set of weights usable for semantic embedding retrieval. This option maps to the /semantic_embed endpoint.
    • instructable: The model has been trained to support any custom instruction for embedding retrieval. This option maps to the /instructable_embed endpoint.

    This flag replaces semantic_embedding_enabled and should always be set. If embedding_type is unset, semantic_embedding_enabled will take control. If both embedding_type and semantic_embedding_enabled are used, implausible combinations are rejected.

    Possible values: [none, raw, semantic, instructable]

    alignedboolean

    Specifies whether the model is aligned s.t. end users can be warned about the model's limitations.

    chatboolean

    True if this model is supported by the chat endpoint.

    prompt_templatestring

    A prompt template that can be used for this model.

  • ]

Authorization: http

name: tokentype: httpscheme: bearerdescription: Can be generated in your [Aleph Alpha profile](https://app.aleph-alpha.com/profile)
var client = new HttpClient();
var request = new HttpRequestMessage(HttpMethod.Get, "https://docs.aleph-alpha.com/model-settings");
request.Headers.Add("Accept", "application/json");
request.Headers.Add("Authorization", "Bearer <token>");
var response = await client.SendAsync(request);
response.EnsureSuccessStatusCode();
Console.WriteLine(await response.Content.ReadAsStringAsync());
Request Collapse all
Auth