AI models are a core resource for inferencing. The models object in the project manifest allows you to easily define models, whether hosted by Hypermode or another host.

hypermode.json
{
  "models": {
    "sentiment-classifier": {
      "task": "classification",
      "sourceModel": "distilbert/distilbert-base-uncased-finetuned-sst-2-english",
      "provider": "hugging-face",
      "host": "hypermode"
    }
  }
}

Each model requires a unique name, specified as a key, containing only alphanumeric characters and hyphens.

Properties

task
string
required

Training intent of the model. Currently, must be one of: classification, embedding, or generation.

sourceModel
string
required

Original relative path of the model within the provider’s repository.

provider
string
required

Organization or directory that provided the source model, such as hugging-face or openai.

host
string
required

Host for the model instance.

  • Specify "hypermode" for models that are automatically deployed by Hypermode.
  • Otherwise, specify a name that matches a host defined in the hosts section of the manifest.

Auto-deployed models

When using hugging-face as the provider and hypermode as the host, Hypermode automatically deploys a dedicated instance of the defined sourceModel when deploying your project. Your project’s functions securely connect to the hosted model, with no further configuration required.