ModelType

The ModelType enum defines the supported AI model providers in the Oblix orchestration system. It's used when hooking models to specify which provider the model belongs to.

Import

from oblix import ModelType

Enum Values

Enum Value	Description
ModelType.OLLAMA	Local models served via Ollama
ModelType.OPENAI	Cloud models provided by OpenAI (GPT series)
ModelType.CLAUDE	Cloud models provided by Anthropic (Claude series)

Orchestration Categories

For orchestration purposes, Oblix groups these model types into two categories:

1. Local Models: OLLAMA models run locally on the device
2. Cloud Models: OPENAI and CLAUDE models run in the cloud

This distinction is important for orchestration decisions, as Oblix intelligently routes prompts between local and cloud models based on system state.

Usage

The ModelType enum is used primarily with the hook_model() method to specify which provider a model belongs to:

from oblix import OblixClient, ModelType

client = OblixClient()

# Hook a local model for orchestration
await client.hook_model(
    model_type=ModelType.OLLAMA,
    model_name="llama2"
)

# Hook a cloud model for orchestration
await client.hook_model(
    model_type=ModelType.OPENAI,
    model_name="gpt-3.5-turbo",
    api_key="your_openai_api_key"
)

# Hook another cloud model for orchestration
await client.hook_model(
    model_type=ModelType.CLAUDE,
    model_name="claude-3-opus-20240229",
    api_key="your_anthropic_api_key"
)

Orchestration Model Selection

When multiple models of different types are hooked, Oblix selects the appropriate model for execution based on:

1. Resource availability: CPU, memory, and GPU usage (from ResourceMonitor)
2. Connectivity status: Network latency, bandwidth, and reliability (from ConnectivityAgent)
3. Model capabilities: Features and performance characteristics of each model

You can override automatic selection by explicitly specifying a model ID in the format {model_type}:{model_name} when executing prompts.