[FEATURE] Model Construct

[chandralegend]

Describe the idea
Model construct allows nice way of initializing models in your jac program

Traditionally, You have to have a class that specifies the Model loading and forward function. This will bring methods specifically designed for inference __infer__ , __train_step__, __train__ etc.

For the Case of With LLM,

import:plugin from "@jaclang/llm", OpenAI;

model llm: OpenAI {
    has temperature = 0.7;
}

can genai_ability(input: str) -> int by llm(reason=True);

will translate into

# Import the OpenAI Base Class

llm = _Jac.Model(OpenAI(temperature=0.7))

# genai_ability uses llm.__infer__