Documentation

Learn how to use Nexa Forge to orchestrate your AI workflows.

API Reference

Complete reference for the Nexa Forge REST API. Endpoints for jobs, workers, and artifacts.

Python SDK

Official Python client for interacting with Nexa Forge programmatically.

Quickstart

1. Install the SDK

pip install nexa-forge

2. Initialize Client

from nexa_forge import NexaForgeClient

client = NexaForgeClient(api_key="YOUR_API_KEY")

Common Workflows

Data Generation

Generate synthetic data for a specific domain.

job = client.generate(
    domain="medical_imaging",
    num_samples=1000,
    params={"resolution": "1024x1024"}
)
print(f"Job ID: {job['job_id']}")

Model Distillation

Distill a large teacher model into a smaller student model.

job = client.distill(
    teacher_model="gpt-4",
    student_model="llama-3-8b",
    dataset_uri="s3://my-bucket/dataset.parquet"
)

Evaluation

Run benchmarks on a trained model.

job = client.evaluate(
    model_id="my-finetuned-model-v1",
    benchmark="mmlu"
)