GAD-1 Instruct

Overview

GAD-1 Instruct is an instruction-tuned version of the original GAD-1 model. This model is designed to follow natural language instructions more effectively, making it suitable for tasks like text generation, email drafting, summarization, and other instruction-based applications.

Unlike the original GAD-1, which focused on general text prediction, GAD-1 Instruct has been fine-tuned specifically to understand and respond to prompts in a more aligned and human-readable manner.

Features

  • Follows natural language instructions reliably.
  • Generates coherent, context-aware text.
  • Suitable for a variety of instruction-driven tasks.
  • Compatible with standard text generation APIs.

Usage

The model can be used via your preferred API or inference endpoint. Example usage for generating text:

import requests

url = "https://rnevo2016--gad-1-instruct-fastapi-app.modal.run/v1/chat/completions"
data = {
    "model": "gad-1-instruct",
    "messages": [{"role": "user", "content": "Greetings > Hello GAD-1!"}],
    "max_tokens": 64,
    "temperature": 0.8
}

resp = requests.post(url, json=data)
print(resp.json()["choices"][0]["message"]["content"])
Downloads last month
14
Safetensors
Model size
77.1M params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Raziel1234/GAD-1-77.1M-Instruct

Base model

Raziel1234/GAD-1
Finetuned
(2)
this model

Collection including Raziel1234/GAD-1-77.1M-Instruct