🚀 GOAT-70B-Storytelling Model

The GOAT-70B-Storytelling model is trained by the GOAT.AI lab as the core model for an autonomous story - writing agent. It can generate high - quality, cohesive, and captivating narratives, such as books, novels, and movie scripts.

🚀 Quick Start

Self - Hosted via transformers

import torch

from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "GOAT-AI/GOAT-70B-Storytelling"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16
)

Using with GOAT - Storytelling - Agent

from goat_storytelling_agent.storytelling_agent import StoryAgent

backend_uri = # Text generation endpoint
writer = StoryAgent(backend_uri, form='novel')
novel_scenes = writer.generate_story('treasure hunt in a jungle')

✨ Features

• High - Quality Narrative Generation: Generate books, novels, movie scripts, etc. with high quality, cohesion, and captivation.
• Flexible Input Utilization: Utilize inputs like plot outlines, character profiles, and their interrelationships to generate stories.

📦 Installation

The README does not provide specific installation steps, so this section is skipped.

💻 Usage Examples

Basic Usage

import torch

from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "GOAT-AI/GOAT-70B-Storytelling"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16
)

Advanced Usage

from goat_storytelling_agent.storytelling_agent import StoryAgent

backend_uri = # Text generation endpoint
writer = StoryAgent(backend_uri, form='novel')
novel_scenes = writer.generate_story('treasure hunt in a jungle')

📚 Documentation

Model Description

Training Details

Training was performed on a GPU cluster of 64xH100s. FSDP ZeRO - 3 sharding is employed for efficient training. We instruction finetune on a dataset of 18K examples for one epoch with batch size of 336, AdamW optimizer with learning rate 1e - 5.

Learn More

• Blogpost: GOAT - Storytelling: Arbitrarily Long Story Writing Agent
• GitHub: [here](https://github.com/GOAT - AI - lab/GOAT - Storytelling - Agent)
• Generated examples: [here](https://huggingface.co/datasets/GOAT - AI/generated - novels/tree/main/generated - books)

Uses

The main purpose of GOAT - 70B - Storytelling is to generate books, novels, movie scripts, etc. as an agent in coping with our GOAT - Storytelling - Agent. It is specifically designed for storywriters.

🔧 Technical Details

Training was carried out on a GPU cluster of 64xH100s. To ensure efficient training, FSDP ZeRO - 3 sharding was adopted. The model was instruction finetuned on a dataset of 18K examples for one epoch, with a batch size of 336. The AdamW optimizer was used with a learning rate of 1e - 5.

📄 License

GOAT - 70B - Storytelling model is based on [Meta's LLaMA - 2 - 70b - hf](https://huggingface.co/meta - llama/Llama - 2 - 70b - hf), and uses its own datasets. The model weights are available under the LLAMA - 2 license.

⚠️ Important Note

The GOAT - 70B - Storytelling model can produce factually incorrect output and should not be relied on to deliver factually accurate information. Therefore, the model could possibly generate wrong, biased, or otherwise offensive outputs.

Open LLM Leaderboard Evaluation Results

Detailed results can be found [here](https://huggingface.co/datasets/open - llm - leaderboard/details_GOAT - AI__GOAT - 70B - Storytelling)