NeuralStar_AlphaWriter_4x7b Open-source Language Model - Free Support for Creative Writing in Multiple Fields

Neuralstar AlphaWriter 4x7b

Developed by OmnicromsBrain

NeuralStar_AlphaWriter_4x7b is a 7B-parameter language model built on Mixture of Experts (MoE) technology, specifically optimized for creative writing tasks by integrating four expert models specialized in different writing domains.

Large Language Model

Transformers

Open Source License:Apache-2.0 #Creative Writing Expert #Mixture of Experts Fusion #Multitask Text Generation

Downloads 21

Release Time : 4/15/2024

Model Overview

This model merges four 7B models, each excelling in distinct writing areas such as creative writing, text revision, story scene generation, and plot structuring, demonstrating outstanding text generation capabilities on the Open LLM Leaderboard.

Model Features

Mixture of Experts Technology

Integrates four specialized models, each responsible for different writing domains, enabling more professional text generation.

Creative Writing Optimization

Specifically optimized for creative writing tasks such as novel creation and story scene generation.

Multi-Domain Collaboration

Combines expert capabilities in dialogue processing, text revision, story creation, and plot structuring.

Quantization Support

Offers various quantization versions like GGUF and iMatrix for easy deployment across different hardware environments.

Model Capabilities

Creative Text Generation

Story Scene Creation

Character Development

Text Grammar Checking

Dialogue Generation

Code Generation Assistance

Use Cases

Creative Writing

Novel Creation

Generates coherent story plots and character dialogues.

Can produce logical long-form story frameworks.

Screenplay Writing

Assists screenwriters in crafting scene dialogues and plot development.

Generates natural dialogues that align with character traits.

Content Editing

Text Revision

Checks and improves grammar and fluency of existing texts.

Provides professional-level text polishing suggestions.

🚀 NeuralStar_AlphaWriter_4x7b

NeuralStar_AlphaWriter_4x7b is a Mixture of Experts (MoE) model. It combines the strengths of multiple base models, aiming to provide excellent performance in text - generation tasks, especially in creative writing scenarios.

image/png

🚀 Quick Start

I was deeply impressed by the writing results of mlabonne/Beyonder - 4x7B - v3 when writing in NovelCrafter. Inspired by the [LLM Course](https://github.com/mlabonne/llm - course) and using LazyMergeKit, I created NeuralStar - AlphaWriter - 4x7b, wondering what a writing model would be like if all 4 “experts” were good at creative writing.

✨ Features

NeuralStar_AlphaWriter_4x7b is a Mixture of Experts (MoE) made with the following models using LazyMergekit:

[mlabonne/AlphaMonarch - 7B](https://huggingface.co/mlabonne/AlphaMonarch - 7B)
FPHam/Karen_TheEditor_V2_STRICT_Mistral_7B
[SanjiWatsuki/Kunoichi - DPO - v2 - 7B](https://huggingface.co/SanjiWatsuki/Kunoichi - DPO - v2 - 7B)
[OmnicromsBrain/NeuralStar - 7b - Lazy](https://huggingface.co/OmnicromsBrain/NeuralStar - 7b - Lazy)

📦 Installation

No specific installation steps are provided in the original README. If you want to use the model, you can refer to the usage code example below.

💻 Usage Examples

Basic Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "OmnicromsBrain/NeuralStar_AlphaWriter_4x7b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
)

messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])

📚 Documentation

🧩 Configuration

base_model: mlabonne/AlphaMonarch-7B
experts:  
  - source_model: mlabonne/AlphaMonarch-7B
    positive_prompts: 
    - "chat"
    - "assistant"
    - "tell me"
    - "explain"
    - "I want"
  - source_model: FPHam/Karen_TheEditor_V2_STRICT_Mistral_7B
    positive_prompts:
    - "edit"
    - "rewrite"
    - "evaluate"
    - "spelling"
    - "grammer"
  - source_model: SanjiWatsuki/Kunoichi-DPO-v2-7B
    positive_prompts:
    - "storywriting"
    - "write"
    - "scene"
    - "prose"
    - "character"
  - source_model: OmnicromsBrain/NeuralStar-7b-Lazy
    positive_prompts:
    - "codex"
    - "plot"
    - "outline"
    - "scenebeat"
    - "count"

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Property	Details
Model Type	Mixture of Experts (MoE)
Base Models	mlabonne/AlphaMonarch - 7B, FPHam/Karen_TheEditor_V2_STRICT_Mistral_7B, SanjiWatsuki/Kunoichi - DPO - v2 - 7B, OmnicromsBrain/NeuralStar - 7b - Lazy

Metric	Value
Avg.	73.31
AI2 Reasoning Challenge (25 - Shot)	70.22
HellaSwag (10 - Shot)	88.31
MMLU (5 - Shot)	64.60
TruthfulQA (0 - shot)	71.70
Winogrande (5 - shot)	82.00
GSM8k (5 - shot)	63.00

📄 License

This project is licensed under the Apache - 2.0 license.

⚡ Quantized Models

Special thanks to MRadermacher for the Static and iMatrx quantized models

.GGUF https://huggingface.co/mradermacher/NeuralStar_AlphaWriter_4x7b - GGUF

iMatrix https://huggingface.co/mradermacher/NeuralStar_AlphaWriter_4x7b - i1 - GGUF

Q4_K_M and Q5_K_M .gguf [Here](https://huggingface.co/OmnicromsBrain/NeuralStar_AlphaWriter_4x7b - GGUF) created with mlabonne/Autogguf

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご