Noro-Hermes-3x7B Open-Source AI Model - Free for Intelligent Assistant, Role-Playing, and Task Processing

Noro Hermes 3x7B

Developed by ThomasComics

Noro-Hermes-3x7B is a Mixture of Experts (MoE) model built using the lazy merge toolkit, combining three 7B-parameter Mistral variant models with capabilities in intelligent assistance, creative role-playing, and general task processing.

Large Language Model

Transformers

Open Source License:Apache-2.0 #Multi-expert fusion #Instruction optimization #Role-play support

Downloads 16

Release Time : 3/27/2024

Model Overview

This model integrates three expert models—Nous-Hermes-2-Mistral-7B-DPO, Noromaid-7B-0.4-DPO, and Mistral-7B-Instruct-v0.2—to achieve versatile processing capabilities suitable for text generation tasks across different scenarios.

Model Features

Mixture of Experts architecture

Integrates three specialized 7B-parameter models, dynamically selecting the most suitable expert for processing inputs via a gating mechanism

Versatile processing

Combines intelligent Q&A, creative writing, and general instruction-following capabilities

Efficient inference

Optimized based on the Mistral architecture, supporting 4-bit quantized inference

Model Capabilities

Intelligent Q&A

Creative writing

Role-playing

Instruction following

Text generation

Use Cases

Intelligent assistant

Knowledge Q&A

Answers various knowledge-based questions

Accuracy close to the original Nous-Hermes model

Creative writing

Story generation

Generates creative stories and character dialogues based on prompts

Exhibits the creative traits of the Noromaid model

General tasks

Instruction execution

Handles various daily task instructions

Maintains the robustness of the base Mistral model

🚀 Noro-Hermes-3x7B

Noro-Hermes-3x7B is a Mixture of Experts (MoE) model created by merging multiple base models using LazyMergekit, offering diverse capabilities in text generation.

🚀 Quick Start

Noro-Hermes-3x7B is a Mixture of Experts (MoE) made with the following models using LazyMergekit:

✨ Features

This model combines the strengths of multiple base models:

NousResearch/Nous-Hermes-2-Mistral-7B-DPO: An intelligent assistant.
NeverSleep/Noromaid-7B-0.4-DPO: A creative role - playing assistant.
mistralai/Mistral-7B-Instruct-v0.2: A general - purpose assistant.

📦 Installation

To use this model, you need to install the necessary libraries:

!pip install -qU transformers bitsandbytes accelerate

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer
import transformers
import torch

model = "ThomasComics/Noro-Hermes-3x7B"

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: mistralai/Mistral-7B-Instruct-v0.2
dtype: bfloat16
gate_mode: hidden
experts:
  - source_model: NousResearch/Nous-Hermes-2-Mistral-7B-DPO
    positive_prompts: ["You are an intelligent assistant."]
  - source_model: NeverSleep/Noromaid-7B-0.4-DPO
    positive_prompts: ["You are a creative roleplaying assistant."]
  - source_model: mistralai/Mistral-7B-Instruct-v0.2
    positive_prompts: ["You are a general purpose assistant."]

📄 License

This project is licensed under the Apache - 2.0 license.

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