metadata

license: cc-by-nc-4.0
tags:
  - merge
  - lazymergekit
dataset:
  - mlabonne/truthy-dpo-v0.1
  - mlabonne/distilabel-intel-orca-dpo-pairs
base_model:
  - mlabonne/Monarch-7B
language:
  - en

👑 NeuralMonarch-7B

Update 14/02/24: NeuralMonarch-7B is the new best-performing 7B model on Nous' benchmark suite! 🎉

NeuralMonarch-7B is a DPO fine-tuned of mlabonne/Monarch-7B using the jondurbin/truthy-dpo-v0.1 and argilla/distilabel-intel-orca-dpo-pairs preference datasets.

It is based on a merge of the following models using LazyMergekit:

Special thanks to Jon Durbin, Intel, and Argilla for the preference datasets.

🔍 Applications

This model uses a context window of 8k. It is compatible with the following chat templates (tested with LM Studio): Alpaca, ChatML, and Mistral Instruct.

Compared to other 7B models, it displays good performance in instruction following and reasoning tasks. It can also be used for RP and storytelling.

⚡ Quantized models

GGUF: https://huggingface.co/mlabonne/NeuralMonarch-7B-GGUF

🏆 Evaluation

The evaluation was performed using LLM AutoEval on Nous suite. See the entire leaderboard here.

Model	Average	AGIEval	GPT4All	TruthfulQA	Bigbench
NeuralMonarch-7B 📄	62.73	45.31	76.99	78.35	50.28
Monarch-7B 📄	62.68	45.48	77.07	78.04	50.14
teknium/OpenHermes-2.5-Mistral-7B 📄	52.42	42.75	72.99	52.99	40.94
mlabonne/NeuralHermes-2.5-Mistral-7B 📄	53.51	43.67	73.24	55.37	41.76
mlabonne/NeuralBeagle14-7B 📄	60.25	46.06	76.77	70.32	47.86
eren23/dpo-binarized-NeuralTrix-7B 📄	62.5	44.57	76.34	79.81	49.27
CultriX/NeuralTrix-7B-dpo 📄	62.5	44.61	76.33	79.8	49.24

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "mlabonne/NeuralMonarch-7B"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

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"])