Quantization made by Richard Erkhov.

NeuralPipe-7B-slerp - GGUF

Model creator: https://huggingface.co/mlabonne/
Original model: https://huggingface.co/mlabonne/NeuralPipe-7B-slerp/

Name	Quant method	Size
NeuralPipe-7B-slerp.Q2_K.gguf	Q2_K	2.53GB
NeuralPipe-7B-slerp.IQ3_XS.gguf	IQ3_XS	2.81GB
NeuralPipe-7B-slerp.IQ3_S.gguf	IQ3_S	2.96GB
NeuralPipe-7B-slerp.Q3_K_S.gguf	Q3_K_S	2.95GB
NeuralPipe-7B-slerp.IQ3_M.gguf	IQ3_M	3.06GB
NeuralPipe-7B-slerp.Q3_K.gguf	Q3_K	3.28GB
NeuralPipe-7B-slerp.Q3_K_M.gguf	Q3_K_M	3.28GB
NeuralPipe-7B-slerp.Q3_K_L.gguf	Q3_K_L	3.56GB
NeuralPipe-7B-slerp.IQ4_XS.gguf	IQ4_XS	3.67GB
NeuralPipe-7B-slerp.Q4_0.gguf	Q4_0	3.83GB
NeuralPipe-7B-slerp.IQ4_NL.gguf	IQ4_NL	3.87GB
NeuralPipe-7B-slerp.Q4_K_S.gguf	Q4_K_S	3.86GB
NeuralPipe-7B-slerp.Q4_K.gguf	Q4_K	4.07GB
NeuralPipe-7B-slerp.Q4_K_M.gguf	Q4_K_M	4.07GB
NeuralPipe-7B-slerp.Q4_1.gguf	Q4_1	4.24GB
NeuralPipe-7B-slerp.Q5_0.gguf	Q5_0	4.65GB
NeuralPipe-7B-slerp.Q5_K_S.gguf	Q5_K_S	4.65GB
NeuralPipe-7B-slerp.Q5_K.gguf	Q5_K	4.78GB
NeuralPipe-7B-slerp.Q5_K_M.gguf	Q5_K_M	4.78GB
NeuralPipe-7B-slerp.Q5_1.gguf	Q5_1	5.07GB
NeuralPipe-7B-slerp.Q6_K.gguf	Q6_K	5.53GB
NeuralPipe-7B-slerp.Q8_0.gguf	Q8_0	7.17GB

Original model description:

license: apache-2.0 tags: - merge - mergekit model-index: - name: NeuralPipe-7B-slerp results: - task: type: text-generation name: Text Generation dataset: name: AI2 Reasoning Challenge (25-Shot) type: ai2_arc config: ARC-Challenge split: test args: num_few_shot: 25 metrics: - type: acc_norm value: 67.75 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=mlabonne/NeuralPipe-7B-slerp name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: HellaSwag (10-Shot) type: hellaswag split: validation args: num_few_shot: 10 metrics: - type: acc_norm value: 86.15 name: normalized accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=mlabonne/NeuralPipe-7B-slerp name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MMLU (5-Shot) type: cais/mmlu config: all split: test args: num_few_shot: 5 metrics: - type: acc value: 63.94 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=mlabonne/NeuralPipe-7B-slerp name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: TruthfulQA (0-shot) type: truthful_qa config: multiple_choice split: validation args: num_few_shot: 0 metrics: - type: mc2 value: 59.8 source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=mlabonne/NeuralPipe-7B-slerp name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: Winogrande (5-shot) type: winogrande config: winogrande_xl split: validation args: num_few_shot: 5 metrics: - type: acc value: 79.64 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=mlabonne/NeuralPipe-7B-slerp name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: GSM8k (5-shot) type: gsm8k config: main split: test args: num_few_shot: 5 metrics: - type: acc value: 69.75 name: accuracy source: url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=mlabonne/NeuralPipe-7B-slerp name: Open LLM Leaderboard

NeuralPipe-7B

This model is a merge of the following models made with mergekit:

⚡ Quantized models

Thanks to TheBloke for the quantized models:

GGUF: https://huggingface.co/TheBloke/NeuralPipe-7B-slerp-GGUF
AWQ: https://huggingface.co/TheBloke/NeuralPipe-7B-slerp-AWQ
GPTQ: https://huggingface.co/TheBloke/NeuralPipe-7B-slerp-GPTQ

🧩 Configuration

slices:
  - sources:
      - model: OpenPipe/mistral-ft-optimized-1218
        layer_range: [0, 32]
      - model: mlabonne/NeuralHermes-2.5-Mistral-7B
        layer_range: [0, 32]
merge_method: slerp
base_model: OpenPipe/mistral-ft-optimized-1218
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "mlabonne/NeuralPipe-7B-slerp"
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"])

Output:

A large language model is an AI system that uses deep learning techniques to process and understand vast amounts of natural language data. It is designed to generate human-like text, perform complex language tasks, and understand the context, nuance, and meaning of textual data. These models are trained on large datasets, often including billions of words, to learn the patterns and relationships in language. As a result, they can generate coherent and contextually relevant text, answer questions, and perform a variety of other language-related tasks. Some well-known large language models include OpenAI's GPT-3, Google's BERT, and Facebook's RoBERTa.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	71.17
AI2 Reasoning Challenge (25-Shot)	67.75
HellaSwag (10-Shot)	86.15
MMLU (5-Shot)	63.94
TruthfulQA (0-shot)	59.80
Winogrande (5-shot)	79.64
GSM8k (5-shot)	69.75