πŸ€– NeuralMaths-Experiment-7b πŸ€–
πŸ” Number One in GSM8K LeaderBoard! πŸ†

image/png

NeuralMaths-Experiment-7b is a merge of the following models using LazyMergekit:

🧩 Configuration

models:
  - model: Kukedlc/NeuralSirKrishna-7b
    # No parameters necessary for base model
  - model: WizardLM/WizardMath-7B-V1.1
    parameters:
      density: 0.66
      weight: 0.2
  - model: mlabonne/NeuralDaredevil-7B
    parameters:
      density: 0.55
      weight: 0.2
  - model: Kukedlc/Neural4gsm8k
    parameters:
      density: 0.55
      weight: 0.2
  - model: Eric111/Mayo
    parameters:
      density: 0.44
      weight: 0.2
  - model: Kukedlc/NeuralSirKrishna-7b
    parameters:
      density: 0.66
      weight: 0.2
merge_method: dare_ties
base_model: Kukedlc/NeuralSirKrishna-7b
parameters:
  int8_mask: true
dtype: bfloat16

🌳 Model Family Tree

image/png

πŸ’» Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Kukedlc/NeuralMaths-Experiment-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"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 73.95
AI2 Reasoning Challenge (25-Shot) 69.71
HellaSwag (10-Shot) 87.48
MMLU (5-Shot) 65.01
TruthfulQA (0-shot) 63.83
Winogrande (5-shot) 82.48
GSM8k (5-shot) 75.21
Downloads last month
90
Safetensors
Model size
7.24B params
Tensor type
BF16
Β·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for Kukedlc/NeuralMaths-Experiment-7b

Spaces using Kukedlc/NeuralMaths-Experiment-7b 6

Evaluation results