README.md

metadata

license: cc-by-nc-4.0
tags:
  - merge
  - mergekit
  - lazymergekit
  - weezywitasneezy/BenchmarkEngineering-7B-slerp
  - senseable/WestLake-7B-v2
base_model:
  - weezywitasneezy/BenchmarkEngineering-7B-slerp
  - senseable/WestLake-7B-v2
model-index:
  - name: BenchmarkEngineering-F2-7B-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: 73.46
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=weezywitasneezy/BenchmarkEngineering-F2-7B-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: 88.88
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=weezywitasneezy/BenchmarkEngineering-F2-7B-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: 64.5
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=weezywitasneezy/BenchmarkEngineering-F2-7B-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: 72.37
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=weezywitasneezy/BenchmarkEngineering-F2-7B-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: 86.11
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=weezywitasneezy/BenchmarkEngineering-F2-7B-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: 69.29
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=weezywitasneezy/BenchmarkEngineering-F2-7B-slerp
          name: Open LLM Leaderboard

BenchmarkEngineering-F2-7B-slerp

This merge seeks to further improve on the original BenchmarkEngineering by integrating the Westlake-7B-v2 model. It does boost the Winogrande score but at the cost of the other benchmarks.

BenchmarkEngineering-F2-7B-slerp is a merge of the following models using LazyMergekit:

• weezywitasneezy/BenchmarkEngineering-7B-slerp
• senseable/WestLake-7B-v2

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	75.77
AI2 Reasoning Challenge (25-Shot)	73.46
HellaSwag (10-Shot)	88.88
MMLU (5-Shot)	64.50
TruthfulQA (0-shot)	72.37
Winogrande (5-shot)	86.11
GSM8k (5-shot)	69.29

🧩 Configuration

slices:
  - sources:
      - model: weezywitasneezy/BenchmarkEngineering-7B-slerp
        layer_range: [0, 32]
      - model: senseable/WestLake-7B-v2
        layer_range: [0, 32]
merge_method: slerp
base_model: weezywitasneezy/BenchmarkEngineering-7B-slerp
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 = "weezywitasneezy/BenchmarkEngineering-F2-7B-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"])