jsfs11's picture
Adding Evaluation Results (#1)
0b56353 verified
metadata
license: apache-2.0
tags:
  - merge
  - mergekit
  - lazymergekit
  - Kukedlc/NeuTrixOmniBe-7B-model-remix
  - PetroGPT/WestSeverus-7B-DPO
  - vanillaOVO/supermario_v4
base_model:
  - Kukedlc/NeuTrixOmniBe-7B-model-remix
  - PetroGPT/WestSeverus-7B-DPO
  - vanillaOVO/supermario_v4
model-index:
  - name: MoEv4Config-TestWeightedTIES-7b
    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: 71.59
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=jsfs11/MoEv4Config-TestWeightedTIES-7b
          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: 88.19
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=jsfs11/MoEv4Config-TestWeightedTIES-7b
          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: 65.07
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=jsfs11/MoEv4Config-TestWeightedTIES-7b
          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: 70.87
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=jsfs11/MoEv4Config-TestWeightedTIES-7b
          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: 83.82
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=jsfs11/MoEv4Config-TestWeightedTIES-7b
          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: 72.78
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=jsfs11/MoEv4Config-TestWeightedTIES-7b
          name: Open LLM Leaderboard

MoEv4Config-TestWeightedTIES-7b

MoEv4Config-TestWeightedTIES-7b is a merge of the following models using LazyMergekit:

🧩 Configuration

models:
  - model: Kukedlc/NeuTrixOmniBe-7B-model-remix
    # No parameters necessary for base model
  - model: Kukedlc/NeuTrixOmniBe-7B-model-remix
    parameters:
      density: [1, 0.7, 0.1]
      weight: [0, 0.3, 0.7, 1]
  - model: PetroGPT/WestSeverus-7B-DPO
    parameters:
      density: [1, 0.7, 0.3]
      weight: [0, 0.25, 0.5, 1]
  - model: vanillaOVO/supermario_v4
    parameters:
      density: 0.33
      weight:
        - filter: mlp
          value: 0.5
        - value: 0
merge_method: ties
base_model: Kukedlc/NeuTrixOmniBe-7B-model-remix
parameters:
  int8_mask: true
  normalize: true
  sparsify:
    - filter: mlp
      value: 0.5
    - filter: self_attn
      value: 0.5
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "jsfs11/MoEv4Config-TestWeightedTIES-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. 75.39
AI2 Reasoning Challenge (25-Shot) 71.59
HellaSwag (10-Shot) 88.19
MMLU (5-Shot) 65.07
TruthfulQA (0-shot) 70.87
Winogrande (5-shot) 83.82
GSM8k (5-shot) 72.78