Gille's picture
Adding Evaluation Results (#1)
979ddfc verified
metadata
license: apache-2.0
tags:
  - moe
  - frankenmoe
  - merge
  - mergekit
  - lazymergekit
  - Gille/StrangeMerges_9-7B-dare_ties
  - Gille/StrangeMerges_8-7B-slerp
base_model:
  - Gille/StrangeMerges_9-7B-dare_ties
  - Gille/StrangeMerges_8-7B-slerp
model-index:
  - name: MoE-StrangeMerges-2x7B
    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: 70.82
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Gille/MoE-StrangeMerges-2x7B
          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: 87.83
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Gille/MoE-StrangeMerges-2x7B
          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: 65.04
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Gille/MoE-StrangeMerges-2x7B
          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: 65.86
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Gille/MoE-StrangeMerges-2x7B
          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: 82.79
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Gille/MoE-StrangeMerges-2x7B
          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: 67.7
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Gille/MoE-StrangeMerges-2x7B
          name: Open LLM Leaderboard

MoE-StrangeMerges-2x7B

MoE-StrangeMerges-2x7B is a Mixure of Experts (MoE) made with the following models using LazyMergekit:

🧩 Configuration

base_model: Gille/StrangeMerges_9-7B-dare_ties
gate_mode: cheap_embed
dtype: float16
experts:
  - source_model: Gille/StrangeMerges_9-7B-dare_ties
    positive_prompts: ["science, logic, math"]
  - source_model: Gille/StrangeMerges_8-7B-slerp
    positive_prompts: ["reasoning, numbers, abstract"]

💻 Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Gille/MoE-StrangeMerges-2x7B"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
)

messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
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. 73.34
AI2 Reasoning Challenge (25-Shot) 70.82
HellaSwag (10-Shot) 87.83
MMLU (5-Shot) 65.04
TruthfulQA (0-shot) 65.86
Winogrande (5-shot) 82.79
GSM8k (5-shot) 67.70