KnutJaegersberg's picture
Adding Evaluation Results (#1)
68aa0bd verified
metadata
license: other
license_name: qwen
license_link: LICENSE
model-index:
  - name: Qwen-1_8B-Chat-llama
    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: 36.95
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=KnutJaegersberg/Qwen-1_8B-Chat-llama
          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: 54.34
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=KnutJaegersberg/Qwen-1_8B-Chat-llama
          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: 44.55
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=KnutJaegersberg/Qwen-1_8B-Chat-llama
          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: 43.7
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=KnutJaegersberg/Qwen-1_8B-Chat-llama
          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: 58.88
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=KnutJaegersberg/Qwen-1_8B-Chat-llama
          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: 19.26
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=KnutJaegersberg/Qwen-1_8B-Chat-llama
          name: Open LLM Leaderboard

Their non-commercial research license applies.

I used this script to make the model and used the tokenizer of CausalLM, as suggested in the comments of the script.

https://github.com/hiyouga/LLaMA-Factory/blob/main/tests/llamafy_qwen.py

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 42.94
AI2 Reasoning Challenge (25-Shot) 36.95
HellaSwag (10-Shot) 54.34
MMLU (5-Shot) 44.55
TruthfulQA (0-shot) 43.70
Winogrande (5-shot) 58.88
GSM8k (5-shot) 19.26