jaskier-7b-dpo-v6.1 / README.md
leaderboard-pr-bot's picture
Adding Evaluation Results
1895014 verified
metadata
language:
  - en
license: cc-by-4.0
library_name: transformers
tags:
  - llm
  - 7b
datasets:
  - jondurbin/truthy-dpo-v0.1
model-index:
  - name: jaskier-7b-dpo-v6.1
    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: 73.29
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=bardsai/jaskier-7b-dpo-v6.1
          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: 88.89
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=bardsai/jaskier-7b-dpo-v6.1
          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: 64.39
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=bardsai/jaskier-7b-dpo-v6.1
          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: 77.47
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=bardsai/jaskier-7b-dpo-v6.1
          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: 84.69
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=bardsai/jaskier-7b-dpo-v6.1
          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: 69.45
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=bardsai/jaskier-7b-dpo-v6.1
          name: Open LLM Leaderboard

Jaskier-7b-dpo-v5.6

Jaskier

This is work-in-progress model, may not be ready for production use

Model based on bardsai/jaskier-7b-dpo-v5.6 (downstream version of Mistral7B) finetuned using Direct Preference Optimization on argilla/distilabel-math-preference-dpo.

How to use

You can use this model directly with a Hugging Face pipeline:


from transformers import pipeline, Conversation
import torch

base_model_name = "bardsai/jaskier-7b-dpo-v6.1"
chatbot = pipeline("conversational", model=base_model_name, torch_dtype=torch.float16, device_map="auto")
conversation = Conversation("Can Poland into space?")
conversation = chatbot(conversation)
print(conversation.messages[-1]["content"])

Output

"Poland, as a nation, doesn't physically travel to space. However, Poland has contributed to the field of space exploration through its scientists, engineers, and collaborations with international space agencies. The Polish Space Agency, established in 2016, aims to promote and coordinate the country's space activities."

Changelog

  • 2024-02-20: Initial release

About bards.ai

At bards.ai, we focus on providing machine learning expertise and skills to our partners, particularly in the areas of nlp, machine vision and time series analysis. Our team is located in Wroclaw, Poland. Please visit our website for more information: bards.ai

Let us know if you use our model :). Also, if you need any help, feel free to contact us at info@bards.ai

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 76.36
AI2 Reasoning Challenge (25-Shot) 73.29
HellaSwag (10-Shot) 88.89
MMLU (5-Shot) 64.39
TruthfulQA (0-shot) 77.47
Winogrande (5-shot) 84.69
GSM8k (5-shot) 69.45