Orca-SOLAR-4x10.7b / README.md
macadeliccc's picture
Adding Evaluation Results (#2)
f3f6964 verified
metadata
language:
  - en
license: apache-2.0
library_name: transformers
tags:
  - code
datasets:
  - Intel/orca_dpo_pairs
model-index:
  - name: Orca-SOLAR-4x10.7b
    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: 68.52
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/Orca-SOLAR-4x10.7b
          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: 86.78
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/Orca-SOLAR-4x10.7b
          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: 67.03
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/Orca-SOLAR-4x10.7b
          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: 64.54
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/Orca-SOLAR-4x10.7b
          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: 83.9
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/Orca-SOLAR-4x10.7b
          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: 68.23
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=macadeliccc/Orca-SOLAR-4x10.7b
          name: Open LLM Leaderboard

πŸŒžπŸš€ Orca-SOLAR-4x10.7_36B

Merge of four Solar-10.7B instruct finetunes.

solar

🌟 Usage

This SOLAR model loves to code. In my experience, if you ask it a code question it will use almost all of the available token limit to complete the code.

However, this can also be to its own detriment. If the request is complex it may not finish the code in a given time period. This behavior is not because of an eos token, as it finishes sentences quite normally if its a non code question.

Your mileage may vary.

🌎 HF Spaces

This 36B parameter model is capabale of running on free tier hardware (CPU only - GGUF)

  • Try the model here

πŸŒ… Code Example

Example also available in colab

from transformers import AutoModelForCausalLM, AutoTokenizer

def generate_response(prompt):
    """
    Generate a response from the model based on the input prompt.

    Args:
    prompt (str): Prompt for the model.

    Returns:
    str: The generated response from the model.
    """
    # Tokenize the input prompt
    inputs = tokenizer(prompt, return_tensors="pt")
    
    # Generate output tokens
    outputs = model.generate(**inputs, max_new_tokens=512, eos_token_id=tokenizer.eos_token_id, pad_token_id=tokenizer.pad_token_id)

    # Decode the generated tokens to a string
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)

    return response


# Load the model and tokenizer
model_id = "macadeliccc/Orca-SOLAR-4x10.7b"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, load_in_4bit=True)

prompt = "Explain the proof of Fermat's Last Theorem and its implications in number theory."


print("Response:")
print(generate_response(prompt), "\n")

Llama.cpp

GGUF Quants available here

llama.cpp-screenshot

Evaluations

https://huggingface.co/datasets/open-llm-leaderboard/details_macadeliccc__Orca-SOLAR-4x10.7b

πŸ“š Citations

@misc{kim2023solar,
      title={SOLAR 10.7B: Scaling Large Language Models with Simple yet Effective Depth Up-Scaling}, 
      author={Dahyun Kim and Chanjun Park and Sanghoon Kim and Wonsung Lee and Wonho Song and Yunsu Kim and Hyeonwoo Kim and Yungi Kim and Hyeonju Lee and Jihoo Kim and Changbae Ahn and Seonghoon Yang and Sukyung Lee and Hyunbyung Park and Gyoungjin Gim and Mikyoung Cha and Hwalsuk Lee and Sunghun Kim},
      year={2023},
      eprint={2312.15166},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 73.17
AI2 Reasoning Challenge (25-Shot) 68.52
HellaSwag (10-Shot) 86.78
MMLU (5-Shot) 67.03
TruthfulQA (0-shot) 64.54
Winogrande (5-shot) 83.90
GSM8k (5-shot) 68.23