README.md

metadata

license: apache-2.0
library_name: peft
tags:
  - code
  - instruct
  - code-llama
datasets:
  - cognitivecomputations/dolphin-coder
base_model: codellama/CodeLlama-7b-hf
model-index:
  - name: codellama_7b_DolphinCoder
    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: 41.98
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Zangs3011/codellama_7b_DolphinCoder
          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: 65.5
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Zangs3011/codellama_7b_DolphinCoder
          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: 38.11
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Zangs3011/codellama_7b_DolphinCoder
          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: 35.45
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Zangs3011/codellama_7b_DolphinCoder
          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: 63.61
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Zangs3011/codellama_7b_DolphinCoder
          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: 9.7
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Zangs3011/codellama_7b_DolphinCoder
          name: Open LLM Leaderboard

Finetuning Overview:

Model Used: codellama/CodeLlama-7b-hf

Dataset: cognitivecomputations/dolphin-coder

Dataset Insights:

Dolphin-Coder dataset – a high-quality collection of 100,000+ coding questions and responses. It's perfect for supervised fine-tuning (SFT), and teaching language models to improve on coding-based tasks.

Finetuning Details:

With the utilization of MonsterAPI's no-code LLM finetuner, this finetuning:

• Was achieved with great cost-effectiveness.
• Completed in a total duration of 15hr 31mins for 1 epochs using an A6000 48GB GPU.
• Costed $31.31 for the entire 1 epoch.

Hyperparameters & Additional Details:

• Epochs: 1
• Total Finetuning Cost: $31.31
• Model Path: codellama/CodeLlama-7b-hf
• Learning Rate: 0.0002
• Data Split: 100% train
• Gradient Accumulation Steps: 128
• lora r: 32
• lora alpha: 64

---

license: apache-2.0

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	42.39
AI2 Reasoning Challenge (25-Shot)	41.98
HellaSwag (10-Shot)	65.50
MMLU (5-Shot)	38.11
TruthfulQA (0-shot)	35.45
Winogrande (5-shot)	63.61
GSM8k (5-shot)	9.70