README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - text-generation-inference
  - transformers
  - unsloth
  - llama
  - trl
base_model: unsloth/llama-3-8b-Instruct-bnb-4bit
datasets:
  - LimYeri/LeetCode_Python_Solutions
pipeline_tag: text-generation
model-index:
  - name: CodeMind-Llama3-8B-unsloth_v2-merged
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: IFEval (0-Shot)
          type: HuggingFaceH4/ifeval
          args:
            num_few_shot: 0
        metrics:
          - type: inst_level_strict_acc and prompt_level_strict_acc
            value: 69.46
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=LimYeri/CodeMind-Llama3-8B-unsloth_v2-merged
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: BBH (3-Shot)
          type: BBH
          args:
            num_few_shot: 3
        metrics:
          - type: acc_norm
            value: 26.66
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=LimYeri/CodeMind-Llama3-8B-unsloth_v2-merged
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MATH Lvl 5 (4-Shot)
          type: hendrycks/competition_math
          args:
            num_few_shot: 4
        metrics:
          - type: exact_match
            value: 5.74
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=LimYeri/CodeMind-Llama3-8B-unsloth_v2-merged
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GPQA (0-shot)
          type: Idavidrein/gpqa
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 2.01
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=LimYeri/CodeMind-Llama3-8B-unsloth_v2-merged
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MuSR (0-shot)
          type: TAUR-Lab/MuSR
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 2.22
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=LimYeri/CodeMind-Llama3-8B-unsloth_v2-merged
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU-PRO (5-shot)
          type: TIGER-Lab/MMLU-Pro
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 27.84
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=LimYeri/CodeMind-Llama3-8B-unsloth_v2-merged
          name: Open LLM Leaderboard

Uploaded model

• Developed by: LimYeri
• License: apache-2.0
• Finetuned from model : unsloth/llama-3-8b-Instruct-bnb-4bit

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

Training Setting

The following hyperparameters are used during SFT:

• num_epochs: 3
• learning_rate: 1e-4
• max_seq_length: None
• optimizer: adamw_8bit
• lr_scheduler_type: linear
• warmup_ratio: 0.03
• weight_decay: 0.01
• lora_rank: 16
• lora_alpha: 16
• lora_dropout: 0
• gradient_checkpointing: true
• fp16: not is_bfloat16_supported()
• bf16: is_bfloat16_supported()

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	22.32
IFEval (0-Shot)	69.46
BBH (3-Shot)	26.66
MATH Lvl 5 (4-Shot)	5.74
GPQA (0-shot)	2.01
MuSR (0-shot)	2.22
MMLU-PRO (5-shot)	27.84