README.md

metadata

license: apache-2.0
library_name: transformers
tags:
  - mergekit
  - merge
base_model:
  - unsloth/Qwen2.5-3B-Instruct
  - bunnycore/Qwen-2.5-3b-R1-lora_model-v.1
model-index:
  - name: Q2.5-R1-3B
    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: 42.14
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/Q2.5-R1-3B
          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: 27.2
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/Q2.5-R1-3B
          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: 26.74
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/Q2.5-R1-3B
          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: 7.94
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/Q2.5-R1-3B
          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: 12.73
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/Q2.5-R1-3B
          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: 31.26
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Triangle104/Q2.5-R1-3B
          name: Open LLM Leaderboard

Merge

Added chat template and eos_token to original model

This is a merge of pre-trained language models created using mergekit.

Merge Details

Qwen2.5 Instruct tied to Reasoning LORA.

Merge Method

This model was merged using the Passthrough merge method using unsloth/Qwen2.5-3B-Instruct + bunnycore/Qwen-2.5-3b-R1-lora_model-v.1 as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

base_model: unsloth/Qwen2.5-3B-Instruct+bunnycore/Qwen-2.5-3b-R1-lora_model-v.1
dtype: bfloat16
merge_method: passthrough
models:
  - model: unsloth/Qwen2.5-3B-Instruct+bunnycore/Qwen-2.5-3b-R1-lora_model-v.1
tokenizer_source: unsloth/Qwen2.5-3B

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	24.67
IFEval (0-Shot)	42.14
BBH (3-Shot)	27.20
MATH Lvl 5 (4-Shot)	26.74
GPQA (0-shot)	7.94
MuSR (0-shot)	12.73
MMLU-PRO (5-shot)	31.26