README.md

metadata

library_name: transformers
tags:
  - mergekit
  - merge
base_model:
  - Qwen/Qwen2.5-7B-Instruct-1M
  - bunnycore/Qwen-2.5-7B-1M-RRP-v1-lora
model-index:
  - name: Qwen2.5-7B-RRP-1M
    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: 74.81
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=bunnycore/Qwen2.5-7B-RRP-1M
          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: 35.65
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=bunnycore/Qwen2.5-7B-RRP-1M
          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: 28.17
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=bunnycore/Qwen2.5-7B-RRP-1M
          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: 7.05
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=bunnycore/Qwen2.5-7B-RRP-1M
          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: 15.8
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=bunnycore/Qwen2.5-7B-RRP-1M
          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: 36.29
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=bunnycore/Qwen2.5-7B-RRP-1M
          name: Open LLM Leaderboard

LoRA trained on a thinking/reasoning and roleplaying dataset and then merged with the Qwen2.5-7B-Instruct-1M model, which supports up to 1 million token context lengths.

What this Model Can Do:

• Roleplay: Engage in creative conversations and storytelling!
• Reasoning: Tackle problems and answer your questions in a logical way (thanks to the LoRA layer).
• Thinking: Use the <think> tag in your system prompts to activate the model's thinking abilities.

Merge Method

This model was merged using the Passthrough merge method using Qwen/Qwen2.5-7B-Instruct-1M + bunnycore/Qwen-2.5-7B-1M-RRP-v1-lora 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: Qwen/Qwen2.5-7B-Instruct-1M+bunnycore/Qwen-2.5-7B-1M-RRP-v1-lora
dtype: bfloat16
merge_method: passthrough
models:
  - model: Qwen/Qwen2.5-7B-Instruct-1M+bunnycore/Qwen-2.5-7B-1M-RRP-v1-lora
tokenizer_source: Qwen/Qwen2.5-7B-Instruct-1M

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	32.96
IFEval (0-Shot)	74.81
BBH (3-Shot)	35.65
MATH Lvl 5 (4-Shot)	28.17
GPQA (0-shot)	7.05
MuSR (0-shot)	15.80
MMLU-PRO (5-shot)	36.29