TheTsar1209/qwen-carpmuscle-r-v0.3

A Fishy Model

qwen-carpmuscle-r-v0.3 was made using Rombodawg's Shared Continuous Finetuning method.

qwen-carpmuscle-v0.3 was made using Unsloth's continuous pretraining on the ChatML format with 24k context on (unsloth/Qwen2.5-14B-bnb-4bit)[https://huggingface.co/unsloth/Qwen2.5-14B-bnb-4bit]. Then qwen-carpmuscle-v0.3 was merged with Qwen/Qwen2.5-14B-Instruct and Qwen/Qwen2.5-14B using TIES to create this model.

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

Configuration

The following YAML configuration was used to produce this model:

models:
    - model: qwen-carpmuscle-v0.3
      parameters:
        weight: 1
        density: 1
    - model: Qwen/Qwen2.5-14B-Instruct
      parameters:
        weight: 1
        density: 1
merge_method: ties
base_model: Qwen/Qwen2.5-14B
parameters:
    weight: 1
    density: 1
    normalize: true
    int8_mask: true
tokenizer_source: qwen-carpmuscle-v0.3
dtype: bfloat16

• Developed by: TheTsar1209
• License: apache-2.0
• Finetuned from model : unsloth/Qwen2.5-14B-bnb-4bit

This qwen2 model was trained 2x faster with Unsloth and Huggingface's TRL library.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	31.52
IFEval (0-Shot)	44.55
BBH (3-Shot)	46.38
MATH Lvl 5 (4-Shot)	27.19
GPQA (0-shot)	13.42
MuSR (0-shot)	12.00
MMLU-PRO (5-shot)	45.59