metadata

base_model: []
library_name: transformers
tags:
  - mergekit
  - merge

1226

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

Merge Details

Merge Method

This model was merged using the linear merge method.

Models Merged

The following models were included in the merge:

runs/qwen2.5-1.5b-step3-general
runs/qwen2.5-1.5b-step3-extend
runs/qwen2.5-1.5b-step3-json-v2
runs/qwen2.5-1.5b-step3-json-no-schema
runs/qwen2.5-1.5b-step3-verify
runs/qwen2.5-1.5b-step3-markdown-v2
runs/qwen2.5-1.5b-step3-qa-json
runs/qwen2.5-1.5b-step3-json
runs/qwen2.5-1.5b-step3-qa-markdown
runs/qwen2.5-1.5b-step3-contentaware-extract
runs/qwen2.5-1.5b-step3-markdown
runs/qwen2.5-1.5b-step3-refine

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: runs/qwen2.5-1.5b-step3-markdown
    parameters:
      weight: 0.2
  - model: runs/qwen2.5-1.5b-step3-markdown-v2
    parameters:
      weight: 0.4
  - model: runs/qwen2.5-1.5b-step3-qa-markdown
    parameters:
      weight: 0.25
  - model: runs/qwen2.5-1.5b-step3-contentaware-extract
    parameters:
      weight: 0.25
  - model: runs/qwen2.5-1.5b-step3-json
    parameters:
      weight: 0.15
  - model: runs/qwen2.5-1.5b-step3-json-v2
    parameters:
      weight: 0.30
  - model: runs/qwen2.5-1.5b-step3-json-no-schema
    parameters:
      weight: 0.25
  - model: runs/qwen2.5-1.5b-step3-qa-json
    parameters:
      weight: 0.25
  - model: runs/qwen2.5-1.5b-step3-general
    parameters:
      weight: 0.15
  - model: runs/qwen2.5-1.5b-step3-verify
    parameters:
      weight: 0.25
  - model: runs/qwen2.5-1.5b-step3-refine
    parameters:
      weight: 0.25
  - model: runs/qwen2.5-1.5b-step3-extend
    parameters:
      weight: 0.35
merge_method: linear
dtype: bfloat16