sometimesanotion/Qwenvergence-14B-v2-Prose

Notes

This is not intended for end users, but as a base for subsequent merges. This model was merged using the Model Stock merge method using Qwen/Qwen2.5-14B as a base.

Model_stock is most effective when integrating finetunes instead of other merges. Its quality decays when used with merged mods unless other steps (finetuning, well-tested evolutionary merges) are taken.

I use model_stock as a foundation and adjust from there. Because this is a bulky merge to have to repeat, I'm making it available to others.

Configuration

The following YAML configuration was used to produce this model:

name:                Qwenvergence-v2-Prose
merge_method:        model_stock
base_model:          Qwen/Qwen2.5-14B
tokenizer_source:    base
parameters:
  int8_mask:         true
  normalize:         true
  rescale:           false
models:
  - model:           EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2
  - model:           allura-org/TQ2.5-14B-Sugarquill-v1
  - model:           oxyapi/oxy-1-small
  - model:           sthenno-com/miscii-14b-1028
  - model:           underwoods/medius-erebus-magnum-14b
  - model:           v000000/Qwen2.5-Lumen-14B
  - model:           arcee-ai/Virtuoso-Small
  - model:           huihui-ai/Qwen2.5-14B-Instruct-abliterated-v2
dtype:               bfloat16
out_dtype:           bfloat16