merge
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the DARE TIES merge method using CultriX/Qwen2.5-14B-Wernickev3 as a base.
Models Merged
The following models were included in the merge:
- djuna/Q2.5-Veltha-14B-0.5
- CultriX/Qwen2.5-14B-Brocav7
- CultriX/SeQwence-14Bv1
- CultriX/Qwen2.5-14B-Broca
- qingy2019/Qwen2.5-Math-14B-Instruct
- CultriX/Qwen2.5-14B-Brocav3
- allknowingroger/QwenSlerp6-14B
Configuration
The following YAML configuration was used to produce this model:
merge_method: dare_ties
base_model: CultriX/Qwen2.5-14B-Wernickev3
dtype: bfloat16
parameters:
epsilon: 0.0075
lambda: 2.1
normalize: true
adaptive_merge_parameters:
task_weights:
tinyArc: 2.1
tinyHellaswag: 1.9
tinyMMLU: 2.1
tinyTruthfulQA: 2.5
tinyTruthfulQA_mc1: 2.4
tinyWinogrande: 2.4
IFEval: 2.4
BBH: 2.3
MATH: 2.6
GPQA: 2.4
MUSR: 2.5
MMLU-PRO: 2.2
smoothing_factor: 0.05
gradient_clipping:
CultriX/Qwen2.5-14B-Wernickev3: 0.82
djuna/Q2.5-Veltha-14B-0.5: 0.84
qingy2019/Qwen2.5-Math-14B-Instruct: 0.87
CultriX/SeQwence-14Bv1: 0.8
CultriX/Qwen2.5-14B-Broca: 0.85
CultriX/Qwen2.5-14B-Brocav3: 0.83
CultriX/Qwen2.5-14B-Brocav7: 0.84
allknowingroger/QwenSlerp6-14B: 0.8
models:
- model: CultriX/Qwen2.5-14B-Wernickev3
parameters:
weight: 0.21
density: 0.78
- model: djuna/Q2.5-Veltha-14B-0.5
parameters:
weight: 0.2
density: 0.76
- model: qingy2019/Qwen2.5-Math-14B-Instruct
parameters:
weight: 0.18
density: 0.75
- model: CultriX/SeQwence-14Bv1
parameters:
weight: 0.17
density: 0.74
- model: CultriX/Qwen2.5-14B-Broca
parameters:
weight: 0.17
density: 0.74
- model: CultriX/Qwen2.5-14B-Brocav3
parameters:
weight: 0.15
density: 0.73
- model: CultriX/Qwen2.5-14B-Brocav7
parameters:
weight: 0.14
density: 0.72
- model: allknowingroger/QwenSlerp6-14B
parameters:
weight: 0.13
density: 0.72
tokenizer_source: CultriX/Qwen2.5-14B-Wernickev3
- Downloads last month
- 49
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for CultriX/Qwen2.5-14B-partialmergept1
Merge model
this model