ToDo: Fill the card with more info.
merge
This is a merge of pre-trained language models created using mergekit.
Merge Details
It's a bit of a test merge to dip my toes into merging Gemma 2. Sadly, however, it seems like 8B is my PC's tolerable limit before performance becomes painstakingly and infuriatingly slow, so after this, I might have to sit out on Gemma 2
Merge Method
This model was merged using the Model Stock merge method using Casual-Autopsy/Gemma-Rad-RP as a base.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
models:
- model: crestf411/gemma2-9B-sunfall-v0.5.2
- model: crestf411/gemma2-9B-daybreak-v0.5
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.5, 0.13, 0.5, 0.13, 0.3]
- model: crestf411/gemstone-9b
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.13, 0.5, 0.13, 0.5, 0.13]
merge_method: dare_ties
base_model: crestf411/gemma2-9B-sunfall-v0.5.2
parameters:
normalize: false
int8_mask: true
dtype: bfloat16
models:
- model: UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3
- model: nldemo/Gemma-9B-Summarizer-QLoRA
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.0625, 0.25, 0.0625, 0.25, 0.0625]
- model: SillyTilly/google-gemma-2-9b-it+rbojja/gemma2-9b-intent-lora-adapter
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.0625, 0.25, 0.0625, 0.25, 0.0625]
- model: nbeerbower/gemma2-gutenberg-9B
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.25, 0.0625, 0.25, 0.0625, 0.25]
merge_method: ties
base_model: UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3
parameters:
normalize: false
int8_mask: true
dtype: bfloat16
models:
- model: IlyaGusev/gemma-2-9b-it-abliterated
- model: TheDrummer/Smegmma-9B-v1
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.5, 0.13, 0.5, 0.13, 0.3]
- model: TheDrummer/Tiger-Gemma-9B-v1
parameters:
density: [0.7, 0.5, 0.3, 0.35, 0.65, 0.35, 0.75, 0.25, 0.75, 0.35, 0.65, 0.35, 0.3, 0.5, 0.7]
weight: [0.13, 0.5, 0.13, 0.5, 0.13]
merge_method: dare_ties
base_model: IlyaGusev/gemma-2-9b-it-abliterated
parameters:
normalize: false
int8_mask: true
dtype: bfloat16
models:
- model: Casual-Autopsy/Gemma-Rad-RP
- model: Casual-Autopsy/Gemma-Rad-Uncen
- model: Casual-Autopsy/Gemma-Rad-IQ
merge_method: model_stock
base_model: Casual-Autopsy/Gemma-Rad-RP
dtype: bfloat16