Gemma
Collection
Gemma models finetuned to improve performance in terms of code generation
•
4 items
•
Updated
Gemmixtral is a merge of the following models using mergekit:
models:
- model: unsloth/gemma-2b-it-bnb-4bit
# no parameters necessary for base model
- model: akameswa/gemma2b_code_Javascript_4bit
parameters:
density: 0.25
weight: 0.25
- model: akameswa/gemma2b_code_python_4bit
parameters:
density: 0.25
weight: 0.25
- model: akameswa/gemma2b_code_java_4bit
parameters:
density: 0.25
weight: 0.25
- model: akameswa/gemma2b_code_cpp_4bit
parameters:
density: 0.25
weight: 0.25
merge_method: ties
base_model: unsloth/gemma-2b-it-bnb-4bit
parameters:
normalize: true
dtype: float16