README.md

metadata

tags:
  - merge
  - mergekit
  - lazymergekit
  - datatab/Yugo45-GPT
  - FlexingD/yarn-mistral-7B-64k-instruct-alpaca-cleaned-origin
base_model:
  - datatab/YugoGPT-Alpaca-v1
  - FlexingD/yarn-mistral-7B-64k-instruct-alpaca-cleaned-origin
license: cc-by-4.0
datasets:
  - datatab/alpaca-cleaned-serbian-full
language:
  - sr

Yugo45-GPT *(7b)

This Yugo45-GPT (7b) model has been fine-tuned on the Alpaca dataset using the gordicaleksa/YugoGPT as the zero ground base model.

• Finetune performed by: datatab
• License: CC-BY-4.0
• Original model author: gordicaleksa/YugoGPT

Yugo45-GPT is a merge of the following models using LazyMergekit:

• datatab/YugoGPT-Alpaca-v1
• FlexingD/yarn-mistral-7B-64k-instruct-alpaca-cleaned-origin

📌 Note

Special thanks for idea Stopwolf and this X post @TheStopwolf

🧩 Configuration

slices:
  - sources:
      - model: datatab/YugoGPT-Alpaca-v1
        layer_range: [0, 32]
      - model: FlexingD/yarn-mistral-7B-64k-instruct-alpaca-cleaned-origin
        layer_range: [0, 32]
merge_method: slerp
base_model: datatab/YugoGPT-Alpaca-v1
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16

🏋🏼 Benchmarks

# TBD

💻 Usage

# TBD