Qwen2.5 14B It Restore

🚀 Model Merging Project

This project combines three methods to merge instruction models with base models, aiming to solve the problems of declined instruction - following and mathematical abilities.

🚀 Quick Start

Combine the three methods of della, ties, and model stock to merge the instruction model with the base model. The goal is to address the issues of the decline in instruction - following ability and mathematical ability that occur when using only the ties merging method or only the della merging method.

💻 Usage Examples

Basic Usage

The following is an example of using the della method for model merging:

models:
  - model: Qwen/Qwen2.5-14B-instruct
    parameters:
      density: 1 
      weight: 1
      lambda: 0.9
merge_method: della
base_model: Qwen/Qwen2.5-14B
parameters:
  density: 1
  weight: 1
  lambda: 0.9
  normalize: true
  int8_mask: true
dtype: bfloat16
name: Qwen2.5-14B-della

Advanced Usage

Here are examples of using the ties and model stock methods for model merging:

Ties Method

models:
  - model: Qwen/Qwen2.5-14B-instruct
    parameters:
      density: 1 
      weight: 1
merge_method: ties
base_model: Qwen/Qwen2.5-14B
parameters:
  density: 1
  weight: 1
  normalize: true
  int8_mask: true
dtype: bfloat16
name: Qwen2.5-14B-ties

Model Stock Method

merge_method: model_stock
base_model: Qwen/Qwen2.5-14B-instruct
models:
  - model: Qwen2.5-14B-della
  - model: Qwen2.5-14B-ties
dtype: bfloat16
tokenizer_source: base
int8_mask: true
normalize: true
name: Qwen2.5-14B-it-restore

📄 License

This project is licensed under the Apache-2.0 license.

📚 Documentation