gugarosa / mh_fine_tuning

📄 Official implementation regarding the paper "Improving Pre-Trained Weights Through Meta-Heuristic Fine-Tuning".

Improving Pre-Trained Weights Through Meta-Heuristic Fine-Tuning

This repository holds all the necessary code to run the very-same experiments described in the paper "Improving Pre-Trained Weights Through Meta-Heuristic Fine-Tuning".

---

References

If you use our work to fulfill any of your needs, please cite us:

@inproceedings{deRosa:21,
  author={De Rosa, Gustavo H. and Roder, Mateus and Papa, João Paulo and Dos Santos, Claudio F.G.},
  booktitle={2021 IEEE Symposium Series on Computational Intelligence (SSCI)}, 
  title={Improving Pre- Trained Weights through Meta - Heuristics Fine- Tuning}, 
  year={2021},
  volume={},
  number={},
  pages={1-8},
  doi={10.1109/SSCI50451.2021.9659945}
}

---

Structure

• core
  • model.py: Defines the base Machine Learning architecture;
• models
  • cnn.py: Defines the Residual Network (ResNet18);
  • mlp.py: Defines the Multi-Layer Perceptron;
  • rnn.py: Defines the Long Short-Term Memory;
• outputs: Folder that holds the saved models and optimization histories, such as .pth and .pkl;
• utils
  • attribute.py: Re-writes getters and setters for nested attributes;
  • loader.py: Utility to load datasets and split them into training, validation and testing sets;
  • objects.py: Wraps objects instantiation for command line usage;
  • optimizer.py: Wraps the optimization task into a single method;
  • targets.py: Implements the objective functions to be optimized.

---

Package Guidelines

Installation

Install all the pre-needed requirements using:

pip install -r requirements.txt

Data configuration

In order to run the experiments, you can use torchvision and torchtext to load pre-implemented datasets.

---

Usage

Model Training

The first step is to pre-train a Machine Learning architecture. To accomplish such a step, one needs to use the following script:

python image_model_training.py -h

or

python text_model_training.py -h

Note that line 74 (for image-based) and 75 (for text-based) should be adjusted on core/model.py according to the used script.

Model Optimization

After conducting the training task, one needs to optimize the weights over the validation set. Please, use the following script to accomplish such a procedure:

python image_model_optimization.py -h

or

python text_model_optimization.py -h

Note that -h invokes the script helper, which assists users in employing the appropriate parameters.

Model Evaluation

Finally, with an optimized models in hands, it is now possible to evaluate the model over the testing set. Please, use the following script to accomplish such a procedure:

python image_model_evaluation.py -h

or

python text_model_evaluation.py -h

Bash Script

Instead of invoking every script to conduct the experiments, it is also possible to use the provided shell script, as follows:

./image_pipeline.sh

or

./text_pipeline.sh

Such a script will conduct every step needed to accomplish the experimentation used throughout this paper. Furthermore, one can change any input argument that is defined in the script.

---

Support

We know that we do our best, but it is inevitable to acknowledge that we make mistakes. If you ever need to report a bug, report a problem, talk to us, please do so! We will be available at our bests at this repository.

---