nn-dataset 1.0.4

Neural Network Dataset

Neural Network Dataset

LEMUR - Learning, Evaluation, and Modeling for Unified Research

The original version of the LEMUR dataset was created by Arash Torabi Goodarzi, Roman Kochnev and Zofia Antonina Bentyn at the Computer Vision Laboratory, University of Würzburg, Germany.

Overview 📖

The primary goal of NN Dataset project is to provide flexibility for dynamically combining various deep learing tasks, datasets, metrics, and neural network models. It is designed to facilitate the verification of neural network performance under various combinations of training hyperparameters and data transformation algorithms, by automatically generating performance statistics. It is primarily developed to support the NN Gen project.

Installation or Update of NN Dataset

Remove old version of the LEMUR Dataset and its database:

source .venv/bin/activate
pip uninstall nn-dataset -y
rm -rf db

Installing the stable version:

source .venv/bin/activate
pip install nn-dataset --upgrade --extra-index-url https://download.pytorch.org/whl/cu124

Installing from GitHub to get the most recent code and statistics updates:

source .venv/bin/activate
pip install git+https://github.com/ABrain-One/nn-dataset --upgrade --force --extra-index-url https://download.pytorch.org/whl/cu124

Adding functionality to export data to Excel files and generate plots for analyzing neural network performance:

source .venv/bin/activate
pip install git+https://github.com/ABrain-One/nn-stat --upgrade --force --extra-index-url https://download.pytorch.org/whl/cu124
pip uninstall nn-dataset -y

and export:

source .venv/bin/activate
python -m ab.plot.export

Environment for NN Dataset Contributors

Pip package manager

Create a virtual environment, activate it, and run the following command to install all the project dependencies:

source .venv/bin/activate
python -m pip install --upgrade pip
pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu124

Docker

All versions of this project are compatible with AI Linux and can be run inside a Docker image:

docker run -v /a/mm:. abrainone/ai-linux bash -c "PYTHONPATH=/a/mm python -m ab.nn.train"

Usage

Standard use cases:

1. Add a new neural network model into the ab/nn/nn directory.
2. Run the automated training process for this model (e.g., a new ComplexNet training pipeline configuration):

source .venv/bin/activate
python -m ab.nn.train -c img-classification_cifar-10_acc_ComplexNet

or for all image segmentation models using a fixed range of training parameters and transformer:

source .venv/bin/activate
python run.py -c img-segmentation -f echo --min_learning_rate 1e-4 -l 1e-2 --min_momentum 0.8 -m 0.99 --min_batch_binary_power 2 -b 6

To reproduce the previous result, set the minimum and maximum to the same desired values:

source .venv/bin/activate
python run.py -c img-classification_cifar-10_acc_AlexNet --min_learning_rate 0.0061 -l 0.0061 --min_momentum 0.7549 -m 0.7549 --min_batch_binary_power 2 -b 2 -f norm_299

To view supported flags:

python run.py -h

Contribution

To contribute a new neural network (NN) model to the NN Dataset, please ensure the following criteria are met:

1. The code for each model is provided in a respective ".py" file within the /ab/nn/nn directory, and the file is named after the name of the model's structure.
2. The main class for each model is named Net.
3. The constructor of the Net class takes the following parameters:
   • in_shape (tuple): The shape of the first tensor from the dataset iterator. For images it is structured as (batch, channel, height, width).
   • out_shape (tuple): Provided by the dataset loader, it describes the shape of the output tensor. For a classification task, this could be (number of classes,).
   • prm (dict): A dictionary of hyperparameters, e.g., {'lr': 0.24, 'momentum': 0.93, 'dropout': 0.51}.
   • device (torch.device): PyTorch device used for the model training
4. All external information required for the correct building and training of the NN model for a specific dataset/transformer, as well as the list of hyperparameters, is extracted from in_shape, out_shape or prm, e.g.:
   batch = in_shape[0]
channel_number = in_shape[1]
image_size = in_shape[2]
class_number = out_shape[0]
learning_rate = prm['lr']
momentum = prm['momentum']
dropout = prm['dropout'].
5. Every model script has function returning set of supported hyperparameters, e.g.:
   def supported_hyperparameters(): return {'lr', 'momentum', 'dropout'}
   The value of each hyperparameter lies within the range of 0.0 to 1.0.
6. Every class Net implements two functions:
   train_setup(self, prm)
   and
   learn(self, train_data)
   The first function initializes the criteria and optimizer, while the second implements the training pipeline. See a simple implementation in the AlexNet model.
7. For each pull request involving a new NN model, please generate and submit training statistics for 100 Optuna trials (or at least 3 trials for very large models) in the ab/nn/stat directory. The trials should cover 5 epochs of training. Ensure that this statistics is included along with the model in your pull request. For example, the statistics for the ComplexNet model are stored in files <epoch number>.json inside folder img-classification_cifar-10_acc_ComplexNet, and can be generated by:
   

python run.py -c img-classification_cifar-10_acc_ComplexNet -t 100 -e 5

See more examples of models in /ab/nn/nn and generated statistics in /ab/nn/stat.

Available Modules

The nn-dataset package includes the following key modules:

1. Dataset:

   • Predefined neural network architectures such as AlexNet, ResNet, VGG, and more.
   • Located in ab.nn.nn.

2. Loaders:

   • Data loaders for datasets such as CIFAR-10 and COCO.
   • Located in ab.nn.loader.

3. Metrics:

   • Common evaluation metrics like accuracy and IoU.
   • Located in ab.nn.metric.

4. Utilities:

   • Helper functions for training and statistical analysis.
   • Located in ab.nn.util.

Citation

If you find the LEMUR Neural Network Dataset to be useful for your research, please consider citing:

@misc{ABrain-One.NN-Dataset,
  author       = {Goodarzi, Arash Torabi and Kochnev, Roman and Khalid, Waleed and Qin, Furui and Kathiriya, Yash Kanubhai and Dhameliya, Yashkumar Sanjaybhai and Ignatov, Dmitry and Timofte, Radu},
  title        = {Neural Network Dataset: Towards Seamless AutoML},
  howpublished = {\url{https://github.com/ABrain-One/nn-dataset}},
  year         = {2024},
}

Licenses

This project is distributed under the following licensing terms:

• for neural network models adopted from other projects
  • Python code under the legacy MIT or BSD 3-Clause license
  • models with pretrained weights under the legacy DeepSeek LLM V2 license
• all neural network models and their weights not covered by the above licenses, as well as all other files and assets in this project, are subject to the MIT license

The idea of Dr. Dmitry Ignatov