https://github.com/cori4nder/vae-meandering-rivers

Implementation of a Variational Autoencoder (VAE) for meandering river images using PyTorch
https://github.com/cori4nder/vae-meandering-rivers

deep-learning image-generation machine-learning meandering-rivers pytorch vae variational-autoencoder

Last synced: about 3 hours ago
JSON representation

Implementation of a Variational Autoencoder (VAE) for meandering river images using PyTorch

• Host: GitHub
• URL: https://github.com/cori4nder/vae-meandering-rivers
• Owner: cori4nder
• License: mit
• Created: 2024-08-15T03:52:50.000Z (about 1 month ago)
• Default Branch: main
• Last Pushed: 2024-08-16T03:23:10.000Z (about 1 month ago)
• Last Synced: 2024-09-22T06:02:17.641Z (5 days ago)
• Topics: deep-learning, image-generation, machine-learning, meandering-rivers, pytorch, vae, variational-autoencoder
• Language: Jupyter Notebook
• Homepage:
• Size: 4.33 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# 🏞️ Variational Autoencoder for Meandering Rivers

This repository contains an implementation of a Variational Autoencoder (VAE) model for generating 128x128 images of meandering rivers.



  

  


  



## 🛠️ Requirements

[![PyPI - PyTorch](https://img.shields.io/pypi/v/torch?label=pytorch&logo=pypi)](https://pypi.org/project/torch/)

[![PyPI - NumPy](https://img.shields.io/pypi/v/numpy?label=numpy&logo=pypi)](https://pypi.org/project/numpy/)

[![PyPI - Pandas](https://img.shields.io/pypi/v/pandas?label=pandas&logo=pypi)](https://pypi.org/project/pandas/)

[![PyPI - Scikit-Learn](https://img.shields.io/pypi/v/scikit-learn?label=scikit-learn&logo=pypi)](https://pypi.org/project/scikit-learn/)

[![PyPI - Matplotlib](https://img.shields.io/pypi/v/matplotlib?label=matplotlib&logo=pypi)](https://pypi.org/project/matplotlib/)

[![PyPI - Pillow](https://img.shields.io/pypi/v/Pillow?label=pillow&logo=pypi)](https://pypi.org/project/Pillow/)

## 🚀 Usage

### 📓 Notebooks

> #### Fully Connected 

- **Training the VAE Fully Connected**: [train_vae_fconnected_notebook](./notebooks/1-train_vae_fconnected_notebook.ipynb)

- **Evaluating the VAE Fully Connected**: [eval_vae_fconnected_notebook](./notebooks/1-eval_vae_fconnected_notebook.ipynb)

> #### Fully Convolutional 

- **Training the VAE Fully Convolutional**: [train_vae_fconv_notebook](./notebooks/2-train_vae_fconv_notebook.ipynb)

- **Evaluating the VAE Fully Convolutional**: [eval_vae_fconv_notebook](./notebooks/2-eval_vae_fconv_notebook.ipynb)

> #### GIFs & Videos

- **Generating GIFs/Videos**: [video_gif_generator.ipynb](./video-gif/video_gif_generator.ipynb)

## 📦 Installing Dependencies

You can install all the necessary dependencies listed in the `requirements.txt` file using one of the following methods:

### 1. **Using pip from the terminal**

If you are in the root directory of the project, where the `requirements.txt` file is located, run:

```bash

$ pip install -r requirements.txt

```

## 🤖 Train 

If you prefer to run the model directly without using the notebooks, you can execute the training script from the terminal:

> Fully Connected

```bash

!python train.py --path "./train_images.h5" --model fconnected  --batch_size 128 --epochs 100

```

> Fully Convolutional

```bash

!python train.py --path "./train_images.h5" --model fconv --batch_size 128 --epochs 100

```

This will start the training process using the `train.py` script, which is configured to load the dataset, prepare the model, and begin training.