Skip to content

evry-paris-saclay/2023-m2cns-deepLearning-BetaVAE

 
 

Repository files navigation

PyTorch VAE

A collection of Variational AutoEncoders (VAEs) implemented in pytorch with focus on reproducibility. The aim of this project is to provide a quick and simple working example for many of the cool VAE models out there. All the models are trained on the CelebA dataset for consistency and comparison.

Requirements

  • Python >= 3.5
  • PyTorch >= 1.3
  • Pytorch Lightning >= 0.6.0 (GitHub Repo)
  • CUDA enabled computing device

Installation

$ git clone https://github.com/GauthierEd/DeepLearningProject.git
$ cd DeepLearningProject
$ pip install -r requirements.txt

Usage

$ cd DeepLearningProject
$ python run.py -c configs/<config-file-name.yaml>

Config file template

model:
  load: false
  path: "<path to the model save>"

model_params:
  name: "<name of VAE model>"
  in_channels: 3
  latent_dim: 
    .         # Other parameters required by the model
    .
    .

data_params:
  data_path: "<path to the celebA dataset>"
  train_batch_size: 64 # Better to have a square number
  val_batch_size:  64
  patch_size: 64  # Models are designed to work for this size
  num_workers: 4
  
exp_params:
  manual_seed: 1265
  LR: 0.005
  weight_decay:
    .         # Other arguments required for training, like scheduler etc.
    .
    .

trainer_params:
  gpus: 1         
  max_epochs: 100
  gradient_clip_val: 1.5
    .
    .
    .

logging_params:
  save_dir: "logs/"
  name: "<experiment name>"

View TensorBoard Logs

$ cd logs/<experiment name>/version_<the version you want>
$ tensorboard --logdir .

License

Apache License 2.0

Permissions Limitations Conditions
✔️ Commercial use ❌ Trademark use ⓘ License and copyright notice
✔️ Modification ❌ Liability ⓘ State changes
✔️ Distribution ❌ Warranty
✔️ Patent use
✔️ Private use

Citation

@misc{Subramanian2020,
  author = {Subramanian, A.K},
  title = {PyTorch-VAE},
  year = {2020},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/AntixK/PyTorch-VAE}}
}

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%