AntixK / PyTorch-VAE

A Collection of Variational Autoencoders (VAE) in PyTorch.
Apache License 2.0
6.62k stars 1.06k forks source link

Dataset not found (running in a google colab sheet) #27

Open A7F opened 3 years ago

A7F commented 3 years ago

Hi, I'm new to pytorch and VAEs in general. I attempted to run your VanillaVAE but I can't figure out how to reference the dataset inside the config file. Specifically, this is my folder structure:

immagine

and my config looks like this:

exp_params:
  dataset: celeba
  data_path: "/content/drive/MyDrive/Colab/celeba"
  img_size: 64
  batch_size: 144 # Better to have a square number
  LR: 0.005
  weight_decay: 0.0
  scheduler_gamma: 0.95

but still the error remains the same: RuntimeError: Dataset not found or corrupted. You can use download=True to download it

Apart from the fact that using download=True doesn't work (looks like it attempts to download an invalid dataset), is there anything to take into account like unzipping the archive img_align_celeba.zip or stuff like that?

breAchyz commented 3 years ago

I have met the same problem and solved it. I extracted the files in 'img_align_celeba.zip' to the current folder. Hope you will solve it too.

jackransomlovell commented 3 years ago

I'm having a similar issue, where I would like to train on a completely different dataset located in my drive. Is this possible or do we have to use the celeba dataset?

MajesticSoftware commented 2 years ago

I'm having a similar issue, where I would like to train on a completely different dataset located in my drive. Is this possible or do we have to use the celeba dataset?

Were you ever able to train on a different dataset?

peacej commented 2 years ago

This seems to work in Colab

!git clone https://github.com/peacej/PyTorch-VAE # I cloned it to change the requirements.txt to make it work 
%cd PyTorch-VAE
!mkdir -p Data
!mkdir -p logs
!pip install -r requirements.txt
!unzip ./Data/celeba.zip -d ./Data/
!unzip ./Data/celeba/img_align_celeba.zip -d ./Data/celeba/ # celeba folder structure defined by https://pytorch.org/vision/main/_modules/torchvision/datasets/celeba.html#CelebA
1619307446 commented 1 year ago

I cannot open the webpage link. If convenient, could you send these two compressed files? thank you

NevilleMthw commented 1 year ago

I'm having a similar issue, where I would like to train on a completely different dataset located in my drive. Is this possible or do we have to use the celeba dataset?

Were you able to work with it on a different dataset? I am guessing the dataset structure needs to be similar is all.