403 on trying to get CelebA dataset C1W4B_Controllable_Generation

Course Q&A Generative Adversarial Networks (GANS) Build Basic Generative Adversarial Networks
1

Hi!

I’m getting an error when trying to train my own classifier in assignment about controllable generation.

HTTPSConnectionPool(host='docs.google.com', port=443): Max retries exceeded with url: /uc?export=download&id=0B7EVK8r0v71pZjFTYXZWM3FlRnM (Caused by ProxyError('Cannot connect to proxy.', OSError('Tunnel connection failed: 403 Forbidden',)))

Looks like there’s not enough rights for the doc in google docs to be accessed from the notebook.

Also I got an error about the seaborn package being not available, so I had to add a cell with !pip3 install seaborn to install it. It can be confusing for other students, who are not proficient with python and jupiter. Also this makes grader to fail submission.

2

Hi @aberezkin ,
Welcome to the community.

I was trying to download the dataset in my local machine using torchvision, but I was getting the error that “the daily quota of file download has exceeded”. It seems to me that the dataset is present in Google Drive and they have a maximum per day quota (resets every 24 hrs) to download a file.

However, I have come up with a temporary solution for downloading the file in your local. Check this post in the torchvision’s github issue to find more details about the dataset. Also, if you have a GPU device to train the network then you can download the dataset from this link.

Currently, I don’t know how to download that dataset from Google Drive in Coursera’s notebook environment server. If I will find a direct download link then I’ll let you know in this thread.

Hopefully, this solves your problem with the dataset.

3

Hi @aberezkin !
As stated in the notebook, training your own classifier with train_classifier("filename") is not possible in the Coursera platform.

# Uncomment the last line to train your own classifier - this line will not work in Coursera.
# If you'd like to do this, you'll have to download it and run it, ideally using a GPU 
# train_classifier("filename")

You can download the assignment notebook at File > Download as > Notebook (.ipynb), then run it on your computer (preferably with a GPU) and download the dataset as @28utkarsh explained.

Cheers, and welcome to the community!