torch-gqn icon indicating copy to clipboard operation
torch-gqn copied to clipboard

PyTorch Implementation of Generative Query Network

Results 6 torch-gqn issues
Sort by recently updated
recently updated
newest added

how many steps does it take to generate such imgs from maze data, shown in readme?

the convert2torch.py takes too much time for data generation, and the generated *.pt.gz couldn't be taken by the GQNDataset function, I guess gzip is needed here

Has this code been tried on the 'rooms_ring_camera' data set? Or is it specific to 'Shepard-Metzler' data set?

Hello, is this possible if you can upload the trained model on the Shepard-Metzler and mazes dataset ? I would like to help to create a test script.

Hi I just wanted to know what lines 68-70 of train.py mean? # Pixel standard-deviation sigma_i, sigma_f = 2.0, 0.7 sigma = sigma_i The images have 3 channels (R,G,B) so...

Whether I use `2xNVIDIA GeForce RTX 2080 Ti` or `4xTeslaV100-SXM2-32GB`or `8xTeslaV100-SXM2-32GB`, it takes about 20 days to train `shepard_metzler_5_parts`. I want to know how long it usually takes to train...