akbokha / image-colorization

Image colorization framework which facilitates the training and evaluation of various neural network architectures

akbokha/image-colorization Issues

Implement Unet
Closed 5 years ago1
Distribute sampling of saved images across batches and classes
Closed 5 years ago1
Run GAN (Dominik's implementation) on places365
Closed 5 years ago1
Run GAN (Abdel's implementation) on places365
Closed 5 years ago1
Run UNET on places365
Closed 5 years ago1
Run RESNET on places365
Closed 5 years ago1
Make script to colorize test set from saved model weights
Closed 5 years ago1
Decide what GAN to work on
Closed 5 years ago1
Implement Nazeri GAN
Closed 5 years ago1
Train places365 classifier
Closed 5 years ago
Finish implementing Semantic interpretability
Closed 5 years ago
Finish GANs (make sure we just need to submit a longjob)
Closed 5 years ago1
Implement GAN
Closed 5 years ago1
Save Classifier Model selectively
Closed 6 years ago
Expand Placeholder dataset
Closed 6 years ago
Implement classifier for Semantic Interpretability
Closed 6 years ago
Use with torch.no_grad() for validation fprop
Closed 6 years ago1
Support copying of datasets from home to scratch disk prior to experiment
Closed 6 years ago1
Validation batch sizes
Closed 6 years ago
Create Places365 Data Loaders
Closed 6 years ago
Implement Places365 dataloaders
Closed 6 years ago1
Implement CIFAR-10 dataloaders
Closed 6 years ago1
Make number of sample images saved configurable
Closed 6 years ago1
copy images to scratch disk on experiment startup
Closed 6 years ago
Save regular model checkpoints
Closed 6 years ago