Glow.md

Glow

Glow is a simple type of generative flow using an invertible 1x1 convolution. Although it is a generative model optimized towards the plain log-likelihood objective, it is capable of efficient realistic-looking synthesis and manipulation of large images.

Parameters

Argument	Description	Default	Choices
--train	Train model	False
--sample	Sample from model	False
--outlier_detection	Outlier detection	False
--dataset	Dataset name	mnist	mnist, cifar10, fashionmnist, chestmnist, octmnist, tissuemnist, pneumoniamnist, svhn, tinyimagenet, cifar100, places365, dtd, imagenet
--no_wandb	Disable Wandb	False
--out_dataset	Outlier dataset name	fashionmnist	mnist, cifar10, fashionmnist, chestmnist, octmnist, tissuemnist, pneumoniamnist, svhn, tinyimagenet, cifar100, places365, dtd, imagenet
--batch_size	Batch size	128
--n_epochs	Number of epochs	100
--lr	Learning rate	0.0002
--hidden_channels	Hidden channels	64
--K	Number of layers per block	8
--L	Number of blocks	3
--actnorm_scale	Act norm scale	1.0
--flow_permutation	Flow permutation	invconv	invconv, shuffle, reverse
--flow_coupling	Flow coupling	affine	additive, affine
--LU_decomposed	Train with LU decomposed 1x1 convs	False
--learn_top	Learn top layer (prior)	False
--y_condition	Class Conditioned Glow	False
--y_weight	Weight of class condition	0.01
--num_classes	Number of classes	10
--sample_and_save_freq	Sample and save frequency	5
--checkpoint	Checkpoint path	None
--n_bits	Number of bits	8
--max_grad_clip	Max Grad clip	0.0
--max_grad_norm	Max Grad Norm	0.0
--num_workers	Number of workers for Dataloader	0
--warmup	Number of warmup epochs	10
--decay	weight decay of learning rate	0

You can find out more about the parameters by checking util.py or by running the following command on the example script:

python GLOW.py --help

Training

You can train this model with the following command:

python GLOW.py --train --dataset octmnist

Sampling

To sample, please provide the checkpoint:

python GLOW.py --sample --dataset octmnist --checkpoint ./../../models/Glow/Glow_octmnist.pt

Outlier Detection

Outlier Detection is performed by using the NLL scores generated by the model:

python GLOW.py --outlier_detection --dataset octmnist --out_dataset mnist --checkpoint ./../../models/Glow/Glow_octmnist.pt