guided-diffusion

This is the codebase for Diffusion Models Beat GANS on Image Synthesis.

This repository is based on openai/improved-diffusion, with modifications for classifier conditioning and architecture improvements.

Usage

Training diffusion models is described in the parent repository. Training a classifier is similar. We assume you have put training hyperparameters into a TRAIN_FLAGS variable, and classifier hyperparameters into a CLASSIFIER_FLAGS variable. Then you can run:

mpiexec -n N python scripts/classifier_train.py --data_dir path/to/imagenet $TRAIN_FLAGS $CLASSIFIER_FLAGS

Make sure to divide the batch size in TRAIN_FLAGS by the number of MPI processes you are using.

Here are flags for training the 128x128 classifier. You can modify these for training classifiers at other resolutions:

TRAIN_FLAGS="--iterations 300000 --anneal_lr True --batch_size 256 --lr 3e-4 --save_interval 10000 --weight_decay 0.05"
CLASSIFIER_FLAGS="--image_size 128 --classifier_attention_resolutions 32,16,8 --classifier_depth 2 --classifier_width 128 --classifier_pool attention --classifier_resblock_updown True --classifier_use_scale_shift_norm True"

For sampling from a 128x128 classifier-guided model, 25 step DDIM:

MODEL_FLAGS="--attention_resolutions 32,16,8 --class_cond True --image_size 128 --learn_sigma True --num_channels 256 --num_heads 4 --num_res_blocks 2 --resblock_updown True --use_fp16 True --use_scale_shift_norm True"
CLASSIFIER_FLAGS="--image_size 128 --classifier_attention_resolutions 32,16,8 --classifier_depth 2 --classifier_width 128 --classifier_pool attention --classifier_resblock_updown True --classifier_use_scale_shift_norm True --classifier_scale 1.0 --classifier_use_fp16 True"
SAMPLE_FLAGS="--batch_size 4 --num_samples 50000 --timestep_respacing ddim25 --use_ddim True"
mpiexec -n N python scripts/classifier_sample.py \
    --model_path /path/to/model.pt \
    --classifier_path path/to/classifier.pt \
    $MODEL_FLAGS $CLASSIFIER_FLAGS $SAMPLE_FLAGS

To sample for 250 timesteps without DDIM, replace --timestep_respacing ddim25 to --timestep_respacing 250, and replace --use_ddim True with --use_ddim False.

This is the codebase for Diffusion Models Beat GANS on Image Synthesis.

Related tags

Overview

guided-diffusion

Usage

Owner

OpenAI

Milano is a tool for automating hyper-parameters search for your models on a backend of your choice.

This repository contains the accompanying code for Deep Virtual Markers for Articulated 3D Shapes, ICCV'21

The official codes of "Semi-supervised Models are Strong Unsupervised Domain Adaptation Learners".

A toolkit for Lagrangian-based constrained optimization in Pytorch

Pytorch Implementation of "Contrastive Representation Learning for Exemplar-Guided Paraphrase Generation"

PyTorch implementations of algorithms for density estimation

Implementation of: "Exploring Randomly Wired Neural Networks for Image Recognition"

LETR: Line Segment Detection Using Transformers without Edges

Code for Environment Dynamics Decomposition (ED2).

Semi-SDP Semi-supervised parser for semantic dependency parsing.

Facilitates implementing deep neural-network backbones, data augmentations

Text-to-SQL in the Wild: A Naturally-Occurring Dataset Based on Stack Exchange Data

Train the HRNet model on ImageNet

Deep Text Search is an AI-powered multilingual text search and recommendation engine with state-of-the-art transformer-based multilingual text embedding (50+ languages).

Libraries, tools and tasks created and used at DeepMind Robotics.

Collection of generative models in Pytorch version.

On Nonlinear Latent Transformations for GAN-based Image Editing - PyTorch implementation

Contrastive unpaired image-to-image translation, faster and lighter training than cyclegan (ECCV 2020, in PyTorch)

Code for the paper "Jukebox: A Generative Model for Music"

A big endian Gentoo port developed on a Pine64.org RockPro64