Navigating StyleGAN2 w latent space using CLIP

Last update: Dec 6, 2022

Related tags

Overview

Navigating StyleGAN2 w latent space using CLIP

an attempt to build sth with the official SG2-ADA Pytorch impl kinda inspired by Generating Images from Prompts using CLIP and StyleGAN based on the og projector.py

things learned:

it's better to generate initial w values from a well converged sample rather than starting with random or median ones
optimizing w and noise inputs works better than w alone
default values of 0.02 for LR/noise work fine with portraits

Quick start

clone SG2 repo, copy clip dir from CLIP repo, install pytorch 1.7.1 and stuff
pick a suitable SG2 PKL (eg FFHQ)
pick a seed
run python3 approach.py --network network-snapshot-ffhq.pkl --outdir project --num-steps 100 --text 'an image of a girl with a face resembling Paul Krugman' --psi 0.8 --seed 12345
alternatively, one can start from a w vector stored as .npz python3 approach.py --network network-snapshot-ffhq.pkl --outdir project --num-steps 100 --text 'an image of a girl with a face resembling Paul Krugman' --w w-7660ca0b7e95428cac94c89459b5cebd8a7acbd4.npz

FFHQ test

python3 approach.py --network stylegan2-ffhq-config-f.pkl --outdir ffhq --num-steps 100 --text 'an image of an Instagram influencer girl' --psi 0.7 --seed 32

You might also like...

[CVPR 2020] Interpreting the Latent Space of GANs for Semantic Face Editing

InterFaceGAN - Interpreting the Latent Space of GANs for Semantic Face Editing Figure: High-quality facial attributes editing results with InterFaceGA

GenForce: May Generative Force Be with You

1.3k Dec 29, 2022

PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models

PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models Code accompanying CVPR'20 paper of the same title. Paper lin

7k Dec 30, 2022

MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space

Update (20 Jan 2020): MODALS on text data is avialable MODALS MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space Table of Conte

38 Dec 15, 2022

PyTorch implementation of the WarpedGANSpace: Finding non-linear RBF paths in GAN latent space (ICCV 2021)

Authors official PyTorch implementation of the "WarpedGANSpace: Finding non-linear RBF paths in GAN latent space" [ICCV 2021].

100 Dec 6, 2022

InterFaceGAN - Interpreting the Latent Space of GANs for Semantic Face Editing

InterFaceGAN - Interpreting the Latent Space of GANs for Semantic Face Editing Figure: High-quality facial attributes editing results with InterFaceGA

1.3k Jan 9, 2023

Code for "SRHEN: Stepwise-Refining Homography Estimation Network via Parsing Geometric Correspondences in Deep Latent Space"

SRHEN This is a better and simpler implementation for "SRHEN: Stepwise-Refining Homography Estimation Network via Parsing Geometric Correspondences in

1 Oct 28, 2022

Code to reproduce the results in the paper "Tensor Component Analysis for Interpreting the Latent Space of GANs".

Tensor Component Analysis for Interpreting the Latent Space of GANs [ paper | project page ] Code to reproduce the results in the paper "Tensor Compon

4 Jun 17, 2022

Implementation based on Paper - Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling

3 Jul 8, 2022

Official Pytorch implementation of the paper "MotionCLIP: Exposing Human Motion Generation to CLIP Space"

MotionCLIP Official Pytorch implementation of the paper "MotionCLIP: Exposing Human Motion Generation to CLIP Space". Please visit our webpage for mor

173 Dec 26, 2022

Comments

$Guessing the best seed using {autoseed_samples} samples$

Guessing the best seed using {autoseed_samples} samples

I am no longer confused by your excellent work,. I have recently been studying this most common method of text-driven text generation.May I ask what is your idea to get the best seed？

opened by 49xxy 0

Navigating StyleGAN2 w latent space using CLIP

Related tags

Overview

Navigating StyleGAN2 w latent space using CLIP

Quick start

FFHQ test

You might also like...

[CVPR 2020] Interpreting the Latent Space of GANs for Semantic Face Editing

PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models

MODALS: Modality-agnostic Automated Data Augmentation in the Latent Space

PyTorch implementation of the WarpedGANSpace: Finding non-linear RBF paths in GAN latent space (ICCV 2021)

InterFaceGAN - Interpreting the Latent Space of GANs for Semantic Face Editing

Code for "SRHEN: Stepwise-Refining Homography Estimation Network via Parsing Geometric Correspondences in Deep Latent Space"

Code to reproduce the results in the paper "Tensor Component Analysis for Interpreting the Latent Space of GANs".

Implementation based on Paper - Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling

Official Pytorch implementation of the paper "MotionCLIP: Exposing Human Motion Generation to CLIP Space"

Comments

Guessing the best seed using {autoseed_samples} samples

Owner

Mike K.

Non-Official Pytorch implementation of "Face Identity Disentanglement via Latent Space Mapping" https://arxiv.org/abs/2005.07728 Using StyleGAN2 instead of StyleGAN

Cartoon-StyleGan2 🙃 : Fine-tuning StyleGAN2 for Cartoon Face Generation

Minimal PyTorch implementation of Generative Latent Optimization from the paper "Optimizing the Latent Space of Generative Networks"

Generating images from caption and vice versa via CLIP-Guided Generative Latent Space Search

Feed forward VQGAN-CLIP model, where the goal is to eliminate the need for optimizing the latent space of VQGAN for each input prompt

Visualizer using audio and semantic analysis to explore BigGAN (Brock et al., 2018) latent space.

CLIP-GEN: Language-Free Training of a Text-to-Image Generator with CLIP

Working demo of the Multi-class and Anomaly classification model using the CLIP feature space

Face Identity Disentanglement via Latent Space Mapping [SIGGRAPH ASIA 2020]

Disentangled Face Attribute Editing via Instance-Aware Latent Space Search, accepted by IJCAI 2021.