https://github.com/imkett/waes-torch
[Tool] Wasserstein Auto-Encoders for Text and Image generation PyTorch Implementation
https://github.com/imkett/waes-torch
generation image-generation text-generation transformer wae
Last synced: 3 months ago
JSON representation
[Tool] Wasserstein Auto-Encoders for Text and Image generation PyTorch Implementation
- Host: GitHub
- URL: https://github.com/imkett/waes-torch
- Owner: ImKeTT
- Created: 2022-03-08T09:00:26.000Z (about 3 years ago)
- Default Branch: main
- Last Pushed: 2022-03-08T09:04:14.000Z (about 3 years ago)
- Last Synced: 2025-01-25T21:55:24.437Z (4 months ago)
- Topics: generation, image-generation, text-generation, transformer, wae
- Language: Python
- Homepage:
- Size: 19.5 KB
- Stars: 5
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## Introduction
This repo is the Wasserstein Auto-Encoders for Text and Image Generation from [this paper](https://arxiv.org/abs/1711.01558) in Pytorch.
### Feature
- Transformer based (from scratch / from Huggingface) Wasserstein Auto-Encoder with **MMD metric** (WAE-MMD) for **text** generation
- Transformer based (from scratch / from Huggingface) Wasserstein Auto-Encoder with **adversarial metric** (WAE-adv) for **text** generation
- Transformer based (from scratch / from Huggingface) Wasserstein Auto-Encoder with **MMD metric and gradient penalty** (WAE-GP-MMD) for **text** generation
- Transformer based (from scratch / from Huggingface) Wasserstein Auto-Encoder with **adversarial metric and gradient penalty** (WAE-GP-adv) for **text** generation
- Wasserstein Auto-Encoder with adversarial metric (WAE-adv) for **image** generation (based on [https://github.com/schelotto/Wasserstein-AutoEncoders](https://github.com/schelotto/Wasserstein-AutoEncoders))
- Wasserstein Auto-Encoder with MMD metric (WAE-MMD) for **image** generation (based on [https://github.com/schelotto/Wasserstein-AutoEncoders](https://github.com/schelotto/Wasserstein-AutoEncoders))### in progress...