Machine Learning Library
5.6k 2026-04-18
lucidrains/DALLE-pytorch
An open-source PyTorch implementation of OpenAI's DALL-E, a text-to-image transformer, including CLIP for generation ranking.
Core Features
Replication of DALL-E text-to-image generation
Integration with CLIP for ranking generated images
Support for training custom DALL-E models
Modular architecture for components like DiscreteVAE
Community-driven training examples and checkpoints
Quick Start
pip install dalle-pytorchDetailed Introduction
This project provides an open-source, PyTorch-based replication of OpenAI's groundbreaking DALL-E model, a transformer capable of generating images from textual descriptions. It also integrates CLIP for improved ranking of generated outputs. By offering a modular and accessible implementation, DALLE-pytorch empowers researchers and developers to experiment with, train, and deploy their own text-to-image generative AI models, fostering innovation in the field of artificial intelligence and creative content generation.