GLIDE is a research model from OpenAI that generates images from text descriptions. You type a description like "a corgi wearing a red hat" and the model produces a photo-realistic image matching that description. It also supports inpainting, which means you can take an existing image, mask out a region of it, and ask the model to fill in that region based on a text prompt. The model works using a technique called diffusion, which starts with random noise and gradually refines it into a coherent image guided by the text input. The version released in this repository is a smaller, filtered version of the full model described in OpenAI's research paper. OpenAI released this reduced version publicly while keeping the full model internal, citing concerns about potential misuse. The repository includes three example notebooks that walk through the main use cases: generating images from text, filling in masked regions of images, and an alternative generation approach that uses a separate model called CLIP to guide the image quality. Each notebook can be run directly in a browser using Google Colab without any local setup. Installation requires cloning the repository and running a single install command. The README is brief and points primarily to the notebooks for usage details.
← openai on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.