GLIDE

tags

Diffusion models, NLP, Computer vision

paper

(Nichol et al. 2022)

Architecture

This model uses joint textual and visual embedding diffusion model followed by some upsampling.

Parameter count

3.5B

Bibliography

1. Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, Mark Chen. March 8, 2022. "GLIDE: Towards Photorealistic Image Generation and Editing with Text-guided Diffusion Models". arXiv. DOI.