James Tompkin

See you at CVPR in the AI for Content Creation Workshop @cvprconference.bsky.social ai-for-content-creation.github.io Weds June 3rd AM, Room 610/612 Great speakers on generative image, video, 3D @ranahanocka.bsky.social, Alan Yuille, Christian Theobalt, Taesung Park, @saining.bsky.social 🖼️🎥🧊

+ 6 orals and 25 posters! Everything from better generative techniques to new control methods to applications in digital humans, 4D worlds, sketching, 3D layout, art, and fashion! Organizers: @krishna-cv.bsky.social @junyanz.bsky.social @lingjieliu.bsky.social Yuheng Li, Thao Nguyen, Yiqing Liang

🎨 Work on generative AI or creative content? AI for Content Creation Workshop @ CVPR! @cvprconference.bsky.social ⏰ Submit by March 23! ai-for-content-creation.github.io • Image, video, 3D, world • Multi-modal • Generation, editing, transfer • VFX, games, fashion, design, architecture & more

The AI for Content Creation workshop #CVPR2025 is accepting paper submissions. ai4cc.net Deadline March 21st 2025 midnight PST. 4 page extended abstracts, 8 pagers, and previously published work (ECCV, NeurIPS, even CVPR)! Many topics 📷📹🎬🎲✒️📃🖼️👗👔🏢 - come spend the day with us!

Can GANs compete in 2025? In 'The GAN is dead; long live the GAN! A Modern GAN Baseline', we show that a minimalist GAN w/o any tricks can match the performance of EDM with half the size and one-step generation - github.com/brownvc/r3gan - work of Nick Huang, @skylion.bsky.social, Volodymyr Kuleshov

We prioritize simplicity and performance over functionality. As a minimal baseline, our model does only basic image generation, lacking many features required for downstream tasks. Think of it as DCGAN in 2025 rather than something feature-rich like StyleGAN. We hope this helps further GAN research!

Can GANs compete in 2025? In 'The GAN is dead; long live the GAN! A Modern GAN Baseline', we show that a minimalist GAN w/o any tricks can match the performance of EDM with half the size and one-step generation - github.com/brownvc/r3gan - work of Nick Huang, @skylion.bsky.social, Volodymyr Kuleshov

To further GAN research, we first improve the GAN loss to alleviate mode dropping and non-convergence. This makes GAN optimization sufficiently easy that we can now discard existing GAN tricks w/o training failure. The dependence on outdated GAN-specific architectures is also eliminated.

Can GANs compete in 2025? In 'The GAN is dead; long live the GAN! A Modern GAN Baseline', we show that a minimalist GAN w/o any tricks can match the performance of EDM with half the size and one-step generation - github.com/brownvc/r3gan - work of Nick Huang, @skylion.bsky.social, Volodymyr Kuleshov