Generative Representation Learning for Scalable Latent Tokenization
A generative tokenizer avoids representation collapse and reaches FID-50K 2.18 with JiT-L.
46
Product
ImageNet
- Mentions · 7d0
- Mentions · 30d3
- All mentions3
- First seenSep 29, 2026
- Last mentioned11d ago
Timeline
3 articles · newest firstFew-Step Generation via Data-Space Iteration
Data-space iteration enables few-step image generation without tuning a flow timestep schedule.
48
I Have a Stream: Making Self-Supervised Learning Work on Continuous Video
StreamMAE adapts masked autoencoders so pretraining works on continuous video streams without global shuffling.
48
Appears with
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.