Scaling Laws for Autoregressive Generative Modeling
TM
Trevor McFedries
@trevvyboi
We identify empirical scaling laws for the cross-entropy loss in four domains: generative image modeling, video modeling, multimodal image$\leftrightarrow$text models, and mathematical problem solving. In all cases autoregressive Transformers smoothly impro...
- Uploaded
- Uploaded Jul 10, 2026
- Queried
- Queried 0 times
No preview text is available for this document yet.
Want to learn more?
Ask a question