Cite
Notes
Only stored in your browser.
Attribution
SmolVLM: Redefining small and efficient multimodal models
arXiv 2025
The Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text
Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations
arXiv 2024
from 3 papers
Loubna Ben allal
Leandro von Werra
A. Feder Cooper
Aaron Gokaslan
Alexander Hägele
Alon Albalak
Andres Marafioti
Anton Lozhkov
Atli Kosson
Aviya Skowron