#k2horizon

Le K2 Horizon d'IFM livre six modèles Apache 2.0 (de 0,9B à 375B) avec tout le cycle d'entraînement ouvert et une nouvelle architecture d'attention sparse

IFM's K2 Horizon ships six Apache 2.0 models (0.9B to 375B) with the full training lifecycle opened and a new MoVA sparse-attention architecture.

IFM released K2 Horizon as six Apache 2.0 models spanning 375 billion down to 0.9 billion parameters that share architecture, vocabulary and training methodology, publishing intermediate checkpoints, data-construction recipes, training code and logs