Falcon-Perception and Falcon-OCR model: early-fusion, natively multimodal, dense Autoregressive Transformer models.
Create short AI videos from text or first and last frames, with synchronized dialogue, effects, and ambience.