Models, methods, and the machinery of intelligence.
arXiv
Verdict: RoutineWe present Kandinsky 6.0 Video, a family of foundation diffusion models for synchronized text-to-audio-video generation, comprising Kandinsky 6.0 Video Lite (3B parameters) and Kandinsky 6.0 Video Pro (29B parameters).
State tracking requires an update at every input, but the depth a Transformer applies to each token is fixed regardless of sequence length.
Reinforcement learning (RL) for post-training large language models (LLMs) incurs substantial computation and memory overhead during rollout generation, which motivates low-precision rollout for efficient RL training.
Real-time robot control demands enough visual history to infer motion and task progress, but processing that history can delay action.
On-Policy Distillation (OPD) trains a student on its own generations using teacher feedback.
Reinforcement learning (RL) has greatly advanced the capabilities of large language models (LLMs), but its memory demands remain a barrier to broader adoption.
Assistants from 2011 answered and waited, and agents from 2023 did a task and stopped.
Highly compressed video autoencoders offer an effective way to accelerate video diffusion models, as the Diffusion Transformer (DiT) operates on far fewer tokens.
Tool-using agents make consequential changes to external state, yet correct outcomes do not guarantee that their actions were supported by evidence established beforehand.