Search papers, labs, and topics across Lattice.
3
0
4
Memory overload in autoregressive video generation can be tackled by absorbing historical context into model weights, achieving up to 50% cache reduction with minimal quality loss.
Achieve better compression in low-bit quantization by considering not just numerical sensitivity, but also the structural role of each layer.
Autoregressive video models can now generate 4-minute videos without retraining, thanks to a clever inference-time hack that fixes positional embedding bias and injects dynamic priors.
