DeltaQuant: 4-bit Video Diffusion Models with Spatiotemporal Delta Smoothing

Prompt: The camera follows a white vintage SUV with a black roof rack speeding up a steep dirt road surrounded by redwoods on a mountain slope. Dust kicks up as the sunlight creates a warm glow, emphasizing the rugged, serene landscape. * Equal contribution tokens) to 4 bits while keeping core tokens in FP8. This decomposition substantially reduces quantization error with minimal overhead. For weight quantization, DeltaQuant incorporates SVDQuant's low-rank decomposition to further reduce quantization error. We also implement an efficient kernel that translates DeltaQuant's computational benefits into real-world speedups. Extensive experiments on Wan2.2 I2V, Wan2.2 T2V, and LTX-Video T2V demonstrate that DeltaQuant maintains high generation fidelity. On Wan2.2, it compresses model size by 2.9× and reduces memory footprint by 2.3×. DeltaQuant is compatible with efficient attention mechanisms and few-step distillation. When integrated with these techniques, it achieves an additional 3.0× acceleration, for a total 111.8× end-to-end speedup.