Model Detail
Wan2.2-NVFP4-Sparse
▲ 12.5%Wan2.2-NVFP4-Sparse is an AI model with 14B parameters released by lightx2v. Distributed under the permissive apache-2.0 license.
Wan2.2-NVFP4-Sparse ships with 14B parameters. The apache-2.0 license is permissive, allowing commercial deployment and derivative work without per-seat fees, though attribution requirements still apply.
Downloads of Wan2.2-NVFP4-Sparse have moved +12.5% over the past 24 hours. That is a slight downtrend, consistent with normal cooling as newer models compete for the same workloads. These numbers are signal, not guarantee — week-over-week download counts on Hugging Face also reflect mirror traffic, CI scrapes, and one-off benchmarking runs.
Wan2.2-NVFP4-Sparse is best fit for general-purpose AI workloads. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.
Collaborative Few-Step Distillation and Low-Bit Quantization for Wan2.2 Dual-Expert Video Diffusion Models
arXiv:2606.00658v1 Announce Type: cross Abstract: Large video diffusion models achieve strong visual quality but remain expensive to deploy because each sample requires many denoising steps and a large resident parameter footprint. This paper studies a deployment-oriented compression pipeline for Wa
Timestep-Aware SVDQuant-GPTQ for W4A4 Quantization of Wan2.2-I2V
arXiv:2605.27003v1 Announce Type: cross Abstract: W4A4 quantization of large video diffusion Transformers offers substantial memory savings but is hindered by two main challenges: sparse large-magnitude activation outliers, and strongly timestep-dependent activation distributions across the multi-st
Tail-Aware HiFloat4: W4A4 Post-Training Quantization for Wan2.2
arXiv:2605.26628v1 Announce Type: new Abstract: This report describes Tail-Aware HiFloat4, our submission to the low-bit text-to-video generation quantization challenge. Our method adapts the public ViDiT-Q post-training quantization pipeline to Wan2.2 under the HiFloat4 numerical format. We quantiz
