NVIDIA DLSS (Deep Learning Super Sampling) is a deep learning neural network that boosts frame rates and generates beautiful, sharp images for your games. It gives you the performance headroom to maximize ray tracing settings and increase output resolution. DLSS is powered by dedicated AI processors on RTX GPUs called Tensor Cores.

One of the key advantages of a deep learning approach to super sampling is that the AI model can continuously improve through ongoing training on NVIDIA’s supercomputer. In fact, each major production release of DLSS has delivered better image quality across wider ranges of games and applications.

We are inviting the developer community to test the latest experimental DLSS models straight off the supercomputer, and provide us with your feedback. Your early input is important to helping us push the state of the art in AI graphics technology.

Experimental DLSS Models

These models may contain new features or image quality improvements that may end up in official DLSS releases. However, they are not fully validated and may contain regressions.

Download Synopsis
White Collie 1Improved object detail in motion and particle visibility. Also improved the handling of moderate native instability and minimized ghosting.
White Collie 2 Alternate improvements for ghosting.

Please send feedback to your NVIDIA Developer Relations representative, email DLSS-Support@nvidia.com or post at forums.developer.nvidia.com.

Note: Research DLLs do not support signed Window Store applications