Search papers, labs, and topics across Lattice.
2
0
4
0
Training 3D Gaussian Splatting models on edge devices is now practical: this method slashes peak memory consumption by 80% without sacrificing visual quality.
Mobile GPUs can now run large DNNs and multi-DNN workloads efficiently thanks to FlashMem, which slashes memory consumption by up to 8.4x and accelerates inference by up to 75x.
