open-edge-platform/geti
Build, train, optimize, and run computer vision models locally, from raw images to live inference. Open source, optimized for Intel® XPU (CUDA also supported).
A star delta has been computed for 52,827 of 63,708 tracked repositories in this period, and only those can be ranked in it.
A repository missing here has not been measured far enough back yet. That does not mean it did not move.
Star counts measured between 10:02 and 00:47 (BKK).
Build, train, optimize, and run computer vision models locally, from raw images to live inference. Open source, optimized for Intel® XPU (CUDA also supported).
All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.
Official Code for "GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training" (ICML 2026) https://arxiv.org/pdf/2602.20399
DIPY is the paragon 3D/4D+ medical imaging library in Python. Contains generic methods for spatial normalization, signal processing, machine learning, statistical analysis and visualization of medical images. Additionally, it contains specialized methods for computational anatomy including diffusion, perfusion and structural imaging.
A python library for self-supervised learning and latent world models on images.
That's all 5 repositories in this period.