RF-DETR is a production-grade computer vision library by Roboflow for real-time object detection, instance segmentation, and keypoint estimation. It implements a DETR-family architecture (LW-DETR, Deformable DETR) with a DINOv2 backbone, providing a complete training-to-deployment pipeline. The library is published as the `rfdetr` PyPI package and serves as the open-source core of Roboflow's detection platform, with larger models available under a separate license. The system is structured as a Python library with a CLI entry point (`rfdetr`). The main model class `RFDETR` (and variants like `RFDETRNano`) provides `predict()`, `train()`, `export()`, and `evaluate()` methods. Training is built on PyTorch Lightning with custom `LightningModule` and `LightningDataModule`, plus callbacks for EMA, COCO evaluation, GPU memory monitoring, and drop-path scheduling. The export subsystem uses a registry pattern with format-specific exporters (ONNX, TensorRT, TFLite, CoreML, OpenVINO, ExecuTorch, LiteRT) that share preprocessing and decoding utilities. Dataset loaders support COCO, YOLO, WebDataset, and Roboflow formats with configurable augmentation backends (torchvision, Albumentations, Kornia). The primary data flow: input images → preprocessing (resize, normalize) → DINOv2 backbone → multi-scale feature projector → deformable attention transformer encoder-decoder → task-specific heads (detection, segmentation, keypoints) → post-processing (top-k selection, NMS-free decoding). Training uses COCO-style datasets with Hungarian matching for label assignment. External dependencies include PyTorch, PyTorch Lightning, HuggingFace Transformers, and optional backends for export. The library integrates with Roboflow's platform for dataset management and deployment. The codebase is actively maintained with extensive test coverage (pytest, CI workflows for CPU/GPU/XLA), backward compatibility testing for legacy checkpoints, and a formal changelog. It demonstrates high craftsmanship in areas like safe checkpoint loading, atomic file operations, custom fused optimizers (Triton), and export robustness. The large optional dependency surface and reliance on some private PyTorch APIs introduce maintenance risk, but overall it is a well-engineered library suitable for production use.
Brush is a cross-platform 3D Gaussian splatting training and rendering engine written in Rust. It implements a fully differentiable renderer with forward and backward passes on GPU via WebGPU (wgpu/cubecl), supporting multiple camera models (pinhole, Kannala-Brandt fisheye, radial-tangential, thin-prism fisheye) and spherical harmonics for view-dependent appearance. The system ingests COLMAP, Nerfstudio, and RealityCapture datasets, trains Gaussian splat representations with configurable losses (L1, SSIM, LPIPS), and exports results in PLY or safetensors format. It provides four frontends: a headless CLI for batch training, a desktop application with an egui-based UI for interactive training and visualization, a WebAssembly web demo with React/Vite, and an Android NDK build. The architecture is modular, with separate crates for rendering, training, dataset loading, serialization, GPU kernels (sort, scan, render), and a virtual filesystem abstraction that unifies local, HTTP, and WASM file access. External dependencies include the Burn deep learning framework (tracel-ai/burn) for autodiff and GPU backend, and Rerun for real-time visualization of training metrics and splat geometry.