tamserve
Live, multi-camera object tracking with SAM3, SAM 2 and EdgeTAM, on one GPU or many. One model per GPU serves every camera: faster than the original code, about 4x less GPU memory, identical masks.
- 225
- Python
- Apache-2.0
- 20 h ago
Live, multi-camera object tracking with SAM3, SAM 2 and EdgeTAM, on one GPU or many. One model per GPU serves every camera: faster than the original code, about 4x less GPU memory, identical masks.
Real-time SAM2 segmentation on edge devices - 40x faster C++ inference with ONNX Runtime for iOS/Android deployment
Free, open-source Viggle alternative for Windows and Linux: swap a person in a 4–30 s video for a character from one image with Wan 2.2 Animate on your own NVIDIA GPU (12 GB VRAM recommended). $0 per video, no cloud, no account. Rust and ONNX Runtime prep, ComfyUI render, MCP server. Unofficial.
Meta's EdgeTAM (SAM 2 for the edge) as four Core ML packages: click once, track the object through video at ~48 ms a frame on the Apple Neural Engine.
Click once, track the object through video: EdgeTAM and SAM 2.1 tiny in Rust on ONNX Runtime's CUDA provider. ~42 ms a frame, mean mask IoU 0.97–0.998 vs PyTorch, no Python at run time. Windows and Linux.