If you've worked with pose estimation models from MMPose, you know the drill - clone the repo, install PyTorch, install mmcv, install mmpose, pray that the CUDA versions align, and then figure out the config system. For deployment or quick prototyping, that's way too much friction. I went through the same thing with CIGPose , a recent pose estimator that uses causal intervention to handle occlusion better than anything else out there right now (67.5 Whole AP on COCO-WholeBody

RTMO-ORT inference — RTMO-M (COCO, 640×640). I put together rtmo-ort , a tiny wrapper that runs RTMO (MMPose) with pure ONNX...

This video highlights my agent in action that utilises a custom stock price retrieval tool to obtain stock prices from NSE - using the...