vitpose

Real-time pose estimation pipeline with 🤗 Transformers

VitPose without MMCV dependencies

Human pose estimation within one line

Easy inference tool for ViTPose using ONNX

Repository to evaluate Deep Learning Models for funscript generation

Simple pipeline using Yolov5 and ViTPose to annotate human pose in videos.

Smooth and stabilise OpenPose in ComfyUI: removes pose jitter, hand flicker and dropped keypoints before ControlNet

14-keypoint car pose estimation on CarFusion — YOLO26-pose main + ViTPose-S baseline, FastAPI serving, HF Hub model

An annotation tool that combines TCFormer and ViTPose as an ensemble setup for COCO Wholebody annotations.

專題產生訓練資料的儲存庫

Controlled benchmark of Vision Transformers for cervical-vertebra landmark localisation in videofluoroscopy. 169 archived runs collapsed into a 9-model x 6-configuration grid, 95% CI on every number, each re-derivable from a single run hash. 243 tests; no GPU or dataset needed to verify.