| license: apache-2.0 | |
| pipeline_tag: keypoint-detection | |
| tags: | |
| - model_hub_mixin | |
| - pytorch_model_hub_mixin | |
| # EdgeCrafter: ECPose | |
| EdgeCrafter is a unified framework for compact Vision Transformers (ViTs) designed for high-performance dense prediction on resource-constrained edge devices. This specific checkpoint is for **ECPose**, a model specialized for pose estimation (keypoint detection). | |
| - **Project Page:** [EdgeCrafter](https://intellindust-ai-lab.github.io/projects/EdgeCrafter/) | |
| - **Paper:** [EdgeCrafter: Compact ViTs for Edge Dense Prediction via Task-Specialized Distillation](https://arxiv.org/abs/2603.18739) | |
| - **Code:** [GitHub Repository](https://github.com/Intellindust-AI-Lab/EdgeCrafter) | |
| ## Model Description | |
| The EdgeCrafter framework addresses the challenge of deploying Vision Transformers on edge devices through task-specialized distillation and edge-aware encoder-decoder designs. ECPose models achieve a strong accuracy-efficiency trade-off, often outperforming CNN-based architectures like YOLO while maintaining low latency on edge hardware. | |
| This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration. | |
| ## Citation | |
| If you find this work useful, please consider citing: | |
| ```bibtex | |
| @article{liu2026edgecrafter, | |
| title={EdgeCrafter: Compact ViTs for Edge Dense Prediction via Task-Specialized Distillation}, | |
| author={Liu, Longfei and Hou, Yongjie and Li, Yang and Wang, Qirui and Sha, Youyang and Yu, Yongjun and Wang, Yinzhi and Ru, Peizhe and Yu, Xuanlong and Shen, Xi}, | |
| journal={arXiv}, | |
| year={2026} | |
| } | |
| ``` |