9 models · ranked by HuggingFace downloads
A DETR-based object detection model from Microsoft Research trained to locate tables in document images. It is the detection stage in a two-step pipeline — a separate structure recognition model then parses the detected table's rows and columns.
table-transformer-structure-recognition has no registered pipeline_tag. It likely serves as a pretraining base or a specialized evaluation model — review the model card before use.
yolos-small has no registered pipeline_tag. It likely serves as a pretraining base or a specialized evaluation model — review the model card before use.
table-transformer-structure-recognition-v1.1-all is released without a specific pipeline. Common uses include feature extraction, encoder probing, and domain-specific fine-tuning.
rtdetr_v2_r50vd is a Real-Time DEtection TRansformer v2 built on a ResNet-50vd backbone, trained on COCO. RT-DETRv2 improves over RT-DETRv1 with flexible denoising training and faster convergence, achieving real-time detection without NMS post-processing. The ResNet-50vd variant targets the speed-accuracy balance point for production deployment.
yolos-fashionpedia is released without a specific pipeline. Common uses include feature extraction, encoder probing, and domain-specific fine-tuning.
PP-DocLayoutV3 is PaddleOCR's third-generation document layout detection model, converted to safetensors format for HuggingFace compatibility. It performs object detection to identify layout regions — text blocks, tables, figures, formulas, headings — in document images using a transformer-based backbone. The model is a building block in PaddleOCR's full document parsing pipeline.
detr-resnet-50 is an open-source object-detection model available on HuggingFace. Details are sourced from the public model registry.
detr-doc-table-detection is an open-source object-detection model available on HuggingFace. Details are sourced from the public model registry.