Table Transformer (DETR) - PubTables1M is a Transformer-based object detection model designed for table detection in unstructured documents. Fine-tuned on the PubTables1M dataset, it is based on DETR (DEtection TRansformer) and employs a "normalize before" approach, where layer normalization is applied before self- and cross-attention. This model enables accurate table extraction, making it useful for document processing tasks such as OCR, data extraction, and information retrieval.
MIT
Sector Agnostic
Microsoft
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
Open
12/03/25 06:34:55
N.A.
0
© 2026 - Copyright AIKosh. All rights reserved.