nemotron-table-structure-v1

Model Overview

Description

The Nemotron Table Structure v1 model is a specialized object detection model designed to identify and extract the structure of tables in images. Based on YOLOX, an anchor-free version of YOLO (You Only Look Once), this model combines a simpler architecture with enhanced performance. While the underlying technology builds upon work from Megvii Technology, we developed our own base model through complete retraining rather than using pre-trained weights.

The model excels at detecting and localizing the fundamental structural elements within tables. Through careful fine-tuning, it can accurately identify and delineate three key components within tables:

1. Individual cells (including merged cells)
2. Rows
3. Columns

This specialized focus on table structure enables precise decomposition of complex tables into their constituent parts, forming the foundation for downstream retrieval tasks. This model helps convert tables into the markdown format which can improve retrieval accuracy.

This model is ready for commercial or non-commercial use.

The Nemotron Table Structure v1 Preview API provides hosted inference for images of cropped tables.

License/Terms of use

GOVERNING TERMS: The trial service is governed by the NVIDIA API Trial Terms of Service. Use of this model is governed by the NVIDIA Open Model License Agreement.

Model Architecture

Architecture Type: YOLOX
Network Architecture: DarkNet53 Backbone + FPN Decoupled head (one 1x1 convolution + 2 parallel 3x3 convolutions (one for the classification and one for the bounding box prediction). The YOLOX architecture is a single-stage object detector that improves on Yolo-v3.
Deployment Geography: Global

Use Case:
This model specializes in analyzing images containing tables by:

• Detecting and extracting table structure elements (rows, columns, and cells)
• Providing precise location information for each detected element
• Supporting downstream tasks like table analysis and data extraction

The model is designed to work in conjunction with OCR (Optical Character Recognition) systems to:

1. Identify the structural layout of tables
2. Preserve the relationships between table elements
3. Enable accurate extraction of tabular data from images

Ideal for:

• Document processing systems
• Automated data extraction pipelines
• Digital content management solutions
• Business intelligence applications

Release Date Build.NVIDIA.com: 10/01/2026 via link
Hugging Face: 10/23/2025 via link

Technical Details

Input

Input Types: Image
Input Format: Base64-encoded PNG or JPEG image of a cropped table, supplied as a data URL in the input array; Red, Green, Blue (RGB)
Input Parameters: Two-Dimensional (2D)
Other Properties Related to Input: Image size resized to (1024, 1024)

Output

Output Type(s): Structured table detections
Output Format: JSON-compatible table detections containing bounding boxes, class labels, and confidence scores
Output Parameters: 1D
Other Properties Related to Output: Bounding boxes are normalized to [0, 1]; object classes are cell, row, and column. confidence_threshold defaults to 0.01, and nms_threshold defaults to 0.25 for table structure.

Software Integration

Runtime: NVIDIA hosted Preview API
API Endpoint: https://ai.api.nvidia.com/v1/cv/nvidia/nemotron-table-structure-v1
Supported Hardware Microarchitecture Compatibility: NVIDIA Ampere, NVIDIA Blackwell, NVIDIA Hopper, NVIDIA Lovelace

Model Version(s):

• nemotron-table-structure-v1 (Preview release 2.0-preview; model artifact nemotron-table-structure-v1:2.0.1)

Training Dataset & Evaluation

Training Dataset

Data collection method by dataset: Automated
Labeling method by dataset: Automated
Pretraining: COCO train2017 Finetuning (by NVIDIA): 23,977 images from Digital Corpora dataset, with annotations from Azure AI Document Intelligence. Number of bounding boxes per class: 1,828,978 cells, 134,089 columns and 316,901 rows. The layout model of Document Intelligence was used with 2024-02-29-preview API version.

Evaluation Results

The primary evaluation set: 2,459 digital corpora images with Azure labels. Number of bounding boxes per class: 200,840 cells, 13,670 columns and 34,575 rows. mAP was used as an evaluation metric.
Data collection method by dataset: Hybrid: Automated, Human
Labeling method by dataset: Hybrid: Automated, Human
Properties: We evaluated with Azure labels from manually selected pages, as well as manual inspection on public PDFs and powerpoint slides.

Per-class Performance Metrics:

Inference:

Acceleration Engine: YOLOX engine using cuDNN fused operations, custom CUDA kernels, and CUDA graphs
Test Hardware:

• NVIDIA Ampere (A100)
• NVIDIA Ampere (A10G)
• NVIDIA Blackwell (B200)
• NVIDIA Blackwell (GB200)
• NVIDIA Blackwell (GB10)
• NVIDIA Blackwell (RTX PRO 6000)
• NVIDIA Hopper (H100)
• NVIDIA Hopper (H200)
• NVIDIA Lovelace (L40S)

Ethical Considerations

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. Developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

For more detailed information on ethical considerations for this model, please see the Model Card++ Bias, Explainability, Safety & Security, and Privacy Subcards.

Please report security vulnerabilities or NVIDIA AI Concerns here.