[Models](/models)/[Document Analysis](/models?category=document)/microsoft/table-transformer-detection

HFTable Extractionmit

# table-transformer-detection

by microsoft

Detect and extract tables from document images

719Kdl/month

428likes

29Mparams

[HuggingFace](https://huggingface.co/microsoft/table-transformer-detection)[Run it on your own data, free](https://studio.mixpeek.com/namespaces/create?mode=managed&utm_source=homepage&utm_medium=model_page&utm_campaign=mvs_onramp)

Identifiers

Model ID

`microsoft/table-transformer-detection`

## Deploy table-transformer-detection

Single-tenant

Mixpeek has no managed extractor for this model. On a single-tenant deployment you upload the weights and a custom plugin serves them next to the rest of your pipeline.

[Read the customer-hosted guide](https://mixpeek.com/docs/customer-hosted/overview) [Talk to us about single-tenant](/demo)

## Overview

Table Transformer uses DETR architecture adapted for table detection and structure recognition in document images. It identifies table regions and their internal structure (rows, columns, headers).

On Mixpeek, Table Transformer extracts structured table data from PDFs and scanned documents, enabling queries over tabular content in your document collections.

## Architecture

DETR-based architecture with ResNet-18 backbone, fine-tuned on PubTables-1M dataset. Separate models for table detection (locating tables) and table structure recognition (parsing rows/columns).

## How it runs

.dataflow-R15f6b .df-packets-static{display:none} @media (prefers-reduced-motion: reduce){ .dataflow-R15f6b .df-packets-anim{display:none} .dataflow-R15f6b .df-packets-static{display:inline} .dataflow-R15f6b .df-toggle{display:none} }

Pause

Inference INPUT MODEL OUTPUT Image object table-transformer-detection mp.inference Detections vector Vector store MVS table-transformer-detection → embeddings, indexed for search

table-transformer-detection takes image, and Mixpeek indexes what it emits.

## Mixpeek SDK Integration

IngestRetrieve

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so table-transformer-detection runs
// on your side and the output is upserted through POST
// /v1/namespaces/{namespace\_id}/documents/upsert. On Enterprise the other
// path is to upload the weights instead: POST /v1/namespaces/{id}/models
// accepts the huggingface format and a custom plugin loads them.
const res = await fetch(
  "https://api.mixpeek.com/v1/namespaces/ns\_your\_namespace/documents/upsert",
  {
    method: "POST",
    headers: {
      Authorization: "Bearer API\_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      collection\_id: "col\_your\_collection",
      documents: \[
        {
          document\_id: "asset-00412",
          // The model produces text, so it lands in payload. Give the
          // collection a text vector index and embed that text to make it
          // searchable rather than only filterable.
          payload: { extracted\_text: modelOutput, source\_key: "archive/2026/asset-00412" },
          vectors: { "text-embedding": embeddingOfModelOutput },
        },
      \],
    }),
  },
);

// Managed alternative, if this exact model is not the requirement:
// document\_graph\_extractor@v1 runs intfloat/multilingual-e5-large-instruct
// (1024-d) over a bucket, with no inference of your own.

## Capabilities

*   Table region detection in document images
*   Table structure recognition (rows, columns, headers)
*   Handles complex table layouts
*   Works with scanned and digital documents

## Use Cases on Mixpeek

Financial report analysis, extract data tables

Scientific paper data extraction

Government document processing, parse structured data

## Benchmarks

Dataset

Metric

Score

Source

PubTables-1M (detection)

AP50

96.9

Smock et al., 2022: Table 3

PubTables-1M (structure)

AP50

95.3

Smock et al., 2022: Table 4

## Performance

Input Size800×800 px (max)

GPU Latency~22ms / page (A100)

CPU Latency~280ms / page

GPU Throughput~45 pages/sec (A100)

GPU Memory~0.9 GB

## Common Pipeline Companions

[

microsoft/layoutlmv3-base

Layout analysis for full document

](/model/microsoft/layoutlmv3-base)[

PaddlePaddle/paddleocr

OCR on detected table cells

](/model/PaddlePaddle/paddleocr)

## Explore on Mixpeek

[

More Document Analysis models

Compare alternatives in this category

](/models?category=document)[

Best Document AI Platforms

Hand-picked tools & platforms compared

](/curated-lists/best-document-ai-platforms)[

OCR & Document AI Internals

Deep-dive technical guide

](/guides/ocr-document-ai-internals)[

Feature Extractors

See how Mixpeek runs models as extractors

](/extractors)[

Mixpeek Vector Store

Store & search embeddings at scale

](/mvs)[

Pricing

Usage-based pricing for pipelines

](/pricing)[

All Curated Lists

Compare models, APIs & infrastructure

](/curated-lists)

### Specification

FrameworkHF

Organizationmicrosoft

FeatureTable Extraction

Outputtable JSON

Modalitiesdocument

RetrieverTable Filter

Parameters29M

Licensemit

Downloads/mo719K

Likes428

[

### Research Paper

PubTables-1M: Towards comprehensive table extraction from unstructured documents

arxiv.org](https://arxiv.org/abs/2110.00061)

### Build a pipeline with table-transformer-detection

Add this model to a processing pipeline alongside other extractors. Combine with retrieval stages for end-to-end search.

[Run it on your own data, free](https://studio.mixpeek.com/namespaces/create?mode=managed&utm_source=homepage&utm_medium=model_page&utm_campaign=mvs_onramp)

### Alternative Models

[

microsoft/table-transformer-structure-recognition-v1.1-all

Table Extraction

](/model/microsoft/table-transformer-structure-recognition-v1.1-all)[

poloclub/UniTable

Table Extraction

](/model/poloclub/UniTable)[

microsoft/table-transformer-structure-recognition-v1.1-pub

Table Extraction

](/model/microsoft/table-transformer-structure-recognition-v1.1-pub)[

docling-project/docling-models

Table Extraction

](/model/docling-project/docling-models)

### Related in Document Analysis

[

microsoft/layoutlmv3-base

Document Structure

](/model/microsoft/layoutlmv3-base)[

naver-clova-ix/donut-base

Document Structure

](/model/naver-clova-ix/donut-base)[

ibm-granite/granite-docling-258M

Document Structure

](/model/ibm-granite/granite-docling-258M)[

fastino/gliner2-base-v1

Document Structure

](/model/fastino/gliner2-base-v1)