https://github.com/jaywalnut310/linear-transformer-for-table-recognition
code for participation in ICDAR2021 Table Recognition track (Team Name: LTIAYN = Kaen Context)
https://github.com/jaywalnut310/linear-transformer-for-table-recognition
table-recognition
Last synced: 3 months ago
JSON representation
code for participation in ICDAR2021 Table Recognition track (Team Name: LTIAYN = Kaen Context)
- Host: GitHub
- URL: https://github.com/jaywalnut310/linear-transformer-for-table-recognition
- Owner: jaywalnut310
- Created: 2021-04-30T10:05:15.000Z (about 5 years ago)
- Default Branch: main
- Last Pushed: 2021-06-16T21:19:43.000Z (about 5 years ago)
- Last Synced: 2025-04-18T23:59:53.729Z (about 1 year ago)
- Topics: table-recognition
- Language: Python
- Homepage:
- Size: 963 KB
- Stars: 21
- Watchers: 5
- Forks: 6
- Open Issues: 5
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Linear Transformer for Table Recognition
## Introduction
This is the code repository for participation in [ICDAR2021 Competition on scientific literature parsing - Task B: Table recognition](https://icdar2021.org/competitions/competition-on-scientific-literature-parsing/) (Team Name: LTIAYN = Kaen Context).
- Dataset: [PubTabNet](https://github.com/ibm-aur-nlp/PubTabNet)
- Metric: [Tree-Edit-Distance-based Similarity(TEDS)](https://github.com/ibm-aur-nlp/PubTabNet/tree/master/src)
- Baseline: [Image-based table recognition: data, model, and evaluation](https://arxiv.org/abs/1911.10683)
## 0. Before Training
1. change the prefined data directory '/data/private/datasets/pubtabnet' to your own data directory in 'processing_pubtabnet.py', 'configs/linear_transformer.yaml'
2. `python processing_pubtabnet.py`
## 1. Training
``` bash
python train.py model_dir=base
```
## 2. After Training
1. inference
```bash
python inference.py -m "./outputs/base/" -i "/data/private/datasets/pubtabnet/val/" -o "./results/val1" -nt 16 -ni 0 -na 20
python inference.py -m "./outputs/base/" -i "/data/private/datasets/pubtabnet/val/" -o "./results/val1" -nt 16 -ni 1 -na 20
...
python inference.py -m "./outputs/base/" -i "/data/private/datasets/pubtabnet/val/" -o "./results/val1" -nt 16 -ni 15 -na 20
```
2. evalution
```bash
python score.py
```