Multi-Type-TD-TSR
adaptnlp
Our great sponsors
Multi-Type-TD-TSR | adaptnlp | |
---|---|---|
4 | 2 | |
236 | 414 | |
- | 0.0% | |
0.0 | 0.0 | |
over 1 year ago | over 2 years ago | |
Jupyter Notebook | Jupyter Notebook | |
MIT License | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
Multi-Type-TD-TSR
-
[D] Getting super-level table extraction
Recently, I've been researching extracting tables from image documents. First I tried with pdfs, however, the data extraction libraries like camelot are inconsistent. I found a deep learning model called CascadeTabNet. The detection results are okay but cell recognition is poor. I even found Multi-Type-TD-TSR for table extraction. It uses image processing techniques to find the grids. It performs well on structured and bordered tables. However, it messes up if the cell is not properly aligned. Even if extraction is successful, aggregation of multi-line cells, i.e post-processing, is not very obvious.
- Multi-Type-TD-TSR - Extracting Tables from Document Images using a Multi-stage Pipeline for Table Detection and Table Structure Recognition (State of the art approach for table structure recognition published on KI2021 - 44th German Conference on Artificial Intelligence)
-
Multi-Type-TD-TSR - Extracting Tables from Document Images using a Multi-stage Pipeline for Table Detection and Table Structure Recognition: from OCR to Structured Table Representations
Check it out on my Github: https://github.com/Psarpei/Multi-Type-TD-TSR
- Multi-Type-TD-TSR - Extracting Tables from Document Images using a Multi-stage Pipeline for Table Detection and Table Structure Recognition: from OCR to Structured Table Representations (New state-of-the-art approach for table structure recognition)
adaptnlp
-
Tools to use for Semantic-searching Question Answering System
Check out adaptnlp
-
Case Sensitivity using HuggingFace & Google's T5 model (base)
Yes, there are capitals in the tokenizer vocabulary of t5-base and t5-small, so both support capitalization. A few days ago I was using t5-small through adaptnlp for extractive summarization and capitalization was working fine (https://github.com/Novetta/adaptnlp). AdaptNLP is basically just a transformers wrapper, so if you can't figure out a solution, you could just dissect their source code.
What are some alternatives?
donut - Official Implementation of OCR-free Document Understanding Transformer (Donut) and Synthetic Document Generator (SynthDoG), ECCV 2022
Basic-UI-for-GPT-J-6B-with-low-vram - A repository to run gpt-j-6b on low vram machines (4.2 gb minimum vram for 2000 token context, 3.5 gb for 1000 token context). Model loading takes 12gb free ram.
MetalTranslate - Customizable machine translation in C++
keytotext - Keywords to Sentences
Recognition-of-logical-document-structures - First approach for recognizing logical document structures like texts, sentences, segments, words, chars and sentence/segment depth based on recurrent neural network grammars.
fastai - The fastai deep learning library
CascadeTabNet - This repository contains the code and implementation details of the CascadeTabNet paper "CascadeTabNet: An approach for end to end table detection and structure recognition from image-based documents"
browser-ml-inference - Edge Inference in Browser with Transformer NLP model
oemer - End-to-end Optical Music Recognition (OMR) system. Transcribe phone-taken music sheet image into MusicXML, which can be edited and converted to MIDI.
gector - Official implementation of the papers "GECToR – Grammatical Error Correction: Tag, Not Rewrite" (BEA-20) and "Text Simplification by Tagging" (BEA-21)
Real-time-Object-Detection-for-Autonomous-Driving-using-Deep-Learning - My Computer Vision project from my Computer Vision Course (Fall 2020) at Goethe University Frankfurt, Germany. Performance comparison between state-of-the-art Object Detection algorithms YOLO and Faster R-CNN based on the Berkeley DeepDrive (BDD100K) Dataset.
Transformers-Tutorials - This repository contains demos I made with the Transformers library by HuggingFace.