# Financial Document Processing

A vision-language model specifically designed for Thai-English real-world document parsing, based on the Qwen2.5-VL-Instruction framework

Qwen2-VL-2B-OCR is an OCR model fine-tuned based on unsloth/Qwen2-VL-2B-Instruct, specializing in extracting complete text from documents, tables, and payroll images.

A specialized image-to-text model for the financial domain, supporting English text recognition, primarily used for processing image content in financial documents.

Donut is a visual document understanding model based on Swin Transformer, specifically fine-tuned for the CORD dataset, capable of extracting structured text information from images.

DETR architecture model trained on PubTables1M and FinTabNet datasets, specialized for table structure recognition tasks

Multilingual invoice processing model optimized based on Donut architecture, capable of extracting key invoice fields

A line-level document understanding model fine-tuned on the DocLayNet dataset based on the LayoutXLM base model, supporting multilingual document layout analysis and token classification.

This is a document understanding model specifically designed for analyzing document layout and content, performing token classification tasks at the paragraph level.

A line-level document understanding model fine-tuned based on LiLT and DocLayNet dataset, supporting multilingual document layout analysis

This is an image classification model generated based on HuggingPics, specifically designed to recognize and classify images into four categories: receipts, paper, invoices, and documents.

This is a Donut model fine-tuned on the CORD-v2 dataset, designed for image-to-text tasks, achieving an average accuracy of 0.901.

An invoice recognition model fine-tuned based on LayoutLMv3-base, trained on the wild_receipt dataset, excelling in extracting structured information from invoices.

A binary classification model based on ViT architecture for identifying whether an image is a receipt/invoice

This is a fine-tuned multimodal LayoutLM model for document question answering tasks, capable of understanding both text and layout information in documents to answer questions.

A document understanding model fine-tuned based on LayoutLMv3-base, specifically designed for named entity recognition tasks on the CORD dataset