mychen76/mistral_ocr2json_v3_chatml
The mychen76/mistral_ocr2json_v3_chatml is a 7 billion parameter Mistral-based instruction-tuned model, fine-tuned from unsloth/mistral-7b-instruct-v0.2-bnb-4bit, with a 4096 token context length. This model specializes in converting OCR results from receipt images into structured JSON objects. It is specifically designed for accurate parsing and recognition of POS receipt data, making it highly effective for automated data extraction from receipts.
Loading preview...
Overview
This model, mychen76/mistral_ocr2json_v3_chatml, is a 7 billion parameter instruction-tuned variant of the Mistral architecture, specifically fine-tuned from unsloth/mistral-7b-instruct-v0.2-bnb-4bit. It is designed to process raw OCR text data from point-of-sale (POS) receipts and convert it into a structured JSON format. The model operates within a 4096 token context length and utilizes the chatml format for interactions.
Key Capabilities
- OCR to JSON Conversion: Transforms unstructured OCR output from receipts into well-formed JSON objects.
- Receipt Data Extraction: Accurately identifies and extracts critical information such as store name, address, invoice details, itemized lists (item name, quantity, price), subtotal, taxes, total, payment information, and customer details.
- Structured Output: Ensures the output is a valid JSON object, suitable for automated processing and database integration.
Good For
- Automated Receipt Processing: Ideal for applications requiring the digital capture and structuring of receipt data, such as expense management systems or financial tracking tools.
- Data Entry Automation: Reduces manual data entry by converting physical receipt information into a machine-readable format.
- Developers working with OCR: Provides a specialized tool for post-processing OCR results from receipts, enhancing data accuracy and usability.