| | --- |
| | license: apache-2.0 |
| | tags: |
| | - ernie |
| | - ernie-4.5 |
| | - document-ai |
| | - paddlepaddle |
| | - hackathon |
| | --- |
| | ernie |
| | ernie-4.5 |
| | document-ai |
| | ocr |
| | paddlepaddle |
| | paddleocr-vl |
| | multimodal |
| | hackathon |
| | submission |
| |
|
| | from transformers import AutoModelForCausalLM, AutoTokenizer |
| |
|
| | tokenizer = AutoTokenizer.from_pretrained("Zenieverse/DocuMind-ERNIE4.5-Document-Reasoning") |
| | model = AutoModelForCausalLM.from_pretrained("Zenieverse/DocuMind-ERNIE4.5-Document-Reasoning") |
| |
|
| | ## Dataset |
| | https://huggingface.co/datasets/Zenieverse/DocuMind-ERNIE4.5-Dataset |
| |
|
| |
|
| |
|
| | # DocuMind – ERNIE 4.5 Document Reasoning |
| |
|
| | Fine-tuned ERNIE 4.5 model for document semantic understanding, |
| | entity extraction, and structured reasoning from OCR text. |
| |
|
| | ## Use Case |
| | - Contract analysis |
| | - Invoice processing |
| | - Document intelligence |
| |
|
| | ## Training |
| | - Base model: ERNIE 4.5 Open-Source |
| | - Fine-tuning: Unsloth / LLaMA-Factory |
| | --- |
| | license: apache-2.0 |
| | --- |
| |
|
