Downloads · 30 days
7
13% of all-time downloads
OmniParse-v1
ManvithReddy/OmniParse-v1
OmniParse-v1 is a image-to-text model from ManvithReddy. Use it when you need a caption or text from an image. The card lists the license as apache-2.0.
OmniParse-v1 is a powerful Vision-Language Model (VLM) customized for Intelligent Document Processing (IDP). It is designed to understand complex document layouts, such as dense mark sheets, invoices, and unstructured…
All-time downloads
53
Public
Parameters
959M
2.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.9 GB · 89%
At a glance
- Task
- Image-to-Text
- License
- apache-2.0
- Model type
- paddleocr_vl
- Access
- Public
- Created
- Jul 5, 2026
- Updated
- Jul 5, 2026
- SHA
- 410f7fd0
Model card
From the Hugging Face model README
OmniParse-v1
OmniParse-v1 is a powerful Vision-Language Model (VLM) customized for Intelligent Document Processing (IDP). It is designed to understand complex document layouts, such as dense mark sheets, invoices, and unstructured PDFs, and natively extract beautifully structured Markdown.
🚀 Features
- Table Recognition: Perfectly extracts grids and nested tables into native Markdown format.
- Formula & Chart Parsing: Capable of understanding and extracting complex visual data.
- Robust OCR: Reads dense, multi-column text effortlessly.
💻 Usage
This model is intended to be used with the OmniParse Python/Gradio pipeline.
from transformers import AutoModelForCausalLM, AutoProcessor
import torch
model_id = "ManvithReddy/OmniParse-v1"
# Load the processor and model
processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
model_id,
trust_remote_code=True,
torch_dtype=torch.bfloat16
).eval()
📖 Origins
This model is derived from the PaddleOCR-VL architecture and customized for local, offline document extraction tasks.
- Task
- Image-to-Text
- Type
- paddleocr_vl
- License
- apache-2.0
- Languages
- en
- Created
- Jul 5, 2026
- Updated
- Jul 5, 2026