richardyoung/ lightonocr-3-4b:Q6_K

56 1 week ago

LightOnOCR-3-4B: LightOn's most accurate OCR model. Page to Markdown, bounding boxes, charts to tables. Images work.

vision 
ollama run richardyoung/lightonocr-3-4b:Q6_K

Details

1 week ago

70ca464ec201 · 4.1GB

model
qwen35
·
4.21B
·
Q6_K
projector
clip
·
334M
·
F16
params
{ "temperature": 0.2, "top_p": 0.9 }

Readme

LightOnOCR-3-4B

LightOnOCR-3-4B: LightOn’s most accurate OCR model. Page to Markdown, bounding boxes, charts to tables. Images work.

🚀 Overview

GGUF build of lightonai/LightOnOCR-3-4B, LightOn’s most accurate LightOnOCR-3 model (Qwen3.5 vision-language architecture): end-to-end OCR with layout and visual understanding. It ships with its vision projector, so image input works in Ollama.

Prompts: send the image with an empty prompt for plain transcription (Markdown), or with the prompt grounding to get every block with a label and a 0–1000 bounding box, image descriptions and chart data as tables. LightOn notes that other instructions are out of distribution. Recommended sampling: temperature 0.2, top_p 0.9 (set in the Ollama tags).

🎯 Key Features

  • Most accurate of the LightOnOCR-3 family — LightOn’s recommendation for most OCR tasks
  • Two modes: plain transcription, or grounding with labels and bounding boxes for every block
  • Visual understanding: short descriptions for images, charts turned into tables of their data
  • Tables, receipts, forms, multi-column layouts and math notation
  • Apache-2.0: research and commercial use

🏷️ Available Versions

Tag Size (+ 0.67 GB projector) BPW Notes
IQ3_M 2.2 GB 3.66 Smallest
IQ4_XS 2.5 GB 4.25 Great quality-size balance
latest / Q4_K_M 2.7 GB 4.85 Recommended
Q5_K_M 3.1 GB 5.68 Higher quality
Q6_K 3.5 GB 6.56 Very high quality
Q8_0 4.5 GB 8.5 Near-lossless

💻 Quick Start

ollama run richardyoung/lightonocr-3-4b
# then drag an image file into the prompt, or send it in the API "images" field

🛠️ Use Cases

  • Digitizing scanned documents and PDFs locally
  • Layout-aware extraction with bounding boxes
  • Chart and table data extraction for RAG

✅ Verified

Tested through Ollama before upload (Q4_K_M and Q8_0) on a synthetic invoice with an empty prompt; Q4_K_M returned: ‘INVOICE 4471 / Date: 2026-09-28 / Widgets x3 ……. \(42.00 / TOTAL: \)86.20’.

🔧 Technical Details

  • Base Model: lightonai/LightOnOCR-3-4B
  • Parameters: 4.2B
  • Context Length: 256K tokens
  • Quantization: GGUF via llama.cpp with an importance matrix (calibration_datav3); vision projector at F16; converted with --no-mtp
  • License: Apache-2.0 (inherited from the base model)

🙏 Acknowledgments

  • Base Model: LightOn
  • Quantization: llama.cpp

Built & maintained by Richard Young · DeepNeuro