Nexora OCR v0.1
Nexora OCR v0.1 is an experimental OCR model designed for text extraction from images. It is fine-tuned from Alibaba's Qwen3.5 0.8B and trained to analyze document and scene-text images while returning accurate transcriptions.
Nexora OCR v0.1 is the first public release in the Nexora OCR series and focuses on establishing a lightweight, efficient baseline for optical character recognition tasks.
The model was trained on a custom OCR instruction dataset prepared for Nexora OCR.
Table of Contents
- Overview
- The Nexora OCR Family
- What's New
- Model Details
- Capabilities
- Limitations
- Intended Use
- Architecture & Training
- Output Format
- Usage
- Evaluation
- Risks & Considerations
- Community & Support
- License
- Acknowledgements
- About Nexora
- Image Credits
Overview
Nexora OCR v0.1 is an experimental OCR model designed for text extraction from images.
The model receives an image together with an instruction and generates a textual transcription of the visible text content.
Nexora OCR v0.1 was fine-tuned from Qwen3.5 0.8B using a supervised fine-tuning workflow.
This release focuses on establishing a strong baseline for character-level accuracy and word-level recognition across common document and scene-text scenarios.
The Nexora OCR Family
Nexora OCR is an experimental model series focused on practical and efficient OCR models for document understanding, text extraction, and scene-text recognition.
The project explores the use of compact vision-language foundation models for specialized OCR workloads, including document digitization, receipt and form parsing, handwritten text recognition experiments, and lightweight OCR pipelines.
Nexora OCR v0.1 is the initial public release of the project.
Future releases may improve character accuracy, word error rate, exact-match performance, robustness to low-quality images, and inference efficiency.
What's New
Nexora OCR v0.1
Nexora OCR v0.1 introduces the first public version of the Nexora OCR model family with a focus on:
- Establishing a lightweight OCR baseline using Qwen3.5 0.8B
- Achieving reasonable character-level accuracy on a held-out benchmark
- Supporting prototype OCR and document-understanding systems
Specific dataset composition, augmentation strategies, and training hyperparameters are not publicly disclosed at this time.
Model Details
| Property | Details |
|---|---|
| Model Name | Nexora OCR v0.1 |
| Model Family | Nexora OCR |
| Model Type | Multimodal Image-Text Model (OCR) |
| Base Model | Qwen/Qwen3.5-0.8B |
| Fine-tuning Base | Qwen/Qwen3.5-0.8B |
| Task | Optical Character Recognition (OCR) |
| Input | Image + Text Instruction |
| Output | Transcribed Text |
| Fine-tuning Method | Supervised Fine-Tuning (SFT) |
| Training Framework | Transformers / Custom SFT |
| Release Status | Experimental |
| Version | v0.1 |
Capabilities
Nexora OCR v0.1 is designed to:
- Extract printed and handwritten text from images
- Process document pages, forms, receipts, and scene-text images
- Handle multi-line and multi-word text regions
- Preserve basic layout and line breaks where possible
- Support prototype OCR and document-understanding pipelines
- Work with natural-language instructions (e.g., "Transcribe all text in this image")
- Run efficiently on modest hardware due to its 0.8B parameter size
Example output:
INVOICE
Date: 2025-12-10
Total: $42.50
The exact output format may vary depending on the prompt and image content. Applications should normalize, post-process, and validate the transcribed text before using it programmatically.
Limitations
Nexora OCR v0.1 is an experimental model and has important limitations.
Benchmark Limitations
Benchmark results should be interpreted according to the evaluation dataset and methodology used.
A benchmark result does not automatically represent real-world accuracy or generalization to unseen document types, languages, or image conditions.
Future evaluations should use larger, more diverse, independently held-out datasets covering multiple languages, fonts, layouts, and image qualities.
Image Quality Limitations
Performance may degrade on:
- Very low-resolution or heavily compressed images
- Strong motion blur, defocus, or noise
- Extreme perspective distortion or occlusion
- Unusual fonts, stylized text, or decorative typography
Different levels of compression, resizing, cropping, or noise may produce different transcription quality.
Language & Script Limitations
The current release is primarily optimized for English text.
Performance on other languages, scripts, or mixed-language documents is not guaranteed and may be significantly lower.
Layout & Structure Limitations
The model focuses on text transcription rather than full document structure understanding.
Complex tables, multi-column layouts, mathematical formulas, or highly structured forms may not be perfectly preserved in the output.
Applications requiring structured extraction (e.g., key-value pairs, table cells) should add post-processing or dedicated parsing layers.
Output Consistency
The model generates free-form text, so responses may occasionally include:
- Extra explanations or commentary
- Minor formatting differences (spacing, line breaks)
- Occasional hallucinated or missing characters/words
Applications should not assume perfectly clean or deterministic output.
Use text normalization, schema validation (where applicable), fallback handling, and error logging in production-like prototypes.
Confidence & Error Metrics
The model does not natively output calibrated confidence scores per character or word.
Character Error Rate (CER) and Word Error Rate (WER) should be estimated externally using reference transcriptions.
Dataset Limitations
The training dataset may contain:
- Ambiguous or noisy labels
- Limited visual diversity in fonts, layouts, or image conditions
- Uneven representation of document types
- Images that do not fully reflect real-world distributions
Generalization
Performance may vary depending on:
- Image type (scan, photo, screenshot, natural scene)
- Resolution, quality, cropping, framing
- Lighting, shadows, glare, or reflections
- Prompt wording and instruction style
- Differences between training and deployment data
Production Readiness
Nexora OCR v0.1 is not production-ready.
It should not be used as the sole transcription component in safety-critical systems, legal documents, medical records, or high-consequence workflows without additional validation.
Human review and secondary verification are recommended for important decisions.
Intended Use
Supported Use Cases
Nexora OCR v0.1 is intended for:
- AI and OCR research
- Document digitization experiments
- Prototype OCR systems
- Educational experimentation
- Benchmarking compact vision-language models for OCR
- Development of text-extraction pipelines
- Research into lightweight OCR inference
- Testing OCR workflows on common document types
- Exploring structured and free-form text outputs from multimodal models
- Researching model robustness against image quality variations
Out-of-Scope Use Cases
The model is not intended for:
- Fully automated transcription without additional safeguards
- Safety-critical or high-stakes document processing
- Applications requiring guaranteed character-level accuracy
- Systems where incorrect transcriptions could cause significant harm
- Automatic enforcement actions without human review or secondary validation
- Medical, legal, financial, or law-enforcement document processing without strong additional controls
Developers should combine model predictions with:
- Policy rules and validation checks
- Uncertainty handling and threshold tuning
- Human review for uncertain or critical cases
- Periodic evaluation on representative deployment data
Architecture & Training
Nexora OCR v0.1 is based on Qwen3.5 0.8B, a compact multimodal vision-language model designed to process image and text inputs.
Training Configuration
| Property | Details |
|---|---|
| Base Model | Qwen/Qwen3.5-0.8B |
| Fine-tuning Base | Qwen/Qwen3.5-0.8B |
| Fine-tuning Method | Supervised Fine-Tuning |
| Training Framework | Transformers / Custom SFT |
| Task | Optical Character Recognition (OCR) |
| Input Modality | Image + Text Instruction |
| Output | Transcribed Text |
| Training Dataset | Custom Nexora OCR instruction dataset |
| Model Version | v0.1 |
Nexora OCR v0.1 was fine-tuned from Qwen/Qwen3.5-0.8B.
A supervised fine-tuning workflow was used with a custom multimodal instruction dataset prepared for Nexora OCR.
The dataset focuses on common document and scene-text scenarios to establish a practical OCR baseline.
Output Format
The intended output is plain text containing the transcribed content, for example:
RECEIPT
Store: Example Mart
Date: 2025-11-03
Items:
- Coffee $3.50
- Sandwich $6.00
Total: $9.50
Because the model generates free-form text, responses may occasionally include additional explanation, formatting differences, or minor errors.
A safer integration should:
- Extract the generated response
- Apply text normalization (trimming, whitespace cleanup)
- Optionally parse into structured fields if needed
- Apply fallback behavior when output is clearly malformed
- Route uncertain or critical cases for additional review
Usage
The recommended way to use Nexora OCR v0.1 is via the nexora-ai Python package.
Installation
pip install nexora-ai
Basic Example
from nexora.ocr import NexoraOCR
ocr = NexoraOCR()
result = ocr.read("image.png")
print(result)
result is a string containing the transcribed text. You can further process it (e.g., split into lines, parse fields, or store in a database) depending on your application.
Notes
- Ensure the image path is correct and the file is readable.
- For best results, use clear, well-lit images with minimal distortion.
- The model is experimental; validate outputs before using them in critical workflows.
Evaluation
Nexora OCR v0.1 should be evaluated using both regression testing and independent held-out evaluation.
Internal Regression Benchmark
| Metric | Result |
|---|---|
| Total Samples | 100 |
| Total Character Accuracy | 80.77% |
| Total CER (Character Error Rate) | 19.23% |
| Exact Match (EM) | 27.00% |
| Average WER (Word Error Rate) | 43.49% |
| Evaluation Type | OCR transcription |
| Benchmark Purpose | Internal regression testing |
| Release | v0.1 |
These results are intended for release-to-release comparison and should not be interpreted as independent estimates of real-world performance across all document types and languages.
Future releases should be evaluated using:
- Larger and more diverse held-out datasets
- Per-document-type metrics (receipts, forms, IDs, scene text, etc.)
- Additional metrics such as precision/recall for key fields, line-level accuracy, and calibration error where applicable
Risks & Considerations
Developers integrating Nexora OCR v0.1 should account for the possibility of character-level transcription errors, word-level misrecognitions, and insertion or deletion errors. Performance may be affected by image quality, compression, font style, text size, layout complexity, lighting, blur, glare, and partially occluded text. The model may also perform differently on unseen image types or scripts, and its output formatting may not always be consistent. Dataset bias toward certain document types or languages and sensitivity to prompt wording may also affect results.
The model should not be used as the sole transcription mechanism in an application. Robust deployments should include appropriate validation and rule-based checks, explicit transcription policies and thresholds, logging and monitoring of error patterns, safe handling of malformed or unexpected outputs, and periodic evaluation on representative deployment data. Human review should be considered for uncertain or critical cases.
Community & Support
Nexora OCR is an experimental project.
Community feedback, testing, and benchmark contributions are welcome.
For information about future releases, visit the Nexora OCR model family on Hugging Face.
License
Nexora OCR v0.1 is based on Qwen3.5 0.8B.
The model is released under the Apache 2.0 license, subject to the applicable terms and conditions of the underlying Qwen model.
Acknowledgements
Nexora OCR v0.1 builds upon Qwen3.5 0.8B by the Qwen team at Alibaba.
We thank the Qwen team for their open foundation models and the broader open-source AI and machine-learning community for the tools, research, and resources that make projects like Nexora OCR possible.
About Nexora
Nexora OCR is an experimental OCR model project focused on developing practical and efficient multimodal AI systems for document understanding and text extraction.
Nexora OCR v0.1 represents the first public release of the project and establishes a baseline for future improvements in character accuracy, word error rate, exact-match performance, and robustness.
Image Credits
The background image used in the Nexora OCR v0.1 model artwork was photographed by J Blueberry and is sourced from Pixabay.
- Photographer: J Blueberry
- Image: Landscape, Rural, Agriculture, Field
- Source: Pixabay
- License: Pixabay Content License
- Downloads last month
- -