en_PP-OCRv4_mobile_rec
PaddlePaddle/en_PP-OCRv4_mobile_rec
A popular open image-to-text model, with 24.6K downloads a month. gigarouter benchmarks and hosts it as an OpenAI-compatible API.
about this model
en_PP-OCRv4_mobile_rec is an English-specific text line recognition model from the PP-OCRv4_rec series, developed by PaddleOCR. It performs image-to-text conversion, outputting recognized text strings from input images containing English and numeric characters.
Key Strengths
- Ultra-lightweight: only 6.8 MB storage size, suitable for mobile and edge deployment.
- Line-level evaluation: any incorrect character marks the entire line as wrong, ensuring practical accuracy.
- Trained specifically for English and numeric character recognition.
Benchmark Performance
| Model | Recognition Avg Accuracy (%) | Model Storage (M) | Description |
|---|---|---|---|
| en_PP-OCRv4_mobile_rec | 70.39 | 6.8 | Ultra-lightweight English recognition model, supports English and numeric characters. |
Accuracy measured under strict line-level scoring.
Use Cases
Best suited for English text extraction from images where model size and inference speed are critical, such as mobile OCR, edge devices, or real-time document scanning pipelines.
Visual Results
Example recognition output:
Pipeline example (OCR with detection and recognition):
For further details, see the PaddleOCR repository and documentation.
We're benchmarking and onboarding en_PP-OCRv4_mobile_rec as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.