For the fastest local setup of this model, Docker is the best choice.
Simply follow the directions outlined below.
>
The loader auto-caches the model archive (several GBs included).
To guarantee smooth performance, the installation process auto-selects the best possible options for your PC.
The **chandra-ocr-2** model delivers *state-of-the-art* optical character recognition with unprecedented accuracy across diverse document types. It leverages a deep convolutional neural network architecture combined with attention mechanisms to capture both fine-grained character shapes and contextual layout cues. The model supports a wide range of languages and scripts, making it suitable for global enterprise workflows. Performance benchmarks show a character error rate below 0.5% on standard benchmarks, outperforming previous generations by over 15%. Integration is streamlined via a lightweight API that processes images in *real-time* with minimal hardware requirements.
| Specification | Value |
|---|---|
| Model size | 210 MB |
| Supported languages | 100 |
| Input resolution | 2048 × 3072 px |
| Processing speed | > 30 fps |
