OpenOCR Data: Scene Text Recognition Dataset for SVTRv2 Model

A dataset associated with the SVTRv2 model, which uses Connectionist Temporal Classification (CTC) for scene text recognition. The dataset was uploaded by author 'topdu' to Hugging Face and was last updated on September 6, 2025. Its specific contents, scale, and license are not detailed in the provided metadata.

Use Cases

• Training scene text recognition models based on the CTC architecture mentioned in the description
• Benchmarking the performance of CTC-based STR models against encoder-decoder models
• Fine-tuning visual models for optical character recognition in scene images

Strengths

• Dataset is associated with a published model (SVTRv2) and research paper, suggesting a research-grade purpose
• Last update timestamp is explicitly provided: 2025-09-06 02:08:07

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download
• Column-level documentation is absent; field semantics must be inferred after download
• Row count, file formats, and license are unknown, which may limit suitability assessment

Provenance

Source

Hugging Face user 'topdu'

Collection Method

Likely collected or generated for training and benchmarking the SVTRv2 model, as referenced in the description.

Time Range

null

Freshness

Last updated 2025-09 02:08:07; freshness should be verified

Geography

null