SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion
Open paper
Benchmark
Model
Rank
Results
optical-character-recognition-ocr-on-ocrbench
SmolDocling-256M-preview
#70
Score: 338
vision-language-understanding-on-mmstar
SmolDocling
#98
Accuracy: 17.0
Rank counts only results with a code link.