TableVLM: Multi-modal Pre-training for Table Structure Recognition
38719Knowledge Triples
| Source Entity | Relationship | Target Entity |
|---|---|---|
|
TableVLM
Method |
trainedOn |
ComplexTable
Material |