Data-to-Text Generation on ToTTo
52.28BLEURe-Table-7B-rerank
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Re-Table-7B-rerankLLM=Mistral 7B, Method Variant=rerank2026.02 | 52.28 | — | — | — | — | — | — | — | — | — | |
| Re-Table-7B-retrievalLLM=Mistral 7B, Method Variant=retrieval2026.02 | 50.3 | — | — | — | — | — | — | — | — | — | |
| T5-3BModel size=3B2020.05 | 49.5 | — | — | — | — | — | — | — | — | 58.4 | |
| BERT-to-BERT2020.05 | 44 | — | — | — | — | — | — | — | — | 52.6 | |
| PGen2020.05 | 41.6 | — | — | — | — | — | — | — | — | 51.6 | |
| Table-LLaVA 7BLLM=Vicuna-1.5 7B2026.02 | 23 | — | — | — | — | — | — | — | — | — | |
| stnd2021.02 | 21.27 | 56.6 | 25.16 | 29.71 | 55.9 | 26.5 | 53 | 20.6 | 2.824 | — | |
| TableLlama+textLLM=Llama-2 7B, Representation=OCR textual table2026.02 | 20.77 | — | — | — | — | — | — | — | — | — | |
| stnd_filteredfiltered=true2021.02 | 19.48 | 56.69 | 22.31 | 27.18 | 29.4 | 8.8 | 70.6 | 50 | 2.706 | — | |
| MBD2021.02 | 18.35 | 50.44 | 25.25 | 28.25 | 91.2 | 50 | 85.3 | 55.9 | 3.613 | — | |
| halwo2021.02 | 17.06 | 77.64 | 22.65 | 29.38 | 61.7 | 38.2 | 61.8 | 32.4 | 2.725 | — | |
| LLaVA v1.5LLM=Vicuna-1.5 7B2026.02 | 6.4 | — | — | — | — | — | — | — | — | — | |
| Llama2+textLLM=Llama-2 7B, Representation=OCR textual table2026.02 | 6.2 | — | — | — | — | — | — | — | — | — | |
| LLaVA v1.6LLM=Vicuna-1.5 7B2026.02 | 5.87 | — | — | — | — | — | — | — | — | — | |
| BLIP2LLM=Flan-T5 7B2026.02 | 4.3 | — | — | — | — | — | — | — | — | — | |
| MonkeyLLM=Qwen 7B2026.02 | 3.5 | — | — | — | — | — | — | — | — | — | |
| Vary-toyLLM=Qwen 1.8B2026.02 | 0.7 | — | — | — | — | — | — | — | — | — | |
| MiniGPT-4LLM=Vicuna 7B2026.02 | 0.2 | — | — | — | — | — | — | — | — | — | |
| Gold (noisy)2021.02 | — | — | — | — | 97.1 | 97.1 | 91.2 | 79.4 | 3.618 | — |