DePlot: One-shot visual language reasoning by plot-to-table translation

Open paper
Benchmark Model Rank Results
chart-question-answering-on-plotqaDePlot+FlanPaLM+Codex (PoT Self-Consistency)#21:1 Accuracy: 66.6
factual-inconsistency-detection-in-chart-captioning-on-chocolate-llmDePlot + GPT-4#2Kendall's Tau-c: 0.117
factual-inconsistency-detection-in-chart-captioning-on-chocolate-lvlmDePlot + GPT-4#2Kendall's Tau-c: 0.129
factual-inconsistency-detection-in-chart-captioning-on-chocolate-ftDePlot + GPT-4#3Kendall's Tau-c: 0.109
chart-question-answering-on-chartqaDePlot+FlanPaLM+Codex (PoT Self-Consistency)#721:1 Accuracy: 79.3
chart-question-answering-on-chartqaDePlot+Codex (PoT Self-Consistency)#781:1 Accuracy: 76.7
chart-question-answering-on-chartqaDePlot+FlanPaLM (Self-Consistency)#941:1 Accuracy: 70.5
chart-question-answering-on-chartqaDePlot+FlanPaLM (CoT)#1011:1 Accuracy: 67.3
chart-question-answering-on-chartqaDePlot+GPT3 (Self-Consistency)#1131:1 Accuracy: 42.3
chart-question-answering-on-chartqaDePlot+GPT3 (CoT)#1141:1 Accuracy: 36.9

Rank counts only results with a code link.