ResearchBenchmarksText-to-3D Generation on 43 prompts and 50 views (evaluation set)Text-to-3D Generation on 43 prompts and 50 views (evaluation set)Follow33.47CLIP ScoreSDI29.663630.651831.6432.6282May 24, 2024Evaluation resultsMethodMethodLinksCLIP ScoreCLIP IQA QualityCLIP IQA SharpnessCLIP IQA RealismIRDiversity ScoreGeneration Time (min)VRAM Usage (GB)SDISteps=10kSteps=10k2024.0533.47829897-1.184.7119—VSDSteps=25kSteps=25k2024.0533.31779896-1.1723.2334—HIFASteps=25kSteps=25k2024.0532.8819897-1.164.7235—ESDSteps=25kSteps=25k2024.0532.79779897-1.214331—SJCSteps=10kSteps=10k2024.0530.39769998-1.7611.613—SDSSteps=10kSteps=10k2024.0529.81769998-1.5118.666—