ResearchBenchmarksSubject-driven Text-to-Image Generation on 3D Icons (test)Subject-driven Text-to-Image Generation on 3D Icons (test)Follow33.14CLIP I2T ScoreBi-LoRA32.401632.593332.78532.9767Aug 27, 2025Evaluation resultsMethodMethodLinksCLIP I2T ScoreCLIP T2T ScoreBi-LoRABackbone Model=SDXL, F...Backbone Model=SDXL, Finetuning Protocol=Dreambooth2025.0833.1446.79LoRABackbone Model=SDXL, F...Backbone Model=SDXL, Finetuning Protocol=Dreambooth2025.0832.4342.27