List your company
ResearchFormerly WizWand

Object Shape Prediction on ViComTe zero-shot

67.3Accuracy (zero-shot)LaMI
19.66832.03444.456.766Jun 19, 2024
Updated 6mo ago

Evaluation results

MethodLinks
LaMILLM Category=Masked Language Models, Image Augmentation Strategy=Generation2024.0667.3
LaMILLM Category=Causal Language Models, Image Augmentation Strategy=Generation2024.0666.8
Z-LaVILLM Category=Causal Language Models, Image Augmentation Strategy=Retrieval2024.0664.4
VaLMLLM Category=Causal Language Models, Image Augmentation Strategy=Retrieval, k=42024.0662.8
VaLMLLM Category=Causal Language Models, Image Augmentation Strategy=Retrieval, k=82024.0659.4
GPT-2LLM Category=Causal Language Models, Evaluation Protocol=Zero-shot2024.0644.5
VokenizationLLM Category=Masked Language Models, Image Augmentation Strategy=Retrieval2024.0643.2
LIVELLM Category=Causal Language Models, Image Augmentation Strategy=Generation2024.0641.5
GPT-2 (FT)LLM Category=Causal Language Models, Evaluation Protocol=Fine-tuned2024.0639.3
BERTLLM Category=Masked Language Models, Evaluation Protocol=Zero-shot2024.0628.1
BERT (FT)LLM Category=Masked Language Models, Evaluation Protocol=Fine-tuned2024.0621.5