Open Lwt-diamond opened 4 months ago
hello, i have ever tested in CHAMELEON dataset, there is the result: CHAMELEON : {'Smeasure': 0.9073958667751745, 'wFmeasure': 0.875725191127773, 'MAE': 0.019581592995290175, 'meanEm': 0.9666163166907054} Since the CHAMELEON dataset contains only 76 images, the evaluation results are not very stable, so many papers currently no longer include this evaluation metric.
Thank you versy much! Could I get the prediction of CHAMELEON dataset?Your work are better than others' result of CHAMELEON dataset in fact.
I'm impressed with your work CamoDiffusion. I'm curious why you didn't include results for the CHAMELEON dataset in your experiments. I think it would be valuable to see how the model would perform on this dataset, as it's a challenging benchmark for named entity recognition. Have you ever test on CHAMELEON ?