cpy0029
cpy0029
您好，读了您Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection 一文后，我有一个疑问。 请问如果将BERT换成CLIP，计算所消耗的资源是否能够下降？如果换成CLIP之后，精度也下降了，那么大约下降多少？ 谢谢
Reading PH2 ['/home/chen/datasets/PH2/images/IMD226.png', '/home/chen/datasets/PH2/images/IMD156.png', '/home/chen/datasets/PH2/images/IMD384.png', '/home/chen/datasets/PH2/images/IMD048.png', '/home/chen/datasets/PH2/images/IMD133.png', '/home/chen/datasets/PH2/images/IMD436.png', '/home/chen/datasets/PH2/images/IMD360.png', '/home/chen/datasets/PH2/images/IMD390.png', '/home/chen/datasets/PH2/images/IMD331.png', '/home/chen/datasets/PH2/images/IMD144.png', '/home/chen/datasets/PH2/images/IMD063.png', '/home/chen/datasets/PH2/images/IMD423.png', '/home/chen/datasets/PH2/images/IMD152.png', '/home/chen/datasets/PH2/images/IMD365.png', '/home/chen/datasets/PH2/images/IMD389.png', '/home/chen/datasets/PH2/images/IMD356.png', '/home/chen/datasets/PH2/images/IMD041.png', '/home/chen/datasets/PH2/images/IMD198.png', '/home/chen/datasets/PH2/images/IMD427.png', '/home/chen/datasets/PH2/images/IMD418.png', '/home/chen/datasets/PH2/images/IMD150.png', '/home/chen/datasets/PH2/images/IMD058.png', '/home/chen/datasets/PH2/images/IMD061.png', '/home/chen/datasets/PH2/images/IMD020.png', '/home/chen/datasets/PH2/images/IMD008.png', '/home/chen/datasets/PH2/images/IMD371.png', '/home/chen/datasets/PH2/images/IMD169.png', '/home/chen/datasets/PH2/images/IMD240.png',...
是否能像SAM或CLIP那样直接调用 例如：from sam import SAM model = SAM(....) .....
RuntimeError: Pretrained weights (checkpoints) not found for model EVA02-CLIP-B-16.Available pretrained tags (['eva', 'eva02', 'eva_clip', 'eva02_clip'].
Can you provide BoxSup-C2_Lbase_CLIP_R5021k_640b64_4x_mlr.pth? Thanks!
Can you provide the code for directly applying the Multi-modal MLR trained in SIC-CADS to CORA?