what if i don't have reliable datasets with label

[anewusername77]

as the title said, can i train this model and lable those data?they are completely without any class labels, just want to assign them with a specific class(no need for meaning of classes)

[niuchuangnn]

Your data are images? SPICE is just to label/cluster images without using any class labels. SPICE may be also extended to other kinds of data with well representation features.

[anewusername77]

thanks, i'll try

[anewusername77]

Your data are images? SPICE is just to label/cluster images without using any class labels. SPICE may be also extended to other kinds of data with well representation features.

dear author,
you mean i can train this model with my own images, but they need to organized as imagenet or cifar?
for example, imagenet_dog need : imagenet_dog_npy , imagenet_dog_lmdb and imagenet_dog_lmdb_meta_info.pkl?

[niuchuangnn]

Hi @scarletteshu, you can write a dataset Class to load your own images, e.g., you can refer to the DatasetFolder, depending on how you organize and load your images. Note that you need to make sure that the outputs of your own data loader are the same as those of the loader in this project. In the current version, the outputs of the data loader in different stages are slightly different.