Why there are so many duplicated images?
Hi, I created this sample cache dataset simply by concating the images of some random data sources, where MME is one of it. As shown here https://huggingface.co/datasets/lmms-lab/MME, the MME originally have different question but I did not perform the deduplication of the images. You can prepare a much more diverse and curated cache dataset in actual implementation to achieve better result.
Hi, I created this sample cache dataset simply by concating the images of some random data sources, where MME is one of it. As shown here https://huggingface.co/datasets/lmms-lab/MME, the MME originally have different question but I did not perform the deduplication of the images. You can prepare a much more diverse and curated cache dataset in actual implementation to achieve better result.
OK got it! Thanks for your answer!
