Letian2003/Oasis
Oasis: One Image is All You Need for Multimodal Instruction Data Synthesis This dataset contains Oasis-500k dataset. [Read the Paper] | [Github Repo] All images come from Cambrian-10M. Instructions and responses are generated by MLLM.
5308
images_part_aa.tar.gzdownload
images_part_ab.tar.gzdownload
images_part_ac.tar.gzdownload
images_part_ad.tar.gzdownload
images_part_ae.tar.gzdownload
images_part_af.tar.gzdownload
images_part_ag.tar.gzdownload
images_part_ah.tar.gzdownload
images_part_ai.tar.gzdownload
images_part_aj.tar.gzdownload
images_part_ak.tar.gzdownload
images_part_al.tar.gzdownload
images_part_am.tar.gzdownload
