search

mlpc-ucsd / BLIVA

(AAAI 2024) BLIVA: A Simple Multimodal LLM for Better Handling of Text-rich Visual Questions
https://arxiv.org/abs/2308.09936
BSD 3-Clause "New" or "Revised" License
257 stars 26 forks source link

how to split triain, val and test set of ocr-vqa dataset? #12

Open adda1221 opened 10 months ago

adda1221 commented 10 months ago

Hi, thanks for your amazing work! how do you split train, val and test set of ocr-vqa dataset, there is not split method on official webset.

gordonhu608 commented 10 months ago

Thank you for your interest in our work. After loading their dataset.json, I believe you can find the split key where it has a value of integer indicating the split.

adda1221 commented 10 months ago

Thank you for your interest in our work. After loading their dataset.json, I believe you can find the split key where it has a value of integer indicating the split. got it, thanks!

jjt076 commented 1 month ago

Thank you for your interest in our work. After loading their dataset.json, I believe you can find the split key where it has a value of integer indicating the split.

Can you give the code to test top1 accuracy?