search

Alibaba-MIIL / ML_Decoder

Official PyTorch implementation of "ML-Decoder: Scalable and Versatile Classification Head" (2021)
MIT License
311 stars 53 forks source link

Nus-wide data set for zsl giving good mAP value but f1 score is less #47

Open aliman80 opened 2 years ago

aliman80 commented 2 years ago

Hi, thanks for your continuous support, for zero-shot learning with the Nus-Wide data set I got mAp values similar to what you mentioned in your papers but f1-score values are less. Can you guide what may be the cause for this or am i missing something.

Best Regards

ShangDongChu commented 1 year ago

Hello,I'm trying to run the ZSL in NUS-WIDE dataset,but I find that the dataset they used seems different from my version,for example,the image filename in my NUS dataset is like '0596_391097803.jpg',but the image filename in the data.csv they provided is like '80002_2058923115_a0e2798162_m.jpg'.Would you tell how to solve this problem ? Best Regards