tankche1 / Semantic-Feature-Augmentation-in-Few-shot-Learning

Semantic Feature Augmentation in Few-shot Learning
87 stars 11 forks source link

question about the paper's result #9

Open flexibility2 opened 5 years ago

flexibility2 commented 5 years ago

Hi, I notice in your original paper's experimental results part, you have written that the RELATION NET's result is 57.02 for 1-shot 71.07 for 5-shot however, in the original paper "Learning to compare: Relation network for few-shot learning", the result is 50.44 and 65.32 respectively. So I wonder how you get the better results? Could you share this part of code? I will be very appreciate for your generous. Thanks a lot! (my email is 729406307@qq.com)