cyyhundred

i have the same issue,have you solved that problem?

no，when i evaluated the model ,the maps were 0(all categories) .i thought maybe i didn't get the proper pretrained weight. i am trying to get it through ------------------ 原始邮件 ------------------ 发件人: "ucbdrive/few-shot-object-detection"...

ok, whats your wechat id?

i encountered the same question,have you solved the problem？ my contrastive loss fluctuates around 8.5

查看yaml文件中是不是有images，会替换成labels，只要换个文件夹名字就行