Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ResNet50_ibn_a在VeRi-766上的推理精度 #26

Open
jbq123 opened this issue Mar 24, 2021 · 3 comments
Open

ResNet50_ibn_a在VeRi-766上的推理精度 #26

jbq123 opened this issue Mar 24, 2021 · 3 comments

Comments

@jbq123
Copy link

jbq123 commented Mar 24, 2021

您好,
我在VeRi-766上复现了基于ResNet50_ibn_a基础网络的训练结果,mAP达到了81.7%。
image
之后运行test.sh,对模型进行推理时,mAP只有65.6%。请问可能是什么原因造成?
谢谢!

@jbq123
Copy link
Author

jbq123 commented Mar 24, 2021

test.sh脚本内容:
python tools/test.py --config_file='configs/veri.yml'
MODEL.DEVICE_ID "('0')"
MODEL.NAME "('resnet50_ibn_a')"
MODEL.MODEL_TYPE "baseline"
DATASETS.TRAIN "('veri',)"
DATASETS.TEST "('veri',)"
DATASETS.ROOT_DIR "('/home/vip/datasets')"
MODEL.PRETRAIN_CHOICE "('self')"
INPUT.SIZE_TRAIN '([320, 320])'
INPUT.SIZE_TEST '([320, 320])'
TEST.WRITE_RESULT True
TEST.WEIGHT "('./output/veri/0412-search/size320/final.pth')"

@Xiangyu-CAS
Copy link
Owner

没加 MODEL.POOLING_METHOD 'GeM' 吧,确保训练和测试的模型结构是一样的,参考scripts/veri/test.sh

@jbq123
Copy link
Author

jbq123 commented Mar 24, 2021

谢谢大佬!确实是这样,之前没注意到。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants