deepinsight / deepinsight/insightface
why verfication accuracy is high,but train_data accuracy is so low?
Open
- Dominant language
- Python
- Stars
- 29.7k
- Forks
- 6.1k
- PR merge metrics
- No merged PRs in 30d
Description
because of the memory limit of GPU,so i only use one verfication data LFW，and batchsize is 128， and the network is resnet34, loss is arcface, GPU is one Tesla P100, after epoch 16, as you can see in the picture, so what's going on? can anybody give a hand? Thanks in advance.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.