Li Tao

## Accuracy check ### Train from scratch Tag | Max acc@1 | Last acc@1 | Last acc@5 -- | -- | -- | -- Swin-T (official*) | 81.29 | 81.286...

I think it is. But it is equal to the current code because sizes are calculated carefully.

@JinyangGuo I think in C3D paper, the results are not directly from the results using C3D model, they use SVM to classify and get 82.3% accuracy. I think in C3D...

@ALL @amm040341 I think it is normal that the test accuracy for official split 01 is lower than 50% if trained from scratch using similar setting. I tried reimplement C3D,...

Hi, @skyqwe123 Sure, I have. But the accuracy you reported is strange because for ucf101, the accuracy should be around 1% for random guess. Therefore, I think it's better to...

Hi, @hhc1997 Yes, I think this is right. You may find better results (usually more than 80%) in some papers but they did not trained from scratch. In the scratch...

@hhc1997 I have tried some networks such as 3D-ResNet-18 and R(2+1)D-18 pretrained on Kinetics-400. These models are from torchvision and usually the performance can be better than 90% easily. In...

@hhc1997 I think the 70% is not normal. If you just use R(2+1)D from torchvision, I remebered the accucacy can be at least over 80. You might need to check...

@fsiar It is normal. I have reported several scratch training results are reported in [this paper](https://arxiv.org/pdf/2001.05661.pdf) and [this paper](https://arxiv.org/pdf/2006.13017.pdf). Usually R3D/R(2+1)D/I3D/S3D can not get better performance than 60% for scratch...

Hi @Mrbishuai , thank you for your interest. It sounds not reasonable because settings are different. Could you provide more detailed information, such as 1. Have you make sure that...