Hello, I used your deepDA code and then replaced the feature extraction network with shufflenetv2, but the accuracy rate was very low, only 0.4. But I use the shufflenetv2 network to classify the source domain dataset with very high accuracy, what is the problem

jindongwang / transferlearning

Transfer learning / domain adaptation / domain generalization / multi-task learning etc. Papers, codes, datasets, applications, tutorials.-迁移学习

http://transferlearning.xyz/

MIT License

13.42k stars 3.81k forks source link

Closed ww0929lib closed 1 year ago

ww0929lib commented 1 year ago

Before you open an issue If it's code running error, maybe you want to check the python or pytorch version before submitting an issue.

Describe the bug A clear and concise description of what the bug is.

To Reproduce Where is this bug happen?

Screenshots If applicable, add screenshots to help explain your problem.

jindongwang commented 1 year ago

Our code is tuned using Resnet and Alexnet. You need to tune the hyperparameters using your own network.