CSer-Tang-hao / BSFA-FSFG

[TCSVT23, Highly Cited Paper] Boosting Few-shot Fine-grained Recognition with Background Suppression and Foreground Alignment
MIT License
18 stars 0 forks source link

Unable to Achieve Matching Prediction Accuracy on Stanford Cars Dataset #5

Open chenjiale12345 opened 5 months ago

chenjiale12345 commented 5 months ago

Dear Author,

I am trying to reproduce the results from your paper using the Stanford Cars dataset, but I am unable to achieve the reported prediction accuracy. I have followed the instructions in your repository and have ensured that I am using the correct dataset and model configuration.

Could you please provide me with the specific hyperparameters that you used during your experiments on the Stanford Cars dataset? This would greatly help me in ensuring that I am fully replicating your setup and achieving the desired results.

Specifically, I am interested in:

I understand that hyperparameters can vary depending on the specific setup and hardware used, but having the ones you used as a starting point would be extremely valuable.

Thank you for your time and assistance in this matter. I greatly appreciate your work and the efforts you have put into making your code publicly available.