Hi, Thank you for your amazing work on AQA!
I want to follow your work but there are some confusion:
I follow the config FineDiving_TSA.yaml, train and test the TSA model, but my result is lower than the result in your paper?
[TEST] tIoU_5: 0.722296, tIoU_75: 0.257677 [TEST] correlation: 0.904310, L2: 45.539504, RL2: 0.004178
Should I run with this config file? or how can I run the code for performance in your paper?
Hi, Thank you for your amazing work on AQA! I want to follow your work but there are some confusion: I follow the config FineDiving_TSA.yaml, train and test the TSA model, but my result is lower than the result in your paper?
[TEST] tIoU_5: 0.722296, tIoU_75: 0.257677 [TEST] correlation: 0.904310, L2: 45.539504, RL2: 0.004178Should I run with this config file? or how can I run the code for performance in your paper?Thank you!