I have a question about performance using different dataset.
I have trained your model using different subset in UCF101 for cross validation.
It seems that value of PSNR seems quite much different. (10 ~ 20psnr and it does not reach ~30 wrote in paper. <model difference i think) and not stable.
@dlwtojd26 Which dataset did you use in your case? Did you replace the current one with image triplets from new dataset? Could you please also share some further insights on your inference.
thank you for sharing your implementation
I have a question about performance using different dataset.
I have trained your model using different subset in UCF101 for cross validation.
It seems that value of PSNR seems quite much different. (10 ~ 20psnr and it does not reach ~30 wrote in paper. <model difference i think) and not stable.
Is this common ?