Thanks for your great work: HRDecoder: High-Resolution Decoder Network for Fundus Image Lesion Segmentation. I follow the Readme to intall your code and try to calculate the Flops of your model on DDR by
python tools/get_flops.py configs/lesion/hrdecoder_fcn_hr48_ddr_2048.py --shape 2048 2048. The result is as follows:

However, in your paper, the result is 440.69 GFLOPs. Moreover, I am afraid that I cannot find the code for the calculation of FPS. I would be appreciated if you could share the right way to calculate of Flops and FPS !
Thanks for your great work: HRDecoder: High-Resolution Decoder Network for Fundus Image Lesion Segmentation. I follow the Readme to intall your code and try to calculate the Flops of your model on DDR by

python tools/get_flops.py configs/lesion/hrdecoder_fcn_hr48_ddr_2048.py --shape 2048 2048. The result is as follows:However, in your paper, the result is 440.69 GFLOPs. Moreover, I am afraid that I cannot find the code for the calculation of FPS. I would be appreciated if you could share the right way to calculate of Flops and FPS !