Thank the authors for their excellent work and for sharing the code! I have a few questions about the structural generation part: What the Gradient Fusion (Grad-F) module in your paper denotes to? In the code that you provided, I can see that f0-f8 are summed to get the output. Does this mean the summation operation is Grad-F? And if so, how does the 'adaptively merged' in the paper come into play here?
Thank the authors for their excellent work and for sharing the code! I have a few questions about the structural generation part: What the Gradient Fusion (Grad-F) module in your paper denotes to? In the code that you provided, I can see that f0-f8 are summed to get the output. Does this mean the summation operation is Grad-F? And if so, how does the 'adaptively merged' in the paper come into play here?