Some questions about the similarities?

[Liqq1]

In the gloria_model.py/def get_local_similarities(130), row_sim is computed by "max"

 row_sim, max_row_idx = torch.max(row_sim, dim=1, keepdim=True)  # [48, 1]

and, in the gloria_loss.py / def local_loss(120), row_sim is computed by "sum"/mean

 if agg == "sum":
     row_sim = row_sim.sum(dim=1, keepdim=True)  # [48, 1]
 else:
     row_sim = row_sim.mean(dim=1, keepdim=True)  # [48, 1]

• I would like to know what this [48,1]-dimensional vector represents ?
• Why does it have different operations?(max, sum,mean)
• and what does the subsequent concat out of the [48,48]-dimensional vector represent?

[tzcskys]

Hi Liqq1, I also have the same question about this, do you have a explanation or suspect for this? Thansk a lot.

[Liqq1]

Hi Liqq1, I also have the same question about this, do you have a explanation or suspect for this? Thansk a lot.

Sorry, I haven't caught on yet. Are you working on that too? Maybe we can add a contact information and talk about it.😊