I’m not sure that the D_real and D_fake respectly corresponds to which part of the formula fo Lgan. Should they each gradually become larger or smaller? Can you help me see my picture if the training state is correct? i guess D_real represent E[C(Xr)] and D_fake represent E[C(Xf)]？and l_d_real represents log(D(Xr, Xf)) and l_d_fake represents the log(1-D(Xf, Xr)) ？
I’m not sure that the D_real and D_fake respectly corresponds to which part of the formula fo Lgan. Should they each gradually become larger or smaller? Can you help me see my picture if the training state is correct? i guess D_real represent E[C(Xr)] and D_fake represent E[C(Xf)]？and l_d_real represents log(D(Xr, Xf)) and l_d_fake represents the log(1-D(Xf, Xr)) ？