Remove smooth transition to adversarial loss

map2map/args.py

@@ -1,4 +1,5 @@
 import argparse
+import warnings
 
 from .train import ckpt_link
 
@@ -91,7 +92,7 @@ def add_train_args(parser):
             'e.g. 0.9 for label smoothing and 1 to disable')
     parser.add_argument('--loss-fraction', default=0.5, type=float,
             help='final fraction of loss (vs adv-loss)')
-    parser.add_argument('--loss-halflife', default=20, type=float,
+    parser.add_argument('--loss-halflife', default=20, type=deprecated,
             help='half-life (epoch) to anneal loss while enhancing adv-loss')
     parser.add_argument('--instance-noise', default=0, type=float,
             help='noise added to the adversary inputs to stabilize training')
@@ -141,6 +142,10 @@ def str_list(s):
     return s.split(',')
 
 
+def deprecated(s):
+    warnings.warn("deprecated argument", DeprecationWarning, stacklevel=2)
+
+
 #def int_tuple(t):
 #    t = t.split(',')
 #    t = tuple(int(i) for i in t)

map2map/train.py

@@ -348,11 +348,10 @@ def train(epoch, loader, model, criterion, optimizer, scheduler,
             loss_adv, = adv_criterion(eval_out, real)
             epoch_loss[1] += loss_adv.item()
 
-            r = loss.item() / (loss_adv.item() + 1e-8)
-            f = args.loss_fraction
-            e = epoch - args.adv_start
-            d = 0.5 ** (e / args.loss_halflife)
-            loss = (f + (1 - f) * d) * loss + (1 - f) * (1 - d) * r * loss_adv
+            ratio = loss.item() / (loss_adv.item() + 1e-8)
+            frac = args.loss_fraction
+            if epoch >= args.adv_start:
+                loss = frac * loss + (1 - frac) * ratio * loss_adv
 
         optimizer.zero_grad()
         loss.backward()