Bienqq
/
car-detection-bayes
2
1
Fork 0
Code Issues Pull Requests Releases Wiki Activity

updates

This commit is contained in:
Glenn Jocher 2018-11-16 22:34:44 +01:00
parent d2c5d7a5fd
commit ed1067bfb5
1 changed files with 5 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
models.py
View File

@ -101,7 +101,6 @@ class YOLOLayer(nn.Module):
self.anchor_h = self.scaled_anchors[:, 1:2].view((1, nA, 1, 1)) self.anchor_h = self.scaled_anchors[:, 1:2].view((1, nA, 1, 1))
self.weights = class_weights() self.weights = class_weights()
self.batch_count = 0
self.loss_means = torch.zeros(6) self.loss_means = torch.zeros(6)
def forward(self, p, targets=None, requestPrecision=False): def forward(self, p, targets=None, requestPrecision=False):
@ -170,13 +169,13 @@ class YOLOLayer(nn.Module):
if nM > 0: if nM > 0:
lx = k * MSELoss(x[mask], tx[mask]) lx = k * MSELoss(x[mask], tx[mask])
ly = k * MSELoss(y[mask], ty[mask]) ly = k * MSELoss(y[mask], ty[mask])
lw = k * MSELoss(w[mask], tw[mask]) lw = (k * 0.7) * MSELoss(w[mask], tw[mask])
lh = k * MSELoss(h[mask], th[mask]) lh = (k * 0.7) * MSELoss(h[mask], th[mask])
# lconf = k * BCEWithLogitsLoss(pred_conf[mask], mask[mask].float()) # lconf = k * BCEWithLogitsLoss(pred_conf[mask], mask[mask].float())
lconf = (k * 10) * BCEWithLogitsLoss(pred_conf, mask.float()) lconf = (k * 5) * BCEWithLogitsLoss(pred_conf, mask.float())
lcls = (k / 10) * CrossEntropyLoss(pred_cls[mask], torch.argmax(tcls, 1)) lcls = (k / 5) * CrossEntropyLoss(pred_cls[mask], torch.argmax(tcls, 1))
# lcls = k * BCEWithLogitsLoss(pred_cls[mask], tcls.float()) # lcls = k * BCEWithLogitsLoss(pred_cls[mask], tcls.float())
else: else:
lx, ly, lw, lh, lcls, lconf = FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0]) lx, ly, lw, lh, lcls, lconf = FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0])
@ -185,7 +184,7 @@ class YOLOLayer(nn.Module):
# lconf += k * BCEWithLogitsLoss(pred_conf[~mask], mask[~mask].float()) # lconf += k * BCEWithLogitsLoss(pred_conf[~mask], mask[~mask].float())
# Sum loss components # Sum loss components
balance_losses_flag = True balance_losses_flag = False
if balance_losses_flag: if balance_losses_flag:
loss_vec = torch.FloatTensor([lx.data, ly.data, lw.data, lh.data, lconf.data, lcls.data]) loss_vec = torch.FloatTensor([lx.data, ly.data, lw.data, lh.data, lconf.data, lcls.data])
self.loss_means = self.loss_means * 0.99 + loss_vec * 0.01 self.loss_means = self.loss_means * 0.99 + loss_vec * 0.01