From cdafbf00ae1f95f81c9db647ef23e08eec0813f5 Mon Sep 17 00:00:00 2001
From: nagadomi <nagadomi@nurs.or.jp>
Date: Thu, 23 Feb 2017 08:48:00 +0900
Subject: [PATCH] Add BCE(binary cross entropy) loss for segmentation

Sigmoid() output is required.
---
 lib/settings.lua | 2 +-
 train.lua        | 4 ++++
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/lib/settings.lua b/lib/settings.lua
index 35f271f..fb0b890 100644
--- a/lib/settings.lua
+++ b/lib/settings.lua
@@ -75,7 +75,7 @@ cmd:option("-learning_rate_decay", 3.0e-7, 'learning rate decay (learning_rate *
 cmd:option("-resume", "", 'resume model file')
 cmd:option("-name", "user", 'model name for user method')
 cmd:option("-gpu", 1, 'Device ID')
-cmd:option("-loss", "huber", 'loss function (huber|l1|mse)')
+cmd:option("-loss", "huber", 'loss function (huber|l1|mse|bce)')
 cmd:option("-update_criterion", "mse", 'mse|loss')
 
 local function to_bool(settings, name)
diff --git a/train.lua b/train.lua
index 6379aeb..0286aee 100644
--- a/train.lua
+++ b/train.lua
@@ -322,6 +322,10 @@ local function create_criterion(model)
       return w2nn.L1Criterion():cuda()
    elseif settings.loss == "mse" then
       return w2nn.ClippedMSECriterion(0, 1.0):cuda()
+   elseif settings.loss == "bce" then
+      local bce = nn.BCECriterion()
+      bce.sizeAverage = true
+      return bce:cuda()
    else
       error("unsupported loss .." .. settings.loss)
    end