diff --git a/residual-layers.lua b/residual-layers.lua
index 92f0ac1..716d013 100644
--- a/residual-layers.lua
+++ b/residual-layers.lua
@@ -84,9 +84,9 @@ function addResidualLayer2(input,  nChannels, nOutChannels, stride)
    end
 
    -- Add them together
-   net = cudnn.SpatialBatchNormalization(nOutChannels)(net)
    net = nn.CAddTable(){net, skip}
-   net = cudnn.ReLU(true)(net)
+   net = cudnn.SpatialBatchNormalization(nOutChannels)(net)
+   --net = cudnn.ReLU(true)(net)
    -- ^ don't put a ReLU here! see http://gitxiv.com/comments/7rffyqcPLirEEsmpX
 
    return net
diff --git a/train-cifar.lua b/train-cifar.lua
index e114829..d3b026e 100644
--- a/train-cifar.lua
+++ b/train-cifar.lua
@@ -178,7 +178,9 @@ function forwardBackwardBatch(batch)
     --]]
 
     -- From https://github.com/bgshih/cifar.torch/blob/master/train.lua#L119-L128
-    if sgdState.epochCounter < 80 then
+    if sgdState.nEvalCounter < 400 then
+        sgdState.learningRate = 0.01
+    elseif sgdState.epochCounter < 80 then
         sgdState.learningRate = 0.1
     elseif sgdState.epochCounter < 120 then
         sgdState.learningRate = 0.01