diff --git a/src/kernels/activation_elu.cu b/src/kernels/activation_elu.cu
index b507b5a..a80e3ae 100644
--- a/src/kernels/activation_elu.cu
+++ b/src/kernels/activation_elu.cu
@@ -9,7 +9,7 @@
 __global__
 void activation_elu(value_type *input, value_type *output, int size) {
 
-    int i = threadIdx.x*(blockIdx.x +1);
+    int i = blockDim.x*blockIdx.x + threadIdx.x;
 
     if(i<size) {    
         value_type k0, k1;
@@ -30,6 +30,9 @@ void activation_elu(value_type *input, value_type *output, int size) {
 */
 void activationELUForward(value_type* srcData, value_type* dstData, int size)
 {
-    activation_elu<<<(size+255)/256, 256>>>(srcData, dstData, size);
+    int blocks = (size+255)/256;
+    int threads = 256;
+    
+    activation_elu<<<blocks, threads>>>(srcData, dstData, size);
     checkCuda( cudaDeviceSynchronize() );
 }
\ No newline at end of file
diff --git a/tests/simple_dense.py b/tests/simple_dense.py
index 112b5f4..ab56f03 100644
--- a/tests/simple_dense.py
+++ b/tests/simple_dense.py
@@ -17,18 +17,19 @@ def dense_model():
             batch_input_shape=(1, 100, 100, 4),         # 100by100by2
             output_shape=(100, 100, 4, 1)))                         # 100by100by2
     model.add(Convolution3D(16, kernel_size=(8, 8, 2), subsample=(4, 4, 1), border_mode="valid", 
-                bias_initializer="random_uniform", activation="relu"))
-    #model.add(ELU())
+                bias_initializer="random_uniform"))
+    model.add(ELU())
     model.add(AveragePooling3D(pool_size=(2, 2, 1)))
     model.add(Convolution3D(16, kernel_size=(4, 4, 2), subsample=(2, 2, 1), border_mode="valid", 
-                bias_initializer="random_uniform", activation="relu"))
+                bias_initializer="random_uniform"))
+    model.add(ELU())
     model.add(Convolution3D(24, kernel_size=(3, 3, 2), subsample=(1, 1, 1), border_mode="valid", 
-                bias_initializer="random_uniform", activation="relu"))
-    #model.add(ELU())
+                bias_initializer="random_uniform"))
+    model.add(ELU())
     model.add(Flatten())
-    model.add(Dense(256, activation="relu", bias_initializer="random_uniform"))
+    model.add(Dense(256, bias_initializer="random_uniform"))
+    model.add(ELU())
     model.add(Dense(32, activation="relu", bias_initializer="random_uniform"))
-    #model.add(ELU())
     model.add(Dense(2, bias_initializer="random_uniform"))
 
     sgd = keras.optimizers.Adam(lr=1e-4, decay=1e-8)
diff --git a/tests/test.cpp b/tests/test.cpp
index 09b4c33..d576155 100644
--- a/tests/test.cpp
+++ b/tests/test.cpp
@@ -22,15 +22,15 @@ int main() {
     tkDNN::dataDim_t dim(1, 1, 100, 100, 4);
     tkDNN::MulAdd     m0 (&net, dim, 2, -1);
     tkDNN::Conv3d     c0 (&net, m0.output_dim, 16, 8, 8, 2, 4, 4, 1,    c0_bin, c0_bias_bin);
-    tkDNN::Activation a0 (&net, c0.output_dim, tkDNN::ACTIVATION_RELU);
+    tkDNN::Activation a0 (&net, c0.output_dim, tkDNN::ACTIVATION_ELU);
     tkDNN::Pooling    p0 (&net, a0.output_dim, 2, 2, 2, 2, tkDNN::POOLING_AVERAGE);
     tkDNN::Conv3d     c1 (&net, p0.output_dim, 16, 4, 4, 2, 2, 2, 1,    c1_bin, c1_bias_bin);
-    tkDNN::Activation a1 (&net, c1.output_dim, tkDNN::ACTIVATION_RELU);
+    tkDNN::Activation a1 (&net, c1.output_dim, tkDNN::ACTIVATION_ELU);
     tkDNN::Conv3d     c2 (&net, a1.output_dim, 24, 3, 3, 2, 1, 1, 1,    c2_bin, c2_bias_bin);
-    tkDNN::Activation a2 (&net, c2.output_dim, tkDNN::ACTIVATION_RELU);
+    tkDNN::Activation a2 (&net, c2.output_dim, tkDNN::ACTIVATION_ELU);
     tkDNN::Flatten    f2 (&net, a2.output_dim);
     tkDNN::Dense      d3 (&net, f2.output_dim, 256,                     d3_bin, d3_bias_bin);
-    tkDNN::Activation a3 (&net, d3.output_dim, tkDNN::ACTIVATION_RELU);
+    tkDNN::Activation a3 (&net, d3.output_dim, tkDNN::ACTIVATION_ELU);
     tkDNN::Dense      d4 (&net, a3.output_dim, 32,                      d4_bin, d4_bias_bin);
     tkDNN::Activation a4 (&net, d4.output_dim, tkDNN::ACTIVATION_RELU);
     tkDNN::Dense      d5 (&net, a4.output_dim, 2,                       d5_bin, d5_bias_bin);