diff --git a/src/runtime/CL/functions/CLGEMM.cpp b/src/runtime/CL/functions/CLGEMM.cpp
index a8d7058..1ad8531 100644
--- a/src/runtime/CL/functions/CLGEMM.cpp
+++ b/src/runtime/CL/functions/CLGEMM.cpp
@@ -181,7 +181,6 @@
 
     TensorInfo tmp_a_info{};
     TensorInfo tmp_b_info{};
-    TensorInfo tmp_output_info{};
 
     // Get the GPU target
     const GPUTarget gpu_target = CLScheduler::get().target();
@@ -229,13 +228,12 @@
     }
 
     // Validate matrix multiply
-    auto_init_if_empty(tmp_output_info, matrix_a_info->clone()->set_tensor_shape(compute_mm_shape(*matrix_a_info, *matrix_b_info, run_interleave_transpose, reshape_info)));
-    ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMMatrixMultiplyKernel::validate(matrix_a_info, matrix_b_info, &tmp_output_info, alpha, run_interleave_transpose, reshape_info, gpu_target));
+    ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMMatrixMultiplyKernel::validate(matrix_a_info, matrix_b_info, output, alpha, run_interleave_transpose, reshape_info, gpu_target));
 
     if(beta != 0 && c != nullptr)
     {
         // Validate matrix addition kernel
-        ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMMatrixAdditionKernel::validate(c, &tmp_output_info, beta));
+        ARM_COMPUTE_RETURN_ON_ERROR(CLGEMMMatrixAdditionKernel::validate(c, output, beta));
     }
 
     return Status{};
diff --git a/src/runtime/CL/functions/CLGEMMConvolutionLayer.cpp b/src/runtime/CL/functions/CLGEMMConvolutionLayer.cpp
index ca6157e..26fd906 100644
--- a/src/runtime/CL/functions/CLGEMMConvolutionLayer.cpp
+++ b/src/runtime/CL/functions/CLGEMMConvolutionLayer.cpp
@@ -391,7 +391,7 @@
         const Size2D kernel_dims(kernel_width, kernel_height);
 
         // Output tensor auto initialization if not yet initialized
-        TensorShape expected_output_shape = compute_im2col_conv_shape(input, kernel_dims, conv_info, append_bias, dilation, true);
+        TensorShape expected_output_shape = compute_im2col_conv_shape(input, kernel_dims, conv_info, append_bias, dilation, true /* num_groups == 1, num_groups */);
 
         auto_init_if_empty(im2col_reshaped_info, input->clone()->set_tensor_shape(expected_output_shape));