Added sigmoid and tanh

Add Sigmoid and Tanh implementation (trivial using CuDNN).

include/aidge/backend/cuda/operator/SigmoidImpl.hpp

+/********************************************************************************
+ * Copyright (c) 2023 CEA-List
+ *
+ * This program and the accompanying materials are made available under the
+ * terms of the Eclipse Public License 2.0 which is available at
+ * http://www.eclipse.org/legal/epl-2.0.
+ *
+ * SPDX-License-Identifier: EPL-2.0
+ *
+ ********************************************************************************/
+
+#ifndef AIDGE_BACKEND_CUDA_OPERATOR_SIGMOIDIMPL_H_
+#define AIDGE_BACKEND_CUDA_OPERATOR_SIGMOIDIMPL_H_
+
+#include <array>
+#include <memory>
+#include <tuple>
+#include <vector>
+
+#include <cudnn.h>
+
+#include "aidge/backend/OperatorImpl.hpp"
+#include "aidge/operator/Sigmoid.hpp"
+#include "aidge/utils/Registrar.hpp"
+#include "aidge/utils/Types.h"
+
+#include "aidge/backend/cuda/utils/CudaUtils.hpp"
+
+namespace Aidge {
+class SigmoidImpl_cuda : public OperatorImpl {
+private:
+    // CuDNN specific variables
+    #if CUDNN_VERSION >= 5000
+        cudnnActivationDescriptor_t mSigmoidDesc = nullptr;
+    #else
+        cudnnActivationMode_t mSigmoidDesc = nullptr;
+    #endif
+    std::shared_ptr<Tensor> mInputFallback;
+
+public:
+    SigmoidImpl_cuda(const Sigmoid_Op &op) : OperatorImpl(op, "cuda") {}
+
+    static std::unique_ptr<SigmoidImpl_cuda> create(const Sigmoid_Op &op) {
+        return std::make_unique<SigmoidImpl_cuda>(op);
+    }
+
+public:
+    void forward();
+    ~SigmoidImpl_cuda();
+
+private:
+    template <class T> void forward_(const Tensor& input);
+};
+
+namespace {
+// add cuda backend to Sigmoid_Op implementation registry
+static Registrar<Sigmoid_Op> registrarSigmoidImpl_cuda("cuda", Aidge::SigmoidImpl_cuda::create);
+}  // namespace
+}  // namespace Aidge
+
+#endif /* AIDGE_BACKEND_CUDA_OPERATOR_SIGMOIDIMPL_H_ */