Add Mobilenetv2 SSD Lite post and preprocessing, add mobilenet demo

Signed-off-by: xavier <micaelaverucchi@gmail.com>

include/tkDNN/pluginsRT/FlattenConcatRT.h

@@ -0,0 +1,76 @@
+#include<cassert>
+
+class FlattenConcatRT : public IPlugin {
+
+public:
+	FlattenConcatRT() {
+		stat = cublasCreate(&handle);
+		if (stat != CUBLAS_STATUS_SUCCESS) {
+			printf ("CUBLAS initialization failed\n");
+			return;
+  		}
+	}
+
+	~FlattenConcatRT(){
+
+	}
+
+	int getNbOutputs() const override {
+		return 1;
+	}
+
+	Dims getOutputDimensions(int index, const Dims* inputs, int nbInputDims) override {
+		return DimsCHW{ inputs[0].d[0] * inputs[0].d[1] * inputs[0].d[2], 1, 1};
+	}
+
+	void configure(const Dims* inputDims, int nbInputs, const Dims* outputDims, int nbOutputs, int maxBatchSize) override {
+		assert(nbOutputs == 1 && nbInputs ==1);
+		rows = inputDims[0].d[0];
+		cols = inputDims[0].d[1] * inputDims[0].d[2];
+		c = inputDims[0].d[0] * inputDims[0].d[1] * inputDims[0].d[2];
+		h = 1;
+		w = 1;
+	}
+
+	int initialize() override {
+		return 0;
+	}
+
+	virtual void terminate() override {
+		checkERROR(cublasDestroy(handle));
+	}
+
+	virtual size_t getWorkspaceSize(int maxBatchSize) const override {
+		return 0;
+	}
+
+	virtual int enqueue(int batchSize, const void*const * inputs, void** outputs, void* workspace, cudaStream_t stream) override {
+		dnnType *srcData = (dnnType*)reinterpret_cast<const dnnType*>(inputs[0]);
+		dnnType *dstData = reinterpret_cast<dnnType*>(outputs[0]);
+		checkCuda( cudaMemcpy(dstData, srcData, rows*cols*sizeof(dnnType), cudaMemcpyDeviceToDevice));
+	
+		float const alpha(1.0);
+		float const beta(0.0);
+		checkERROR( cublasSgeam( handle, CUBLAS_OP_T, CUBLAS_OP_N, rows, cols, &alpha, srcData, cols, &beta, srcData, rows, dstData, rows ));
+		return 0;
+	}
+
+
+	virtual size_t getSerializationSize() override {
+		return 5*sizeof(int);
+	}
+
+	virtual void serialize(void* buffer) override {
+		char *buf = reinterpret_cast<char*>(buffer);
+		tk::dnn::writeBUF(buf, c);
+		tk::dnn::writeBUF(buf, h);
+		tk::dnn::writeBUF(buf, w);
+		tk::dnn::writeBUF(buf, rows);
+		tk::dnn::writeBUF(buf, cols);
+	}
+
+	int c, h, w;
+	int rows, cols;
+	cublasStatus_t stat; 
+	cublasHandle_t handle; 
+};

include/tkDNN/pluginsRT/ReshapeRT.h

@@ -0,0 +1,61 @@
+#include<cassert>
+
+class ReshapeRT : public IPlugin {
+
+public:
+	ReshapeRT(dataDim_t new_dim) {
+		n = new_dim.n;
+		c = new_dim.c;
+		h = new_dim.h;
+		w = new_dim.w;
+	}
+
+	~ReshapeRT(){
+
+	}
+
+	int getNbOutputs() const override {
+		return 1;
+	}
+
+	Dims getOutputDimensions(int index, const Dims* inputs, int nbInputDims) override {
+		return DimsCHW{ c,h,w};
+	}
+
+	void configure(const Dims* inputDims, int nbInputs, const Dims* outputDims, int nbOutputs, int maxBatchSize) override {
+	}
+
+	int initialize() override {
+		return 0;
+	}
+
+	virtual void terminate() override {
+	}
+
+	virtual size_t getWorkspaceSize(int maxBatchSize) const override {
+		return 0;
+	}
+
+	virtual int enqueue(int batchSize, const void*const * inputs, void** outputs, void* workspace, cudaStream_t stream) override {
+		dnnType *srcData = (dnnType*)reinterpret_cast<const dnnType*>(inputs[0]);
+		dnnType *dstData = reinterpret_cast<dnnType*>(outputs[0]);
+
+		checkCuda( cudaMemcpy(dstData, srcData, c*h*w*sizeof(dnnType), cudaMemcpyDeviceToDevice));
+		return 0;
+	}
+
+
+	virtual size_t getSerializationSize() override {
+		return 4*sizeof(int);
+	}
+
+	virtual void serialize(void* buffer) override {
+		char *buf = reinterpret_cast<char*>(buffer);
+		tk::dnn::writeBUF(buf, n);
+		tk::dnn::writeBUF(buf, c);
+		tk::dnn::writeBUF(buf, h);
+		tk::dnn::writeBUF(buf, w);
+	}
+
+	int n, c, h, w;
+};

include/tkDNN/pluginsRT/SoftmaxRT.h

@@ -0,0 +1,64 @@
+#include<cassert>
+
+class SoftmaxRT : public IPlugin {
+
+public:
+	SoftmaxRT(const tk::dnn::dataDim_t* dim) {
+		assert(dim != nullptr);
+		this->dim.n = dim->n;
+		this->dim.c = dim->c;
+		this->dim.h = dim->h;
+		this->dim.w = dim->w;
+		this->dim.l = dim->l;
+	}
+
+	~SoftmaxRT(){
+
+	}
+
+	int getNbOutputs() const override {
+		return 1;
+	}
+
+	Dims getOutputDimensions(int index, const Dims* inputs, int nbInputDims) override {
+		return DimsNCHW{this->dim.n,this->dim.c,this->dim.h,this->dim.w };
+	}
+
+	void configure(const Dims* inputDims, int nbInputs, const Dims* outputDims, int nbOutputs, int maxBatchSize) override {		
+	}
+
+	int initialize() override {
+		return 0;
+	}
+
+	virtual void terminate() override {
+	}
+
+	virtual size_t getWorkspaceSize(int maxBatchSize) const override {
+		return 0;
+	}
+
+	virtual int enqueue(int batchSize, const void*const * inputs, void** outputs, void* workspace, cudaStream_t stream) override {
+		dnnType *srcData = (dnnType*)reinterpret_cast<const dnnType*>(inputs[0]);
+		dnnType *dstData = reinterpret_cast<dnnType*>(outputs[0]);
+		
+		
+		return 0;
+	}
+
+
+	virtual size_t getSerializationSize() override {
+		return 5*sizeof(int);
+	}
+
+	virtual void serialize(void* buffer) override {
+		char *buf = reinterpret_cast<char*>(buffer);
+		tk::dnn::writeBUF(buf, this->dim.n);
+		tk::dnn::writeBUF(buf, this->dim.c);
+		tk::dnn::writeBUF(buf, this->dim.h);
+		tk::dnn::writeBUF(buf, this->dim.w);
+		tk::dnn::writeBUF(buf, this->dim.l);
+	}
+
+	dataDim_t dim;
+};