예제 #1
0
void convertProtoToLua(void** handle, const char* lua_name, const char* cuda_package)
{
  const caffe::NetParameter netparam = *(const caffe::NetParameter*)handle[1];

  std::ofstream ofs (lua_name);

  ofs << "require '" << cuda_package << "'\n";
  ofs << "require 'cunn'\n";
  ofs << "model = {}\n";
  if(std::string(cuda_package)=="ccn2")
    ofs<< "table.insert(model, {'torch_transpose_dwhb', nn.Transpose({1,4},{1,3},{1,2})})\n";
  
  int num_output = netparam.input_dim_size();
  for (int i=0; i<netparam.layers_size(); ++i)
  {
    std::vector<std::pair<std::string, std::string>> lines;
    auto& layer = netparam.layers(i);
    switch(layer.type())
    {
      case caffe::LayerParameter::CONVOLUTION:
      {
	auto &param = layer.convolution_param();
	int groups = param.group() == 0 ? 1 : param.group();
	int nInputPlane = layer.blobs(0).channels()*groups;
	int nOutputPlane = param.num_output();
	num_output = nOutputPlane;
	int kW = param.kernel_w();
	int kH = param.kernel_h();
	int dW = param.stride_w();
	int dH = param.stride_h();
	if(kW==0 || kH==0)
	{
	  kW = param.kernel_size();
	  kH = kW;
	}
	if(dW==0 || dH==0)
	{
	  dW = param.stride();
	  dH = dW;
	}
	int pad_w = param.pad_w();
	int pad_h = param.pad_h();
        if(pad_w==0 || pad_h==0)
        {
          pad_w = param.pad();
          pad_h = pad_w;
        }
	if(std::string(cuda_package) == "ccn2")
	{
          if(kW != kH || dW != dH || pad_w != pad_h)
          {
            std::cout << "ccn2 only supports square images!\n";
            break;
          }
	  char buf[1024];
	  sprintf(buf, "ccn2.SpatialConvolution(%d, %d, %d, %d, %d, %d)", 
	      nInputPlane, nOutputPlane, kW, dW, pad_w, groups);
	  lines.emplace_back(layer.name(), buf);
	}
	else
	{
	  char buf[1024];
	  const char* mm_or_not = std::string(cuda_package)=="nn" ? "MM" : "";
	  sprintf(buf, "%s.SpatialConvolution%s(%d, %d, %d, %d, %d, %d, %d, %d)", 
	      cuda_package, mm_or_not, nInputPlane, nOutputPlane, kW, kH, dW, dH, pad_w, pad_h);
	  lines.emplace_back(layer.name(), buf);
	}
	break;
      }
      case caffe::LayerParameter::POOLING:
      {
	auto &param = layer.pooling_param();
	std::string ptype = param.pool() == caffe::PoolingParameter::MAX ? "Max" : "Avg";
	int kW = param.kernel_w();
	int kH = param.kernel_h();
	int dW = param.stride_w();
	int dH = param.stride_h();
	if(kW==0 || kH==0)
	{
	  kW = param.kernel_size();
	  kH = kW;
	}
	if(dW==0 || dH==0)
	{
	  dW = param.stride();
	  dH = dW;
	}

	if(std::string(cuda_package) == "ccn2")
	{
	  char buf[1024];
	  sprintf(buf, "ccn2.Spatial%sPooling(%d, %d)", ptype.c_str(), kW, dW);
	  lines.emplace_back(layer.name(), buf);
	}
	else if(std::string(cuda_package) == "cudnn")
	{
	  char buf[1024];
	  sprintf(buf, "%s.Spatial%sPooling(%d, %d, %d, %d):ceil()", cuda_package, ptype=="Avg" ? "Average" : "Max", kW, kH, dW, dH);
	  lines.emplace_back(layer.name(), buf);
	}
	break;
      }
      case caffe::LayerParameter::RELU:
      {
	lines.emplace_back(layer.name(), "nn.ReLU()");
	break;
      }
      case caffe::LayerParameter::LRN:
      {
	if(std::string(cuda_package) == "ccn2")
	{
	  auto &param = layer.lrn_param();
	  int local_size = param.local_size();
	  float alpha = param.alpha();
	  float beta = param.beta();
	  char buf[1024];
	  sprintf(buf, "ccn2.SpatialCrossResponseNormalization(%d, %.6f, %.4f)", local_size, alpha, beta);
	  lines.emplace_back(layer.name(), buf);
	}
	break;
      }
      case caffe::LayerParameter::INNER_PRODUCT:
      {
	auto &param = layer.inner_product_param();
	int nInputPlane = layer.blobs(0).width();
	int nOutputPlane = param.num_output();
	char buf[1024];
	sprintf(buf, "nn.Linear(%d, %d)", nInputPlane, nOutputPlane);
	if(num_output != nInputPlane)
	{
	  if(std::string(cuda_package) == "ccn2")
	    lines.emplace_back("torch_transpose_bdwh", "nn.Transpose({4,1},{4,2},{4,3})");
	  lines.emplace_back("torch_view", "nn.View(-1):setNumInputDims(3)");
	}
	lines.emplace_back(layer.name(), buf);
	num_output = nOutputPlane;
	break;
      }
      case caffe::LayerParameter::DROPOUT:
      {
	char buf[1024];
	sprintf(buf, "nn.Dropout(%f)", netparam.layers(i).dropout_param().dropout_ratio());
	lines.emplace_back(layer.name(), buf);
	break;
      }
      case caffe::LayerParameter::SOFTMAX_LOSS:
      {
	lines.emplace_back(layer.name(), "nn.SoftMax()");
	break;
      }
      case caffe::LayerParameter::SOFTMAX:
      {
	lines.emplace_back(layer.name(), "nn.SoftMax()");
	break;
      }
      default:
      {
	std::cout << "MODULE " << netparam.layers(i).name() << " UNDEFINED\n";
	break;
      }
    }

    if(!lines.empty())
      for(auto& it: lines)
	ofs << "table.insert(model, {'" << it.first << "', " << it.second << "})\n";
    else
      ofs << "-- module '" << layer.name() << "' not found\n";
  }
}
예제 #2
0
void convertProtoToLuaV1(const caffe::NetParameter &netparam, const char* lua_name, const char* cuda_package)
{
  PACKAGE_TYPE cuda_package_type = CCN2;
  if(std::string(cuda_package) == "ccn2")
    cuda_package_type = CCN2;
  else if(std::string(cuda_package) == "nn")
    cuda_package_type = NN;
  else if(std::string(cuda_package) == "cudnn")
    cuda_package_type = CUDNN;

  std::ofstream ofs (lua_name);

  ofs << "require '" << cuda_package << "'\n";
  ofs << "require 'cunn'\n";
  ofs << "local model = {}\n";
  if(std::string(cuda_package)=="ccn2")
    ofs<< "table.insert(model, {'torch_transpose_dwhb', nn.Transpose({1,4},{1,3},{1,2})})\n";
  else if(std::string(cuda_package)=="nn" || std::string(cuda_package)=="cudnn")
    ofs<< "require 'inn'\n";

  int num_output = netparam.input_dim_size();
  for (int i=0; i<netparam.layers_size(); ++i)
  {
    std::vector<std::pair<std::string, std::string>> lines;
    auto& layer = netparam.layers(i);
    switch(layer.type())
    {
      case caffe::V1LayerParameter::CONVOLUTION:
      {
        auto &param = layer.convolution_param();
        int groups = param.group() == 0 ? 1 : param.group();
        int nInputPlane = layer.blobs(0).channels()*groups;
        int nOutputPlane = layer.blobs(0).num();
        //int nOutputPlane = param.num_output();
        num_output = nOutputPlane;
        int kW = param.kernel_w();
        int kH = param.kernel_h();
        int dW = param.stride_w();
        int dH = param.stride_h();
        if(kW==0 || kH==0)
        {
          kW = param.kernel_size();
          kH = kW;
        }
        if(dW==0 || dH==0)
        {
          dW = param.stride();
          dH = dW;
        }
        int pad_w = param.pad_w();
        int pad_h = param.pad_h();
        if(pad_w==0 || pad_h==0)
        {
          pad_w = param.pad();
          pad_h = pad_w;
        }
        if(cuda_package_type == CCN2)
        {
          if(kW != kH || dW != dH || pad_w != pad_h)
          {
            std::cout << "ccn2 only supports square images!\n";
            break;
          }
          char buf[1024];
          sprintf(buf, "ccn2.SpatialConvolution(%d, %d, %d, %d, %d, %d)",
              nInputPlane, nOutputPlane, kW, dW, pad_w, groups);
          lines.emplace_back(layer.name(), buf);
        }
        else if(cuda_package_type == NN)
        {
          if(groups != 1)
          {
            std::cout << "nn supports no groups!\n";
            break;
          }
          char buf[1024];
          sprintf(buf, "nn.SpatialConvolutionMM(%d, %d, %d, %d, %d, %d, %d, %d)",
              nInputPlane, nOutputPlane, kW, kH, dW, dH, pad_w, pad_h);
          lines.emplace_back(layer.name(), buf);
        }
        else
        {
          char buf[1024];
          sprintf(buf, "cudnn.SpatialConvolution(%d, %d, %d, %d, %d, %d, %d, %d, %d)",
              nInputPlane, nOutputPlane, kW, kH, dW, dH, pad_w, pad_h, groups);
          lines.emplace_back(layer.name(), buf);
        }
        break;
      }
      case caffe::V1LayerParameter::POOLING:
      {
        auto &param = layer.pooling_param();
        int kW = param.kernel_w();
        int kH = param.kernel_h();
        int dW = param.stride_w();
        int dH = param.stride_h();
        int padW = param.pad_w();
        int padH = param.pad_h();
        if(kW==0 || kH==0)
        {
          kW = param.kernel_size();
          kH = kW;
        }
        if(dW==0 || dH==0)
        {
          dW = param.stride();
          dH = dW;
        }

        char buf[1024];
        switch(cuda_package_type)
        {
          case CCN2: // ceil mode by default
            if(param.pool() == caffe::PoolingParameter::MAX)
              sprintf(buf, "ccn2.SpatialMaxPooling(%d, %d)", kW, dW);
            else if(param.pool() == caffe::PoolingParameter::AVE)
              sprintf(buf, "ccn2.SpatialAvgPooling(%d, %d)", kW, dW);
            else if(param.pool() == caffe::PoolingParameter::STOCHASTIC)
              THError("Stochastic pooling is not implemented in DHWB format");
            break;
          case CUDNN:
            if(param.pool() == caffe::PoolingParameter::MAX)
              sprintf(buf, "cudnn.SpatialMaxPooling(%d, %d, %d, %d, %d, %d):ceil()", kW, kH, dW, dH, padW, padH);
            else if(param.pool() == caffe::PoolingParameter::AVE)
              sprintf(buf, "cudnn.SpatialAveragePooling(%d, %d, %d, %d, %d, %d):ceil()", kW, kH, dW, dH, padW, padH);
            else if(param.pool() == caffe::PoolingParameter::STOCHASTIC)
              sprintf(buf, "inn.SpatialStochasticPooling(%d, %d, %d, %d)", kW, kH, dW, dH);
            break;
          case NN:
            if(param.pool() == caffe::PoolingParameter::MAX)
              //sprintf(buf, "nn.SpatialMaxPooling(%d, %d, %d, %d, %d, %d):ceil()", kW, kH, dW, dH, padW, padH);
              sprintf(buf, "nn.SpatialMaxPooling(%d, %d, %d, %d, %d, %d)", kW, kH, dW, dH, padW, padH);
            else if(param.pool() == caffe::PoolingParameter::AVE)
              sprintf(buf, "inn.SpatialAveragePooling(%d, %d, %d, %d)", kW, kH, dW, dH); // padding is not supported yet
            else if(param.pool() == caffe::PoolingParameter::STOCHASTIC)
              sprintf(buf, "inn.SpatialStochasticPooling(%d, %d, %d, %d)", kW, kH, dW, dH);
            break;
        }
        lines.emplace_back(layer.name(), buf);
        break;
      }
      case caffe::V1LayerParameter::RELU:
      {
        switch(cuda_package_type)
        {
          case CUDNN:
            lines.emplace_back(layer.name(), "cudnn.ReLU(true)");
            break;
          default:
            lines.emplace_back(layer.name(), "nn.ReLU(true)");
            break;
        }
        break;
      }
      case caffe::V1LayerParameter::TANH:
      {
        switch(cuda_package_type)
        {
          case CUDNN:
            lines.emplace_back(layer.name(), "cudnn.Tanh(true)");
            break;
          default:
            lines.emplace_back(layer.name(), "nn.Tanh()");
            break;
        }
        break;
      }
      case caffe::V1LayerParameter::SIGMOID:
      {
        switch(cuda_package_type)
        {
          case CUDNN:
            lines.emplace_back(layer.name(), "cudnn.Sigmoid(true)");
            break;
          default:
            lines.emplace_back(layer.name(), "nn.Sigmoid()");
            break;
        }
        break;
      }
      case caffe::V1LayerParameter::LRN:
      {
        auto &param = layer.lrn_param();
        int local_size = param.local_size();
        float alpha = param.alpha();
        float beta = param.beta();
        float k = param.k();
        char buf[1024];
        if(std::string(cuda_package) == "ccn2")
          sprintf(buf, "ccn2.SpatialCrossResponseNormalization(%d, %.6f, %.4f, %f)", local_size, alpha, beta, k);
        else
          sprintf(buf, "inn.SpatialCrossResponseNormalization(%d, %.6f, %.4f, %f)", local_size, alpha, beta, k);
        lines.emplace_back(layer.name(), buf);
        break;
      }
      case caffe::V1LayerParameter::INNER_PRODUCT:
      {
        auto &param = layer.inner_product_param();
        int nInputPlane = layer.blobs(0).width();
        int nOutputPlane = param.num_output();
        char buf[1024];
        sprintf(buf, "nn.Linear(%d, %d)", nInputPlane, nOutputPlane);
        if(num_output != nInputPlane)
        {
          if(std::string(cuda_package) == "ccn2")
            lines.emplace_back("torch_transpose_bdwh", "nn.Transpose({4,1},{4,2},{4,3})");
          lines.emplace_back("torch_view", "nn.View(-1):setNumInputDims(3)");
        }
        lines.emplace_back(layer.name(), buf);
        num_output = nOutputPlane;
        break;
      }
      case caffe::V1LayerParameter::DROPOUT:
      {
        char buf[1024];
        sprintf(buf, "nn.Dropout(%f)", layer.dropout_param().dropout_ratio());
        lines.emplace_back(layer.name(), buf);
        break;
      }
      case caffe::V1LayerParameter::SOFTMAX_LOSS:
      {
        lines.emplace_back(layer.name(), "nn.SoftMax()");
        break;
      }
      case caffe::V1LayerParameter::SOFTMAX:
      {
        lines.emplace_back(layer.name(), "nn.SoftMax()");
        break;
      }
      default:
      {
        std::cout << "MODULE " << layer.name() << " UNDEFINED\n";
        break;
      }
    }
    if(!lines.empty())
      for(auto& it: lines)
        ofs << "table.insert(model, {'" << it.first << "', " << it.second << "})\n";
    else
    {
      ofs << "-- warning: module '" << layer.name() << "' not found\n";
      std::cout << "warning: module '" << layer.name() << "' not found\n";
    }
  }
  ofs << "return model";
}