[NCNN学习笔记]-1

1、前言

本次继续学习NCNN，希望能够坚持，往期学习NCNN的链接如下。

[NCNN学习笔记]-0

2、学习内容

2.1、batchnorm_arm.cpp

这个章节学习NCNN中batchnorm在NEON上的实现。batchnorm的学习可参考链接:https://zhuanlan.zhihu.com/p/93643523

在NCNN中batchnorm的预处理为ncnn-master\src\layer\batchnorm.cpp

int BatchNorm::load_model(const ModelBin& mb)
{slope_data = mb.load(channels, 1);if (slope_data.empty())return -100;mean_data = mb.load(channels, 1);if (mean_data.empty())return -100;var_data = mb.load(channels, 1);if (var_data.empty())return -100;bias_data = mb.load(channels, 1);if (bias_data.empty())return -100;a_data.create(channels);if (a_data.empty())return -100;b_data.create(channels);if (b_data.empty())return -100;// 通过https://zhuanlan.zhihu.com/p/93643523中的公式，可以很推算出下面的过程for (int i = 0; i < channels; i++){float sqrt_var = sqrtf(var_data[i] + eps);if (sqrt_var == 0.f)sqrt_var = 0.0001f; // sanitize divide by zeroa_data[i] = bias_data[i] - slope_data[i] * mean_data[i] / sqrt_var;b_data[i] = slope_data[i] / sqrt_var;}return 0;
}

通过上面的代码，明白了a_data和b_data的作用，下面就正式开始学习在数据处理过程中的batchnorm吧

int BatchNorm_arm::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{int elembits = bottom_top_blob.elembits();int dims = bottom_top_blob.dims;         // 数据维度，最多为4int elempack = bottom_top_blob.elempack; // 每个数据可被分为多少组   例如float32x4_t可被分为4组if (elempack == 4)    // float32x4_t、int32x4_t、float16x4_t{// 对一行进行batchnormif (dims == 1 ){int w = bottom_top_blob.w; #pragma omp parallel for num_threads(opt.num_threads)for (int i = 0; i < w; i++){float* ptr = (float*)bottom_top_blob + i * 4;   // 每间隔4个数据取一次地址float32x4_t _a = vld1q_f32((const float*)a_data + i * 4); float32x4_t _b = vld1q_f32((const float*)b_data + i * 4);float32x4_t _p = vld1q_f32(ptr);_p = vmlaq_f32(_a, _p, _b);   // _a + _p.*b     y = a_data + b_data * xvst1q_f32(ptr, _p);}}// 对每一行batchnormif (dims == 2){int w = bottom_top_blob.w;int h = bottom_top_blob.h;#pragma omp parallel for num_threads(opt.num_threads)for (int i = 0; i < h; i++){float32x4_t _a = vld1q_f32((const float*)a_data + i * 4);float32x4_t _b = vld1q_f32((const float*)b_data + i * 4);float* ptr = bottom_top_blob.row(i);for (int j = 0; j < w; j++)  {float32x4_t _p = vld1q_f32(ptr);_p = vmlaq_f32(_a, _p, _b);vst1q_f32(ptr, _p);ptr += 4;}}}// 针对channel进行batchnormif (dims == 3 || dims == 4){int w = bottom_top_blob.w;int h = bottom_top_blob.h;int d = bottom_top_blob.d;int c = bottom_top_blob.c;int size = w * h * d;#pragma omp parallel for num_threads(opt.num_threads)for (int q = 0; q < c; q++){float32x4_t _a = vld1q_f32((const float*)a_data + q * 4);float32x4_t _b = vld1q_f32((const float*)b_data + q * 4);float* ptr = bottom_top_blob.channel(q);for (int i = 0; i < size; i++){float32x4_t _p = vld1q_f32(ptr);_p = vmlaq_f32(_a, _p, _b);vst1q_f32(ptr, _p);ptr += 4;}}}return 0;}return 0;
}

2.2、bias_arm.cpp

这个章节学习ncnn中的的bias计算

int Bias_arm::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{int w = bottom_top_blob.w;int h = bottom_top_blob.h;int d = bottom_top_blob.d;int channels = bottom_top_blob.c;int size = w * h * d;const float* bias_ptr = bias_data;#pragma omp parallel for num_threads(opt.num_threads)for (int q = 0; q < channels; q++){float* ptr = bottom_top_blob.channel(q);  // 第q个channel的数据float bias = bias_ptr[q];int nn = size >> 2; // 处理4的整数倍的数据int remain = size - (nn << 2);     // 剩余处理向量float32x4_t _bias = vdupq_n_f32(bias);for (; nn > 0; nn--){float32x4_t _p = vld1q_f32(ptr);float32x4_t _outp = vaddq_f32(_p, _bias);  // x + biasvst1q_f32(ptr, _outp);   ptr += 4;}// 剩余向量使用c语言计算for (; remain > 0; remain--){*ptr = *ptr + bias;ptr++;}}return 0;
}