root
/
MNN
mirror of https://github.com/alibaba/MNN.git
1
0
Fork 0
Code Issues Actions 1 Packages Projects Releases Wiki Activity
MNN/docs/inference/python.md

551 lines
33 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Python API使用
## 安装
MNN Python API可以使用[源码安装](../compile/pymnn.md)，也可以直接使用`pip`安装预编译whl包；`pip`安装用法如下：
```bash
# 外部版本安装
pip install MNN==$version
# 公司内部版本安装
pip install -i https://artifacts.antgroup-inc.cn/simple/ MNN-Internal==$version
```
## 概览
MNN在C++的基础上，增加了Python扩展。扩展单元包括两个部分：
- MNN：负责推理，训练，图像处理和数值计算
- MNNTools：对MNN的部分工具进行封装，包括：mnn，mnnconvert和mnnquant
### MNN
[MNN](../pymnn/MNN.md)
- [expr](../pymnn/expr.md)
- [nn](../pymnn/nn.md)
- [loss](../pymnn/loss.md)
- [compress](../pymnn/compress.md)
- [data](../pymnn/data.md)
- [optim](../pymnn/optim.md)
- [cv](../pymnn/cv.md)
- [numpy](../pymnn/numpy.md)
- [linalg](../pymnn/linalg.md)
- [random](../pymnn/random.md)
### MNNTools
MNNTools提供目前主要是2个工具，用法可以参考[mnnconvert](../tools/python.html#mnnconvert)和[mnnquant](../tools/python.html#mnnquant)
## 使用Python Module API
### 数据类型
Python中的`Module API`与C++中的函数名略有区别，用法相似。主要数据类型如下：
- [_Module](../pymnn/_Module.md) 模型实例
- [Var](../pymnn/Var.md) 模型的输入输出
### 推理流程
基本推理流程如下：
- [创建Module](../pymnn/nn.html#load-module-from-file-file-name-input-names-output-names-dynamic-shape-mutable-rearrange-backend-memory-mode-power-mode-precision-mode)
- 创建输入: 使用`expr`或`numpy`函数创建`Var`即可作为输入
- [执行推理](../pymnn/_Module.html#forward-input)
- 获取输出: 输出为`Var`类型，可以通过`expr`或`numpy`函数执行后处理
### 示例
```python
import MNN.nn as nn
import MNN.cv as cv
import MNN.numpy as np
import MNN.expr as expr
# 配置执行后端，线程数，精度等信息；key-value请查看API介绍
config = {}
config['precision'] = 'low' # 当硬件支持（armv8.2）时使用fp16推理
config['backend'] = 0 # CPU
config['numThread'] = 4 # 线程数
rt = nn.create_runtime_manager((config,))
# 加载模型创建_Module
net = nn.load_module_from_file('mobilenet_v1.mnn', ['data'], ['prob'], runtime_manager=rt)
# 读取图片
image = cv.imread('cat.jpg')
# 转换为float32, 形状为[224,224,3]
image = cv.resize(image, (224, 224), mean=[103.94, 116.78, 123.68], norm=[0.017, 0.017, 0.017])
# 增加batch HWC to NHWC
input_var = np.expand_dims(image, 0)
# NHWC to NC4HW4
input_var = expr.convert(input_var, expr.NC4HW4)
# 执行推理
output_var = net.forward(input_var)
# NC4HW4 to NHWC
output_var = expr.convert(output_var, expr.NHWC)
# 打印出分类结果, 282为猫
print("output belong to class: {}".format(np.argmax(output_var)))
# output belong to class: 282
```
其他示例可以参考[示例](../pymnn/RuntimeManager.html#example)；也可以参考[示例工程](../start/demo.html#id5)。
## 使用Python Session API *[deprecated]*
不建议使用该API执行推理，建议使用Module API
### 数据类型
Python中`Session API`的函数名与用法与C++基本一样。使用的主要数据类型如下：
- [Interpreter](../pymnn/Interpreter.md) 解释器，持有模型资源
- [Session](../pymnn/Session.md) 会话，持有推理资源
- [Tensor](../pymnn/Tensor.md) 用来描述输入输出数据
- [CVImageProcess](../pymnn/CVImageProcess.md) 图像处理模块
- [CVMatrix](../pymnn/CVMatrix.md) 用来描述图像的仿射变换
### 推理流程
基本推理流程如下：
- [创建Interpreter](../pymnn/Interpreter.html#interpreter-model-path)
- [创建Session](../pymnn/Interpreter.html#createsession-config-runtime)
- [获取Session的输入输出](../pymnn/Interpreter.html#getsessioninput-session-tensorname)
- [使用ImageProcess/cv进行图像处理（可选）](../pymnn/CVImageProcess.html#example)
- [拷贝数据到输入Tensor](../pymnn/Tensor.html#copyfrom-from)
- [执行resize（可选）](../pymnn/Interpreter.html#resizesession-session)
- [执行Session](../pymnn/Interpreter.html#runsession-session)
- [获取输出Tensor数据](../pymnn/Tensor.html#copytohosttensor-to)
### 示例
```python
import MNN
import MNN.cv as cv
import MNN.numpy as np
import MNN.expr as expr
# 创建interpreter
interpreter = MNN.Interpreter("mobilenet_v1.mnn")
# 创建session
config = {}
config['precision'] = 'low'
config['backend'] = 'CPU'
config['thread'] = 4
session = interpreter.createSession(config)
# 获取会话的输入输出
input_tensor = interpreter.getSessionInput(session)
output_tensor = interpreter.getSessionOutput(session)
# 读取图片
image = cv.imread('cat.jpg')
dst_height = dst_width = 224
# 使用ImageProcess处理第一张图片，将图片转换为转换为size=(224, 224), dtype=float32，并赋值给input_data1
image_processer = MNN.CVImageProcess({'sourceFormat': MNN.CV_ImageFormat_BGR,
'destFormat': MNN.CV_ImageFormat_BGR,
'mean': (103.94, 116.78, 123.68, 0.0),
'filterType': MNN.CV_Filter_BILINEAL,
'normal': (0.017, 0.017, 0.017, 0.0)})
image_data = image.ptr
src_height, src_width, channel = image.shape
input_data1 = MNN.Tensor((1, dst_height, dst_width, channel), MNN.Halide_Type_Float, MNN.Tensor_DimensionType_Tensorflow)
#设置图像变换矩阵
matrix = MNN.CVMatrix()
x_scale = src_width / dst_width
y_scale = src_height / dst_height
matrix.setScale(x_scale, y_scale)
image_processer.setMatrix(matrix)
image_processer.convert(image_data, src_width, src_height, 0, input_data1)
# 使用cv模块处理第二张图片，将图片转换为转换为size=(224, 224), dtype=float32，并赋值给input_data2
image = cv.imread('TestMe.jpg')
image = cv.resize(image, (224, 224), mean=[103.94, 116.78, 123.68], norm=[0.017, 0.017, 0.017])
input_data2 = np.expand_dims(image, 0) # [224, 224, 3] -> [1, 224, 224, 3]
# 合并2张图片到，并赋值给input_data
input_data1 = expr.const(input_data1.getHost(), input_data1.getShape(), expr.NHWC) # Tensor -> Var
input_data = np.concatenate([input_data1, input_data2]) # [2, 224, 224, 3]
input_data = MNN.Tensor(input_data) # Var -> Tensor
# 演示多张图片输入，所以将输入resize到[2, 3, 224, 224]
interpreter.resizeTensor(input_tensor, (2, 3, 224, 224))
# 重新计算形状分配内存
interpreter.resizeSession(session)
# 拷贝数据到输入Tensor
input_tensor.copyFrom(input_data)
# 执行会话推理
interpreter.runSession(session)
# 从输出Tensor拷贝出数据
output_data = MNN.Tensor(output_tensor.getShape(), MNN.Halide_Type_Float, MNN.Tensor_DimensionType_Caffe)
output_tensor.copyToHostTensor(output_data)
# 打印出分类结果: 282为猫，385为象
output_var = expr.const(output_data.getHost(), [2, 1001])
print("output belong to class: {}".format(np.argmax(output_var, 1)))
# output belong to class: array([282, 385], dtype=int32)
```
其他示例可以参考[示例](../pymnn/Interpreter.html#example)；也可以参考[示例工程](../start/demo.html#session)。
## 使用cv/numpy API
### 数据类型
Python的`cv`和`numpy`接口，其中`cv`是对C++中`tools/cv`实现的封装；`numpy`则是对`expr`接口的封装；这两个接口主要为了提高MNN的易用性，与`opencv`与`numpy`做到了再接口上的部分兼容，在用法和思路上基本一致。主要数据类型如下：
- [Var](../pymnn/Var.md) `cv`中的图像，`numpy`中的`ndarray`
### 主要用法
`cv`和`numpy`主要用作模型的前后处理部分，和一些数值计算任务。比如从图片直接读取数据后一般需要执行颜色空间变换，数据类型变换，缩放，裁剪等操作，这些可以用`cv`模块函数实现；模型输出的结果可能需要做一些额外的变换和计算，这些可以用`numpy`模块函数实现。
### 示例
使用`cv`与`numpy`中的函数做前后处理，执行模型推理的例子
```python
import MNN
import MNN.cv as cv
import MNN.numpy as np
# 加载模型
net = MNN.nn.load_module_from_file('mobilenet_v1.mnn', ["data"], ["prob"])
# cv模块图片处理
image = cv.imread('cat.jpg')
image = cv.resize(image, (224, 224))
# 类似ndarray的数值运算
image = image - (103.94, 116.78, 123.68)
image = image * (0.017, 0.017, 0.017)
input_var = np.expand_dims(image, 0)
input_var = MNN.expr.convert(input_var, MNN.expr.NC4HW4)
output_var = net.forward(input_var)
output_var = MNN.expr.convert(output_var, MNN.expr.NHWC)
# 类似numpy操作的后处理
print("output belong to class: {}".format(np.argmax(output_var)))
```
其他示例可以参考[示例](../pymnn/Interpreter.html#example)；也可以参考[示例工程](../start/demo.html#id5)。
### cv能力列表
cv模块提供了与OpenCV相似的接口函数，具备基础的图像处理能力，目前支持的cv函数60个。
#### 图像编解码
| 函数名 | 功能 |
| --- | --- |
| [haveImageReader](../pymnn/cv.html#haveimagereader-filename) | 是否可读（解码） |
| [haveImageWriter](../pymnn/cv.html#haveimagewriter-filename) | 是否可写（编码） |
| [imdecode](../pymnn/cv.html#imdecode-buf-flag) | 从内存解码为Mat |
| [imencode](../pymnn/cv.html#imencode-ext-img-params) | 编码Mat到内存中 |
| [imread](../pymnn/cv.html#imread-filename-flag) | 读图片 |
| [imwrite](../pymnn/cv.html#imwrite-filename-img-params) | 写图片 |
#### 图像滤波
| 函数名 | 功能 |
| --- | --- |
| [blur](../pymnn/cv.html#blur-src-ksize-bordertype) | 均值滤波，平滑模糊 |
| [boxFilter](../pymnn/cv.html#boxfilter-src-ddepth-ksize-normalize-bordertype) | 盒子滤波， |
| [dilate](../pymnn/cv.html#dilate-src-kernel-iterations-bordertype) | 膨胀 |
| [filter2D](../pymnn/cv.html#filter2d-src-ddepth-kernel-delta-bordertype) | 2d卷积 |
| [GaussianBlur](../pymnn/cv.html#gaussianblur-src-ksize-sigmax-sigmay-bordertype) | 高斯模糊 |
| [getDerivKernels](../pymnn/cv.html#getderivkernels-dx-dy-ksize-normalize) | 求导数，实际为Sobel/Scharr |
| [getGaborKernel](../pymnn/cv.html#getgaborkernel-ksize-sigma-theta-lambd-gamma-psi) | 获取Gabor核 |
| [getGaussianKernel](../pymnn/cv.html#getgaussiankernel-ksize-sigma) | 获得高斯核 |
| [getStructuringElement](../pymnn/cv.html#getstructuringelement-shape-ksize) | 获取结构化元素用于形态学操作 |
| [Laplacian](../pymnn/cv.html#laplacian-src-ddepth-ksize-scale-delta-bordertype) | 边缘检测滤波 |
| [pyrDown](../pymnn/cv.html#pyrdown-src-dstsize-bordertype) | 高斯平滑+下采样 |
| [pyrUp](../pymnn/cv.html#pyrup-src-dstsize-bordertype) | 上采样+高斯平滑 |
| [Scharr](../pymnn/cv.html#scharr-src-ddepth-dx-dy-scale-delta-bordertype) | 边缘检测滤波 |
| [sepFilter2D](../pymnn/cv.html#sepfilter2d-src-ddepth-kx-ky-delta-bordertype) | 2个一维kernel做滤波 |
| [Sobel](../pymnn/cv.html#sobel-src-ddepth-dx-dy-ksize-scale-delta-bordertype) | 边缘检测滤波 |
| [spatialGradient](../pymnn/cv.html#spatialgradient-src-ksize-bordertype) | 梯度，实际为Sobel |
| [sqrBoxFilter](../pymnn/cv.html#sqrboxfilter-src-ddepth-ksize-normalize-bordertype) | 平方后滤波 |
#### 图像形变
| 函数名 | 功能 |
| --- | --- |
| [getAffineTransform](../pymnn/cv.html#getaffinetransform-src-dst) | 仿射变换 |
| [getPerspectiveTransform](../pymnn/cv.html#getperspectivetransform-src-dst) | 透视变换 |
| [getRectSubPix](../pymnn/cv.html#getrectsubpix-image-patchsize-center) | 截取矩形区域 |
| [getRotationMatrix2D](../pymnn/cv.html#getrotationmatrix2d-center-angle-scale) | 旋转矩阵 |
| [invertAffineTransform](../pymnn/cv.html#invertaffinetransform-m) | 仿射变换矩阵求逆 |
| [resize](../pymnn/cv.html#resize-src-dsize-fx-fy-interpolation-code-mean-norm) | 图片放缩 |
| [warpAffine](../pymnn/cv.html#warpaffine-src-m-dsize-flag-bordermode-bordervalue-code-mean-norm) | 仿射变换 |
| [warpPerspective](../pymnn/cv.html#warpperspective-src-m-dsize-flag-bordermode-bordervalue) | 透视变换 |
#### 图像转换
| 函数名 | 功能 |
| --- | --- |
| [blendLinear](../pymnn/cv.html#blendlinear-src1-src2-weight1-weight2) | 线性混合2个图像 |
| [threshold](../pymnn/cv.html#threshold-src-thresh-maxval-type) | 逐像素阈值化 |
#### 绘画函数
| 函数名 | 功能 |
| --- | --- |
| [arrowedLine](../pymnn/cv.html#arrowedline-img-pt1-pt2-color-thickness-linetype-shift-tiplength) | 画箭头 |
| [circle](../pymnn/cv.html#circle-img-center-radius-color-thickness-linetype-shift) | 画圆 |
| [drawContours](../pymnn/cv.html#drawcontours-img-contours-contouridx-color-thickness-linetype) | 画轮廓 |
| [fillPoly](../pymnn/cv.html#fillpoly-img-contours-color-linetype-shift-offset) | 填充多边形 |
| [line](../pymnn/cv.html#line-img-pt1-pt2-color-thickness-linetype-shift) | 画线段 |
| [rectangle](../pymnn/cv.html#rectangle-src-pt1-pt2-color-thickness-linetype-shift) | 画矩形 |
#### 色彩空间转换
| 函数名 | 功能 |
| --- | --- |
| [cvtColor](../pymnn/cv.html#cvtcolor-src-code-dstcn) | 颜色空间转换 |
| [cvtColorTwoPlane](../pymnn/cv.html#cvtcolortwoplane-src1-src2-code) | YUV420到RGB的转换 |
#### 结构函数
| 函数名 | 功能 |
| --- | --- |
| [findContours](../pymnn/cv.html#findcontours-image-mode-method-offset) | 轮廓检测 |
| [contourArea](../pymnn/cv.html#contourarea-points-oriented) | 计算轮廓的面积 |
| [convexHull](../pymnn/cv.html#convexhull-points-clockwise-returnpoints) | 计算点集的凸包 |
| [minAreaRect](../pymnn/cv.html#minarearect-points) | 最小外接矩形 |
| [boundingRect](../pymnn/cv.html#boundingrect-points) | 计算点集的最小外接矩形 |
| [connectedComponentsWithStats](../pymnn/cv.html#connectedcomponentswithstats-image-connectivity) | 计算图像的连通域 |
| [boxPoints](../pymnn/cv.html#boxpoints-box) | 计算矩形的四个顶点坐标 |
#### 直方图
| 函数名 | 功能 |
| --- | --- |
| [calcHist](../pymnn/cv.html#calchist-imgs-channels-mask-histsize-ranges-accumulate) | 计算直方图 |
#### 3D
| 函数名 | 功能 |
| --- | --- |
| [Rodrigues](../pymnn/cv.html#rodrigues-src) | 旋转矩阵转换为旋转向量 |
| [solvePnP](../pymnn/cv.html#solvepnp-objectpoints-imagepoints-cameramatrix-distcoeffs-flags) | 计算2d到3d的映射 |
#### 数组操作函数
| 函数名 | 功能 |
| --- | --- |
| [copyTo](../pymnn/cv.html#copyto-src-mask-dst) | 带mask的拷贝 |
| [bitwise_and](../pymnn/cv.html#bitwise-and-src1-src2-dst-mask) | 带mask按位与 |
| [bitwise_or](../pymnn/cv.html#bitwise-or-src1-src2-dst-mask) | 带mask按位或 |
| [bitwise_xor](../pymnn/cv.html#bitwise-xor-src1-src2-dst-mask) | 带mask按位异或 |
| [hconcat](../pymnn/cv.html#hconcat-src) | 水平方向拼接 |
| [vconcat](../pymnn/cv.html#vconcat-src) | 垂直方向拼接 |
| [mean](../pymnn/cv.html#mean-src-mask) | 求均值 |
| [flip](../pymnn/cv.html#flip-src-flipcode) | 翻转 |
| [rotate](../pymnn/cv.html#rotate-src-rotatemode) | 旋转 |
### numpy能力列表
numpy函数170个，函数列表如下:
#### 数组创建
| 函数名 | 功能 |
| --- | --- |
| [empty](../numpy.html#empty-shape-dtype-float32-order-c) | 空数组 |
| [empty_like](https://numpy.org/doc/stable/reference/generated/numpy.empty_like.html#numpy.empty_like) | 空数组like |
| [eye](https://numpy.org/doc/stable/reference/generated/numpy.eye.html#numpy.eye) | 对角线2d数组 |
| [identity](https://numpy.org/doc/stable/reference/generated/numpy.identity.html#numpy.identity) | 对角线2d数组 |
| [ones](https://numpy.org/doc/stable/reference/generated/numpy.ones.html#numpy.ones) | 全1数组 |
| [ones_like](https://numpy.org/doc/stable/reference/generated/numpy.ones_like.html#numpy.ones_like) | 全1数组like |
| [zeros](https://numpy.org/doc/stable/reference/generated/numpy.zeros.html#numpy.zeros) | 全0数组 |
| [zeros_like](https://numpy.org/doc/stable/reference/generated/numpy.zeros_like.html#numpy.zeros_like) | 全0数组like |
| [full](https://numpy.org/doc/stable/reference/generated/numpy.full.html#numpy.full) | 填充 |
| [full_like](https://numpy.org/doc/stable/reference/generated/numpy.full_like.html#numpy.full_like) | 填充like |
| [array](https://numpy.org/doc/stable/reference/generated/numpy.array.html#numpy.array) | 创建数组 |
| [asarray](https://numpy.org/doc/stable/reference/generated/numpy.asarray.html#numpy.asarray) | 创建数组 |
| [asanyarray](https://numpy.org/doc/stable/reference/generated/numpy.asanyarray.html#numpy.asanyarray) | 创建数组 |
| [ascontiguousarray](https://numpy.org/doc/stable/reference/generated/numpy.ascontiguousarray.html#numpy.ascontiguousarray) | 创建数组 |
| [asmatrix](https://numpy.org/doc/stable/reference/generated/numpy.asmatrix.html#numpy.asmatrix) | 创建2d数组 |
| [copy](https://numpy.org/doc/stable/reference/generated/numpy.copy.html#numpy.copy) | 拷贝数组 |
| [arange](https://numpy.org/doc/stable/reference/generated/numpy.arange.html#numpy.arange) | 范围创建 |
| [linspace](https://numpy.org/doc/stable/reference/generated/numpy.linspace.html#numpy.linspace) | 区间创建 |
| [logspace](https://numpy.org/doc/stable/reference/generated/numpy.logspace.html#numpy.logspace) | log区间创建 |
| [geomspace](https://numpy.org/doc/stable/reference/generated/numpy.geomspace.html#numpy.geomspace) | log区间创建 |
| [meshgrid](https://numpy.org/doc/stable/reference/generated/numpy.meshgrid.html#numpy.meshgrid) | 坐标矩阵 |
| [mat](https://numpy.org/doc/stable/reference/generated/numpy.mat.html#numpy.mat) | 矩阵 |
#### 数组操作
| 函数名 | 功能 |
| --- | --- |
| [copyto](https://numpy.org/doc/stable/reference/generated/numpy.copyto.html#numpy.copyto) | 拷贝至 |
| [shape](https://numpy.org/doc/stable/reference/generated/numpy.shape.html#numpy.shape) | 获取形状 |
| [reshape](https://numpy.org/doc/stable/reference/generated/numpy.reshape.html#numpy.reshape) | 改变形状 |
| [ravel](https://numpy.org/doc/stable/reference/generated/numpy.ravel.html#numpy.ravel) | 拉平 |
| [flat](https://numpy.org/doc/stable/reference/generated/numpy.ndarray.flat.html#numpy.ndarray.flat) | 拉平 |
| [flatten](https://numpy.org/doc/stable/reference/generated/numpy.ndarray.flatten.html#numpy.ndarray.flatten) | 拉平 |
| [moveaxis](https://numpy.org/doc/stable/reference/generated/numpy.moveaxis.html#numpy.moveaxis) | 移动维度 |
| [rollaxis](https://numpy.org/doc/stable/reference/generated/numpy.rollaxis.html#numpy.rollaxis) | 轮转维度 |
| [swapaxes](https://numpy.org/doc/stable/reference/generated/numpy.swapaxes.html#numpy.swapaxes) | 交换维度 |
| [T](https://numpy.org/doc/stable/reference/generated/numpy.ndarray.T.html#numpy.ndarray.T) | 转置 |
| [transpose](https://numpy.org/doc/stable/reference/generated/numpy.transpose.html#numpy.transpose) | 转置 |
| [atleast_1d](https://numpy.org/doc/stable/reference/generated/numpy.atleast_1d.html#numpy.atleast_1d) | 至少1维 |
| [atleast_2d](https://numpy.org/doc/stable/reference/generated/numpy.atleast_2d.html#numpy.atleast_2d) | 至少2维 |
| [atleast_3d](https://numpy.org/doc/stable/reference/generated/numpy.atleast_3d.html#numpy.atleast_3d) | 至少3维 |
| [broadcast_to](https://numpy.org/doc/stable/reference/generated/numpy.broadcast_to.html#numpy.broadcast_to) | 广播 |
| [broadcast_arrays](https://numpy.org/doc/stable/reference/generated/numpy.broadcast_arrays.html#numpy.broadcast_arrays) | 数组广播 |
| [expand_dims](https://numpy.org/doc/stable/reference/generated/numpy.expand_dims.html#numpy.expand_dims) | 增加维度 |
| [squeeze](https://numpy.org/doc/stable/reference/generated/numpy.squeeze.html#numpy.squeeze) | 压缩1维度 |
| [asfarray](https://numpy.org/doc/stable/reference/generated/numpy.asfarray.html#numpy.asfarray) | 转浮点 |
| [asscalar](https://numpy.org/doc/stable/reference/generated/numpy.asscalar.html#numpy.asscalar) | 转标量 |
| [concatenate](https://numpy.org/doc/stable/reference/generated/numpy.concatenate.html#numpy.concatenate) | 连接 |
| [stack](https://numpy.org/doc/stable/reference/generated/numpy.stack.html#numpy.stack) | 连接 |
| [vstack](https://numpy.org/doc/stable/reference/generated/numpy.vstack.html#numpy.vstack) | 垂直连接 |
| [hstack](https://numpy.org/doc/stable/reference/generated/numpy.hstack.html#numpy.hstack) | 水平连接 |
| [dstack](https://numpy.org/doc/stable/reference/generated/numpy.dstack.html#numpy.dstack) | 深度连接 |
| [column_stack](https://numpy.org/doc/stable/reference/generated/numpy.column_stack.html#numpy.column_stack) | 列连接 |
| [row_stack](https://numpy.org/doc/stable/reference/generated/numpy.row_stack.html#numpy.row_stack) | 行连接 |
| [split](https://numpy.org/doc/stable/reference/generated/numpy.split.html#numpy.split) | 切分 |
| [array_split](https://numpy.org/doc/stable/reference/generated/numpy.array_split.html#numpy.array_split) | 数组切分 |
| [dsplit](https://numpy.org/doc/stable/reference/generated/numpy.dsplit.html#numpy.dsplit) | 深度切分 |
| [hsplit](https://numpy.org/doc/stable/reference/generated/numpy.hsplit.html#numpy.hsplit) | 水平切分 |
| [vsplit](https://numpy.org/doc/stable/reference/generated/numpy.vsplit.html#numpy.vsplit) | 垂直切分 |
| [tile](https://numpy.org/doc/stable/reference/generated/numpy.tile.html#numpy.tile) | 重复堆叠 |
| [repeat](https://numpy.org/doc/stable/reference/generated/numpy.repeat.html#numpy.repeat) | 重复 |
| [reshape](https://numpy.org/doc/stable/reference/generated/numpy.reshape.html#numpy.reshape) | 变形 |
#### 坐标操作
| 函数名 | 功能 |
| --- | --- |
| [nonzero](https://numpy.org/doc/stable/reference/generated/numpy.nonzero.html#numpy.nonzero) | 非0元素坐标 |
| [where](https://numpy.org/doc/stable/reference/generated/numpy.where.html#numpy.where) | 条件选取 |
| [unravel_index](https://numpy.org/doc/stable/reference/generated/numpy.unravel_index.html#numpy.unravel_index) | 反拉平坐标 |
#### 线性代数
| 函数名 | 功能 |
| --- | --- |
| [dot](https://numpy.org/doc/stable/reference/generated/numpy.dot.html#numpy.dot) | 点乘 |
| [vdot](https://numpy.org/doc/stable/reference/generated/numpy.vdot.html#numpy.vdot) | 点乘 |
| [inner](https://numpy.org/doc/stable/reference/generated/numpy.inner.html#numpy.inner) | 内积 |
| [matmul](https://numpy.org/doc/stable/reference/generated/numpy.matmul.html#numpy.matmul) | 矩阵乘 |
#### 逻辑函数
| 函数名 | 功能 |
| --- | --- |
| [all](https://numpy.org/doc/stable/reference/generated/numpy.all.html#numpy.all) | 全部非0 |
| [any](https://numpy.org/doc/stable/reference/generated/numpy.any.html#numpy.any) | 任意非0 |
| [logical_and](https://numpy.org/doc/stable/reference/generated/numpy.logical_and.html#numpy.logical_and) | 与 |
| [logical_or](https://numpy.org/doc/stable/reference/generated/numpy.logical_or.html#numpy.logical_or) | 或 |
| [logical_not](https://numpy.org/doc/stable/reference/generated/numpy.logical_not.html#numpy.logical_not) | 否 |
| [logical_xor](https://numpy.org/doc/stable/reference/generated/numpy.logical_xor.html#numpy.logical_xor) | 异或 |
| [array_equal](https://numpy.org/doc/stable/reference/generated/numpy.array_equal.html#numpy.array_equal) | 相等 |
| [array_equiv](https://numpy.org/doc/stable/reference/generated/numpy.array_equiv.html#numpy.array_equiv) | 相等 |
| [greater](https://numpy.org/doc/stable/reference/generated/numpy.greater.html#numpy.greater) | 大于 |
| [greater_equal](https://numpy.org/doc/stable/reference/generated/numpy.greater_equal.html#numpy.greater_equal) | 大于等于 |
| [less](https://numpy.org/doc/stable/reference/generated/numpy.less.html#numpy.less) | 小于 |
| [less_equal](https://numpy.org/doc/stable/reference/generated/numpy.less_equal.html#numpy.less_equal) | 小于等于 |
| [equal](https://numpy.org/doc/stable/reference/generated/numpy.equal.html#numpy.equal) | 等于 |
| [not_equal](https://numpy.org/doc/stable/reference/generated/numpy.not_equal.html#numpy.not_equal) | 不等 |
#### 数学函数
| API | 功能 |
| --- | --- |
| [sin](https://numpy.org/doc/stable/reference/generated/numpy.sin.html#numpy.sin) | 正弦 |
| [cos](https://numpy.org/doc/stable/reference/generated/numpy.cos.html#numpy.cos) | 余弦 |
| [tan](https://numpy.org/doc/stable/reference/generated/numpy.tan.html#numpy.tan) | 正切 |
| [arcsin](https://numpy.org/doc/stable/reference/generated/numpy.arcsin.html#numpy.arcsin) | 反正弦 |
| [arccos](https://numpy.org/doc/stable/reference/generated/numpy.arccos.html#numpy.arccos) | 反余弦 |
| [arctan](https://numpy.org/doc/stable/reference/generated/numpy.arctan.html#numpy.arctan) | 反正切 |
| [hypot](https://numpy.org/doc/stable/reference/generated/numpy.hypot.html#numpy.hypot) |
|
| [arctan2](https://numpy.org/doc/stable/reference/generated/numpy.arctan2.html#numpy.arctan2) |
|
| [sinh](https://numpy.org/doc/stable/reference/generated/numpy.sinh.html#numpy.sinh) |
|
| [cosh](https://numpy.org/doc/stable/reference/generated/numpy.cosh.html#numpy.cosh) |
|
| [tanh](https://numpy.org/doc/stable/reference/generated/numpy.tanh.html#numpy.tanh) |
|
| [arcsinh](https://numpy.org/doc/stable/reference/generated/numpy.arcsinh.html#numpy.arcsinh) |
|
| [arccosh](https://numpy.org/doc/stable/reference/generated/numpy.arccosh.html#numpy.arccosh) |
|
| [arctanh](https://numpy.org/doc/stable/reference/generated/numpy.arctanh.html#numpy.arctanh) |
|
| [around](https://numpy.org/doc/stable/reference/generated/numpy.around.html#numpy.around) |
|
| [round_](https://numpy.org/doc/stable/reference/generated/numpy.round_.html#numpy.round_) |
|
| [rint](https://numpy.org/doc/stable/reference/generated/numpy.rint.html#numpy.rint) |
|
| [floor](https://numpy.org/doc/stable/reference/generated/numpy.floor.html#numpy.floor) |
|
| [ceil](https://numpy.org/doc/stable/reference/generated/numpy.ceil.html#numpy.ceil) |
|
| [trunc](https://numpy.org/doc/stable/reference/generated/numpy.trunc.html#numpy.trunc) |
|
| [prod](https://numpy.org/doc/stable/reference/generated/numpy.prod.html#numpy.prod) | 积 |
| [sum](https://numpy.org/doc/stable/reference/generated/numpy.sum.html#numpy.sum) | 和 |
| [nanprod](https://numpy.org/doc/stable/reference/generated/numpy.nanprod.html#numpy.nanprod) | 积 |
| [nansum](https://numpy.org/doc/stable/reference/generated/numpy.nansum.html#numpy.nansum) | 和 |
| [exp](https://numpy.org/doc/stable/reference/generated/numpy.exp.html#numpy.exp) | e指数 |
| [expm1](https://numpy.org/doc/stable/reference/generated/numpy.expm1.html#numpy.expm1) | e指数-1 |
| [exp2](https://numpy.org/doc/stable/reference/generated/numpy.exp2.html#numpy.exp2) | 2指数 |
| [log](https://numpy.org/doc/stable/reference/generated/numpy.log.html#numpy.log) | 对数 |
| [log10](https://numpy.org/doc/stable/reference/generated/numpy.log10.html#numpy.log10) | 10对数 |
| [log2](https://numpy.org/doc/stable/reference/generated/numpy.log2.html#numpy.log2) | 2对数 |
| [log1p](https://numpy.org/doc/stable/reference/generated/numpy.log1p.html#numpy.log1p) | x+1对数 |
| [logaddexp](https://numpy.org/doc/stable/reference/generated/numpy.logaddexp.html#numpy.logaddexp) | exp对数 |
| [logaddexp2](https://numpy.org/doc/stable/reference/generated/numpy.logaddexp2.html#numpy.logaddexp2) | 2指数对数 |
| [sinc](https://numpy.org/doc/stable/reference/generated/numpy.sinc.html#numpy.sinc) | |
| [signbit](https://numpy.org/doc/stable/reference/generated/numpy.signbit.html#numpy.signbit) | |
| [copysign](https://numpy.org/doc/stable/reference/generated/numpy.copysign.html#numpy.copysign) | |
| [frexp](https://numpy.org/doc/stable/reference/generated/numpy.frexp.html#numpy.frexp) | |
| [ldexp](https://numpy.org/doc/stable/reference/generated/numpy.ldexp.html#numpy.ldexp) | |
| [add](https://numpy.org/doc/stable/reference/generated/numpy.add.html#numpy.add) | 加 |
| [reciprocal](https://numpy.org/doc/stable/reference/generated/numpy.reciprocal.html#numpy.reciprocal) | 倒数 |
| [positive](https://numpy.org/doc/stable/reference/generated/numpy.positive.html#numpy.positive) | 取正 |
| [negative](https://numpy.org/doc/stable/reference/generated/numpy.negative.html#numpy.negative) | 取负 |
| [multiply](https://numpy.org/doc/stable/reference/generated/numpy.multiply.html#numpy.multiply) | 乘 |
| [divide](https://numpy.org/doc/stable/reference/generated/numpy.divide.html#numpy.divide) | 除 |
| [power](https://numpy.org/doc/stable/reference/generated/numpy.power.html#numpy.power) | 指数 |
| [subtract](https://numpy.org/doc/stable/reference/generated/numpy.subtract.html#numpy.subtract) | 减 |
| [true_divide](https://numpy.org/doc/stable/reference/generated/numpy.true_divide.html#numpy.true_divide) | 除 |
| [floor_divide](https://numpy.org/doc/stable/reference/generated/numpy.floor_divide.html#numpy.floor_divide) | 除 |
| [float_power](https://numpy.org/doc/stable/reference/generated/numpy.float_power.html#numpy.float_power) | 指数 |
| [fmod](https://numpy.org/doc/stable/reference/generated/numpy.fmod.html#numpy.fmod) | 模 |
| [mod](https://numpy.org/doc/stable/reference/generated/numpy.mod.html#numpy.mod) | 模 |
| [modf](https://numpy.org/doc/stable/reference/generated/numpy.modf.html#numpy.modf) | 模 |
| [remainder](https://numpy.org/doc/stable/reference/generated/numpy.remainder.html#numpy.remainder) | 余 |
| [divmod](https://numpy.org/doc/stable/reference/generated/numpy.divmod.html#numpy.divmod) | 除，余 |
| [convolve](https://numpy.org/doc/stable/reference/generated/numpy.convolve.html#numpy.convolve) | 卷积 |
| [clip](https://numpy.org/doc/stable/reference/generated/numpy.clip.html#numpy.clip) | 缩小范围 |
| [sqrt](https://numpy.org/doc/stable/reference/generated/numpy.sqrt.html#numpy.sqrt) | 平方根 |
| [cbrt](https://numpy.org/doc/stable/reference/generated/numpy.cbrt.html#numpy.cbrt) | 立方根 |
| [square](https://numpy.org/doc/stable/reference/generated/numpy.square.html#numpy.square) | 平方 |
| [absolute](https://numpy.org/doc/stable/reference/generated/numpy.absolute.html#numpy.absolute) | 绝对值 |
| [fabs](https://numpy.org/doc/stable/reference/generated/numpy.fabs.html#numpy.fabs) | 绝对值 |
| [sign](https://numpy.org/doc/stable/reference/generated/numpy.sign.html#numpy.sign) | 符号 |
| [maximum](https://numpy.org/doc/stable/reference/generated/numpy.maximum.html#numpy.maximum) | 取大 |
| [minimum](https://numpy.org/doc/stable/reference/generated/numpy.minimum.html#numpy.minimum) | 取小 |
| [fmax](https://numpy.org/doc/stable/reference/generated/numpy.fmax.html#numpy.fmax) | 取大 |
| [fmin](https://numpy.org/doc/stable/reference/generated/numpy.fmin.html#numpy.fmin) | 取小 |
#### 数组扩充
| 函数名 | 功能 |
| --- | --- |
| [pad](https://numpy.org/doc/stable/reference/generated/numpy.pad.html#numpy.pad) | 扩充 |
#### 随机采样
| 函数名 | 功能 |
| --- | --- |
| random | 随机数 |
| rand | 随机数 |
| randn | 随机数 |
| randint | 随机定点数 |
#### 排序，搜索，计数
| 函数名 | 功能 |
| --- | --- |
| [sort](https://numpy.org/doc/stable/reference/generated/numpy.sort.html#numpy.sort)，[lexsort](https://numpy.org/doc/stable/reference/generated/numpy.lexsort.html#numpy.lexsort)，[argsort](https://numpy.org/doc/stable/reference/generated/numpy.argsort.html#numpy.argsort) | 排序 |
| [argmax](https://numpy.org/doc/stable/reference/generated/numpy.argmax.html#numpy.argmax) | 最大值坐标 |
| [nanargmax](https://numpy.org/doc/stable/reference/generated/numpy.nanargmax.html#numpy.nanargmax) | 最大值坐标 |
| [argmin](https://numpy.org/doc/stable/reference/generated/numpy.argmin.html#numpy.argmin) | 最小值坐标 |
| [nanargmin](https://numpy.org/doc/stable/reference/generated/numpy.nanargmin.html#numpy.nanargmin) | 最小值坐标 |
| [argwhere](https://numpy.org/doc/stable/reference/generated/numpy.argwhere.html#numpy.argwhere) | 非0坐标 |
| [flatnonzero](https://numpy.org/doc/stable/reference/generated/numpy.flatnonzero.html#numpy.flatnonzero) | 非0元素 |
| [count_nonzero](https://numpy.org/doc/stable/reference/generated/numpy.count_nonzero.html#numpy.count_nonzero) | 非0总数 |
#### 统计
| 函数名 | 功能 |
| --- | --- |
| [amin](https://numpy.org/doc/stable/reference/generated/numpy.amin.html#numpy.amin) | 最小值 |
| [amax](https://numpy.org/doc/stable/reference/generated/numpy.amax.html#numpy.amax) | 最大值 |
| [nanmin](https://numpy.org/doc/stable/reference/generated/numpy.nanmin.html#numpy.nanmin) | 最小值 |
| [nanmax](https://numpy.org/doc/stable/reference/generated/numpy.nanmax.html#numpy.nanmax) | 最大值 |
| [ptp](https://numpy.org/doc/stable/reference/generated/numpy.ptp.html#numpy.ptp) | 范围 |
| [average](https://numpy.org/doc/stable/reference/generated/numpy.average.html#numpy.average) | 均值 |
| [mean](https://numpy.org/doc/stable/reference/generated/numpy.mean.html#numpy.mean) | 均值 |
| [std](https://numpy.org/doc/stable/reference/generated/numpy.std.html#numpy.std) | 标准差 |
| [var](https://numpy.org/doc/stable/reference/generated/numpy.var.html#numpy.var) | 方差 |
| [nanmean](https://numpy.org/doc/stable/reference/generated/numpy.nanmean.html#numpy.nanmean) | 均值 |
| [nanstd](https://numpy.org/doc/stable/reference/generated/numpy.nanstd.html#numpy.nanstd) | 标准差 |
| [nanvar](https://numpy.org/doc/stable/reference/generated/numpy.nanvar.html#numpy.nanvar) | 方差 |
Reference in New Issue View Git Blame Copy Permalink