1 2 3[](https://pypi.org/project/mindspore) 4[](https://badge.fury.io/py/mindspore) 5[](https://pepy.tech/project/mindspore) 6[](https://hub.docker.com/r/mindspore/mindspore-cpu) 7[](https://github.com/mindspore-ai/mindspore/blob/master/LICENSE) 8[](https://join.slack.com/t/mindspore/shared_invite/zt-dgk65rli-3ex4xvS4wHX7UDmsQmfu8w) 9[](https://gitee.com/mindspore/mindspore/pulls) 10 11[查看中文](./README_CN.md) 12 13<!-- TOC --> 14 15- [What Is MindSpore](#what-is-mindspore) 16 - [Automatic Differentiation](#automatic-differentiation) 17 - [Automatic Parallel](#automatic-parallel) 18- [Installation](#installation) 19 - [Pip mode method installation](#pip-mode-method-installation) 20 - [Source code compilation installation](#source-code-compilation-installation) 21 - [Docker Image](#docker-image) 22- [Quickstart](#quickstart) 23- [Docs](#docs) 24- [Community](#community) 25 - [Governance](#governance) 26 - [Communication](#communication) 27- [Contributing](#contributing) 28- [Maintenance phases](#maintenance-phases) 29- [Maintenance status](#maintenance-status) 30- [Release Notes](#release-notes) 31- [License](#license) 32 33<!-- /TOC --> 34 35## What Is MindSpore 36 37MindSpore is a new open source deep learning training/inference framework that 38could be used for mobile, edge and cloud scenarios. MindSpore is designed to 39provide development experience with friendly design and efficient execution for 40the data scientists and algorithmic engineers, native support for Ascend AI 41processor, and software hardware co-optimization. At the meantime MindSpore as 42a global AI open source community, aims to further advance the development and 43enrichment of the AI software/hardware application ecosystem. 44 45<img src="https://gitee.com/mindspore/mindspore/raw/master/docs/MindSpore-architecture.png" alt="MindSpore Architecture"/> 46 47For more details please check out our [Architecture Guide](https://www.mindspore.cn/tutorials/en/master/beginner/introduction.html). 48 49### Automatic Differentiation 50 51Currently, there are two automatic differentiation techniques in mainstream deep learning frameworks: 52 53- **Operator Overloading (OO)**: Overloading the basic operators of the programming language to encapsulate their gradient rules. Record the operation trajectory of the network during forward execution in an operator overloaded manner, then apply the chain rule to the dynamically generated data flow graph to implement automatic differentiation. 54- **Source Transformation (ST)**: This technology is evolving from the functional programming framework and performs automatic differential transformation on the intermediate expression (the expression form of the program during the compilation process) in the form of just-in-time compilation (JIT), supporting complex control flow scenarios, higher-order functions and closures. 55 56PyTorch used OO. Compared to ST, OO generates gradient graph in runtime, so it does not need to take function call and control flow into consideration, which makes it easier to develop. However, OO can not perform gradient graph optimization in compilation time and the control flow has to be unfolded in runtime, so it is difficult to achieve extreme optimization in performance. 57 58MindSpore implemented automatic differentiation based on ST. On the one hand, it supports automatic differentiation of automatic control flow, so it is quite convenient to build models like PyTorch. On the other hand, MindSpore can perform static compilation optimization on neural networks to achieve great performance. 59 60<img src="https://gitee.com/mindspore/mindspore/raw/master/docs/Automatic-differentiation.png" alt="Automatic Differentiation" width="600"/> 61 62The implementation of MindSpore automatic differentiation can be understood as the symbolic differentiation of the program itself. Because MindSpore IR is a functional intermediate expression, it has an intuitive correspondence with the composite function in basic algebra. The derivation formula of the composite function composed of arbitrary basic functions can be derived. Each primitive operation in MindSpore IR can correspond to the basic functions in basic algebra, which can build more complex flow control. 63 64### Automatic Parallel 65 66The goal of MindSpore automatic parallel is to build a training method that combines data parallelism, model parallelism, and hybrid parallelism. It can automatically select a least cost model splitting strategy to achieve automatic distributed parallel training. 67 68<img src="https://gitee.com/mindspore/mindspore/raw/master/docs/Automatic-parallel.png" alt="Automatic Parallel" width="600"/> 69 70At present, MindSpore uses a fine-grained parallel strategy of splitting operators, that is, each operator in the figure is split into a cluster to complete parallel operations. The splitting strategy during this period may be very complicated, but as a developer advocating Pythonic, you don't need to care about the underlying implementation, as long as the top-level API compute is efficient. 71 72## Installation 73 74### Pip mode method installation 75 76MindSpore offers build options across multiple backends: 77 78| Hardware Platform | Operating System | Status | 79| :---------------- | :--------------- | :----- | 80| Ascend910 | Ubuntu-x86 | ✔️ | 81| | Ubuntu-aarch64 | ✔️ | 82| | EulerOS-aarch64 | ✔️ | 83| | CentOS-x86 | ✔️ | 84| | CentOS-aarch64 | ✔️ | 85| GPU CUDA 10.1 | Ubuntu-x86 | ✔️ | 86| CPU | Ubuntu-x86 | ✔️ | 87| | Ubuntu-aarch64 | ✔️ | 88| | Windows-x86 | ✔️ | 89 90For installation using `pip`, take `CPU` and `Ubuntu-x86` build version as an example: 91 921. Download whl from [MindSpore download page](https://www.mindspore.cn/versions/en), and install the package. 93 94 ```bash 95 pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.2.0-rc1/MindSpore/cpu/ubuntu_x86/mindspore-1.2.0rc1-cp37-cp37m-linux_x86_64.whl 96 ``` 97 982. Run the following command to verify the install. 99 100 ```python 101 import numpy as np 102 import mindspore.context as context 103 import mindspore.nn as nn 104 from mindspore import Tensor 105 from mindspore.ops import operations as P 106 107 context.set_context(mode=context.GRAPH_MODE, device_target="CPU") 108 109 class Mul(nn.Cell): 110 def __init__(self): 111 super(Mul, self).__init__() 112 self.mul = P.Mul() 113 114 def construct(self, x, y): 115 return self.mul(x, y) 116 117 x = Tensor(np.array([1.0, 2.0, 3.0]).astype(np.float32)) 118 y = Tensor(np.array([4.0, 5.0, 6.0]).astype(np.float32)) 119 120 mul = Mul() 121 print(mul(x, y)) 122 ``` 123 124 ```text 125 [ 4. 10. 18.] 126 ``` 127 128Use pip mode method to install MindSpore in different environments. Refer to the following documents. 129 130- [Using pip mode method to install MindSpore in Ascend environment](https://gitee.com/mindspore/docs/blob/master/install/mindspore_ascend_install_pip_en.md) 131- [Using pip mode method to install MindSpore in GPU environment](https://gitee.com/mindspore/docs/blob/master/install/mindspore_gpu_install_pip_en.md) 132- [Using pip mode method to install MindSpore in CPU environment](https://gitee.com/mindspore/docs/blob/master/install/mindspore_cpu_install_pip_en.md) 133 134### Source code compilation installation 135 136Use the source code compilation method to install MindSpore in different environments. Refer to the following documents. 137 138- [Using the source code compilation method to install MindSpore in Ascend environment](https://gitee.com/mindspore/docs/blob/master/install/mindspore_ascend_install_source_en.md) 139- [Using the source code compilation method to install MindSpore in GPU environment](https://gitee.com/mindspore/docs/blob/master/install/mindspore_gpu_install_source_en.md) 140- [Using the source code compilation method to install MindSpore in CPU environment](https://gitee.com/mindspore/docs/blob/master/install/mindspore_cpu_install_source_en.md) 141 142### Docker Image 143 144MindSpore docker image is hosted on [Docker Hub](https://hub.docker.com/r/mindspore), 145currently the containerized build options are supported as follows: 146 147| Hardware Platform | Docker Image Repository | Tag | Description | 148| :---------------- | :---------------------- | :-- | :---------- | 149| CPU | `mindspore/mindspore-cpu` | `x.y.z` | Production environment with pre-installed MindSpore `x.y.z` CPU release. | 150| | | `devel` | Development environment provided to build MindSpore (with `CPU` backend) from the source, refer to <https://www.mindspore.cn/install/en> for installation details. | 151| | | `runtime` | Runtime environment provided to install MindSpore binary package with `CPU` backend. | 152| GPU | `mindspore/mindspore-gpu` | `x.y.z` | Production environment with pre-installed MindSpore `x.y.z` GPU release. | 153| | | `devel` | Development environment provided to build MindSpore (with `GPU CUDA10.1` backend) from the source, refer to <https://www.mindspore.cn/install/en> for installation details. | 154| | | `runtime` | Runtime environment provided to install MindSpore binary package with `GPU CUDA10.1` backend. | 155 156> **NOTICE:** For GPU `devel` docker image, it's NOT suggested to directly install the whl package after building from the source, instead we strongly RECOMMEND you transfer and install the whl package inside GPU `runtime` docker image. 157 158- CPU 159 160 For `CPU` backend, you can directly pull and run the latest stable image using the below command: 161 162 ```bash 163 docker pull mindspore/mindspore-cpu:1.1.0 164 docker run -it mindspore/mindspore-cpu:1.1.0 /bin/bash 165 ``` 166 167- GPU 168 169 For `GPU` backend, please make sure the `nvidia-container-toolkit` has been installed in advance, here are some install guidelines for `Ubuntu` users: 170 171 ```bash 172 DISTRIBUTION=$(. /etc/os-release; echo $ID$VERSION_ID) 173 curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | apt-key add - 174 curl -s -L https://nvidia.github.io/nvidia-docker/$DISTRIBUTION/nvidia-docker.list | tee /etc/apt/sources.list.d/nvidia-docker.list 175 176 sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit nvidia-docker2 177 sudo systemctl restart docker 178 ``` 179 180 Then edit the file daemon.json: 181 182 ```bash 183 $ vim /etc/docker/daemon.json 184 { 185 "runtimes": { 186 "nvidia": { 187 "path": "nvidia-container-runtime", 188 "runtimeArgs": [] 189 } 190 } 191 } 192 ``` 193 194 Restart docker again: 195 196 ```bash 197 sudo systemctl daemon-reload 198 sudo systemctl restart docker 199 ``` 200 201 Then you can pull and run the latest stable image using the below command: 202 203 ```bash 204 docker pull mindspore/mindspore-gpu:1.1.0 205 docker run -it -v /dev/shm:/dev/shm --runtime=nvidia --privileged=true mindspore/mindspore-gpu:1.1.0 /bin/bash 206 ``` 207 208 To test if the docker image works, please execute the python code below and check the output: 209 210 ```python 211 import numpy as np 212 import mindspore.context as context 213 from mindspore import Tensor 214 from mindspore.ops import functional as F 215 216 context.set_context(mode=context.PYNATIVE_MODE, device_target="GPU") 217 218 x = Tensor(np.ones([1,3,3,4]).astype(np.float32)) 219 y = Tensor(np.ones([1,3,3,4]).astype(np.float32)) 220 print(F.tensor_add(x, y)) 221 ``` 222 223 ```text 224 [[[ 2. 2. 2. 2.], 225 [ 2. 2. 2. 2.], 226 [ 2. 2. 2. 2.]], 227 228 [[ 2. 2. 2. 2.], 229 [ 2. 2. 2. 2.], 230 [ 2. 2. 2. 2.]], 231 232 [[ 2. 2. 2. 2.], 233 [ 2. 2. 2. 2.], 234 [ 2. 2. 2. 2.]]] 235 ``` 236 237If you want to learn more about the building process of MindSpore docker images, 238please check out [docker](https://gitee.com/mindspore/mindspore/blob/master/scripts/docker/README.md) repo for the details. 239 240## Quickstart 241 242See the [Quick Start](https://www.mindspore.cn/tutorials/en/master/beginner/quick_start.html) 243to implement the image classification. 244 245## Docs 246 247More details about installation guide, tutorials and APIs, please see the 248[User Documentation](https://gitee.com/mindspore/docs). 249 250## Community 251 252### Governance 253 254Check out how MindSpore Open Governance [works](https://gitee.com/mindspore/community/blob/master/governance.md). 255 256### Communication 257 258- [MindSpore Slack](https://join.slack.com/t/mindspore/shared_invite/zt-dgk65rli-3ex4xvS4wHX7UDmsQmfu8w) - Communication platform for developers. 259- IRC channel at `#mindspore` (only for meeting minutes logging purpose) 260- Video Conferencing: TBD 261- Mailing-list: <https://mailweb.mindspore.cn/postorius/lists> 262 263## Contributing 264 265Welcome contributions. See our [Contributor Wiki](https://gitee.com/mindspore/mindspore/blob/master/CONTRIBUTING.md) for 266more details. 267 268## Maintenance phases 269 270Project stable branches will be in one of the following states: 271 272| **State** | **Time frame** | **Summary** | 273|-------------|---------------|--------------------------------------------------| 274| Planning | 1 - 3 months | Features are under planning. | 275| Development | 3 months | Features are under development. | 276| Maintained | 6 - 12 months | All bugfixes are appropriate. Releases produced. | 277| Unmaintained| 0 - 3 months | All bugfixes are appropriate. No Maintainers and No Releases produced. | 278| End Of Life (EOL) | N/A | Branch no longer accepting changes. | 279 280## Maintenance status 281 282| **Branch** | **Status** | **Initial Release Date** | **Next Phase** | **EOL Date**| 283|------------|--------------|--------------------------|----------------------------------------|-------------| 284| **r1.8** | Maintained | 2022-07-29 | Unmaintained <br> 2023-07-29 estimated | | 285| **r1.7** | Maintained | 2022-04-29 | Unmaintained <br> 2023-04-29 estimated | | 286| **r1.6** | Maintained | 2022-01-29 | Unmaintained <br> 2023-01-29 estimated | | 287| **r1.5** | Maintained | 2021-10-15 | Unmaintained <br> 2022-10-15 estimated | | 288| **r1.4** | Maintained | 2021-08-15 | Unmaintained <br> 2022-08-15 estimated | | 289| **r1.3** | End Of Life | 2021-07-15 | | 2022-07-15 | 290| **r1.2** | End Of Life | 2021-04-15 | | 2022-04-29 | 291| **r1.1** | End Of Life | 2020-12-31 | | 2021-09-30 | 292| **r1.0** | End Of Life | 2020-09-24 | | 2021-07-30 | 293| **r0.7** | End Of Life | 2020-08-31 | | 2021-02-28 | 294| **r0.6** | End Of Life | 2020-07-31 | | 2020-12-30 | 295| **r0.5** | End Of Life | 2020-06-30 | | 2021-06-30 | 296| **r0.3** | End Of Life | 2020-05-31 | | 2020-09-30 | 297| **r0.2** | End Of Life | 2020-04-30 | | 2020-08-31 | 298| **r0.1** | End Of Life | 2020-03-28 | | 2020-06-30 | 299 300## Release Notes 301 302The release notes, see our [RELEASE](https://gitee.com/mindspore/mindspore/blob/master/RELEASE.md). 303 304## License 305 306[Apache License 2.0](https://gitee.com/mindspore/mindspore/blob/master/LICENSE) 307