Draft 0.1 release plan

[scarlett2018]

The issue is to list out what we would like to mark as 0.1.

[nnfbot]

Thanks for the report @scarlett2018! I will look into it ASAP! (I'm a bot).

[AlisaChen98]

• Build and Installation:

  • Support out-of-box installation with docker image
  • Support source code install on native system and docker
  • Support devices like CUDA GPUs, and ROCm GPUs.

• Models, Framework and Operators:

  • Support DNN model formats including TensorFlow and ONNX
  • Support commonly used models including AlexNet, VGG11, ResNet50, seq2seq, BERT, etc.
  • Support more than 100 commonly used operators.

• Model Compilation and Execution:

  • Provide a full-stack optimization mechanism, including data-flow graph optimizations, model-specific kernel selection, kernel co-scheduling, etc.
  • Provide ahead-of-time and source-to-source(model-to-code) compilation to reduce runtime overhead
  • Remove third-party library or framework dependencies

• Usability:

  • Provide command line tool nnfusion
  • Provide tools for users to freeze TensorFlow and PyTorch models
  • Provide flexible way to customize optimization through direct code modification on generated code

[AlisaChen98]

NNFusion V0.1 发版说明

• 构建及安装
  • 支持使用docker镜像直接安装
  • 支持根据项目源代码在本机系统或docker环境中自行编译安装
  • 支持CUDA GPU以及ROCm GPU等设备
• 模型、框架及运算符
  • 支持目前主流深度神经网络模型格式，包括TensorFlow及ONNX
  • 支持多种常用模型，包括AlexNet、VGG11、ResNet50、seq2seq、BERT等
  • 支持超过100种常用的算子
• 模型的编译与执行
  • 提供丰富的性能优化策略，包括数据流图的优化、模型特定的内核选择、算子协同调度等
  • 支持端到端的模型到源代码的AOT编译，从而有效消除运行时开销
  • 消除对第三方库或框架的依赖
• 易用性：
  • 提供命令行工具 nnfusion 来支持一键模型编译
  • 提供导出TensorFlow及PyTorch模型的工具
  • 通过支持直接修改生成的代码来提供灵活的模型定制化优化

[jlxue]

@AlisaChen98 @scarlett2018 i have updated the release notes, pls take a look.

[AlisaChen98]

@AlisaChen98 @scarlett2018 i have updated the release notes, pls take a look.

Edit to:

• Provide ahead-of-time and source-to-source(model-to-code) compilation to reduce runtime overhead
• Remove third-party library or framework dependencies