TVM Chinese Website Is Officially Launched! The Most Comprehensive Machine Learning Model Deployment "reference Book" Is Here
Contents at a glance: Recently, the TVM Chinese document, translated and proofread by MLC community volunteers, was officially released and is now hosted on Hyper.AI's official website.
Keywords: TVM open source machine learning compiler
This article was first published on WeChat official account: HyperAI
Five years after its launch, TVM has become a popular deep learning compilation stack
In August 2017, Chen Tianqi, then a doctoral student at the University of Washington, and his team members officially released TVM.
This is an open source model compilation framework, the full name of which is Tensor Virtual Machine, which means tensor virtual machine.It aims to automatically compile machine learning algorithms into machine language that can be executed by the underlying hardware, thereby leveraging multiple types of computing power.
TVM system overview: Applicable to CPU, GPU and various machine learning acceleration chips
paper:
TVM: An Automated End-to-End Optimizing Compiler for Deep Learning
In the Zhihu question "How do you evaluate the newly open-source TVM of Chen Tianqi's team?", Chen Tianqi personally explained it, saying:
TVM attempts to summarize the manual optimization experience of deep learning ops at a higher level of abstraction, so that users can quickly explore efficient op implementation space in an automatic or semi-automatic way.
This question has received nearly 260,000 views on Zhihu.
Since its opening to the public nearly 5 years ago,TVM has the advantages of flexibility, high performance, and adaptability to any hardware. It is favored by many major manufacturers, such as NVIDIA, AMD, ARM, AWS, Facebook, etc., all of which use TVM, an open source end-to-end deep learning hardware compilation stack, to enhance support for arbitrary hardware.
It took 3 months to release the Chinese documentation for v0.10.0
Currently, the Chinese learning materials related to TVM on the Internet are relatively scattered. The few Chinese materials available are mostly compiled and output by practitioners or enthusiasts in related industries based on their own study or work needs.
In order to help domestic developers systematically learn TVM and promote the development of machine learning compilation in China,MLC (Machine Learning Compilation) community volunteers started with the documentation and localized the TVM official documentation v0.10.0 in an open source collaborative way.
TVM Chinese documentation is hosted by Hyper.AI
Portal →https://tvm.hyper.ai/
The Chinese documentation is based on the structure of the English documentation on the TVM official website.Adjustments have been made to better suit Chinese people's reading habits, providing a series of practical documents for reference from installing TVM to using development.
There are 3 ways to install TVM: source code installation, Docker image installation, and NNPACK Contrib installation
In the user tutorial section, more than ten tutorials are also provided for developers to quickly search and locate for learning.
Currently, TVM Chinese documentation is still being updated. , We hope that more developers who care about and focus on machine learning compilation will join in, gain value from the community, and create value for the community.
We welcome all forms of contributions, including but not limited to:
- Claim the translation and review of TVM English documents
- Submit an issue or PR to make Chinese documents more standardized
- Promote project tutorials, blog posts, and actively participate in community activities
- More
Open Source Community Everything comes from the community and everything gives back to the community. Visit the TVM Chinese documentation GitHub Repo now. We hope you will become a member of us and contribute to the development of machine learning compilers in the Chinese field!
⭐️ TVM in GitHub:
⭐️ TVM Chinese on GitHub:
https://github.com/hyperai/tvm-cn
Acknowledgements:
The successful launch of TVM Chinese site is inseparable from the strong support of MLC.AI community. We sincerely thank all community members for their active participation and assistance:
-- over--