Skip to content

Commit

Permalink
Merge pull request #3903 from windsonsea/nav
Browse files Browse the repository at this point in the history
[en] update baize/intro
  • Loading branch information
windsonsea authored Feb 23, 2024
2 parents 69b480f + 2f00641 commit 5dba787
Show file tree
Hide file tree
Showing 3 changed files with 31 additions and 16 deletions.
41 changes: 28 additions & 13 deletions docs/en/docs/baize/intro/index.md
Original file line number Diff line number Diff line change
@@ -1,40 +1,55 @@
---
MTPE: windsonsea
date: 2024-02-23
hide:
- toc
---

# What is Cloud Native AI
# What is Baize

Cloud Native AI is an AI computing platform based on a cloud-native operating system introduced by DaoCloud (referred to as the DaoCloud AI Computing Platform). The DaoCloud AI Computing Platform provides a software and hardware integrated AI computing experience, integrating heterogeneous computing power, optimizing GPU performance, achieving unified scheduling and operation of computing resources, maximizing computing efficiency, reducing computing costs, and providing optimized AI development frameworks to simplify AI development and deployment, accelerating the implementation of AI applications in various industries.
Baize is a cloud native AI computing platform developed by DaoCloud that can running on any K8s-based system.
This innovative platform, known as Cloud Native AI Platform, offers a seamless integration of software
and hardware for a superior AI computing experience. Baize combines various computing powers,
optimizes GPU performance, streamlines the scheduling and management of computing resources,
and enhances computing efficiency. With its focus on reducing costs and simplifying AI development,
Baize provides optimized AI development frameworks that accelerate the implementation of AI applications
across different industries.

**Key Features**

- Fully Managed Computing Resources

Leveraging DCE (DaoCloud Enterprise), it provides powerful infrastructure capabilities, supporting super-large-scale computing clusters, heterogeneous GPUs, and one-stop hosting, as well as a series of software and hardware integrated acceleration solutions such as vGPU.
Powered by DCE (DaoCloud Enterprise), Baize offers robust infrastructure capabilities to support
large-scale computing clusters, heterogeneous GPUs, and comprehensive hosting services.
It also includes integrated acceleration solutions like vGPU.

- Data Orchestration
- Dataset Management

Supports data management and orchestration capabilities during model development, providing functions such as dataset management, multi-data source access, dataset preloading, etc. It is optimized from the underlying container storage engine to ensure efficient and stable data processing.
Baize supports efficient dataset management during model development, with features such as
dataset organization, multi-data source access, and dataset preloading.
The underlying container storage engine ensures smooth and effective data processing.

- Development Environment Management

Meets the needs of MLOps and LLMOps engineers for development environments, providing various development environments including JupyterLab, VSCode (in progress), supporting custom development environments, and one-click mounting of various GPU, dataset, and other resources.
Baize caters to the requirements of MLOps and LLMOps engineers by providing a range of
development environments, including JupyterLab and upcoming support for VSCode. It allows
for custom development environments and easy access to GPU, dataset, and other resources.

- Task Management
- Job Management

Supports full lifecycle management of training tasks, providing various ways to quickly create tasks; supports mainstream task frameworks such as Pytorch, TensorFlow, PaddlePaddle, naturally supporting various task scheduling types including single-machine, distributed, multi-node, multi-GPU, etc.
Baize offers comprehensive job lifecycle management, enabling quick creation of jobs and support
for popular job frameworks like Pytorch, TensorFlow, and PaddlePaddle. It includes various
job scheduling options, such as single-machine, distributed, multi-node, and multi-GPU setups.

- GPU Management

Allows viewing of all GPU resources and GPU usage, supports viewing the current and historical running tasks in GPUs, facilitating GPU stress assessment.
Users can monitor GPU resources and usage through Baize, with features for tracking current
and historical job activities on GPUs. This facilitates GPU stress assessment and optimization.

- Queue Management

Supports creating queues and associating them with workspaces to ensure the coordination and isolation of queue resources in various clusters.

**Product Logical Architecture**

Baize allows for the creation of queues and their association with workspaces to ensure
efficient coordination and resource isolation within different clusters.

[Download DCE 5.0](../../download/index.md){ .md-button .md-button--primary }
[Install DCE 5.0](../../install/index.md){ .md-button .md-button--primary }
Expand Down
4 changes: 2 additions & 2 deletions docs/en/navigation.yml
Original file line number Diff line number Diff line change
Expand Up @@ -979,9 +979,9 @@ nav:
- Config Parameters: middleware/minio/user-guide/config.md
- MinIO Identity Management: middleware/minio/user-guide/user-management.md
- Cloud Native AI:
- Cloud Native AI:
- Baize:
- Introduction:
- What is Cloud Native AI: baize/intro/index.md
- What is Baize: baize/intro/index.md
- Installation: baize/intro/install.md
- First Time Usage: baize/intro/first.md
- Development Console:
Expand Down
2 changes: 1 addition & 1 deletion docs/zh/navigation.yml
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ nav:
- 产品文档:
- 首页: dce/index.md
- 云原生启航: dce/get-started.md
- 安装:
- 探索和安装:
- 产品介绍:
- 申请社区免费体验: dce/license0.md
- 功能和优势: dce/features.md
Expand Down

0 comments on commit 5dba787

Please sign in to comment.