DOCSEARCHbyAlgolia
Submit Docs

Alaya Documentation

website·Indexed 32 days ago

https://docs.alayanew.com

Documentation for Alaya's AI computing platform, featuring guides on VKS (Virtual Kubernetes Service), CCI, and HyperTrain. It includes tutorials for LLM fine-tuning with LLaMA Factory, vLLM deployment, and remote development using VS Code, DevPod, and SSH. The platform provides comprehensive storage options including NAS, Object Storage (OSS), and high-performance file storage, alongside API references for storage management and tooling for image acceleration and mirrors.

Tokens
233.1K
Snippets
683
Records
1.5K
Agent score
60%

What's inside Alaya

  1. Overview of Alaya NeW Computing Products

    Alaya NeW provides high-performance computing resources through three primary offerings:

    1. Computing Packages (算力包): Elastic high-performance computing measured in DCU (Data Computing Unit). 1 DCU = 312 TFLOPS × 1 hour. Supports on-demand consumption and dynamic scaling.
    2. Elastic Container Cluster (VKS): A Kubernetes cluster with multi-node, multi-card support and high-speed network interconnects, optimized for large model training.
    3. Cloud Container Instance (CCI): A lightweight, serverless container service based on Kubernetes, suitable for AI inference, rapid model verification, and development.

  2. Overview of the Intelligent Driving Cloud Solution

    The Intelligent Driving Cloud Solution provides a one-stop ecosystem for autonomous driving enterprises. It leverages high-performance computing (HPC) cloud services, cloud infrastructure, and an end-to-end toolchain to accelerate iteration efficiency and reduce the cost and barriers of business deployment. The solution covers the entire lifecycle: data management, model training, cloud simulation verification, and inference deployment.

  3. Overview of Virtual Kubernetes Service (VKS)

    Virtual Kubernetes Service (VKS) is a managed Kubernetes service provided by Alaya AI. It is specifically designed for training and inference workloads, offering native support for GPU scheduling, cross-node RDMA, shared/dynamic storage, and automated service publishing to allow on-demand GPU utilization.

  4. Overview of Jiu Zhang Intelligent Computing Cloud OpenAPI

    The Jiu Zhang Intelligent Computing Cloud OpenAPI provides standardized, programmable RESTful interfaces for GPU computing power, distributed scheduling, container technology, and storage management. It allows developers to integrate resource creation, task scheduling, and monitoring/operations directly into their own systems or automated workflows.

  5. Overview of OpenClaw AI Assistant

    OpenClaw is an open-source personal AI assistant hosted on Alaya NeW Cloud. Unlike standard chat-based LLMs, OpenClaw can be embedded into daily toolchains (Lark, IDEs, and terminals) and is capable of executing tasks directly on a user's machine—such as organizing files, sending emails, and processing data—provided it has the necessary authorization. It operates 24/7 and supports multiple IM platforms including Feishu, DingTalk, WhatsApp, and Telegram.

  6. Overview of System Messages in Customer Center

    The System Messages module in the Customer Center provides a centralized view of all notifications and reminders sent by the platform to the user. It allows users to track platform dynamics and business information across several categories: Product, Service, Activity, Security, and Failure messages.

  7. Overview of OpenClaw AI Agent

    OpenClaw is an open-source AI agent designed as a 'digital employee' capable of executing real-world tasks on a computer, such as organizing files, sending emails, and processing data. It operates 24/7 in the cloud, removing the need for local hardware resources. It supports integration with multiple chat platforms including Feishu, DingTalk, WhatsApp, and Telegram.

  8. Overview of RDMA High-Performance Networking

    RDMA (Remote Direct Memory Access) is a high-performance network communication technology that allows a computer to access memory data from another computer directly, bypassing the CPU, interrupts, and operating system kernel. This results in low latency, high bandwidth, low CPU utilization, and zero-copy data transfers (direct transfer between application buffers).

  9. Overview of LLaMA Factory for LLM Fine-Tuning

    LLaMA Factory is an open-source, low-code framework designed for fine-tuning Large Language Models (LLMs). It provides a zero-code WebUI for model selection, dataset preparation, training, evaluation, and exporting. The framework supports over 100 models and integrates various tuning algorithms and monitoring tools, making it suitable for domain-specific tuning (medical, legal, financial), task optimization (QA, translation), and resource-constrained environments using LoRA/QLoRA.

  10. Overview of Alaya NeW Resource Quotas

    Resource Quotas provide centralized management and allocation of usable resources within an Intelligent Computing Center. Enterprise administrators can monitor resource usage, view quotas, and control the boundaries for resource activation. Quotas are managed independently per Intelligent Computing Center; changes in one center do not affect others.

  11. Overview of the Industry Research Agent Solution

    The Industry Research Agent Solution is a one-stop platform built on Alaya NeW Cloud (DataCanvas). It integrates industry knowledge graphs, generative AI, and intelligent decision-making technologies to automate the entire research workflow—from data collection and cleaning to analysis and report generation. The solution aims to increase research efficiency by up to 90% and ensure professional accuracy through a combination of Knowledge Graphs and RAG (Retrieval-Augmented Generation).

  12. Overview of Manufacturing Vertical Solution

    The Manufacturing Vertical Solution leverages a manufacturing-specific Large Language Model (LLM) base and integrates enterprise IT (Information Technology) and OT (Operational Technology) data to build intelligent agents. It is designed to empower the entire manufacturing lifecycle—including research, production, supply, sales, and service—supporting scenarios such as operational analysis, production process optimization, and equipment failure diagnosis.