Data Management (DMS) offers managed Dify to implement deep integration of Data+AI development. DMS is a one-stop solution that covers the entire development lifecycle, including Notebook development, data processing, model building, and Large Language Model (LLM) application development. DMS allows you to use Dify as a platform to develop LLM applications, helping enterprises implement AI solutions.
Container resources in the Dify instance belong to users.
Background information
Dify is an open-source LLM application development platform that helps developers quickly create production-level generative AI applications. Dify combines Backend as a Service (BaaS) with LLMOps, greatly simplifying the development process of generative AI applications. Regardless of technical expertise, Dify allows you to easily participate in AI projects and data management activities, quickly creating enterprise-level applications. For more information, see Dify documentation.
DMS Dify provides all features of open-source Dify and also integrates with the Alibaba Cloud ecosystem covering Tongyi LLM, ApsaraDB vector engine, RAG retrieval service, and NL2SQL. This helps implement end-to-end production-level AI applications.
Overview
Log on to the DMS console. Create a workspace and go to the Dify environment.
Add members to the workspace for collaborative development.
Configure Dify resources to seamlessly integrate with DMS.
Use Dify to develop LLM applications to accelerate the implementation of enterprise-level AI solutions.
Benefits
In addition to the features of open-source Dify, DMS Dify provides the following benefits.
Benefit | Description |
No need to log on to Dify | You only need to log on to DMS to use Dify. |
Quality models | Quality models developed by Alibaba Cloud are available on DMS. ChatBI provides model fusion capabilities with an accuracy of up to 95%. |
One-stop service | DMS provides one-stop management and development for databases, data warehouses, and multi-modal data. This covers end-to-end data development, traditional model training, LLM fine-tuning, and orchestration of published models in Dify. |
Smooth experience | You can quickly create a Dify environment in DMS. Dify allows you to integrate database resources. For example, you can configure database engines, and use Natural Language to SQL (NL2SQL) and native ApsaraDB database plug-ins. |
Prerequisites
Before using Dify for development, you must perform the following operations:
Create a virtual private cloud (VPC). You need to use this VPC when creating a workspace and Dify resources.
Create an ApsaraDB RDS for PostgreSQL database, Tair (Redis OSS-compatible) database, Object Storage Service (OSS) bucket, and AnalyticDB for PostgreSQL vector engine in advance, and register them with DMS. For more information, see Preparations.
Usage notes
During the public preview, Dify is supported only in the China (Hangzhou), China (Hong Kong), and Singapore regions.
During the public preview, the 4 CUs allocated for Dify are free of charge. Subsequent charges will be announced in advance.