Text to Video Generation Using Alibaba Cloud's Compute Nest Aided by Model Scope
Compute Nest is an all in one service platform which helps the user to create applications by deployment from one console for all the products needed. Compute Nest has some services catering Generative AI solutions powered by Model Scope. This includes applications like RAG with LLM and text to video generation. In this blog we walk through the steps needed to deploy a web application for generating video from the text prompts. The text-video generation model is a model composed of three sub-networks, including: text feature extraction model, text feature-to-video latent space diffusion model, and video space mapping model. The overall model parameters are about 1.7 billion. Currently only English input is supported. The diffusion model uses the Unet3D structure and performs iterative denoising on pure Gaussian noise videos to achieve video generation. Enter the console of Compute Nest.
Click on "Service Catalog". Here you can browse through various services available and get to know more details.
Search for text to video in the keywords and locate the service we need. Here is where we are going to select and create the application.
Click on "View Details".
Note the minimum hardware requirements and click on "Launch Now".
Enter a name for the service and select the region of your convenience. Here the application uses "Singapore".
ECS instance with 16 vCPUs and NVIDIA T4 with PAYG mode is selected.
Enter the password to use for ECS and admin console of the application being incubated. We have chosen Zone A of the VPC in Singapore region.
Since we use already existing VPC, deselect "WhetherCreateVpc". Select an already existing VPC and a vswitch among those in Zone A. Click on "Confirm Order".
Check on the option to agree Computing Nest Service Agreement and click on "Create Now".
Now we reach the list of services being managed upon creation.
The service shows it is deploying and it will take some time to complete.
The deployed service will be identified with the green tick in the status. Now click on details in actions.
In the overview, we can get access to the login URL. This login URL is used to access the application. Enter the username as "admin" and provide the password you entered earlier for the application.
Read previous post:
Tongyi Bailian - Model Studio with Chinese Version of Alibaba Cloud
Read next post:
Tongyi Wanxiang - An Alibaba Cloud SaaS for Multimodal Content Generation
Comments
Related Products
-
Compute Nest
Cloud Engine for Enterprise Applications
Learn More -
Token Plan
Build more, spend less. One plan, every modality.
Learn More -
Alibaba Cloud Model Studio
A one-stop generative AI platform to build intelligent applications that understand your business, based on Qwen model series such as Qwen-Max and other popular models
Learn More -
QwenWork
QwenWork is dedicated to helping employees strengthen their professional competitiveness in the AI era and to enabling enterprises to improve organizational effectiveness.
Learn More
More Posts by ferdinjoe
- Stop Building AI Demos. Start Shipping Enterprise AI. (The SA & FDE Playbook for Qwen)
- Forward Deployed Engineering Using Alibaba Cloud
- Qwen Work for Teachers
- Image and Video Generation from Text Prompts Using Wan 2.5 Preview Model from Model Studio
- Tongyi Wanxiang - An Alibaba Cloud SaaS for Multimodal Content Generation
- Tongyi Bailian - Model Studio with Chinese Version of Alibaba Cloud
- Implementation of Stable Diffusion Web UI Using Heterogeneous ECS Instances
- Setting Alibaba Cloud ECS Instance With NGINX Supporting PHP and MySQL
- Development to Deployment of 3-Tier Architecture of a Simple Web Application Using Alibaba Cloud