Zidong Taichu: Multi-modal large model platform supporting text creation, image generation, 3D understanding, signal analysis and other tasks

General Introduction

Zidong Taichu is a new-generation multimodal large model platform launched by the Institute of Automation of the Chinese Academy of Sciences and the Wuhan Institute of Artificial Intelligence. The platform supports multiple tasks such as multi-round question and answer, text creation, image generation, 3D understanding and signal analysis, with powerful cognitive, comprehension and creation capabilities. Zidong Taichu not only provides API access, but also supports the processing and analysis of multimodal data, which is suitable for AI application development and algorithm research in various industries.

Function List

• Text creation: Supports creative copywriting, text expansion, article continuation, content summarization, multi-language translation, writing poems and lyrics, writing code, solving math problems, and more.
• Q&A: Enhancement of large model retrieval through proprietary knowledge bases and networked search to improve knowledge learning efficiency.
• multimodal understanding: Support for graphic quizzing, visual localization, visual referencing, OCR quizzing, music comprehension, and more.
• Multimodal Content Generation: Provide multiple art styles for AI painting and intelligently generate music clips.
• 3D Understanding: 3D scene understanding and object perception based on point cloud data.
• signal analysis: Support for radar signal identification and knowledge interaction.
• Big Model Open Service PlatformProvide one-stop big model training and push service, support the whole process of data import, cleaning, labeling, fine-tuning, release and so on.
• Intelligent labeling: Provide a variety of annotation task templates to support model intelligent annotation and manual review.
• Service deployment: One-click publishing models as online services with API call support.
• Reasoning Enhancement: Build proprietary knowledge bases through documentation, manage and orchestrate plug-ins, and enable large model inference enhancements.

Using Help

Installation and use

1. Register & Login::
   • Visit the official website of Zidong Taichu (https://taichu-web.ia.ac.cn/), click on the "Register" button and fill in the relevant information to complete the registration.
   • After successful registration, use the registered email and password to log in to the platform.
2. API Access::
   • After logging in, go to the "API Access" page and apply for an API Key.
   • After obtaining the API Key, you can call various API interfaces provided by ZiDong Taichu in the development environment.
3. Data import and management::
   • Enter the "Data Center" module and upload the multimodal data that needs to be processed (supports docx, pdf, txt, md and other formats).
   • Data cleaning and matching using built-in data preprocessing operators.
4. Intelligent labeling::
   • In the "Intelligent Labeling" module, select the appropriate labeling task template for intelligent model labeling and manual review.
5. Model fine-tuning::
   • Enter the "model fine-tuning" module, select the pre-trained large model, and perform zero-code one-click fine-tuning.
   • Visualize the training process, view log output and resource monitoring in real time.
6. Service deployment::
   • In the "Service Deployment" module, you can publish the model as an online service with one click and generate the API call address.
   • Supports authorization requests and model export for easy integration into other applications.
7. Reasoning Enhancement::
   • Import documents to build an exclusive knowledge base, manage and orchestrate plug-ins, and realize large model inference enhancement.
   • Publish Agent applications to enhance the utility and credibility of the model.
8. data return::
   • Inference data is collected online, streamed back into the dataset, and the model is continuously iterated and optimized.

usage example

• Text creationCall the Text Creation API and input text creation requirements, such as "Generate an article about artificial intelligence", and the system will automatically generate text content that meets the requirements.
• Image GenerationCall the Image Generation API and enter a drawing instruction, such as "Generate a Van Gogh-style starry sky picture", and the system will generate an image in the corresponding style.
• 3D Understanding: Upload the point cloud data and call the 3D understanding API, the system will analyze and understand the 3D scene and output the object perception result.

The layman's guide to getting started with AI

Help you learn how to utilize AI tools at a low cost and from a zero base.AI, like office software, is an essential skill for everyone. Mastering AI will give you an edge in your job search and half the effort in your future work and studies.

View Details>