Zidong Taichu – A 100 billion parameter full-modal model and assistant launched by the Chinese Academy of Sciences and the Wuzhi Academy | AI Tool Set

Zidong Taichu is a full-modal big model jointly launched by the Institute of Automation of the Chinese Academy of Sciences and the Wuhan Institute of Artificial Intelligence. It is a 2.0 version upgraded and created based on the 100 billion parameter multi-modal big model “Zidong Taichu 1.0”. The Zidong Taichu big model supports multiple rounds of question-and-answer tasks such as question-and-answer, text creation, image generation, 3D understanding, signal analysis, etc., and has strong cognitive, understanding, and creative capabilities, which can bring a new interactive experience.

The main functions of Zidong Taichu

1. Image capability

1. Image description: Based on the image materials uploaded by the user, accurately understand and answer image recognition questions
2. Target detection: Supports target detection tasks of rich categories, and can determine the target type, quantity and corresponding position information.
3. Image search: Based on a massive high-quality image material library, it can search for users with high correlation exquisite image materials
4. Image generation: generate corresponding exquisite pictures based on user command demands, and can modify the description and fine-tune the picture content
5. Text recognition: Based on image content, it supports multi-scenario, multi-lingual, and high-precision text detection and recognition services.

2. Language ability

1. Chinese Q&A: Accurately understand the context of the problem entered by the user and be able to make accurate knowledgeable Q&A
2. Text continuation: Automatically continue writing rich story content based on user input stories
3. Text creation: accurately understand the user’s input intentions and generate text content with coherent semantic meaning and smooth logic
4. Title generation: Based on the understanding of the article or long text, quickly aggregate and generate streamlined and generalized title documents.
5. Grammar analysis: accurately understand and analyze the grammar of Chinese and English sentences, remind grammar errors and modify them, etc.
6. Machine translation: Helps users translate various types of text materials, including Chinese and English translation, classical Chinese and vernacular translation, etc.
7. Ancient poetry creation: improvise poetry, quatrains, etc. based on the theme or guided content given by the user
8. Code understanding: Understand most programming languages, algorithms and data structures such as C language, Python, and JAVA, and quickly give the required solutions
9. Code writing: It can help users quickly write simple code snippets, such as functions, classes, or loops.
10. Mathematical calculation: It can handle both conventional mathematical calculation problems, as well as mathematical application problems of chickens and rabbits in the same cage recorded in “Sunzi Jing”
11. Logical reasoning: supports the handling of complex logical reasoning problems, including scientific reasoning, common sense reasoning, space-time reasoning, etc.

3. Video capability

1. Video description: Based on the video material uploaded by the user, accurately understand and answer questions about video recognition and video description.
2. Video search: Based on a massive high-quality video material library, it can search for users with high correlation exquisite video materials
3. Video Q&A: Based on the video material uploaded by the user, accurately understand and answer video-related questions, while supporting context information understanding and multiple rounds of Q&A

4. Musical ability

1. Music generation: Controllable generation of high-fidelity music through a given text prompt and supports improvisation of music played by multiple styles and instruments
2. Music multimodal Q&A: Based on the understanding of the music materials uploaded by users, relevant multimodal Q&A tasks can be completed

V. Audio capability

1. Audio fake recognition: Zidong Taichu can determine whether the current audio is a real person speaking or a machine synthesis
2. Audio Event Classification: It can detect the sound event types contained in the current audio, currently supports 11 single sound events and mixed sound events
3. Voice recognition: It can quickly and accurately recognize voice as text, and supports voice interaction and voice content analysis for mobile phone applications.
4. Phonetic synthesis: Provide highly anthropomorphic, smooth and natural speech synthesis services to meet the needs of various scenarios such as text reading and voice broadcasting.

6. 3D capability

1. 3D scene description: Zidong Taichu 2.0 has 3D scene understanding and object perception capabilities based on point cloud data

7. Signal capability

1. Signal recognition: supports radar signal identification and knowledge interaction, and can quickly grasp the basic source and parameters of the signal with the help of models.

How to use Zidong Taichu

1. Visit Zidong Taichu’s official website (taichu-web.ia.ac.cn) and click on the dialogue to experience it
2. Log in/register your account. After the login is successfully applied, it will automatically jump to the dialogue interface.
3. Enter your question or enter a slash to select the recommended prompt command (you can also choose the built-in example to view), and then click Send
4. Zidong Taichu will answer your questions intelligently

Frequently Asked Questions