Today I tried a tool called ChatLongDoc, which can conduct question-and-answer interactions based on multiple documents or web pages. https://github.com/webpilot-ai/ChatLongDoc/blob/main/chatLongDoc.py
According to the official introduction on GitHub, this project breaks through the length limitations of OpenAI Chat-LLMs (such as ChatGPT), allowing users to interact with any long document to assist in content understanding and obtain valuable insights. Compared to ChatPDF, ChatLongDoc supports various file formats including PDFs, Word documents, TXT files, and web pages. The implementation principle is simple and clear, easy to extend, and can be effectively integrated into other applications.
I operated it in the Google Colab environment, following these steps:
%cd /content
from google.colab import drive
drive.mount('/content/drive')
%cd /content/drive/MyDrive/CLD
! git clone https://github.com/webpilot-ai/ChatLongDoc
%cd ChatLongDoc
! pip install -r requirements.txt
Then replace openai_api_key.txt with your own key. That completes the preparation.
parameter. This time, I provided several links related to Worldcoin and asked ChatLongDoc to help summarize them for me.
!python chatLongDoc.py --text_path "https://worldcoin.org/blog/worldcoin/world-app-faqs" --text_path "https://worldcoin.org/blog/worldcoin/introducing-worldcoin-alex-blania-sam-altman" --text_path "https://worldcoin.org/blog/worldcoin/understanding-orb-why-worldcoin-uses-biometrics"
The reference materials this time were these three links:
https://worldcoin.org/blog/worldcoin/world-app-faqs https://worldcoin.org/blog/worldcoin/introducing-worldcoin-alex-blania-sam-altman https://worldcoin.org/blog/worldcoin/understanding-orb-why-worldcoin-uses-biometrics
Below is the answer given by ChatLongDoc based on the webpage content.
Enter your question: What is WorldCoin?
Based on the provided text, WorldCoin is a digital currency system that uses biometric recognition technology. It employs a biometric imaging device called Orb to verify an individual's unique identity. By performing iris scans, Orb generates an iris code to confirm an individual's uniqueness. This biometric recognition technology helps solve the problem of 4.04 billion people globally who lack access to identity verification, enabling them to obtain financial services and participate in the global economy. At the same time, WorldCoin promises to protect user privacy, deleting collected image data immediately after verification without selling any personal data, and all data will be securely encrypted.
Enter your question: What is the relationship between WorldCoin and OpenAI?
The relationship between Worldcoin and OpenAI is not mentioned in the provided passages.
Enter your question: What is the name of the mobile application inside?
The mobile application mentioned in the passages is called the "World App."
Enter your question: What are the functions of the World App?
According to the document fragment, the functions of the World App include:
Biometric verification through Orb Can obtain Worldcoin digital tokens Provides features such as blogs, media centers, etc. Understand the use of Worldcoin and Orb, and explain why biometric technology is used Provide privacy protection measures, including timely deletion of collected images, no sale of personal data, etc. Can perform biometric verification using irises Provides digital identity solutions, serving 4.4 billion people globally who lack legal identity or cannot undergo digital verification The Orb device collects iris images and uses machine learning and computer vision technology to create an iris code for verifying individual uniqueness All operations take place on the Orb device; unless explicitly requested to back up data, images do not leave the device Worldcoin does not use irises to identify personal identities, only for verifying uniqueness Orb-collected data is securely encrypted Provides more information about privacy and security commitments Provides more resources and news subscriptions regarding Worldcoin, Orb, and privacy commitments
Please note, the above answers are inferred based on the provided document fragments and may be incomplete or contain errors.