https://github.com/dvp-git/gemini-information-extractor

A simple single interface information extractor app using the latest gemini-vision-pro model
https://github.com/dvp-git/gemini-information-extractor

gemini-api generative-ai llm streamlit

Last synced: 17 days ago
JSON representation

A simple single interface information extractor app using the latest gemini-vision-pro model

• Host: GitHub
• URL: https://github.com/dvp-git/gemini-information-extractor
• Owner: dvp-git
• License: mit
• Created: 2024-03-30T16:43:04.000Z (9 months ago)
• Default Branch: main
• Last Pushed: 2024-07-23T22:27:58.000Z (5 months ago)
• Last Synced: 2024-08-08T00:43:07.042Z (4 months ago)
• Topics: gemini-api, generative-ai, llm, streamlit
• Language: Python
• Homepage: https://gemini-information-extractor-9hvvirtqtzkvdt7ypmv9ju.streamlit.app/
• Size: 14.6 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• project-awesome - dvp-git/gemini-information-extractor - A simple single interface information extractor app using the latest gemini-vision-pro model (Python)

README

        
## Info-Extractor  

A simple image information extractor app built using Google's gemini-vision-pro API and streamlit as front-end.

Install:

---

1. Clone the directory to your local environment

    ```

    git clone https://github.com/dvp-git/gemini-information-extractor.git

    ```

2. Change directory to *gemini-information-extractor* and install the necessary libraries using requirements.txt. Note: Use the latest version of python, preferably >=3.10

   ```

   pip install -r requirements.txt

   ```

4. Create a folder called *.streamlit* in your project directory and a file called *secrets.toml* inside it. Save your key as *GOOGLE_API_KEY="insert-your-key-here"* inside the toml file . Replace the *'insert-your-key-here'* with your own api key. This key is required to run the inference for generation of content.

5. Run the streamlit app as :

   ```

   streamlit run gemini_.py

   ```

   

For information on getting a key check: https://ai.google.dev/ 

Usage:

---

**NOTE: Do NOT upload any confidential documents as it may be retained by the system.**

The app is hosted on : https://gemini-information-extractor-9hvvirtqtzkvdt7ypmv9ju.streamlit.app/

You can either check the app there or follow the install steps  and use it in your local environment.

Application inputs:

- *Input Prompt*: The area represents the kind of information you'd like to extract from the image. This is optional in which case you'd get a simple description of the image.

- *Browse files* : Upload the file from which you want to extract information. Use jpg, jpeg or png format.

- *Generate* : Click this button to generate the information extracted.

Example 1: No input prompt

![image](https://github.com/dvp-git/gemini-information-extractor/assets/43114889/531b0977-0c37-478f-aef8-c0f13765015c)

If you do give an input prompt, you will receive the information you requested.

![image](https://github.com/dvp-git/gemini-information-extractor/assets/43114889/8d86d3e5-1aa5-4a86-b54f-1ed6694deb7e)

Streamlit is a powerful tool for building frontend in a short amount of time. 

You can create powerful applications using LLM's from google or HuggingFace and even LangChain . Models on HuggingFace can even be downloaded to your local environment , incase you do not wish to send your information to an API. 

Resources :

---

https://ai.google.dev/ 

https://docs.streamlit.io/get-started/fundamentals/main-concepts

https://docs.streamlit.io/library/cheatsheet