Image Analysis and Captioning for Instagram

Deployed @ image-cap.streamlit.app

Using HuggingFace's transformers library for:

• Image Analysis
• Generate captions alinged for social media
• Save audio as a .wav file in working directory

Installation

• Make a new directory in terminal mkdir <new-directory>
• Navigate into the new directory cd <new-directory>
• Create virtual environment python3 -m venv venv
• Activate virtual environment (mac os/linux) source venv/bin/activate Windows: venv\Scripts\activate.bat
• Clone the repository git clone https://github.com/engagepy/hf_img_analysis_captioning.git
• Install the requirements pip install -r requirements.txt
• Create .env file and add the following variables

HF_TOKEN=
OPENAI_API_KEY=

HuggingFace link: https://huggingface.co/ OpenAi link: https://platform.openai.com/

Usage

• Put an image in the working directory
• Assign it the function in app.py generate_story(img2text("<Your-Imag-Name.Extension>"))
• Run the script python app.py

Enjoy

Special thanks to the giants on whose shoulders we progress compoundingly in computing.

• HuggingFace
• OpenAI
• Streamlit
• PyTorch
• TensorFlow
• LangChain