HashtagifyMe

A webapp that generates hashtags based on the image sentiment and trends of instagram.

Concept

• Uses BLIP: Image Captioning Base model from transformers.
• After given an Image the model generates the labels describing the image or given scenario.
• The extracted labels or description are sent to the instagram and trending hashtags relevant to the particular image are fetched and given as output.

Tech Stack

• Languages - Python, HTML, CSS, JavaScript
• Frameworks - Pytorch, Flask, BootStrap
• Libraries - Transformers

Run Locally

Step 1. Clone the project

  git clone https://github.com/Aditya6122/HashtagifyMe.git

Step 2. Go to the project directory

  cd HashtagifyMe

Step 3. Create and activate a python virtual environment

  python -m venv venv
  venv\Scripts\activate

Step 4. Install dependencies

  pip install -r local_env.txt

Step 5. Add the Instagram credentials

• Create InstagramScraping/credentials.py
• Add your Instagram login credentials

  USERNAME = <your-username>
  PASSWORD = <your-password>

Step 6. Creat a folder for model file

• When you are running the program for the first time. It will download the model state dict from remote drive link.

  mkdir model

Step 7. Make sure you have chromedriver installed.

Step 8. Run the server

  python main.py

Step 9. Getting the Inference

• Go to the localhost

http://127.0.0.1:5000/

• Choose an Image
• Click on Submit

🎉🎉 Hurray !!! It's Working...