https://github.com/kr1shnasomani/visionverse

Image Captioner using DenseNet201 and LSTM
https://github.com/kr1shnasomani/visionverse

computer-vision deep-learning keras matplotlib neural-network numpy pandas tensorflow tqdm

Last synced: 6 months ago
JSON representation

Image Captioner using DenseNet201 and LSTM

• Host: GitHub
• URL: https://github.com/kr1shnasomani/visionverse
• Owner: kr1shnasomani
• Created: 2024-12-19T19:53:15.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-03-18T20:03:47.000Z (over 1 year ago)
• Last Synced: 2025-04-06T13:19:02.290Z (about 1 year ago)
• Topics: computer-vision, deep-learning, keras, matplotlib, neural-network, numpy, pandas, tensorflow, tqdm
• Language: Jupyter Notebook
• Homepage:
• Size: 46.9 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
CaptionCraft

The project is an image captioning model using DenseNet201 and LSTM, generating descriptive captions for images from the Flickr8k dataset. Utilized TensorFlow, Keras, and custom data generators for feature extraction and sequence prediction.


## Execution Guide:

1. Clone the repository:

   ```

   git clone https://github.com/kr1shnasomani/CaptionCraft.git

   cd CaptionCraft

   ```

2. Install the dependencies:

   ```

   pip install -r requirements.txt

   ```

3. On running all the cells it outputs a file named `model.keras` (it stores the trained model)

  

## Accuracy & Loss Over Epochs:

![image](https://github.com/user-attachments/assets/f29900b0-40fb-4ca6-a080-0fd0e421eeed)

![image](https://github.com/user-attachments/assets/7989c24f-adf9-4f2c-8512-aa4776a73947)

## Model Prediction:

![image](https://github.com/user-attachments/assets/fde528fa-b830-4097-902c-d9bb53bf49d4)