This project demonstrates object localization using deep learning with PyTorch. The model predicts bounding boxes around objects within images, making it ideal for applications in computer vision and robotics.

• Task: Object localization on images.
• Dataset: Custom dataset with bounding box annotations.
• Model: Fine-tuned EfficientNet from TIMM.
• Augmentations: Applied using Albumentations.

1. Clone the repository:

   git clone https://github.com/Navini11/object-localization.git

2. Install the required packages:

   pip install -r requirements.txt

Download and extract the dataset:

git clone https://github.com/parth1620/object-localization-dataset.git

1. Data Preparation: Load and visualize the dataset.
2. Augmentations: Apply transformations to enhance the dataset.
3. Model Training: Train the EfficientNet model to predict bounding boxes.
4. Inference: Test the model on validation data and visualize predictions.

• Training Loss: [Include metrics]
• Validation Loss: [Include metrics]

Run the training script:

python train.py

Run inference on sample images:

python inference.py

• Custom Dataset Class
• Data Augmentation
• EfficientNet Backbone
• MSE Loss for Bounding Box Regression

• PyTorch
• TIMM (PyTorch Image Models)
• OpenCV
• Albumentations

This project is licensed under the MIT License. This is under the guided project done in coursera course, "deep-learning-with-pytorch--object-localization"