🍖 One Piece Character Classifier using Transfer Learning

This project implements a character classifier for the popular anime One Piece using transfer learning techniques. The classifier can identify 18 different characters from the series with high accuracy, making it a valuable tool for fans and developers working on One Piece related projects.

💡 Key Features

• Utilizes MobileNet as the base model with additional custom layers
• Trained on a dataset of 18 One Piece character classes from Kaggle
• Achieves 96% accuracy on the validation dataset
• Deployed in multiple formats for versatile use:
  • TensorFlow Lite
  • TensorFlow.js
  • SavedModel for TensorFlow Serving

🏗️ Technical Details

• Data Split: 80% training, 20% testing
• Model Architecture: MobileNet base with custom layers
• Inference Methods: Web interface via Streamlit and API endpoints via TensorFlow Serving

✍ Usage

The model can be easily integrated into various applications, including:

• Fan websites and apps
• Character recognition tools
• Content moderation for One Piece-related platforms

🧑🏻‍💻 Deployment using Streamlit Web App

• Available online at: https://onepiececlassifier.streamlit.app/
• Can be run locally using: streamlit run app.py

🧑🏻‍💻 Deployment using TensorFlow Serving with Docker

Step 1: Install the Required Packages

To run this notebook, you will need to install the required packages. You can install them using the following command:

pip install -r requirements.txt

Step 2: Run the TensorFlow Serving Docker Container

You can run the TensorFlow Serving Docker container using the following command:

sudo docker run -d --name tf_serving_predict \
  -v /repository_path/saved_model:/models/predict \
  -p 8501:8501 \
  -e MODEL_NAME=predict \
  tensorflow/serving:latest

Step 3: Perform Inference on a New Image

You can now perform inference on a new image using the inference notebook provided in this repository or by sending a POST request to the TensorFlow Serving API endpoint. The API endpoint for the TensorFlow Serving is:

http://localhost:8501/v1/models/predict:predict

Step 4: Stop the Docker Container

After you are done with the inference, you can stop the docker container using the following command:

sudo docker stop tf_serving_predict
sudo docker rm tf_serving_predict