Fine-tune VGG16 Art Classification Model + Flask Web App + AWS Deployment
Budget: ₹1,500 – ₹12,500 INR
I’ve built a basic image classification model using VGG16 in Keras, trained on a dataset of around 6,000 images across 5 categories: Rococo, Minimalism, Fauvism, Renaissance Art, and Pop Art. The model is giving okay-ish results, but I want it refined, better-optimized, and deployed.
Additionally, I’ve already designed my own HTML, CSS, and JavaScript for the web interface locally, and I’d like the freelancer to integrate it with the Flask web app.
Here’s what I need help with:
Deliverables:
Model Improvement:
Analyze my current training pipeline (early stopping, checkpoints, etc.).
Improve accuracy and generalization (fine-tune, apply data augmentation, adjust learning rate, etc.).
Optionally: Train a fresh model if needed for better performance.
Flask Web App Integration:
Integrate my pre-designed HTML, CSS, and JavaScript into the Flask web app.
The app should allow users to upload an image and get a prediction.
Display predicted category and confidence score
Show top-3 predictions and enhance the user interface.
AWS Deployment:
Host the Flask app on AWS EC2 (or suggest best alternative).
Ensure proper security configurations, port setups, and public access.
What I’ll Provide:
Jupyter notebook with the current model.
Image dataset (or zipped version).
Sample predictions + model checkpoints.
HTML, CSS, and JS files that I’ve designed locally for the web interface.
Since its a college project, I do need the code.
Additionally, I’ve already designed my own HTML, CSS, and JavaScript for the web interface locally, and I’d like the freelancer to integrate it with the Flask web app.
Here’s what I need help with:
Deliverables:
Model Improvement:
Analyze my current training pipeline (early stopping, checkpoints, etc.).
Improve accuracy and generalization (fine-tune, apply data augmentation, adjust learning rate, etc.).
Optionally: Train a fresh model if needed for better performance.
Flask Web App Integration:
Integrate my pre-designed HTML, CSS, and JavaScript into the Flask web app.
The app should allow users to upload an image and get a prediction.
Display predicted category and confidence score
Show top-3 predictions and enhance the user interface.
AWS Deployment:
Host the Flask app on AWS EC2 (or suggest best alternative).
Ensure proper security configurations, port setups, and public access.
What I’ll Provide:
Jupyter notebook with the current model.
Image dataset (or zipped version).
Sample predictions + model checkpoints.
HTML, CSS, and JS files that I’ve designed locally for the web interface.
Since its a college project, I do need the code.