Ashutosh1975270/transfer-learning-resnet

🧠 Transfer Learning with ResNet

 

A professional deep learning implementation from the 13-Project Machine Learning Curriculum. Fully deployed on Hugging Face Spaces and GitHub using custom Docker networking.

🎯 The Problem

Fine-tuning a pre-trained ResNet-50 image classifier for custom datasets.

🛠️ Architecture

• —Data Engineering: Custom DataLoader utilities and artifact tracking
• —Model Framework: PyTorch nn.Module with optimized training loops
• —Inference Pipeline: Real-time prediction tracking inside the GUI

🚀 How to Run Locally

# 1. Clone the master repository
git clone [repo_url]
cd [project_folder]

# 2. Install dependencies
pip install -r requirements.txt

# 3. Launch the AI Dashboard
streamlit run app.py

📁 Repository Structure

📦 transfer_learning_&_fine-tuning
 ┣ 📂 data/        # Raw and processed serialized data (.pkl)
 ┣ 📂 models/      # Saved neural network weights and metrics
 ┣ 📂 charts/      # Evaluative tracking graphs
 ┣ 📂 notebooks/   # Original experimentation files
 ┣ 📜 app.py       # Core Streamlit execution layer
 ┣ 📜 Dockerfile   # Deployment configurations
 ┗ 📜 path_utils.py # Resilient pathing handlers

Maintained by the automated DL Architecture Pipeline.