Deep Learning Practice

A collection of deep learning mini-projects across NLP / Generative AI and Computer Vision, built while studying applied DL. Each folder is a self-contained script with its own requirements.txt.

📦 Consolidated from two practice repositories into one organized collection.

👁️ Computer Vision

Project	Description	Stack
TransferLearning-Zature	Pneumonia detection from chest X-rays via transfer learning	Transfer Learning
vehicle_tracking	Vehicle detection & tracking in video	YOLOv8 · OpenCV
CNN-CicekClassification	Flower classification with a custom CNN	TensorFlow/Keras
MNIST-OnIsleme	MNIST digit recognition + preprocessing (ANN)	TensorFlow/Keras

Pneumonia detection — results

Transfer learning on chest X-rays. Confusion matrix and a sample prediction from the held-out set (outputs/):

🗣️ NLP & Generative AI

Project	Description	Stack
RAG_ile_sozlesme_avukati	RAG "contract lawyer" — FAISS vector DB + LLM Q&A over contract PDFs	FAISS · OpenAI · LangChain
Smart_Asistan_with_Gemini	Gemini-powered assistant with persistent SQLite memory	Google Gemini · SQLite
LSTM_TextProduction	Character/word-level text generation with an LSTM	TensorFlow/Keras
Sentiment_Analysis_RNN	Sentiment analysis with an RNN (train + predict scripts)	TensorFlow/Keras

🔑 Configuration

The LLM projects read API keys from environment variables (never hard-coded). Copy .env.example to .env and fill in your own keys:

cp .env.example .env
# then edit .env

⚠️ Trained model weights (.h5, .pt), datasets (e.g. the chest X-ray set), and .env files are not tracked — install requirements and download datasets locally to run.

▶️ Running a project

cd <category>/<project>
pip install -r requirements.txt
python <script>.py

📄 License

Released under the MIT License.