NVIDIA Deep Learning Institute Certificate
This project was part of an Nvidia Deep Learning Institute course at American University in Cairo. The project involved experimenting with deep learning and transfer learning using Python, TensorFlow: Keras, and deep learning concepts.
- Created a Convolutional Neural Network to classify images of hands showing Asl letters
- Augmented Training Images to increase variety and prevent overfitting
- Created an LSTM model to generate news headlines based on a given text
- Prepared a pre-trained model called ImageNet for transfer learning
- Retrained ImageNet to detect the presence of a specific dog from a photograph
- Trained and fine-tuned hyper-parameters to reach an accuracy of 97%
- Retrained ImageNet to identify rotten fruits from good ones reaching an accuracy of 96%
- Trained and fine-tuned hyper-parameters to reach an accuracy of 96%
- Python
- TensorFlow: Keras
- Deep Learning
- Transfer Learning
The project was completed as part of an Nvidia Deep Learning Institute course at AuC University. Special thanks to the instructors and teaching assistants for their guidance and support.