WSQ Deep Learning (SF) (Synchronous and Asynchronous E-Learning)
About This Course
This module covers building deep learning models and deploying them for inference across five instructional units. Learners design and train deep neural networks, develop convolutional neural networks for image data, and build recurrent neural networks for sequential data. They then tune model performance through hyperparameter optimization, and serialize trained models for deployment and inferencing.
The toolset comprises Python, TensorFlow/Keras, Jupyter Notebook, hyperparameter tuning libraries (Keras Tuner, Optuna), and ONNX, with learning consolidated through a project that moves from building and training deep neural networks to developing convolutional and recurrent architectures, systematically optimizing hyperparameters, and serializing a trained model for inference in a deployed application.
The module suits graduates and working professionals moving into deep learning and AI model development; a working foundation in Python and machine learning fundamentals is recommended. On completion, learners are prepared for job roles such as Deep Learning Engineer, Machine Learning Engineer, AI Developer, and Computer Vision Engineer.
This course is designed to target the audience who would like to gain deep learning skills.
What You'll Learn
This course comprises of following Instructional Units
1. Deep Neural Networks
2. Convolutional Neural Networks
3. Recurrent Neural Networks
4. Hyperparameter Optimization
5. Model Serialization and Inferencing
Knowledge Statements:
Describe the architecture, activation functions, and training process of deep neural networks.
Explain how convolutional neural networks extract spatial features from image data.
Identify how recurrent neural networks model sequential and time-dependent data.
Explain the role of hyperparameters and the strategies used to optimize them.
Describe how trained models are serialized, versioned, and served for inferencing.
Skills Statements:
Build and train deep neural networks for a project.
Develop convolutional neural networks for image classification tasks.
Create recurrent neural networks for sequence-based problems.
Apply hyperparameter optimization techniques to improve model performance.
Serialize trained models and implement inferencing pipelines.
Entry Requirements
Academic Qualification: Minimum 1 credit in O Level or its equivalent
Work Experience – Minimum 1 year experience in any business process