Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Course Outline
Introduction to Deep Learning
- Distinguishing deep learning from traditional machine learning approaches.
- Exploring real-world applications in computer vision, NLP, and other domains.
- An overview of the deep learning ecosystem, including TensorFlow 2.x, Keras, and PyTorch.
- Configuring a GPU-accelerated development environment.
The Mechanics of Deep Learning
- Examining artificial neurons, activation functions, and network layer structures.
- Understanding forward propagation and the computation of predictions.
- Applying loss functions for both classification and regression tasks.
- Mastering gradient descent optimization and backpropagation.
- Training a foundational neural network using the MNIST dataset.
Convolutional Neural Networks for Computer Vision
- Concepts of convolution, filters, and feature maps.
- Utilizing pooling layers for dimensionality reduction.
- Reviewing key CNN architectures such as LeNet, VGG, and ResNet.
- Constructing and training a CNN for image classification.
- Visualizing learned features and intermediate activations.
Data Augmentation and Improving Model Accuracy
- How data augmentation mitigates overfitting and enhances generalization.
- Applying image transformations including rotation, flipping, zooming, and cropping.
- Creating augmentation pipelines using Keras preprocessing layers.
- Incorporating regularization techniques like Dropout and batch normalization.
- Monitoring training progress via validation metrics and early stopping.
Transfer Learning with Pre-Trained Models
- The principles behind transfer learning and its effectiveness.
- Loading pre-trained models from Keras Applications, such as ResNet, EfficientNet, and MobileNet.
- Performing feature extraction by freezing base layers and training new classifiers.
- Fine-tuning models by selectively unfreezing layers for domain adaptation.
- Achieving high accuracy with constrained training data.
Recurrent Networks and Sequence Modeling
- Understanding sequential data and temporal dependencies.
- Exploring Recurrent Neural Networks (RNNs) and the vanishing gradient issue.
- Using LSTM and GRU cells to handle long-range dependencies.
- Training a character-level text generation model.
- Leveraging word embeddings and the Keras Embedding layer.
Natural Language Processing Fundamentals
- Text preprocessing steps: tokenization, padding, and vocabulary construction.
- Developing text classifiers using RNNs and LSTMs.
- Concepts of sequence-to-sequence models for machine translation.
- The role of attention mechanisms in contemporary NLP.
- Practical NLP implementation using TensorFlow 2.x text processing APIs.
Final Project: Image Captioning
- Integrating computer vision and NLP within a multimodal architecture.
- Extracting image features using a pre-trained CNN encoder.
- Building an LSTM-based decoder for generating captions.
- Managing multiple input layers via the Keras functional API.
- Training and evaluating the complete captioning pipeline.
Next Steps and Resources
- Deploying trained models using TensorFlow Serving.
- Investigating transformer architectures and large language models.
- NVIDIA DLI advanced workshops and certification paths.
- Accessing community resources, datasets, and project inspiration.
Requirements
- Foundational skills in Python programming, including functions, loops, dictionaries, and arrays.
- Understanding of core programming concepts such as variables, conditional logic, and data structures.
- No prior experience in deep learning or machine learning is necessary.
Target Audience
- Software developers and engineers looking to transition into AI and machine learning.
- Data analysts and data scientists aiming to expand their skill set with deep learning techniques.
- Technical professionals seeking to comprehend and apply neural network models.
- Students and researchers starting their exploration of deep learning.
8 Hours
Testimonials (2)
The training was organized and well-planned out, and I come out of it with systematized knowledge and a good look at topics we looked at
Magdalena - Samsung Electronics Polska Sp. z o.o.
Course - Deep Learning with TensorFlow 2
I really liked the end where we took the time to play around with CHAT GPT. The room was not set up the best for this- instead of one large table a couple of small ones so we could get into small groups and brainstorm would have helped
