Machine Learning 2.0: Transformer for language processing and image generation
.svg)
.svg)
.svg)
Machine Learning Engineer
1. the path to Transformer technology
- Why are transformers a groundbreaking development?
- Possible applications for Transformer models
- The path to Transformer technology
- Supervised and unsupervised learning
- The connections with deep learning and neural networks
2. prerequisites for understanding transformers
- Multi-layer perception (MLP) and feed-forward networks (FNN)
- Loss functions, batch normalization and encoder-decoder architecture (ResNet)
- Introduction to N-Gram and Word2Vec
- Recurrent Neural Networks (RNN) and Long Short-term Memory (LSTM)
3. the Transformer architecture explained in detail
- Use transformer for large language models (LLM)
- Word embeddings, position encoding, self-attention and multi-head attention
- Connecting the building blocks of the transformer
- Comparison of Transformer models: BERT versus GPT
- Transformer for the recognition and generation of image data
- Architecture and use of vision transformers
4. application of transformer models
- Instruction tuning and strategies for effective prompt engineering
- Integration of Retrieval Augmented Generation (RAG) with Transformer architectures
- Capabilities of transformers to generate different media types
- Scalability and adversarial (hostile) attacks on transformers
You will learn all about the technical and mathematical basics of the new Transformer technology and the models that are being developed with it.
You will get a sound impression of the possibilities offered by transformers and where their limits lie.
You understand how generative AI works in detail and how to develop generative AI in the areas of text and image generation.
With Retrieval Augmented Generation (RAG), you will learn a method that allows you to process external information with transformers.
You will gain insights into the training, evaluation, application and integration of Transformer models.
The content of this training supports the obligation to provide evidence of the promotion of AI competence within the meaning of Art. 4 EU AI Regulation.
This training training is conducted in a group of a maximum of 12 participants using the Zoom video conferencing software.
Individual support from the trainers is guaranteed - in the virtual classroom or individually in break-out sessions.
The case studies are provided in the form of Jupyter notebooks, which you can easily install locally on your own computer. You do not need any prior technical knowledge.
Once you have registered, you will find all the information, downloads and extra services for this training course in your online learning environment.
This training course is aimed at anyone who wants to understand the complex Transformer technology in detail and use it in their own projects.
Developers who want to develop applications with generative AI will learn how the underlying data models are structured and how they work.
Data scientists and data analysts will learn how models are created with Transformer technology and how they can be used for data analysis.
Software architects and designers will learn about the concepts behind generative AI and large language models and gain impetus for integration and deployment.
Basic knowledge of mathematical algebra, statistics and machine learning concepts is required.
This course is a valuable building block in the qualification as a Machine Learning Engineer, Data Engineer and Data Scientist.
.svg)
