Machine Learning & AI Model Optimization Training
Level
AdvancedDuration
24h / 3 daysDate
Individually arrangedPrice
Individually arrangedMachine Learning & AI Model Optimization Training
The Machine Learning & AI Model Optimization course is a practical, 3-day program focused on improving the performance and efficiency of machine learning models during training and inference. Participants will learn techniques such as quantization, pruning, mixed-precision training, and tools for accelerating models, including ONNX, TensorRT, and Triton. The program combines 80% hands-on workshops with 20% theoretical introduction, providing skills to optimize machine learning models in domains such as Computer Vision, Natural Language Processing, and LLMs/SLMs, with a focus on edge device deployment and production environments.
Who is this training for?
Data scientists and machine learning engineers with project experience in ML-
Specialists responsible for deploying models in resource-constrained environments
-
Professionals working with large language models and computer vision who want to optimize performance
-
ML engineers interested in advanced training and inference methods to reduce time and costs
What will you learn during this training?
- Apply advanced optimization techniques for machine learning and deep neural network models
- Accelerate training and inference processes while maintaining model quality
- Compress models and adapt them for edge device deployment
- Work with tools (ONNX, TensorRT, Triton) and distributed frameworks for handling large models
- Design scalable and efficient AI solutions ready for production
Training Program
-
Day 1: Optimization Basics and Advanced Work Environments
-
Module 1: Introduction to ML & AI Model Optimization
- Needs and objectives of optimization across modeling stages
- Overview of techniques: quantization, pruning, mixed-precision training
- Data pipeline optimization – improving loading and augmentation
- Agile model lifecycle management for optimization (benchmarking, evaluation)
-
Module 2: Data Pipelines and Flow Management
- Automating loading, preprocessing, and augmentation for efficiency
- Practical: testing and tuning augmentation pipelines
- Validating data quality and efficiency in model training
-
Module 3: Technologies and Tools for Optimization
- ONNX – model exchange and acceleration standard
- Inference acceleration frameworks: TensorRT, Triton
- Integration with popular libraries (PyTorch, TensorFlow)
- Hands-on workshop: preparing a model for optimization
-
Module 1: Introduction to ML & AI Model Optimization
-
Day 2: Optimization Techniques and Practical Deployment
-
Module 4: Training Process Optimization
- Mixed-precision training – reducing resource requirements
- Gradient accumulation and distributed training
- Hyperparameter selection and tuning for optimization
-
Module 5: Model Compression and Post-Training Adaptation
- Quantization and pruning methods and applications
- Performance- and memory-optimized model formats
- Deployment of optimized models on edge devices
- Case study: optimizing large language models and computer vision models
-
Module 4: Training Process Optimization
-
Day 3: Production Optimization and Scaling
-
Module 6: Applying Optimization in Production Environments
- Specifics of optimizing LLMs and CV models
- Deploying optimized models in cloud and on-premise environments
- Performance monitoring, diagnostics, and inference troubleshooting
- Practical cost and efficiency analysis: case studies
- Scaling and managing AI models at large scale
-
Module 7: Best Practices and Trends in AI Optimization
- Resource usage monitoring and automated scaling
- Diagnostics and troubleshooting after deployment
- Tools for automated benchmarking and post-update testing
- Trends in large model optimization and adaptive algorithms
- Discussion and participant knowledge exchange
-
Module 6: Applying Optimization in Production Environments
Contact us
we will organize training for you tailored to your needs
Error: Contact form not found.