Fine-Tuning Models and Large Language Models (LLMs) Training Course

Introduction to Customisation

• What is customisation?
• Use cases and benefits of customisation
• Overview of pre-trained models and transfer learning

Preparing for Customisation

• Collecting and cleaning datasets
• Understanding task-specific data requirements
• Exploratory data analysis and preprocessing

Customisation Techniques

• Transfer learning and feature extraction
• Customising transformers with Hugging Face
• Customising for supervised vs unsupervised tasks

Customising Large Language Models (LLMs)

• Adapting LLMs for NLP tasks (e.g., text classification, summarisation)
• Training LLMs with custom datasets
• Controlling LLM behaviour with prompt engineering

Optimisation and Evaluation

• Hyperparameter tuning
• Evaluating model performance
• Addressing overfitting and underfitting

Scaling Customisation Efforts

• Customising on distributed systems
• Leveraging cloud-based solutions for scalability
• Case studies: Large-scale customisation projects

Best Practices and Challenges

• Best practices for successful customisation
• Common challenges and troubleshooting
• Ethical considerations in customising AI models

Advanced Topics (Optional)

• Customising multi-modal models
• Zero-shot and few-shot learning
• Exploring LoRA (Low-Rank Adaptation) techniques

Summary and Next Steps