آموزش گواهینامه تخصصی مدل‌های زبانی بزرگ (LLM) - آخرین آپدیت

سرفصل ها و درس ها

Module 1. Introduction to LLMs

• Introduction

• 1.1. Overview of Artificial Intelligence and Natural Language Processing (NLP)

• 1.2. Evolution of Language Models (from N-grams to Transformers)

• 1.2. Evolution of Language Models (from N-grams to Transformers)

• 1.3. What Are Large Language Models?

• 1.4. Key Features and Capabilities of LLMs

• 1.4. Key Features and Capabilities of LLMs 2

• 1.5. Activity: Explore LLMs through interactive sessions (e.g., ChatGPT, Bard, )

• 1.6. Conclusion

Module 2. Core Technologies and Architectures of LLMs

• Module 2. Core Technologies and Architectures of LLMs

• 2.1. Neural Networks and Deep Learning Basics

• 2.2. Attention Mechanisms and Transformers

• 2.3. Pre-training and Fine-tuning Paradigms

• 2.4. Tokenization and Contextual Embeddings

• 2.5. Popular LLM Architectures: GPT, BERT, T5, and PaLM

• 2.5. Popular LLM Architectures: GPT, BERT, T5, and PaLM 2

• 2.6. Activity: Visualize attention maps in transformers using tools like Hugging

• 2.7. Conclusion

Module 3. Training and Scaling LLMs

• Module 3. Training and Scaling LLMs

• 3.1. Data Collection and Preprocessing for LLMs

• 3.2. Compute Requirements and Scaling Challenges

• 3.3. Model Optimization Techniques (e.g., mixed-precision training)

• 3.4. Distributed Training for LLMs

• 3.5. Overview of OpenAI GPT, Meta LLaMA, and Google PaLM Training Practices

• 3.6. Activity: Simulate a small-scale model training using libraries like Tensor

• 3.7. Conclusion

Module 4. Applications of LLMs

• Module 4. Applications of LLMs

• 4.1. Text Generation and Summarization

• 4.1. Text Generation and Summarization 2

• 4.2. Chatbots and Virtual Assistants

• 4.3. Sentiment Analysis and Customer Insights

• 4.4. Question Answering Systems

• 4.5 Code Generation and Automation

• 4.5 Code Generation and Automation 2

• 4.6. Activity: Build a chatbot or text summarization tool using OpenAI's API

• 4.7. Conclusion

Module 5. Fine-Tuning and Customizing LLMs

• Module 5. Fine-Tuning and Customizing LLMs

• 5.1. Techniques for Fine-Tuning Pre-trained Models

• 5.2. Domain-Specific Adaptations of LLMs

• 5.3. Few-Shot and Zero-Shot Learning with LLMs

• 5.4. Case Study: Fine-Tuning for Healthcare, Legal, or E-Commerce Applications

• 5.5. Activity: Fine-tune a pre-trained LLM on a specific dataset

• 5.6. Conclusion

Module 6. Deployment and Optimization of LLMs

• Module 6. Deployment and Optimization of LLMs

• 6.1. Model Inference and Latency Optimization

• 6.2. Edge Deployment vs. Cloud Deployment

• 6.3. Introduction to Model Compression Techniques (e.g., pruning, quantization)

• 6.3. Introduction to Model Compression Techniques (e.g., pruning, quantization)2

• 6.4. APIs and Frameworks for LLM Deployment (OpenAI , Hugging Face, TensorFloW)

• 6.5. Activity: Deploy a fine-tuned model via an API and test its performance.

• 6.6. Conclusion

Module 7. Ethical and Security Considerations

• Module 7. Ethical and Security Considerations

• 7.1. Bias, Fairness, and Responsible AI

• 7.2. Data Privacy Concerns and Mitigation

• 7.3. Risks of Misinformation and Misuse (e.g., deepfakes, fake news)

• 7.4. Regulations and Governance for LLMs

• 7.5. Activity: Analyze an ethical dilemma in LLM usage through group discussion.

• 7.6. Conclusion

Module 8. Future of LLMs

• Module 8. Future of LLMs

• 8.1 Advances in Multimodal Models (e.g., GPT-4 Vision)

• 8.2. Emerging Trends in LLM Efficiency (parse models, memory-efficient )

• 8.3. Cross-Disciplinary Applications of LLMs

• 8.4. Research Frontiers in LLMs

• 8.5. Activity: Research and present on the potential impact of LLMs

• 8.6. Conclusion

Capstone Project.

• Capstone Project.