Hands-On Large Language Models (6th Early Access)

Brief Table of Contents (Not Yet Final)
1. Categorizing Text
   Supervised Text Classification
      Model Selection
      Data
      Classification Head
         Example
      Pre-Trained Embeddings
         Example
   Zero-shot Classification
      Pre-Trained Embeddings
         Example
       Natural Language Inference
         Example
   Classification with Generative Models
      In-Context Learning
         Example
         Zero-shot Classification
         Few-shot Classification
      Named Entity Recognition
         Example
   Summary
2. Semantic Search
   Three Major Categories of Language-Model-based Search Systems
   Dense Retrieval
      Dense Retrieval Example
         Caveats of Dense Retrieval
      Chunking Long Texts
         One vector per document
         Multiple vectors per document
      Nearest Neighbor Search vs. Vector Databases
      Fine-tuning embedding models for dense retrieval
   Reranking
      Reranking Example
      Open Source Retrieval and Reranking with Sentence Transformers
      How Reranking Models Work
   Generative Search
      What is Generative Search?
   Other LLM applications in search
      Evaluation metrics
         Mean Average Precision (MAP)
   Summary
3. Text Clustering and Topic Modeling
   Text Clustering
      Data
      How do we perform Text Clustering?
         1. Embed documents
         2. Reduce dimensionality
         3. Cluster embeddings
   Topic Modeling
      BERTopic
          Code Overview
      Example
          (Interactive) Visualizations
      Representation Models
         KeyBERTInspired
          Part-of-Speech
          Maximal Marginal Relevance
      Text Generation
         Prompting
         HuggingFace
         OpenAI
         Cohere
         LangChain
      Topic Modeling Variations
   Summary
4. Text Generation with GPT Models
   Using Text Generation Models
      Choosing a Text Generation Model
      Loading a Text Generation Model
      Controlling the Model Output
         Temperature
         top_p
   Intro to Prompt Engineering
      The Basic Ingredients of a Prompt
      Instruction-based Prompting
   Advanced Prompt Engineering
      The Potential Complexity of a Prompt
      In-Context Learning: Providing Examples
      Chain Prompting: Breaking up the Problem
   Reasoning with Generative Models
      Chain-of-Thought: Think Before Answering
         Zero-shot Chain-of-Thought
      Tree-of-Thought: Exploring Intermediate Steps
   Output Verification
      Providing Examples
      Grammar: Constrained Sampling
   Summary
5. Multimodal Large Language Models
   Transformers for Vision
   Multimodal Embedding Models
      CLIP: Connecting Text and Images
         How can CLIP generate multimodal embeddings?
         OpenCLIP
   Making Text Generation Models Multimodal
      BLIP-2: Bridging the Modality Gap
      Preprocessing Multimodal Inputs
         Preprocessing Images
         Preprocessing Text
      Use Case 1: Image Captioning
      Use Case 2: Multimodal Chat-based Prompting
    Summary
6. Tokens & Token Embeddings
   LLM Tokenization
      How tokenizers prepare the inputs to the language model
      Word vs. Subword vs. Character vs. Byte Tokens
      Comparing Trained LLM Tokenizers
         bert-base-uncased
         bert-base-cased
         gpt2
         google/flan-t5-xxl
         GPT-4
         bigcode/starcoder
         facebook/galactica-1.3b
      Tokenizer Properties
         Tokenization methods
         Tokenizer Parameters
         The Tokenizer Training Dataset
      A Language Model Holds Embeddings for the Vocabulary of its Tokenizer
      Creating Contextualized Word Embeddings with Language Models
         Code Example: Contextualized Word Embeddings From a Language Model (Like BERT)
   Word Embeddings
      Using Pre-trained Word Embeddings
      The Word2vec Algorithm and Contrastive Training
   Embeddings for Recommendation Systems
      Recommending songs by embeddings
   Summary
7. Creating Text Embedding Models
   Embedding Models
   What is Contrastive Learning?
   SBERT
   Creating an Embedding Model
      Generating contrastive examples
      Train model
      In-depth Evaluation
      Loss Functions
         Cosine Similarity
         Multiple Negatives Ranking Loss
   Fine-tuning an Embedding Model
      Supervised
      Augmented SBERT
   Unsupervised Learning
      Transformer-based Denoising AutoEncoder
      Domain Adaptation
      Generative Pseudo-Labeling
   Summary