Module 1: Introduction to Hugging Face Ecosystem

• Overview of Hugging Face: Model Hub, Datasets, Spaces, and Inference API
• Installing and configuring the Transformers library
• Navigating the Hugging Face Model Hub
• Understanding model cards and selecting the right model

Module 2: Transformer Architecture Fundamentals

• Attention mechanism and self-attention explained
• Encoder-only, decoder-only, and encoder-decoder architectures
• Popular models: BERT, GPT-2, T5, RoBERTa, and DistilBERT
• Tokenisation: BPE, WordPiece, and SentencePiece

Module 3: NLP Pipelines with Transformers

• Using the Hugging Face pipeline API for NLP tasks
• Text classification and sentiment analysis
• Named entity recognition and token classification
• Question answering, summarisation, and translation

Module 4: Fine-Tuning Pre-Trained Models

• When and why to fine-tune vs use out-of-the-box models
• Preparing datasets with the Hugging Face Datasets library
• Fine-tuning BERT for text classification with Trainer API
• Evaluating fine-tuned models with standard NLP metrics

Module 5: Parameter-Efficient Fine-Tuning (PEFT)

• Introduction to PEFT: LoRA, QLoRA, and Prefix Tuning
• Fine-tuning large models on limited hardware
• Using the PEFT library with Hugging Face Transformers
• Practical fine-tuning of Llama and Mistral models

Module 6: Working with Large Language Models

• Loading and running LLMs locally with Transformers
• Quantisation with bitsandbytes for efficient inference
• Text generation parameters: temperature, top-p, and repetition penalty
• Deploying models with Hugging Face Inference Endpoints

Module 7: Hugging Face Spaces and Deployment

• Building and deploying NLP apps with Gradio on Spaces
• Hugging Face Inference API for production use
• Model optimisation with ONNX and TorchScript
• Cost and latency considerations for production NLP

Module 8: Hands-on Projects and Assessment

• Fine-tuning a sentiment classifier for a domain-specific dataset
• Building an NLP pipeline application deployed on HF Spaces
• End-to-end NLP project with evaluation and reporting
• Final assessment and course certification

Experience in the Industry Learn from NLP engineers who have fine-tuned transformer models and deployed Hugging Face pipelines for text classification, summarisation, and named entity recognition at scale.

Backgrounds at the Top Our Hugging Face trainers have contributed to NLP projects at leading AI research organisations and enterprise data science teams.

Clear & Effective Teaching Transformer architecture, tokenisation, model fine-tuning, and the Hugging Face ecosystem are taught progressively from fundamentals to deployment.

Hands-On Learning Focus Students fine-tune pre-trained models, build NLP pipelines, and deploy Hugging Face models using Inference API and Hugging Face Spaces.

Up-to-Date Knowledge Trainers cover the latest Hugging Face releases including Transformers, Datasets, PEFT, and the Model Hub with thousands of pre-trained models.