James Briggs Courses

AlexNet and ImageNet Explained

How ImageNet and AlexNet made deep learning practical, with a PyTorch walkthrough of image preprocessing and classification.

• YouTube
• 36 minutes
• Self-Paced
• Free Video

How to Use Color Histograms for Image Retrieval

Use color histograms and OpenCV to retrieve visually similar images by comparing their color profiles rather than relying on metadata.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

Evaluation Measures for Search and Recommender Systems

Learn to evaluate search and recommender systems offline with Recall@K, MRR, MAP@K, and NDCG@K, including Python implementations and trade-offs.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

Long Form Question Answering in Haystack

Build a long-form question-answering agent that retrieves information from documents using Haystack, embeddings, and a generator.

• YouTube
• 36 minutes
• Self-Paced
• Free Video

Spotify's Podcast Search Explained

Build a Spotify-style semantic podcast search system with sentence transformers, vector search, synthetic queries, fine-tuning, and recall evaluation.

• YouTube
• 50 minutes
• Self-Paced
• Free Video

Is GPL the Future of Sentence Transformers - Generative Pseudo-Labeling Deep Dive

Learn how generative pseudo-labeling turns unstructured text into training data for fine-tuning sentence transformers.

• YouTube
• 53 minutes
• Self-Paced
• Free Video

Train Sentence Transformers by Generating Queries - GenQ

Learn to fine-tune sentence-transformer bi-encoders with synthetically generated queries from unstructured text for asymmetric semantic search.

• YouTube
• 27 minutes
• Self-Paced
• Free Video

Streamlit for ML - MUI Card Component Build

Build a custom Streamlit card component with React and Material UI, adding titles, text, links, buttons, icons, and the Roboto font.

• YouTube
• 27 minutes
• Self-Paced
• Free Video

How to Build a Q&A AI in Python - Open-Domain Question-Answering

Build an open-domain question-answering system in Python by fine-tuning a Sentence Transformers retriever and querying a Pinecone vector database.

• YouTube
• 39 minutes
• Self-Paced
• Free Video

Building Transformer Tokenizers - Dhivehi NLP #1

Build a WordPiece tokenizer for Dhivehi, addressing low-resource data, Thaana script, and integration with BERT-compatible Transformers.

• YouTube
• 33 minutes
• Self-Paced
• Free Video

Question-Answering in NLP - Extractive QA and Abstractive QA

Explore how open-domain semantic search answers natural-language questions with extractive and abstractive models using open-book and closed-book pipelines.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

All You Need to Know on Multilingual Sentence Vectors - 1 Model, 50+ Languages

Learn to build multilingual sentence transformers by transferring knowledge across languages with parallel data and knowledge distillation.

• YouTube
• 40 minutes
• Self-Paced
• Free Video

Fine-Tune High Performance Sentence Transformers With Multiple Negatives Ranking

Learn to fine-tune sentence transformers with multiple negatives ranking loss, from NLI pair preprocessing through PyTorch and Sentence Transformers implementations.

• YouTube
• 37 minutes
• Self-Paced
• Free Video

Fine-Tune Sentence Transformers the OG Way - With NLI Softmax Loss

Fine-tune Sentence-BERT to produce sentence embeddings with the original natural language inference approach using softmax loss.

• YouTube
• 37 minutes
• Self-Paced
• Free Video

HNSW for Vector Search Explained and Implemented with Faiss - Python

Understand HNSW's layered proximity graphs for approximate nearest-neighbor search, then implement and tune an HNSW index with Faiss.

• YouTube
• 35 minutes
• Self-Paced
• Free Video