Martin KeywoodCutting through the quagmire of using RAG with LLM’s — introducing BRAGIf you’re reading this then I assume you’re like me in that you love AI, you love the opportunity it offers and are blown away with what…7 min read·Mar 29, 2024----
Martin KeywoodFine tuning a T5 small model to generate SQL from natural language with 92.3% accuracyI was inspired by https://huggingface.co/cssupport/t5-small-awesome-text-to-sql and the results it was achieving, and wondered if I could…5 min read·Mar 29, 2024--1--1
Martin KeywoodLLM Prompt Engineering — going beyond the basics, and how you can mitigate against confusionLLM’s can sometimes feel brittle, or confusing. How can you help make things a little clearer?9 min read·Apr 11, 2023--1--1
Martin KeywoodA Simplified Explanation of Machine Learning and Deep Learning — with a twistDISCLAIMER: Please don’t judge me too harsh on the simplistic tone and content here. Please read to the end. I believe you’ll be impressed…7 min read·Dec 24, 2022--1--1
Martin KeywoodWhat do the Machine Learning Training and Validation graphs tell us?Do you look at the Training and Validation graphs in Machine Learning and see a mess or a thing of beauty? Either way, they tell us more…8 min read·Jun 7, 2022----
Martin KeywoodinTowards Data ScienceA Simple Maths Free PyTorch Model FrameworkLove Data Science and Deep Learning but get confused by all the maths and formulas and just want a simple explanation and set of examples?11 min read·Aug 29, 2021----
Martin KeywoodinTowards Data ScienceWould it really be possible to have a single model that could do custom NER and QA on the same data?Having been hugely impressed so far with Transformers and my experimentation with BERT and GPT2, I came across T5 and thought it just…10 min read·Dec 20, 2020----
Martin KeywoodinTowards Data ScienceHow could I do simple inference on my Fine-Tuned Transformers NER models?“NER, NER, na NER, NER !!” — Understanding NER in Transformers had been mocking me for a while. It was time for me to get the last laugh.7 min read·Dec 15, 2020--1--1