A simple (and incomplete) collection of content I have created and shared over time. Mostly about NLP, LLMs, Information Retrieval, and Vector Search.
Title | ≈ Date | type |
---|---|---|
🗂️⛏️ Structured data extraction with Small Language Models | 2024-08-02 | post + notebook |
🌉 Introduction to mechanistic interpretability of LLMs | 2024-08-01 | post |
🔎 BM42: a new ranking algorithm for hybrid RAG | 2024-07-05 | post + notebook |
🎤 yo-Llama 🦙: altering the behavior of a LLM by amplifying a feature direction in the activation space | 2024-07-01 | post + model + notebook |
🌌 Creating adventures with local LLMs: llamafile + Character Codex | 2024-06-24 | post + notebook |
RAG Evaluation with 🔥 Prometheus 2 | 2024-06-17 | blog post + notebook |
⚙️ Prompt Optimization with Haystack + DSPy | 2024-06-05 | post + notebook |
LLaMantino 3: a good 🇮🇹 Italian Language Model | 2024-05-20 | post |
🧑🏫 AutoQuizzer: create a quiz from a URL and play/let the LLM play | 2024-05-16 | post + demo |
🐍 PyCon Italy 2024 - LLM/NLP compilation | 2024-05-10 | post |
🔎 Sparse Embedding Retrieval in Haystack | 2024-04-29 | post + notebook |
Playing with 🦙 Llama 3 (RAG about Oscar night 🎬) | 2024-04-19 | post + notebook |
🦙📱 Running Small Language Models on a cheap smartphone | 2024-04-09 | post |
💎 gemma-2b-orpo-GGUF. Thoughts on quantization. | 2024-04-08 | post + model |
💎 gemma-2b-orpo: a Small Language Model trained with ORPO | 2024-03-26 | post + model + notebooks |
🧪📑 From raw text to structured data with LLMs and function calling | 2024-03-20 | post + notebook |
🧭 Choosing an embedding inference solution | 2024-03-10 | post |
LLMs 4 Devs: from 0 to your 1st LLM application @ Open Source Day | 2024-03-08 | talk + repository |
Haystack - an open framework for LLM applications: 🎙️ interview @ Intervista Pythonista with Massimiliano Pippi, in Italian 🇮🇹 | 2024-03-01 | podcast |
Zero-day experiments with Google Gemma 💎 | 2024-02-21 | post + notebook |
🗺️🧭 What's the best LLM inference solution for me? | 2024-02-20 | post |
🧩🧩 Merging Language Models: what I learned | 2024-02-05 | post |
Can Language Models self-improve? 🏋️📈 - Self-Rewarding Language Models paper | 2024-01-22 | post |
🦙 Ollama - Haystack integration | 2024-01-09 | post |
🦙 Ollama - beyond the surface (unpolished notes) 📝 | 2024-01-05 | post |
🇮🇹🇬🇧 Multilingual RAG from a 🎧 podcast | 2024-01-03 | post + notebook |
🧪🦍 Information Extraction with open LLMs and Function Calling | 2023-12-21 | post + notebook |
LLM, Haystack and open-source: 🎙️ interview @ Pointer Podcast with Sara Zanzottera, in Italian 🇮🇹 | 2023-12-15 | podcast |
📌 Collection of notebooks: using Mistral models with the Haystack LLM framework | 2023-12-14 | post + repository |
⛵ Navigating the LLM frameworks landscape: a comprehensive survey | 2023-11-22 | post |
⚗️ How to distill the capabilities of GPT-4 into smaller models? (Zephyr report) | 2023-11-13 | post |
Using 🪁 Zephyr models + Haystack to generate answers on your data | 2023-11-06 | article + notebook |
🔍 Improve RAG by embedding metadata 🏷️ - pt. 2 | 2023-10-25 | post + notebook |
🔍 Improve RAG by embedding metadata 🏷️ - pt. 1 | 2023-10-23 | post + notebook |
Zephyr 7B Alpha: how I sharded a Large Language Model | 2023-10-16 | post |
Loading unstructured data into your LLM application (unstructured.io + Haystack) | 2023-10-02 | post |
Mistral + Haystack: how I built a 🎸 Rock RAG pipeline | 2023-09-29 | post + notebook |
⚡ vLLM: how this fast LLM serving engine works (PagedAttention) | 2023-09-26 | post |
Strategic Passage Ranking for RAG: how to overcome the Lost in the Middle problem | 2023-08-14 | post |
🚀 Load LLMs in Colab 💻 using quantization | 2023-07-31 | post |
Chat with your documents using Haystack | 2023-07-31 | post |
Llama2 🦙 on Haystack | 2023-07-20 | post + notebook |
Hatch: the Python 🐍 project manager | 2023-07-12 | post |
What would mother 👩 say? (by Tuana Çelik): an agent for generating tweets | 2023-07-07 | post |
Fact checking rocks 🎸: how to build a fact-checking system @ Berlin Buzzwords | 2023-06-20 | talk + repository |
💊 Open-source Neural Search/LLM frameworks: an overview | 2023-05-09 | post |
Experimenting with Haystack Agents 🕵️ | 2023-04-05 | post + notebook |
💊 Retrieval Augmented Generation (RAG) | 2023-03-14 | knowledge pill |
💊 Extractive Question Answering Evaluation | 2023-03-13 | knowledge pill |
Celebrating open source 🥳 | 2023-02-28 | post |
New Question Answering model for the 🇮🇹 Italian language! | 2023-02-07 | post |
💊 Machine Reading Models (= Readers) | 2023-02-05 | knowledge pill |
💊 Combining document retrieval and machine comprehension for Question Answering | 2023-03-02 | knowledge pill |
💊 What is Question Answering? | 2023-01-17 | knowledge pill |
💊 Retrieval Evaluation | 2023-01-06 | knowledge pill |
💊 Retrieve and Re-Rank | 2022-12-17 | knowledge pill |
💊 SentenceTransformers for Dense Retrieval | 2022-12-04 | knowledge pill |
💊 Dense Passage Retrieval | 2022-12-03 | knowledge pill |
💊 From sparse representations to Language Models | 2022-11-26 | knowledge pill |
💊 Sparse retrieval: Bag-of-words, TF-IDF | 2022-11-24 | knowledge pill |
💊 Sparse retrieval: BM25 | 2022-11-24 | knowledge pill |
💊 What is Neural Search? | 2022-11-20 | knowledge pill |
Reduce fastText models size | 2022-08-23 | post |
Who killed Laura Palmer? How to implement a question answering system, based on a TV series wiki @ PyCon Italia 2022 | 2022-06-03 | talk + repository |
Monkey patching in Python | 2021-11-29 | post |
Fine-tuning transformers easily with Ludwig | 2021-10-25 | post |
Non-invasive Python profiling with py-spy | 2021-05-23 | post |
Compare sentences with a list of fixed words, using fastText word embeddings | 2021-05-03 | post |
Text and metadata extraction from files with Apache Tika | 2021-04-09 | post |
Table extraction from PDF with Camelot | 2021-03-26 | post |
Simple Docker healthcheck | 2021-03-19 | post |
How to detect political memes? A multimodal approach | 2020-12-17 | post |