davanstrien
Machine Learning Librarian @huggingface
Built from davanstrien's public GitHub profile, United Kingdom. Not on Take My Repo yet, so there are no posts or updates here.
Hugging Face recipes: run OCR, transcription, classification, embeddings and more over a Hub dataset in one command, locally or on HF Jobs. Built for people and coding agents.
Using open source LLMs to build synthetic datasets for direct preference optimization
Batch inference for datasets: rows in, any OpenAI-compatible endpoint, resumable parquet out. Congestion-aware concurrency — no tuning.
Command Line Interface for running 🤗 Transformers Image Classification locally
AI Design Patterns for Galleries, Libraries, Archives, and Museums
Experimental tl;dr summaries for datasets on the Hugging Face Hub!
How the data, model and evals for a small curation decision model were made (ultrabert-380m-curation)
Workshop materials on Python as part of a series of Library Carpentry workshops at the British Library
Wouldn't it be nice to generate parts of our dataset card automagically?
A minimal, toolz-style helper for batched, bounded-memory access to Hugging Face bucket files.
Convert Data Lifeboats from Flickr Foundation to HuggingFace datasets and Spaces
Read and write Hugging Face Hub buckets with Polars (pure-Python IO plugin)
Structured extraction from GLAM documents: benchmark harness, typed-KIE scorer and the National Library of Scotland index-cards leaderboard
POC: automatic bucket-backed snapshots of agent working directories — a safety net for when agents do something weird
Minimal hf CLI extension for the Hugging Face datasets library: no-download inspect, streaming preview, footgun-free folder->Hub publish (exploratory)