Archive
All the posts on the Gaia Lab blog, from newest to oldest.
Closing the series with what is not a research line but makes them possible: open models served as Slurm jobs for others to use, hardware validation, the tests prior to enabling containers …
Predicting the chlorophyll of a coastal lagoon from Sentinel-2 imagery with spatiotemporal graph neural networks, and answering the question a reviewer asks: whether the graph contributes …
Detecting grooming, harassment or threats in a minor's conversations without the conversation leaving the phone. The line generates half a million synthetic examples, fine-tunes a …
Segmenting structures in dental images with convolutional networks and transformers, comparing six model families against a classical baseline and, before declaring a winner, training the …
To estimate a forest's biomass you need to know where each tree begins and ends. The line compares three deep learning architectures on point clouds, with ten seeds, and measures what …
An open diffusion model, from the AlphaFold 3 family, applied to thousands of protein pairs to predict the structure of the complex and the confidence of the binding. Fifteen jobs, two …
Almost 300 GGUF files from dozens of families, served with llama.cpp and scored on function calling, code and long context, with speed and energy per card. The line that reserves the most …
Several small open models deliberate to build prompts that make gpt-oss-120b give in. The line measures how much deliberating adds over attacking alone, and devotes as much effort to the …
A 3-billion-parameter Llama 3.2, LoRA fine-tuning and one question: what dose of safety data is needed for it to refuse what is harmful without refusing everything. The line tries fixed …
AI-generated translation
Everything on this site so far was written in Spanish and translated into English by an AI. The translation has not been reviewed by a person: the Spanish version is the one that counts.
