Specialize a model: datasets, LoRA, evaluations
Datasets, LoRA configs and evaluations to adapt a model to your domain.
AI dataset on the population of U.S. states
43 dialogues for an AI assistant on U.S. state populations: 2025 estimates,…
AI evaluation of English text anonymization
Checks that an anonymization model or context lets no U.S. personal data…
AI dataset on the population of Dutch municipalities
37 dialogues for an AI assistant on Dutch municipalities and provinces:…
AI dataset on the population of Irish counties
35 dialogues for an AI assistant on Ireland's 26 counties: 2022 census, change…
Configuration QLoRA pour anonymiser avec Mistral 7B
Configuration QLoRA de départ pour spécialiser Mistral 7B Instruct dans…
28 échanges d'appels d'outils fr/en pour un agent sur les noms de lieux du…
20 examples to anonymise UK texts (UK GDPR): NHS numbers, +44 phones, GB IBANs,…
What does adapting a model mean?
Adapting a model means retraining it on your own examples so it masters your domain, vocabulary or formats: that is fine-tuning. It takes a dataset, training settings (often LoRA) and evaluations to check what it learned.