Tag
6 articles
NVIDIA introduces Structured Data Models, a new library that enables zero-training tabular predictions using pretrained foundation models. The tool simplifies data science workflows by leveraging in-context learning for rapid, accurate predictions.
Knowledgator's new GLiFormer encoder achieves a 91.10 F1 score on nested JSON extraction, nearly matching GPT-5.6-luna's performance while avoiding token generation.
SAP has acquired Prior Labs for over €1 billion, focusing on AI that understands structured data like spreadsheets, signaling a shift from chatbots to data reasoning in enterprise AI.
Learn how to use Lift to extract structured data from research PDFs, create controlled evaluation benchmarks, and build a queryable knowledge base from extracted information.
Datalab introduces lift, a 9B open-weight vision model that accurately extracts structured JSON from PDFs using schema-constrained decoding and abstention training, achieving 90.2% field accuracy.
Learn how to build AI pipelines that produce reliable, structured outputs using Outlines and Pydantic. Understand how type-safe and schema-constrained LLM pipelines work and why they matter for real-world AI applications.