Sign in to view source links and access this dataset
Description
124,974 high-quality instruction-following examples in English, formatted for AI training. The dataset covers IFC 4.3.x-devel standard entities, properties, relationships, and best practices for Building Information Modeling. It was created by Dietmar2020 and last updated on July 5, 2025.
Use Cases
Fine-tuning language models for IFC schema queries based on the described instruction-following format.
Training AI assistants for BIM software based on coverage of IFC entities and properties.
Developing educational tools for BIM best practices based on the instructional examples.
Creating benchmarks for evaluating model performance on construction domain knowledge.
Strengths
Contains 124,974 total examples, providing a substantial corpus for training.
Includes a defined split of 112,476 examples for training and 12,498 for validation.
Focuses on the IFC 4.3.x-devel standard, suggesting up-to-date technical coverage.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
huggingface user Dietmar2020
Freshness
Last updated 2025-07-05 08:00:58
License is unknown; terms of use must be verified before application.