Doab Nuextract3 Smoke Output: 50 Document Samples Processed by a Vision-Language Model

50 document samples were processed on 2026-05-21 using the NuExtract3 vision-language model for structured extraction. The output dataset is derived from the 'davanstrien/doab-nuextract3-smoke-input' source and was created by the user davanstrien. Processing was completed in 5.3 minutes using the 'structured-extraction' mode.

Use Cases

• Benchmarking structured extraction models based on the described vision-language processing.
• Training downstream NLP models on the extracted structured information from documents.
• Analyzing the performance of the NuExtract3 model on a specific document corpus.

Strengths

• Contains 50 processed samples, providing a defined test set.
• Processing details are explicitly documented, including the model used (numind/NuExtract3) and processing date (2026-05-21).

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is a processed output, so its quality is dependent on the source data and model performance.

Provenance

Source

davanstrien/doab-nuextract3-smoke-input

Collection Method

Processed with the numind/NuExtract3 vision-language model in structured-extraction mode.

Freshness

Last updated 2026-05-21 09:35:00; freshness should be verified.