Files
siemens_ragas/scenarios/siemens_build/siemens-pdf-build.yaml

18 lines
491 B
YAML
Raw Normal View History

job_name: siemens-pdf-question-bank
input:
path: ../../datasets/siemens-pdfs
glob: "*.pdf"
parser:
provider: aliyun_docmind
failure_mode: skip
generation:
output_type: online_question_bank
review_mode: draft_with_manual_review
max_questions_per_document: 10
max_source_chunks_per_question: 3
output:
dataset_path: ../../datasets/raw/generated/siemens-pdf-question-bank.csv
artifact_dir: ../../outputs/dataset-builds/siemens-pdf-question-bank
runtime:
max_documents: 17