Method _process_doc

lm-eval-harness/lm_eval/tasks/scrolls/task.py:451–452 · view source on GitHub ↗

(self, doc)

Source from the content-addressed store, hash-verified

449	DATASET_NAME = "qmsum"
450
451	def _process_doc(self, doc):
452	return [_process_doc_prepended_question(doc)]
453
454	def doc_to_text(self, doc):
455	return f"{doc['text']}\n\nQuestion: {doc['question']}\nAnswer:"

training_docsMethod · 0.45

validation_docsMethod · 0.45

_get_prune_textMethod · 0.45

_process_doc_prepended_questionFunction · 0.85

no test coverage detected