You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
DataConnectors should have a common interface to interact with AutoLLM platform.
TODOs:
Define the interface.
Define the placeholders each task demands.
For example: QA task would demand {question}, {context} placeholders in the input template. Classification task would demand {possible labels}, {input} as placeholders.
Implement DataConnectors for selected tasks and datasets. (use the exiting builder for PICO and PubMedQA for this)
DataConnectors should finally export datasets to LLM4KMU HF org or store them as DatasetDict locally.
DataConnectors should have a common interface to interact with AutoLLM platform.
TODOs:
For example: QA task would demand {question}, {context} placeholders in the input template. Classification task would demand {possible labels}, {input} as placeholders.
builderfor PICO and PubMedQA for this)DatasetDictlocally.