stanfordnlp / pyvene

Stanford NLP Python Library for Understanding and Improving PyTorch Models via Interventions
http://pyvene.ai
Apache License 2.0
545 stars 46 forks source link

[P1] Disable base output by default in fwd() and gen() #118

Closed frankaging closed 4 months ago

frankaging commented 4 months ago

Description

The forward() and generate() call by default returns the output of unintervened model. This could cause slow down during training and inference. We disable them by default.

Testing Done

notebook is updated.

Checklist: