Skip to content

Conversation

rvashurin
Copy link
Collaborator

  • Adds new config group similarity_model which contains parameters for NLI and CrossEncoder models.
  • Moves default values for batch size and models paths to a new config group
  • References new configs in all existing configs
  • Allows support for V2 and V3 deberta models
  • Removes unused hf_cache parameter (can be regulated globally with HF_HOME)

@rvashurin rvashurin requested review from IINemo, ArtemVazh and cant-access-rediska0123 and removed request for IINemo August 15, 2025 10:40
)
self._deberta.to(self.device)

if "v2" in self.deberta_path or "v3" in self.deberta_path:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It seems that the if-else is unnecessary, since everything should work with AutoModelForSequenceClassification and AutoTokenizer

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants