Open QuangTQV opened 2 months ago
Is the instruction model in your project a bi-encoder? Can the cross-encoder be fine-tuned in the direction of 'Instruction-Finetuned Text' like that?
Is the instruction model in your project a bi-encoder? Can the cross-encoder be fine-tuned in the direction of 'Instruction-Finetuned Text' like that?