Parrot


Révision datée du 10 juin 2024 à 09:28 par Pitpitt (discussion | contributions) (Page créée avec « == Définition == XXXXXXXXX == Français == ''' Parrot''' == Anglais == ''' Parrot''' Existing methods mainly focus on aligning vision encoders with LLMs through supervised fine-tuning (SFT) to endow LLMs with multimodal abilities, making MLLMs' inherent ability to react to multiple languages progressively deteriorate as the training process evolves.  We empirically find that the imbalanced SFT datasets, primarily composed of English-centric image-text pairs... »)
(diff) ← Version précédente | Voir la version actuelle (diff) | Version suivante → (diff)

Définition

XXXXXXXXX

Français

Parrot

Anglais

Parrot


Existing methods mainly focus on aligning vision encoders with LLMs through supervised fine-tuning (SFT) to endow LLMs with multimodal abilities, making MLLMs' inherent ability to react to multiple languages progressively deteriorate as the training process evolves.  We empirically find that the imbalanced SFT datasets, primarily composed of English-centric image-text pairs, lead to significantly reduced performance in non-English languages.


Source

Source : huggingface



Contributeurs: wiki