Nikhil Mehta

A new method enhances hyperparameter tuning efficiency using previous model data.

Unlocking the potential of few-shot image manipulation for all.

Unpacking the key elements driving video understanding in large multimodal models.