5 papers
Crisp Attention: Regularizing Transformers via Structured Sparsity
Sagar Gandhi, Vishal Gandhi
The quadratic computational cost of the self-attention mechanism is a primary challenge in scaling Transformer models. While attention sparsity is widely studied as a technique to…
Personalized Image Generation from an Author Writing Style
Sagar Gandhi, Vishal Gandhi
Translating nuanced, textually-defined authorial writing styles into compelling visual representations presents a novel challenge in generative AI. This paper introduces a pipeline…
Fine-Tuning Without Forgetting: Adaptation of YOLOv8 Preserves COCO Performance
Vishal Gandhi, Sagar Gandhi
The success of large pre-trained object detectors hinges on their adaptability to diverse downstream tasks. While fine-tuning is the standard adaptation method, specializing these…
ELSA: A Style Aligned Dataset for Emotionally Intelligent Language Generation
Vishal Gandhi, Sagar Gandhi
Advancements in emotion aware language processing increasingly shape vital NLP applications ranging from conversational AI and affective computing to computational psychology and c…
Prompt Sentiment: The Catalyst for LLM Change
Vishal Gandhi, Sagar Gandhi
The rise of large language models (LLMs) has revolutionized natural language processing (NLP), yet the influence of prompt sentiment, a latent affective characteristic of input tex…