Pinned Loading

1. content-based-movie-recommender12 content-based-movie-recommender12 Public

   Implementation and evaluation of a Visual Riddles benchmark to analyze commonsense and world-knowledge reasoning capabilities of Vision-Language Models (VLMs) using zero-shot and few-shot learning.

   Jupyter Notebook

2. Visual-Riddles-Commonsense-Reasoning-Benchmark-for-Vision-Language-Models Visual-Riddles-Commonsense-Reasoning-Benchmark-for-Vision-Language-Models Public

   Visual Riddles evaluates commonsense reasoning in vision-language models beyond object recognition. Using synthetic images and reasoning-based questions, I built an evaluation pipeline for open-end…

   Jupyter Notebook