FreedomIntelligence/MLLM-Bench

Features

• Evaluation Benchmarks - Evaluating multimodal models using automated scoring.
• Multimodal Benchmarks - Evaluation framework using GPT-4V with per-sample criteria.