VLM Performance - a takara-ai Collection

takara-ai 's Collections

Medical

Synthetic Data Generation

LLM Performance

Foundational Vision

VLM Performance

Autonomous Agents

Audio

VLM Performance

updated Jul 10

Vision language models are blind

Paper • 2407.06581 • Published Jul 9 • 82

Note Use the BlindTest Eval benchmark for vision tasks that are easy for humans.