OpenFlamingo: An Open-Source Framework for Training Large Autoregressive Vision-Language Models

Benchmark Model Rank Results
visual-question-answering-on-mm-vetOpenFlamingo-9B (MPT-7B)#158GPT-4 score: 24.8±0.2Params: 9B
visual-question-answering-on-mm-vetOpenFlamingo-9B (LLaMA-7B)#165GPT-4 score: 21.8±0.1Params: 9B
visual-question-answering-on-mm-vet-v2OpenFlamingo-9B#17GPT-4 score: 17.6±0.2Params: 9B
visual-question-answering-vqa-on-core-mmOpenFlamingo-v2#14Overall score: 6.82Deductive: 8.88Abductive: 5.3Analogical: 1.11