BLIP-2 ViT-L FlanT5 XL (zero-shot)

Reported on 4 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing4 results

Visual Question Answering (VQA)onOK-VQA
Accuracy· 2023-01-30
39.4
best: 66.8 (PaLI-X-VPD)
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597
Visual Question Answering (VQA)onVQA v2 val
Accuracy· 2023-01-30
62.6
best: 82.19 (BLIP-2 ViT-G OPT 6.7B (fine-tuned))
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597
Visual Question Answering (VQA)onGQA test-dev
Accuracy· 2023-01-30
44.4
best: 72.1 (CFR)
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597
Visual Question Answering (VQA)onVQA v2 test-dev
Accuracy· 2023-01-30
62.3
best: 84.3 (PaLI)
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597