List your company

Image Captioning on NoCaps

127CIDErPALI-17B
52.6471.94591.25110.555Apr 13, 2020Apr 16, 2021Apr 19, 2022Apr 22, 2023Apr 24, 2024Apr 27, 2025May 1, 2026

Evaluation results

MethodLinks
PALI-17Bmode=specialist2023.08127—
LyricsBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12126.8—
PALI-X-55Bmode=per-task finetuning2023.07126.3—
InstructBLIPzero-shot=true, backbone=Vicuna-7B2023.05123.1—
InstructBLIPBackbone LLM=Vicuna-7B2024.06123.1—
InstructBLIPLLM=Vicuna-7B2023.12123.1—
InstructBLIPzero-shot=true, backbone=Vicuna-13B2023.05121.9—
InstructBLIPzero-shot=true, backbone=Vicuna-13B2023.08121.9—
InstructBLIPBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12121.9—
InstructBLIPBackbone LLM=Vicuna-13B2024.06121.9—
InstructBLIPLLM=Vicuna-13B2023.12121.9—
Qwen-VLzero-shot=true, backbone=Qwen-7B2023.08121.4—
Qwen-VL-Chatzero-shot=true2023.08120.2—
Qwen-VLBackbone=Qwen-7B, Evaluation protocol=0-shot2023.12120.2—
Qwen-VLBackbone LLM=Qwen-7B2024.06120.2—
Qwen-VLLLM=Qwen-7B2023.12120.2—
InstructBLIPzero-shot=true, backbone=FlanT5-XXL2023.05120—
InstructBLIPzero-shot=true, backbone=FlanT5-XL2023.05119.9—
MM1-30BShot=16, Model Scale=30B2024.03116—
MM1-30BShot=162024.03116—
OneLLM-7BBackbone LLM=LLaMA2-7B2024.06115.9—
OneLLM-7BLLM=LLAMA2-7B2023.12115.9—
ChatBridge-13BBackbone LLM=Vicuna-13B2024.06115.7—
ChatBridge-13BLLM=Vicuna-13B2023.12115.7—
MiCo-Chat-7BBackbone LLM=Vicuna-7B2024.06111.8—
MM1-30BShot=8, Model Scale=30B2024.03111.6—
MM1-30BShot=82024.03111.6—
MM1-7BShot=162024.03111.1—
SimVLMModel Size=Huge2021.08110.314.5
Emu-Izero-shot=true, instruction_tuned=true2023.07108.8—
SimVLMModel Size=Large2021.08108.514.2
MM1-3BShot=162024.03107.6—
BLIP-2zero-shot=true, backbone=Vicuna-7B2023.05107.5—
BLIP-2Backbone LLM=Vicuna-7B2024.06107.5—
BLIP-2LLM=Vicuna-7B2023.12107.5—
IDEFICS-80BShot=16, Model Scale=80B2024.03107—
IDEFICS-80BShot=162024.03107—
MM1-7BShot=8, Model Scale=7B2024.03106.6—
MM1-7BShot=82024.03106.6—
IDEFICS-80BShot=8, Model Scale=80B2024.03105.7—
IDEFICS-80BShot=82024.03105.7—
MM1-3BShot=8, Model Scale=3B2024.03104.7—
MM1-3BShot=82024.03104.7—
BLIP-2zero-shot=true, backbone=FlanT5-XL2023.05104.5—
BLIP-2zero-shot=true, backbone=Vicuna-13B2023.05103.9—
BLIP-2zero-shot=true, backbone=Vicuna-13B2023.08103.9—
BLIP-2Backbone=Vicuna-13B, Evaluation protocol=0-shot2023.12103.9—
BLIP-2Backbone LLM=Vicuna-13B2024.06103.9—
BLIP-2LLM=Vicuna-13B2023.12103.9—
MM1-30BShot=42024.03103.8—
AIMv2architecture=ViT-3B/14, #patches=576, resolution=336px2024.11102.9—
AIMv2architecture=ViT-H/14, #patches=576, resolution=336px2024.11102.7—
OriginalBit=162025.03102.5—
QuIP#Bit=22025.03102.1—
CASP_QuIP#Bit=22025.03102.1—
AIMv2architecture=ViT-1B/14, #patches=576, resolution=336px2024.11101.5—
SimVLMhugezero-shot=true2021.10101.4—
Unified-IO-XLzero-shot=true2023.08100—
AIMv2architecture=ViT-L/14, #patches=576, resolution=336px2024.1199.9—
MM1-3BShot=42024.0399.7—
IDEFICS-80BShot=42024.0399.6—
BLIP-2zero-shot=true, backbone=FlanT5-XXL2023.0598.4—
Emuzero-shot=true, prompt_strategy=Chain-of-Thought2023.0796.5—
MM1-7BShot=42024.0396.2—
SimVLMModel Size=Base2021.0894.813.1
SigLIParchitecture=ViT-So400M/14, #patches=7522024.1194.5—
OpenAI CLIParchitecture=ViT-L/14, #patches=5762024.1193.3—
SigLIParchitecture=ViT-L/14, #patches=5762024.1192.7—
VinVLModel Size=Large2021.0892.513.1
CASP_GPTQBit=2.22025.0392.4—
CASP_AQLMBit=22025.0391.5—
IDEFICS-9BShot=162024.0389.4—
DINOv2architecture=ViT-g/14, #patches=30342024.1189.1—
GenLIPArch=g/16, Data=8.0B2026.0588.3—
GenLIPArch=So/16, Data=8.0B2026.0587.5—
IDEFICS-9BShot=8, Model Scale=9B2024.0386.8—
IDEFICS-9BShot=82024.0386.8—
SigLIP2Arch=g/16, Data=40.0B2026.0584.4—
OVision2Arch=L/16, Data=12.8B2026.0584.3—
SigLIP2Arch=So/16, Data=40.0B2026.0584.3—
SigLIPArch=L/16, Data=40.0B2026.0584—
SigLIP2Arch=L/16, Data=40.0B2026.0582.9—
GenLIPArch=L/16, Data=8.0B2026.0582.6—
IDEFICS-9BShot=42024.0381.3—
OSCARLModel scale=Large2020.0480.911.3
UNITERModel Size=Large2021.0880.911.3
AIMv2Arch=L/14, Data=12.0B2026.0580.1—
OSCARBModel scale=Base2020.0478.811.7
SoTABModel scale=Base2020.0473.111.2
AQLMBit=22025.0373.1—
IDEFICS-80BBackbone=LLaMA-65B, Evaluation protocol=0-shot2023.1265—
IDEFICS-80BShot=0, Text-only demonstrations=True, Model Scale=80B2024.0365—
IDEFICS-80BShot=0, Includes text-only demonstrations=true2024.0365—
SoTASModel scale=Small2020.0461.59.2
MM1-7BShot=0, Model Scale=7B2024.0361—
MM1-7BShot=02024.0361—
MetaLMzero-shot=true2023.0758.7—
MM1-3BShot=0, Model Scale=3B2024.0355.6—
MM1-3BShot=02024.0355.6—
CLIPArch=L/14, Data=12.8B2026.0555.5—