List your company

Image Captioning on COCO (test) (BLEU@4, METEOR, CIDEr, SPICE)

149.2CIDErPaLI
4.1241.78579.45117.115Nov 27, 2023Apr 17, 2024Sep 7, 2024Jan 28, 2025Jun 19, 2025Nov 9, 2025Apr 1, 2026

Evaluation results

MethodLinks
PaLITraining Data=1.6B, Parameters=17B, Backbone/LLM=mT5-XXL2023.11149.2———
CogVLMTraining Data=1.5B, Parameters=6.5B, Backbone/LLM=Vicuna-7B2023.11149.1———
Qwen-VLTraining Data=1.4B, Parameters=9.6B, Backbone/LLM=Qwen-7B2023.11148.7———
REVEALTraining Data=1.3B, Parameters=2.1B, Memory Bank=true, Backbone/LLM=T52023.11145.4———
BLIP-2Training Data=129M, Parameters=1.2B, Backbone/LLM=FlanT5-XL2023.11144.542.4——
EVCAPTraining Data=COCO, Parameters=3.97M, Memory Bank=true, Backbone/LLM=Vicuna-13B2023.11140.141.531.224.7
BLIPTraining Data=129M, Parameters=446M2023.11136.740.4——
Qwen3-VL-Instruct RGBPScale=8B2026.04136.638.1828.922.24
MiniGPT4Training Data=5M, Parameters=3.94M, Backbone/LLM=Vicuna-13B2023.11129.63829.623.4
VinVLTraining Data=8.9M, Parameters=110M2023.11129.338.230.323.6
RCA-NOCTraining Data=COCO, Parameters=110M, Memory Bank=true2023.11128.437.429.623.1
OSCARTraining Data=4.1M, Parameters=338M2023.11127.837.430.723.5
AoANet+MATraining Data=COCO, Memory Bank=true2023.111213828.721.8
SmallCapTraining Data=COCO, Parameters=7M, Memory Bank=true, Backbone/LLM=GPT22023.11119.73727.921.3
ClipCapTraining Data=COCO, Parameters=43M, Backbone/LLM=GPT22023.11113.133.527.521.1
ViECapTraining Data=COCO, Parameters=110M, Backbone/LLM=GPT22023.1192.927.224.818.2
ProCapBackbone=Vicuna-1.5-7B2026.0486.2626.222.0814.88
ProCapBackbone=OpenLLaMA-3B2026.0466.0319.921.2214.9
ProCapBackbone=OPT-2.7B2026.0463.419.717.7810.96
ProCapBackbone=TinyLlama-1.1B2026.0446.913.1317.9212.6
Qwen3-VL-InstructScale=8B2026.0414.347.5220.5814.68
Qwen3-VL-InstructScale=2B2026.0414.110.422.1816.14
Qwen3-VL-InstructScale=4B2026.0412.866.8220.4414.32
FastVLMScale=0.5B2026.0411.546.2816.289.38
FastVLMScale=1.5B2026.04106.215.388.92
FastVLMScale=7B2026.049.76.4416.369.38